Optimization of polypropylene splats using the flame spray

Optimization of Polypropylene Splats Using the

Flame Spray Process

KADHIM AL AMARA

A thesis submitted in partial fulfillment of the requirement for the

degree of

Doctor of Philosophy

Industrial Research Institute Swinburne (IRIS)

Faculty of Engineering and Industrial Science

Swinburne University of Technology

Melbourne, Australia

2012

Abstract

KADHIM AL AMARA Page i

Abstract

Thermal spray is a well-established and widely-employed technology for

producing engineered coatings used in various industrial applications. Thermal spraying

is a solvent-less and low-Volatile Organic Compounds (VOCs) technique with the

capability to process a wide range of coating materials and substrates including metals,

high melting point ceramics and low processing temperature polymers. Thermal

spraying of polymer powders is a 100% solids process used to form a protective coating

for different applications. The relatively low temperature required, simplicity and

portability of the flame spray technique promote the process to extensive use for

deposition of polymer coating applications.

A thermal spray coating is established by customizing the spray parameters

based on the intended application. The coating is formed as molten or semi-molten

individual particles impact, flatten and solidify to form overlapping splats. Thus, a

complete and thorough understanding of the intrinsic building blocks of a thermal spray

coating and individual splats is essential to comprehend the characteristics of the

ensemble of splats that constitute a coating. The effects of spraying conditions on splat

geometries and morphologies of polypropylene (PP) deposited onto glass and mild steel

substrates using the flame spray process must be understood so that these conditions

may be optimized.

Glass substrates were sprayed using different stand-off distances to examine the

influence of stand-off distance (SOD) on PP splat morphology at room temperature. The

studies indicated that decreasing the SOD from 350 to 150 mm produced coherent,

integral disc-shaped splats that exhibited minimum splashing behaviour. Splat diameter

and thickness were shown to be influenced by the particle SOD. A longer SOD

produced larger diameter PP splats that were thin; however, a shorter SOD yielded

unmelted or partially molten splats. The splat diameter to thickness ratio of PP was

found to be about 5 and a SOD of 150–200 mm is recommended to produce a dense

coating with acceptable mechanical properties.

Abstract

KADHIM AL AMARA Page ii

Several statistical concepts were employed to measure the splat metrics of

formation, including an estimate of splash area. Measurements were performed of

equivalent diameter, degree of splashing, spreading factor, deposition efficiency and

circularity of PP splats deposited onto a glass substrate at room temperature using the

flame spray process. The results indicated that as the SOD increased from 100 mm to

250 mm, the degree of splashing, flattening ratio and spread factor increased, while the

deposition efficiency and splat thickness decreased. Splat circularity was steady at

around 0.9 and indicated that the splats were close to circular for all SODs.

The effect of substrate surface topology and chemistry were demonstrated by

spraying PP powder particles onto flat and grit-blasted mild steel substrates at

predetermined temperatures of room temperature (RT), 70°C, 120°C and 170°C.

Splat formation and morphology were also investigated using optical

microscopy and scanning electron microscopy. The splat-substrate interface was

investigated using focused ion beam. The degree of crystallinity of the PP particles was

analysed by Raman spectroscopy. The geometrical merits of the splat qualities were

calculated via “imageJ” software.

Single solidified splats were collected from polished substrates. The study found

that increasing the substrate temperature from room temperature to 170ºC produced PP

splats with larger diameter and improved the splat-substrate contact. Substrate

preheating was shown to be a dominant factor in the splat formation and splat shape.

Raman spectra showed that generally the degree of crystallinity of the PP particles rose

with increasing substrate temperature. Comparison of the spectra of the feedstock

powder and the deposited PP particles confirmed that there was no thermal degradation

of the material under these processing conditions. The intensity of the band at 809 cm-1

rose in relation to that at 841 cm-1.

The experimental examination of polymer splat morphologies produced under

different spray conditions has permitted understanding of the most effective parameters

that contribute to optimal splats. Knowledge on the behaviour of individual splats and

the phenomena related to the thermal spray splat formation of polymer materials have

been enhanced.

Acknowledgements

KADHIM AL AMARA Page iii

Acknowledgements

This research project would not have been possible without the assistance of

many individuals. I would like to show my gratitude to many people in recognition of

all the help and support that I received throughout this project. First and foremost I

would like to thank my primary supervisor, Professor Christopher Berndt for his

guidance, instructions and encouragement and for pushing me sometimes well outside

my comfort zone to complete this project, which launches my research and development

skills along with instilling confidence in my academic abilities and my research

independence.

From Swinburne University of Technology, I’d also like to thank Dr. Yat Choy

Wang for being supportive and cooperative throughout this research. I would like to

express my special thanks and appreciation to Dr. Saeed Saber-Samandari for the

cooperation, discussions and sharing of scientific conversations as well as opening

windows and networking through my research work. Special thanks to Dr. Paul

Stoddart for his help, support and discussion concerning use of micro Raman and

personally for his inputs and helpful attitude. Thanks to Dr. Igor Sbarski for his help and

informative resources regarding polymer behaviour. Thanks to Mr. Arnold Rowntree

for his help with sample preparations and Dr. James Wang for using the SEM. I would

also like to thank Dr. Magdi Morks for his help. I kindly thank the senior technical

officers Mr. Brian Dempster and Mr. Andrew Moore for their help throughout the

progress of my project.

I also have to thank some people from outside Swinburne University who helped

to facilitate various necessary equipment and tools in experiments, testing,

characterization and analysis. I would like to acknowledge Dr. Nick Birbilis of Monash

University for facilitating the use of 3D profilometry, Dr Sharath Sriram of RMIT

University for using the 2D profilometry, Mr. David Vowles from Monash University

(MCEM) for helping with the FIB. Thanks to Dr. Ute Schmidt from WITec, Germany

Acknowledgements

KADHIM AL AMARA Page iv

for her help on micro Raman analysis. Thanks to Mr. Donald McAuley from Struers,

Australia for his help on sample preparation.

I am grateful to the international experts in thermal spray technology during

their visits to Swinburne University of Technology. Thanks to Prof. Ghislain Montavon

from LERMPS, France; Prof. Sanjay Sampath from Stony Brook University, USA; Dr.

Shrikant Joshi from ARCI, India; and Associate Prof. Christophe Chazelas and

Associate Prof. Gilles Mariaux both from Limoges University, France.

From my deep heart, I thank my family at large including father, mother,

brothers and sisters for their prayers, encouragement and push towards completion of

this work. Thanks again, especially, to my Mum and Dad. During my postgraduate

years, this thesis would certainly not have existed without my “little” family

contribution. My wife who committed herself with comfort and a study environment, as

well as taking care of my other responsibilities and obligations including my dear

children. My wife has been always my pillar, support and inspiration by all means, and I

sincerely thank her. I would also like to thank my children (Ahmed, Jennat, Ali and

Abrar) for always being encouraging in a practical sense by saying “Go dad go!” and

emotionally for always being good and well behaved with pleasant school records that

make me proud of them. And with the completion of this PhD, I wish they will be more

proud and educationally motivated.

Declaration

KADHIM AL AMARA Page v

Declaration

I declare that this thesis contains no material that has been accepted for the

award of any other degree or diploma, except where due reference is made in the text.

To the best of my knowledge, this thesis contains no material previously published or

written by another person, except where due reference is made in the text.

Kadhim Al amara

Table of Contents

KADHIM AL AMARA Page vi

Table of Contents

OPTIMIZATION OF POLYPROPYLENE SPLATS USING THE FLAME SPRAY PROCESS

ABSTRACT .................................................................................................................................................... I

ACKNOWLEDGEMENTS .................................................................................................................................. III

DECLARATION .............................................................................................................................................. V

TABLE OF CONTENTS .................................................................................................................................... VI

LIST OF FIGURES ........................................................................................................................................... X

LIST OF TABLES .......................................................................................................................................... XVI

LIST OF ABBREVIATIONS ............................................................................................................................. XVIII

CHAPTER 1 ....................................................................................................................................... 1

1 INTRODUCTION ......................................................................................................................... 1

1.1 PREFACE ........................................................................................................................................... 1

1.2 OBJECTIVES OF THE RESEARCH PROJECT .................................................................................................. 4

1.3 STRUCTURE OF THESIS ......................................................................................................................... 5

CHAPTER 2 ....................................................................................................................................... 7

2 LITERATURE REVIEW .................................................................................................................. 7

2.1 INTRODUCTION ................................................................................................................................... 7

2.2 POLYMERIC THERMAL SPRAY COATING (PTSC) ...................................................................................... 13

2.2.1 Polymer Characteristics in Thermal Spray Coatings ........................................................... 17

2.2.2 Thermal Spray Variables of Polymers ................................................................................. 20

2.2.3 Microstructure of Thermal Spray Polymer Coatings ........................................................... 22

2.2.4 Advantages of Polymer Use in Thermal Spray .................................................................... 25

2.2.5 Applications of Thermal Spray Polymer Coatings ............................................................... 26

2.2.6 Polymer Matrix Composites in Thermal Spray ................................................................... 29

2.3 FLAME SPRAY PROCESS ...................................................................................................................... 31

2.4 SINGLE SPLATS IN THERMAL SPRAY ....................................................................................................... 33

2.4.1 Splat Morphologies in Thermal Spray ................................................................................ 35

2.4.2 Splat Splashing in Thermal Spray ....................................................................................... 36

2.4.3 Splats of Thermal Spray Polymeric Coatings ...................................................................... 38

2.5 INFLUENCE OF STAND-OFF DISTANCE ON SPLAT FORMATION ..................................................................... 39

Table of Contents

KADHIM AL AMARA Page vii

2.6 INFLUENCE OF SUBSTRATE FACTORS ON SPLAT FORMATION ...................................................................... 42

2.6.1 Substrate Surface Temperature .......................................................................................... 42

2.6.2 Surface Topography ........................................................................................................... 46

2.7 SUBSTRATE MATERIAL: MILD STEEL ..................................................................................................... 47

2.8 FEEDSTOCK MATERIAL: POLYPROPYLENE ............................................................................................... 48

2.8.1 Polypropylene Morphology ................................................................................................ 49

2.8.2 Thermal Properties of Polypropylene ................................................................................. 50

2.8.3 Polypropylene Applications ................................................................................................ 51

CHAPTER 3 ..................................................................................................................................... 52

3 EXPERIMENTAL PROCEDURES .................................................................................................. 52

3.1 POWDER PROCESSING ....................................................................................................................... 52

3.2 POWDER INJECTION PORT .................................................................................................................. 54

3.3 CHARACTERISATION TECHNIQUES ......................................................................................................... 56

3.3.1 Scanning Electron Microscopy (SEM) ................................................................................. 56

3.3.2 Energy Dispersive X-ray Spectroscopy (EDS) ....................................................................... 57

3.3.3 Two-dimensional Profilometry ........................................................................................... 58

3.3.4 Three-dimensional Non-contact Profilometry .................................................................... 59

3.3.5 Focused Ion Beam (FIB) ...................................................................................................... 60

3.3.6 Raman Spectroscopy .......................................................................................................... 61

3.3.7 Image Processing ................................................................................................................ 63

CHAPTER 4 .................................................................................................................................... 64

4 INFLUENCE OF STAND-OFF DISTANCE ON POLYPROPYLENE SPLATS DEPOSITED ONTO A FLAT

SURFACE ........................................................................................................................................... 64

4.1 INTRODUCTION ................................................................................................................................. 64

4.2 METHODOLOGY ................................................................................................................................ 65

4.3 RESULTS AND DISCUSSION .................................................................................................................. 66

4.3.1 Particle Size and Morphology Analysis ............................................................................... 66

4.3.2 Splat Analysis ...................................................................................................................... 67

4.3.3 Statistical Analysis of Splat Morphologies .......................................................................... 72

4.3.4 Spreading Factor ................................................................................................................ 75

4.3.5 Molecular Structure (Micro–Raman Spectroscopy) ............................................................ 76

4.3.6 Crystallization of Polypropylene Splats in Thermal Spray ................................................... 77

4.4 CONCLUSIONS .................................................................................................................................. 82

Table of Contents

KADHIM AL AMARA Page viii

CHAPTER 5 .................................................................................................................................... 84

5 INFLUENCE OF STAND-OFF DISTANCE ON POLYPROPYLENE SPLATS DEPOSITED ONTO A ROUGH

SURFACE ........................................................................................................................................... 84

5.1 INTRODUCTION ................................................................................................................................. 84

5.2 METHODOLOGY ................................................................................................................................ 85

5.3 RESULTS AND DISCUSSION .................................................................................................................. 86

5.3.1 Mild Steel Surface Roughening ........................................................................................... 86

5.3.2 Splat Formation on a Rough Surface .................................................................................. 88

5.3.3 Splat Cracking and Delamination ....................................................................................... 93

5.3.4 Energy Dispersive X-ray Analysis ........................................................................................ 97

5.4 CONCLUSIONS ................................................................................................................................ 101

CHAPTER 6 ................................................................................................................................... 102

6 EFFECT OF SUBSTRATE TEMPERATURE ON POLYPROPYLENE SPLATS DEPOSITED ONTO A FLAT

SURFACE ......................................................................................................................................... 102

6.1 INTRODUCTION ............................................................................................................................... 102

6.2 METHODOLOGY .............................................................................................................................. 103

6.2.1 Substrate and Particle Preparation .................................................................................. 103

6.2.2 Microstructural Analysis ................................................................................................... 104

6.3 RESULTS AND DISCUSSION ................................................................................................................ 105

6.3.1 Splat Characterization ...................................................................................................... 105

6.3.2 Splat-Substrate Interface .................................................................................................. 110

6.3.3 Structural Characterisation .............................................................................................. 115

6.4 CONCLUSIONS ................................................................................................................................ 117

CHAPTER 7 .................................................................................................................................. 119

7 EFFECT OF SUBSTRATE TEMPERATURE ON THE SPLAT FORMATION OF FLAME SPRAYED

POLYPROPYLENE ONTO A ROUGH SURFACE .................................................................................... 119

7.1 INTRODUCTION ............................................................................................................................... 119

7.2 METHODOLOGY .............................................................................................................................. 120


7.4 CONCLUSIONS ................................................................................................................................ 127

CHAPTER 8 .................................................................................................................................. 128

8 SPLAT TAXONOMY OF THERMALLY SPRAYED POLYPROPYLENE .............................................. 128

8.1 INTRODUCTION ............................................................................................................................... 128

Table of Contents

KADHIM AL AMARA Page ix

8.2 EXPERIMENTAL PROCEDURES ............................................................................................................ 129


8.3.1 Particle Size Analysis ......................................................................................................... 132

8.3.2 Splat Morphology ............................................................................................................. 134

8.3.3 Splat Equivalent Diameters .............................................................................................. 135

8.3.4 Degree of Splashing (DS) .................................................................................................. 137

8.3.5 Circularity ......................................................................................................................... 138

8.3.6 Feret Diameter ................................................................................................................. 139

8.3.7 Splat Perimeter ................................................................................................................. 141

8.3.8 Spreading Factor .............................................................................................................. 142

8.3.9 Deposition Efficiency (DE) ................................................................................................. 143

8.3.10 Splat Thickness ................................................................................................................. 143

8.3.11 Effect of Particle Size on Deposition Efficiency ................................................................. 145

8.4 CONCLUSIONS ................................................................................................................................ 148

CHAPTER 9 ................................................................................................................................... 149

9 CONCLUSIONS, RECOMMENDATIONS AND FUTURE WORK ..................................................... 149

9.1 CONCLUSIONS ................................................................................................................................ 149

9.2 RECOMMENDATIONS ....................................................................................................................... 150

9.3 FUTURE WORK ............................................................................................................................... 151

CHAPTER 10 ................................................................................................................................. 153

10 REFERENCES .......................................................................................................................... 153

APPENDIX A ................................................................................................................................. 170

SAMPLE-SIZE DETERMINATION ............................................................................................................. 170

APPENDIX B ................................................................................................................................. 179

PARTICLE SHAPE READING VIA COMPUTER IMAGES .............................................................................. 179

APPENDIX C .................................................................................................................................. 180

TRANSFORMATION FUNCTION OF SPREADING FACTOR ......................................................................... 180

LIST OF PUBLICATIONS ............................................................................................................... 183

List of Figures

KADHIM AL AMARA Page x

List of Figures Fig. 2-1. Surface coating techniques used for surface modification. The position of

thermal spray technology in the field is highlighted. ................................................... 7

Fig. 2-2. Typical thermal spray coating processes, modified from modified from [27]. ............ 9

Fig. 2-3. A flow chart that relates technological and industrial classifications of thermal spray to the character of the splat. .............................................................................. 11

Fig. 2-4. Numbers of polymer thermal spray journal publications with respect to year. Note that data has been selected in two-year intervals. ............................................. 13

Fig. 2-5. Fishnet diagram showing parameters involved in splat formation for thermal spray technology. ....................................................................................................... 16

Fig. 2-6. A schematic of the wetting contact angle of a droplet on a substrate with all surface tension forces acting on it. ............................................................................. 22

Fig. 2-7. Schematic of general microstructural features of the major constituents recognised in thermal spray coatings. ........................................................................ 24

Fig. 2-8. A schematic of the typical flame spray process. ........................................................ 31

Fig. 2-9. Schematic of a typical flame spray process illustrating the location of the external powder injection. .......................................................................................... 32

Fig. 2-10. A two-step mechanism of single splat formation in thermal spray deposition. ......... 34

Fig. 2-11. Effect of spray distance of PEEK coatings onto aluminium and low carbon steel substrate on (a) top-surface roughness, (b) adhesion index and (c) hardness [63]. ... 41

Fig. 2-12. Nylon-11 splats deposited onto a glass slide substrate (a) at room temperature and (b) preheated to 190°C [11]. ................................................................................ 43

Fig. 2-13. PEEK splats deposited by plasma spray onto a polished substrate at (a) 23°C and (b) 323°C [14]. .................................................................................................... 44

Fig. 2-14. (a) Morphologies of UHMWPE deposited by flame spray onto a glass slide and (b) the same splat after being oven-heated for 5 minutes at 200°C [47]. ................... 45

Fig. 2-15. The relationship between substrate transition temperature and the thermal conductivity of feedstock materials deposited onto a steel substrate. ........................ 45

Fig. 3-1. Substrate mounting fixture used in this study designed to perform one spray run at multiple stand-off distances. .................................................................................. 54

Fig. 3-2. Schematic of the external powder feed port used in this study, dimensions in millimetres. ................................................................................................................ 55

List of Figures

KADHIM AL AMARA Page xi

Fig. 3-3. Photos of the spray torch and the location of the external powder port used in this study. ................................................................................................................... 55

Fig. 3-4. The ZEISS Supra 40 VP field emission scanning electron microscope used for imaging in this study, and an image of a typical splat obtained using this system. ....................................................................................................................... 57

Fig. 3-5. Photograph of the two-dimensional Ambios XP-2 surface profilometry technique used for surface profile measurements, and an image of the splat thickness profile obtained using the system. .............................................................. 58

Fig. 3-6. Photograph of the non-contact optical Veeco WYKO NT1100 three-dimensional profilometry technique used for surface profile measurements, and an image of a splat thickness profile obtained using the system. ............................... 60

Fig. 3-7. Photograph of the Quanta 3D FEG high-resolution, low vacuum SEM / FIB used for splat sectioning, and an image of a splat cut obtained using the system. ..... 61

Fig. 3-8. Photograph of the Renishaw inVia Raman spectrometry system and an example of a spectrum obtained using the system. .................................................................. 63

Fig. 4-1. SEM images showing the irregular morphology PP feedstock powder: (a) at low magnification (200x) and (b) at high magnification (600x). ...................................... 66

Fig. 4-2. Particle size distribution of polypropylene feedstock powder. .................................. 66

Fig. 4-3. Relationship between aspect ratio and frequency of PP particles for the approximated elliptical shape. .................................................................................... 67

Fig. 4-4. Microscopic image of coalesced splats. ..................................................................... 68

Fig. 4-5. (a) SEM image of a PP “fried-egg” splat on a glass substrate at room temperature; (b) Profile of splat cross section along A–B with splat core and rim contact angles. (Note that the X and Y axes are different scales). ............................. 69

Fig. 4-6. SEM images of single polypropylene splats on a glass substrate at different stand-off distances at room temperature: (a and b) a “fried-egg” splat with few splashes, (c and d) a “fried-egg” splat with little splashing, and (e and f) an overheated fully spread splat...................................................................................... 70

Fig. 4-7. 3D view showing thickness and diameter of PP splat scanned by WYKO surface profilometer at 150 mm stand-off distance. ................................................... 71

Fig. 4-8. 3D view showing thickness and diameter of PP splat scanned by WYKO surface profilometer: (a) splat at 200 mm stand-off distance and (b) splat at 350 mm stand-off distance. ............................................................................................... 71

Fig. 4-9. 2D Profile of splat cross section (along X-Y shown in inset) using data obtained from XP-2 surface profilometry; note that the X and Y axes are different scales, and (b) SEM image of the splat from side view. ........................................................ 72

Fig. 4-10. Frequency of splat shape formations for different spray distances and powder injection distances with N = 100 for each measurement. ........................................... 73

List of Figures

KADHIM AL AMARA Page xii

Fig. 4-11. Relationship between the degree of particle melting and stand-off distance of flame sprayed PP on a glass substrate at room temperature. Data obtained from SEM observations. ..................................................................................................... 74

Fig. 4-12. Schematic of the three splat zones identified with respect to SOD. .......................... 75

Fig. 4-13. Particle size (diameter) and splat size distributions of polypropylene. ...................... 76

Fig. 4-14. Raman spectroscopy spectra for flame sprayed PP splats and bulk PP. .................... 77

Fig. 4-15. Description of a polypropylene splat behaviour during melting and solidification. ............................................................................................................. 78

Fig. 4-16. SEM images of single polypropylene splats onto a glass substrate at 250 mm: (a) the whole splat with different degree of PP crystallizations indicated by dotted circles, (b) magnification of crystallization nuclei and spherulites, and (c) splat edge zoom depicting a region of two joining spherulites. ................................. 79

Fig. 4-17. SEM images of single polypropylene splats on a glass substrate at SOD of 250 mm: (a) the whole splat with different-sized bough-like spherulites, (b) the spherulites’ shape and size at the splat centre and (c) the spherulites’ shape and size at the splat edge. ................................................................................................. 80

Fig. 4-18. SEM images of bonding failure for a PP splat deposited onto a glass substrate: (a) full splat, (b) magnification of areas with differing adhesion, (c) higher magnification of splat edge ripped from the substrate, and (d) area of PP crystallization with higher splat adhesion. ................................................................. 81

Fig. 5-1. Mild steel grit blasted substrate. ................................................................................ 86

Fig. 5-2. Surface roughness of mild steel substrates obtained by optical surface profilometry: (a) contour plot of the mild steel substrate surface with lines X and Y indicating the locations of profile measurements; (b) SEM image of the rough surface of mild steel substrate; (c and d) plot of depth vs. horizontal distance of X and Y profiles, respectively. ................................................................ 87

Fig. 5-3. Three-dimensional image of mild steel surface roughness using 3D profilometry of the mild steel roughened surface (SEM shown in inset figure). ............................ 88

Fig. 5-4. Schematic of mechanical adhesion of a splat onto substrate asperities with indication of a lower surface peak at A and higher surface peak at B. ...................... 89

Fig. 5-5. SEM images of splat formation of PP on a rough surface of mild steel, where (a) is a full splat and (b and c) are two areas of the substrate where PP flow is downwards into grooves. ........................................................................................... 89

Fig. 5-6. SEM images of polypropylene single splats deposited onto a mild steel rough substrate at room temperature with different stand-off distances: (a and b) splats at 100-150 mm, (c and d) at 200-250 mm, and (e and f) at 250-300 mm. ................. 91

Fig. 5-7. Schematic of a droplet deposited onto a rough surface with SEM images showing the effect of stand-off distance: (a) polyproplene splat deposited at low SOD, and (b) polyproplene splat deposited at high SOD. ......................................... 92

List of Figures

KADHIM AL AMARA Page xiii

Fig. 5-8. SEM images of splats at a SOD of 150 mm deposited onto (a) a flat glass substrate and (b) a rough mild steel substrate. ........................................................... 93

Fig. 5-9. Schematic of splat deposited on a rough surface with an air bubble trapped in substrate valleys. ........................................................................................................ 94

Fig. 5-10. SEM images of a polypropylene splat deposited onto a rough surface of mild steel, showing a trapped gas bubble leaking out of the splat leaving a hole or crack; (a) the full splat and (b) a higher magnification image of the crack. .............. 94

Fig. 5-11. SEM images of a polypropylene splat deposited onto a substrate at SOD of 150 mm showing bonding failure: (a) full splat, (b) higher magnification of splat crack and delamination, and (c) higher magnification of the delaminated splat edge. ........................................................................................................................... 96

Fig. 5-12. SEM image of a PP splat deposited onto a rough mild steel surface at room temperature at SOD of 100 mm, showing the points examined by EDX across line AB. ...................................................................................................................... 97

Fig. 5-13. EDX spectra results of eight points indicated by Fig. 5-12 and PP powder. ............. 98

Fig. 5-14. SEM images of PP single splats on a glass substrate at SODs of 100 mm, 150 mm, 200 mm and 250 mm; a, b, c and d respectively showing the eight points of the EDX reading across the splat diameters (between A and B). ............. 100

Fig. 5-15. Chart depicting the weight percentage of carbon in the PP feedstocks and eight points along the splat diameters (between A and B as shown in Fig. 5-13) for PP particles deposited onto a mild steel substrate at four SODs. .................................. 100

Fig. 6-1. Particle size distribution of polypropylene powder feedstock. ................................ 104

Fig. 6-2. Particle and splat size distribution of polypropylene splats deposited onto a mild steel substrate preheated to various substrate temperatures. .................................... 105

Fig. 6-3. SEM images of flame sprayed polypropylene particles deposited onto a polished mild steel substrate at different preheating temperatures; showing the effect of substrate temperature on the splat morphology, size, surface texture and adhesion to the substrate. Figs. a, b, c and d are a top view of full single splats at different preheating temperatures (i.e., room temperature, 70, 120 and 170ºC, respectively). Figs ac, bc, cc and dc show the surface texture at the splat centre for the same splats. Figs ae, be, ce and de show the splat edge in contact with the polished substrates at higher magnification. .............................................. 106

Fig. 6-4. Schematic of a droplet deposited onto a smooth surface showing the forces acting on droplet flattening and the effect of substrate temperature, with SEM images of a polyproplene splat deposited onto a mild steel substrate held at (a) a low temperature of 70ºC and (b) a high temperature of 170ºC. ............................... 109

Fig. 6-5. SEM images of polypropylene splats deposited onto polished mild steel substrates at different preheat temperatures: (a) a splat deposited at room temperature and the side view at higher magnification after FIB milling showing delamination across the splat edge in contact with the substrate; (b)a splat deposited at 120ºC and the side view indicating good adhesion between the splat edge and the preheated mild steel substrate. .............................................. 111

List of Figures

KADHIM AL AMARA Page xiv

Fig. 6-6. SEM images of PP splats deposited onto mild steel substrates showing the degree of melting deformation at two different conditions where the substrate was preheated to (a) 70ºC and (b) 120ºC. ................................................................ 112

Fig. 6-7. SEM images of many polypropylene splats deposited onto a mild steel substrate at preheat temperatures of (a) 70°C, (b) 120°C and (c) 170°C. ............................... 113

Fig. 6-8. FIB images of a PP splat deposited on mild steel: (a) the full splat and (b) a high magnification depicts a good adhesion with no defects at the splat-substrate interface.................................................................................................................... 115

Fig. 6-9. Raman spectra for PP powder and flame sprayed PP on mild steel substrates preheated to various temperatures with the two significant PP peaks highlighted. . 117

Fig. 7-1. Schematic of droplets before and at impact onto smooth and rough substrates....... 121

Fig. 7-2. Three-dimensional surface roughness image of mild steel substrates used in this study. Obtained by ImageJ software. The inset figure shows the SEM image of this surface. .............................................................................................................. 122

Fig. 7-3. SEM images of flame sprayed polypropylene particles impacted onto a grit blasted mild steel substrate. Figures a, b, c and d are the top view of full single splats at preheating temperatures of room temperature, 70, 120 and 170ºC, respectively. ............................................................................................................. 123

Fig. 7-4. SEM images of flame sprayed polypropylene particles impacted onto a grit blasted mild steel substrate. Figures a, b, c, and d show the splat edge in contact with the rough substrate at preheating temperatures of room temperature, 70, 120 and 170ºC, respectively. .................................................................................... 123

Fig. 7-5. SEM images of PP splats impacted on a grit blasted mild steel substrate at two preheat temperatures; (a) 120ºC, (b) 170ºC, (c) higher magnification of image “a” and (d) higher magnification of image “b”. ....................................................... 124

Fig. 7-6. Polypropylene splats deposited onto a mild steel substrate preheated at 170°C, showing a crystalline structure and induced porosity; (a) layer of splats and (b) higher magnification indicating a porous structure. ................................................. 126

Fig. 7-7. Spherulite adhesion between splats of particles deposited onto a mild steel substrate preheated at 170 ° C; (a) two splats in close contact and (b) three splats in close contact. .............................................................................................. 126

Fig. 8-1. SEM images of the same splat showing the two different measurement techniques using image analysis: a) measuring splat particle size without splash features, and b) measuring actual splat area with splash features. ........................... 130

Fig. 8-2. Feret diameter definitions. ....................................................................................... 131

Fig. 8-3. Schematic of the top view of particle and splat shape factors. ................................ 132

Fig. 8-4. SEM image of polypropylene showing the particle size used in this study. ............ 133

Fig. 8-5. Particle size distribution of polypropylene feedstock. ............................................. 133

List of Figures

KADHIM AL AMARA Page xv

Fig. 8-6. SEM images of PP splats deposited onto a glass slide, with SODs of 100 mm, 150 mm, 200 mm and 250 mm for a, b, c and d respectively. ................................. 134

Fig. 8-7. Comparison of particle and splat size distributions analysed without splashes, for different SODs. ................................................................................................... 135

Fig. 8-8. Comparison of particle and splat size distributions, analysed with splash features included, for different SODs. ..................................................................... 136

Fig. 8-9. Comparison between the average splat particle size with and without splash features for different SODs. ..................................................................................... 136

Fig. 8-10. Relationship between the degree of splashing and SOD. ........................................ 137

Fig. 8-11. Relationship between average splats circularity and SOD, for analysis with and without splash features. ............................................................................................ 139

Fig. 8-12. Comparison of the feret diameter for splats with and without splash features at all SODs of 100 splats. ............................................................................................. 140

Fig. 8-13. Feret diameter definition applied to a thermal spray splat. ...................................... 140

Fig. 8-14. Comparison of the splat perimeter with and without splash features at all SODs averaged over 100 splats. ......................................................................................... 141

Fig. 8-15. Relationship between the splat spreading factor and SOD. ..................................... 142

Fig. 8-16. Comparison of the spreading factors for splats with and without splash features at all SODs, including calculation of the splash area for 100 splats. ....................... 142

Fig. 8-17. Relationship between the number of splats per unit area (5 mm2) and SOD. .......... 143

Fig. 8-18. Relationship between splat thickness and SOD. ...................................................... 144

Fig. 8-19. Relationship between splat size and splat thickness for different SOD. .................. 144

Fig. 8-20. Comparison of number of splats for two mean particle sizes at different SOD of 100 splats. ................................................................................................................ 145

Fig. 8-21. Schematic of small and large polymer particles fed externally into the flame. The inset figure shows a small particle ‘bouncing off’ the flame. ........................... 147

List of Tables

KADHIM AL AMARA Page xvi

List of Tables Table 1-1. Summary of thermal spray experiments conducted. .................................................... 6

Table 2-1. Typical characteristics of the major thermal spray processes, adapted from [3, 32 and 33] .................................................................................................................. 12

Table 2-2. Coating technique limitations. ................................................................................... 14

Table 2-3. Major polymers used in thermal spraying with their related attribute. ...................... 19

Table 2-4. Definitions of the dimensionless quantities acting on impacted particles of polymers and metal. ................................................................................................... 22

Table 2-5. Advantages of polymeric thermal spray coating ........................................................ 26

Table 2-6. Polymeric thermal spray coating applications. .......................................................... 28

Table 2-7. Common polymer matrices used in thermal spray and their characteristics. ............. 30

Table 3-1. Thermo-physical properties of polypropylene used in this study. ............................. 52

Table 4-1. Flame spraying parameters used in this study............................................................ 65

Table 4-2. Description of the splat formation zones identified in the thermal spraying of polymers. .................................................................................................................... 75

Table 5-1. Spraying parameters used in this study. ..................................................................... 85

Table 6-1. Flame spray parameters used in this study. .............................................................. 103

Table 6-2. Characteristics of splats impacted onto a substrate held at different temperatures. ............................................................................................................ 109

Table 6-3. Average ratio of Raman spectra peak intensities at 809 cm-1 and 841 cm-1 for polypropylene under different substrate conditions, each value was based on measurements of four different splats, with an average standard deviation of ±0.06. ....................................................................................................................... 116

Table 8-1. Flame spraying parameters used in this study.......................................................... 129

Table 8-2. Equivalent diameter of splats* with and without splash features for different stand-off distances. .................................................................................................. 136

Table 8-3. Comparison of the splat equivalent and feret diameter of 100 splats at each SOD with and without splash features. ............................................................................. 141

Table 8-4. Number of splats for different particle sizes and the relative deposition efficiency between them. ......................................................................................... 146

List of Tables

KADHIM AL AMARA Page xvii

Table 8-5. Relative deposition efficiency (achieved at lower SOD of 100 mm) for two different mean particle sizes. ................................................................................... 148

Table 8-6. Results summary of all splat metrics relative to stand-off distance increase. .......... 148

List of Abbreviations

KADHIM AL AMARA Page xviii


ASTM American Society for Testing and Materials

APS Air Plasma Spray

Bi Biot Number

CA Contact Angle

CV Coefficient of Variation

CVD Chemical Vapour Deposition

DE Deposition Efficiency

DS Degree of Splashing

EAA Ethylene-Acrylic Acid

ED Equivalent Diameter

EDX Energy-dispersive X-ray spectroscopy

EMI Electromagnetic Interference

EVA Ethylene Vinyl Acetate

Ft Feret Diameter

FIB Focused Ion Beam

HVOF High Velocity Oxygen Fuel

K Sommerfeld parameter

LCP Liquid Crystalline Polymer

LDPE Low Density Polyethylene

MDPE Medium Density Polyethylene

Pe Peclet Number

PE Polyethylene

PEEK Polyetheretherketone

PET Polyethylene Terephthalate

PMC Polymer Matrix Composite

PMMA Polymethylmethacrylate

PP Polypropylene

PPS Polyphenylene Sulfide


KADHIM AL AMARA Page xix

PTFE Polytetrafluoroethylene

PTS Polymeric Thermal Spray

PTSC Polymeric Thermal Spray Coating

PVD Physical Vapour Deposition

PVDF Polyvinylidene Difluoride

Re Reynolds number

RFI Radio Frequency Interference

RT Room Temperature

SD Standard Deviation

SEM Scanning Electron Microscopy

SOD Stand-off Distance

Tg Transition temperature

TS Thermal Spray

UHMWPE Ultra High Molecular Weight Polyethylene

VAc Vinyl Acetate

VOCs Volatile Organic Compounds

We Weber number

Chapter 1: Introduction

KADHIM AL AMARA Page 1

Chapter 1

1 Introduction

1.1 Preface

The application of polymers in thermal spray technology has increased because

of their numerous advantages; including competitive cost and low weight, useful

physical properties and higher specific strength than other materials [1]. Thermal

spraying of polymers represents an alternative method to process polymer powders to

produce polymeric coatings. Thermal spraying of polymers can be traced back to the

1940’s [2]. Polymer thermal spray was commercially practiced in the 1980’s [3] with its

development initiated for infrastructure applications and later, many industries

(including automotive, aircraft and petrochemical) began using thermally sprayed

polymer coatings to provide surface protection against humidity, corrosion, and

aggressive chemical environments. Since the mid-1990s, polymer spraying has gained

significant commercial attention and acceptance as a replacement for other coating

processes [4].

Polymers are now used as surface coatings for many different applications

including corrosion protection, wear resistance and decoration using coating techniques

[5-6]. The polymer thermal spray (PTS) process is capable of coating nonconductive

components (e.g,. concrete, plastic, and fibreglass) [7] and heat-sensitive materials (e.g.,

wood products) [8]. Thermoplastic coating use has widened to include applications that

use the toughness, strength and flexibility of the coatings [9]. Thermoplastic materials

can always be remelted and will fuse with new material in the molten state to make it

possible for coatings to be repaired by remelting or applying additional material to a

damaged area [7].

The PTS process uses a heat source to melt the polymeric powder. The molten or

semi-molten particles are propelled towards a prepared surface. The droplets strike the

substrate, flatten, and form thin platelets (also called “splats”) that conform and adhere



to the irregularities of the prepared substrate and to each other. As the sprayed polymer

droplets impact upon the surface, they quench rapidly, solidifying and building up, splat

by splat, into a lamellar structure [3]; thereby forming the polymer thermal spray

coating (PTSC).

There are a large number of interrelated variables in the thermal spray process

that can lead to significant variations in the splat formation and, thus, in the

microstructure of the splats and coating. Coating characteristics such as porosity,

roughness, adhesive (splat-substrate bonding) and cohesive (splat-splat bonding)

strengths depend on the morphology of splat formation. The study of splat formation

and layering initiated in the 1970s [10]. The research in polymer splat formation and

analysis is very limited and focused on the HVOF spray process for the polyimide of

Nylon-11 [11-13] and using HVAF and plasma thermal spray methods for PEEK [14].

Stand-off distance (SOD) was demonstrated to influence the splat morphology

including splat size, degree of splashing, circularity, spreading factor and thickness. The

disc-shaped splat is more desirable to manufacture a dense coating with minimum

porosity and enhanced mechanical properties. Changing SOD leads to changes in the

temperature and velocity of particles that control the cohesion between splats.

Therefore, the SOD could be directly related to the deposition efficiency (DE) of the

feedstock (i.e., the coating material) powder.

Substrate preparation and conditions, including surface morphology, are

important variables that influence splat spreading, splat adhesion and the coating-

substrate interfacial bonding [12]. Since the thermal spray coating reliability is

contingent primarily upon the mechanical bonding of coating materials to the substrate,

it is critical for the substrate to be properly prepared to ensure maximum bond strength.

Mechanical interlocking and chemical bonding are the two main bonding mechanisms

theorised for thermal spray coatings. Good mechanical bonding can be achieved when

splats penetrate into the asperities of a roughened surface. Substrate preparations for

thermal spray include cleaning, drying and substrate roughening using the grit blast

process [15].



Grit blasting the substrate surface is implemented in a number of ways: (i)

mechanically via surface roughening, (ii) chemically by removing any oxide layers and

(iii) a cleaning process by removing rust, scale, sand, burrs and any other contamination

that might affect the coating quality. The scale of the interface roughness is a significant

fraction of the coating thickness [16].

Flame spray is one thermal spray process capable of depositing a wide range of

materials and material forms [17]. Powder flame spraying is the simplest of all the spray

processes, in which a powder is fed through the central bore of a nozzle, where it melts

and is carried by combusting oxy-fuel gases towards the substrate. Flame spraying is a

relatively low cost coating technique, with low energy required to melt the particles, and

particles being deposited at low velocity and low temperature [3]. Any material that

does not decompose, vaporize, sublime or dissociate on heating can be flame sprayed,

including refractory metals, ceramics, polymers, cermets (ceramic and metal

composites), polycers (composites of polymers and ceramics) as well as alloys. These

materials can be deposited onto a variety of substrates. The flame spray deposition

capability is excellent and it is ideal when large spray coverage areas are involved.

HVOF and plasma thermal spray techniques are also used in manufacturing polymeric

coatings.

Polypropylene (PP) is one thermoplastic polymer that can be flame sprayed [18].

Polypropylene demonstrates high tensile strength, compressive strength and impact

resistance, as well as superior dielectric properties and a high strength-to weight ratio.

Furthermore, PP has a low density, provides high resistance to chemical attack, and is

economically attractive for high volume applications such as bridges, ships and

infrastructure, which are primarily mild steel products.

Mild steel is used in a variety of fields, including construction, infrastructure,

automotive, marine, equipment manufacture and buildings. Mild steel is one of the most

widely used metals in industries [19]. Steel manufacture in the United States totals

approximately 90 million metric tons annually [11]. One of the most problematic

conditions that must be overcome is the poor corrosion resistance of mild steel. The low

corrosion resistance of mild steel contributes greatly to financial losses as well as social

consequences such as loss of life [20]. Most mild steel industrial applications are



exposed to aqueous media, including ships, bridges, pipelines, cooling water systems

and heat exchangers. Maintenance or redundancy in existing industrial structures is a

viable option. Despite an increasing interest in the replacement of mild steel with

stainless steel or polymer, the cost of such replacement would rise to billions of dollars.

The primary intent of this research work is to provide an alternate solution to

protect steel from corrosion using a polymer deposited via a portable flame spray

process. The introduction of strict environmental pollution regulations regarding VOC

usage make it necessary for surface coating technology to be efficient, economically

viable and it also should be environmentally safe.

1.2 Objectives of the Research Project

This thesis examines the effects of spray conditions on splat geometries and

splat morphology of polypropylene (PP) single splats deposited onto glass and mild

steel substrates using the flame spray process. The effects of substrate surface

conditions on splat formation and morphology are also demonstrated. The experimental

aim within this focused study is to examine how single splats can be created, analysed,

and then more fully characterized. Ultimately, the aim concerning an enhancement in

knowledge is to examine specifically how building blocks are stacked together as an

ensemble to confer certain attributes.

The objectives of this research work were as follows:

To establish a fundamental understanding of flame spray processing variables on

microstructural characteristics by examining the influence of stand-off distance

on polypropylene splat morphology.

To understand the rapid solidification process of thermal spray deposition by

accurate characterization of splat shapes by means of a quantified metric.

To examine the splat formation and morphology of polypropylene powders

flame sprayed onto mild steel substrates that have been polished and preheated

to different temperatures.

To investigate and consolidate information related to the splat geometry of

polymer feedstocks using several geometric quantities with the aid of imaging

software to determine the splashing behaviour of polymer particles sprayed at

different stand-off distances.



1.3 Structure of Thesis

This thesis is organised as follows:

Chapter 1 provides a summary introduction covering the technology; the process;

the feedstock; and the substrate and parameters changed throughout this study; i.e.,

SOD, substrate chemistry and substrate topology. The objectives of the research and the

structure of the thesis are included in this chapter.

Chapter 2 contains a focused literature review. This encompasses an overview of

the thermal spray technology and where it is positioned with regard to other surface

coating techniques; and the characteristics of polymers use in thermal spray, including

spray variables, microstructure advantages and applications. Flame spraying is the

process used throughout this study. Therefore, literature was gathered that related to the

use of flame spray methods in manufacturing polymer coatings. Substrate factors

influencing the splat morphology, including substrate chemistry and topology, are

covered.

Chapter 3 outlines the experimental methods used for the preparation of the

feedstocks and substrates. This chapter describes the characterization techniques used

throughout the study.

Chapter 4 contains the sections for equipment, experimental methods, results, and

discussion. The conclusion to the study summarises the influence of SOD on the splat

morphology of PP powder particles deposited onto a flat glass substrate surface.

Recommendations are provided for the optimum stand-off distance to spray PP for the

given spray parameters.

Chapter 5 is structured similarly to Chapter 4, with the focus being on the

influence of the stand-off distance on the splat morphology of PP particles deposited

onto a rough mild steel substrate surface.

Chapter 6 examines the influence of the substrate preheating temperature on the

splat morphology of PP powder particles deposited onto a polished mild steel substrate.

Details of the equipment and experimental methods along with results, discussion,



conclusion and recommendations for the optimum preheating substrate to spray PP for

the given spray parameter are presented.

Chapter 7 is structured similarly to Chapter 6. However, it concentrates on the

influence of the substrate preheat temperature with regard to the deposition of PP onto a

rough substrate of mild steel. The splat morphology of PP powder particles is examined.

Chapter 8 details a qualitative study on the rapid solidification process of thermal

spray deposition by means of a quantified metric, using different SODs. This includes

measurement of the splash area, equivalent diameter, and degree of splashing. The

spreading factor, deposition efficiency and circularity with relation to the process

efficiency are also discussed.

Finally, the overall conclusions, recommendations and future work are presented

in Chapter 9. Table 1-1 summarises the parameters studied in each chapter along with

the aim of the study.

Table 1-1. Summary of thermal spray experiments conducted.

Parameter Aim Substrate condition Chapter

SOD The effect on splat morphology Glass, smooth 4

SOD The effect on splat morphology Mild steel, rough 5

Substrate

temperature

The effect on splat morphology,

microstructure and bonding

Mild steel, polished 6

Substrate

temperature

The effect on splat morphology,

microstructure and bonding

Mild steel, rough 7

SOD The effect on splat metric

geometries

Glass 8

Chapter 2: Literature Review


Chapter 2

2 Literature Review

2.1 Introduction

A new challenge in coating technologies is to provide surface improvements

without compromising the environment. Surface coating techniques enable

modifications to surface properties, achieving improvements in the performance and

reliability of the coated components. The currently employed solvent-based systems

cause the emission of Volatile Organic Compounds (VOCs). Therefore, alternative

coating solutions that result in less environmental pollution are being sought, using

general requirements and considerations to determine the suitability of the coating

systems [21]. Thermal spray (TS) is recognised as an important technical process for

coating applications amongst the industrial coating techniques such as physical vapour

deposition (PVD), chemical vapour deposition (CVD), fluidized bed coating,

electrostatic spray coating and painting, Fig. 2-1 [22].

TECHNICAL COATING

ATOMIC DEPOSITION PARTICULATE DEPOSITION BULK COATING SURFACE

MODIFICATION

Thermal Spraying

Impact Plating

Fusion Coatings Mechanical

Electroplating

Spin Coating

Diffusion

Wetting Processes

Chemical conversion

Thermal

Sputtering

Electrolytic

Chemical (vapor)

Chemical (liquid)

Ion Implantation

Leaching

Surface Enrichment

Mechanical

Chemical Vapor Deposition

Electrolytic Environment

Plasma Environment

Vacuum Environment

Overlaying

Electrostatic Spraying

Fig. 2-1. Surface coating techniques used for surface modification. The position of thermal spray technology in the field is highlighted.



A thermal spray coating is 100% solid and involves minimal release of

VOCs [23]. Thermal spray technology is capable of processing a variety of coating

materials and substrates such as metals, high melting point ceramics and polymers that

are conventionally processed at low temperatures. The potential range of TS coating

materials includes alloys, conventional composites, polycers (polymer and ceramic

composites), cermets (ceramic and metal composites), carbides, oxides and silica. The

coating application may be performed under environmental conditions such as high

humidity, as well as at temperatures below freezing, or underwater.

TS could be an ideal alternative technology that can increase the quality of

surfaces and meet the demands of less pollution and non-usage of solvent-based coating

systems. Thermal spray process sales are increasing regularly, by almost 10% each year

since 1990 [24]. The world market of thermal spray was about 0.8 billion US$ in 1991

[25], while in 1997 the global thermal spray coating industry represented US$1.35

billion [18]. In 2004, the thermal spray coating industry represents a global market of

about US$5 billion; 30% of which is European based. Fifty percent of this activity is

devoted to plasma spray processing [26]. Generally the global powder coatings market

in 2002 was $3.3 billion and dominated by Europe and Pacific markets in Asia [27].

The thermal spray process was invented by Swiss engineer M. U. Schoop in 1906

and later patented in 1912 [28]. The wire-arc spray process was patented by the same

inventor [15]. The innovative process of producing coatings by means of a metallic

powder is generally carried out by forcing a mixture of compressed air and metallic

powder through a concentric oxy-hydrogen or other gas flame, whereby the particles are

heated, melted and accelerated to impact a substrate, upon where rapid solidification

and coating build-up occurs [29]. The feedstocks were initially limited to low melting-

point materials such as tin and lead, and then progressively extended to steels. Later, in

about 1912, Schoop developed a device for spraying metal wires. The development of

wire-flame spray was carried out to avoid unmelted particles in the spray jet, where the

wire is drawn by drive rolls into the rear of the gun, with a restriction on material

melting temperature (Tm) of 1500 to 1600°C [24].

Since the development of the first engineering polymers in the 1940s, polymeric

coatings have rapidly developed in industry because they offer improved quality,



increased productivity, as well as economic benefits. There are several techniques for

applying polymeric coatings such as air or airless spray, brushing, electrostatic

deposition, fluidized-bed dipping and thermal spray polymer coating: which consists of

thermoplastic and thermoset materials. Plasma spray torches were introduced in 1958

[30]. These were an attractive option to the aerospace and aircraft industries to spray

polymers via plasma spray methods. High velocity oxy fuel (HVOF), commercially

introduced in 1974 [31], was also used to spray polymers.

Several thermal spray processes (Fig. 2-2 modified from [27]) use thermal and

kinetic energy, in the form of either a combustion flame (flame spray and HVOF) or

electricity (arcs or plasma). This energy is used to melt and accelerate the injected

feedstock materials, which can be in powder, wire or rod form. The resulting molten or

nearly molten particles are accelerated and propelled towards the surface, where they

impact and form splats. Upon impact, splats are consolidated, cooled and then solidify

to form the coating.

TS PROCESSES

COMBUSTION

ELECTRIC ARC

PLASMA

Other

Flame wire

HVOF

Flame rod

D-gun®

Flame powder

Nitrogen

Vacuum

High velocity

Low velocity

Helium

Shroud

Underwater

Inert

Vacuum

Shroud

Inert

Cold spray

Induction,RF Plasmas

Controlled atmosphere (chamber)

Air

Controlled atmosphere (chamber)

Air

High velocity

Low velocity

HVIF

HVAF

Fig. 2-2. Typical thermal spray coating processes, modified from modified from [27].



TS process approaches differ in the jet temperature (T) and particle velocity (V)

(the main parameters employed to generate the coating), and also in the process

characteristics and equipment used. These processing differences generate the different

properties of the coating. The specific process implemented in any situation depends on

related factors including splat and coating formation, the application and the industry,

Fig. 2-3. The capabilities and coating characteristics vary across TS processes, as shown

in Table 2-1, adapted from [3, 32-33].

TS coating processes offer a solution to many shortcomings associated with other

coating techniques, with proven capability to save money through extending part life,

reducing maintenance and reconditioning of components. TS processes can manufacture

intelligent composites that are used in wide-ranging applications. They are used to

produce polymeric and polymeric reinforced composite coatings using plasma, HVOF

and flame spray processes. Thermal spray polymers have been used widely in many

areas and are showing great success. Recycling is one advantage of polymer usage that

should be beneficial to the thermal spray technology field because the recycling rate of

polymers is 30-40 wt. %. The subject of thermal spraying polymeric coatings is a

promising technical field with much potential.



Applications IndustriesSplat formation

Disc shape

Fragmented

“fried egg” shape

Splash shape Printing

Medical

Paper making

Agriculture

Power generation

Defence

Aerospace

AutomotiveWear resistance

Food processing equipment

Electrically insulating/conducting

Hard chromium replacement

EMI/RFI shielding

Carbon fibre composites

Corrosion/oxidation protection

Thermal barrier

Clearance control

Bio-compatibility

Electronics

Food processing

Coating formation

Surface macro/microstructure

Bond strength

Wear and friction

Mechanical properties

Residual stress

Microstructure

Porosity

Density

Fig. 2-3. A flow chart that relates technological and industrial classifications of thermal spray to the character of the splat.



Table 2-1. Typical characteristics of the major thermal spray processes, adapted from [3, 32 and 33]

Spraying process

Flame powder spray

HVOF Plasma Spray

Detonation Gun

Electric arc wire

Heat source Oxy-acetylene, Oxy-hydrogen

Oxy-propylene, H2, Propane,

LPG

Plasma arc O2, Acetylene, Nitrogen, Gas

Detonation

Arc between

electrodes Jet Temp. (°C) 2,500 2,500-5,300 6,000-

15,000 3900 5500

Jet velocity (m / sec)

50-100 200-1,000 300-1,000

Particle Temp. (°C)

2,000 610-3,000 >3,500

Particle velocity (m / sec)

50-100 200-1,000 200-800 910 240

Propellant Air Combustion Jet inert gas Detonation shock waves

Air

Material feed type

Powder Powder, Wire Powder Powder Wire

Coating material

Metallic, Ceramic

Metallic, Ceramic

Metallic, Ceramic, Plastic,

Metallic, Ceramic, Plastic,

Ductile material

Relative cost low =1

1 3 5 unknown 2

Relative bond strength

Fair Excellent Very good to

excellent

Excellent Good

Adhesion (MPa)

7-18 >70 4 to >70 >70 12

Oxide content (%)

10-15 1-5 1-3 1-5 10-20

Porosity (%) 10-15 1-2 2-5 1-2 8-15 Max. spray rate (kg / hr.)

7 14 5 1 16

Gas flow (m3 / hr.)

11 28-57 4.2 11 71

Typical deposit thickness (mm)

0.05-5 0.1 To >2 0.1-1 0.05-0.3 0.1 to >5

Atmosphere around particles

CO, CO2, H2O N2, CO, CO2, H2O

N2, Ar, H2, O2

N2, CO, CO2, H2O

N2, O2

Energy required to melt (kW / kg)

11-22 22-200 13-22 220 0.2-0.4



The research and publications on thermal spraying in the polymeric field is

relatively small compared with those for metals and ceramics. Figure 2-4 shows the

number of journal publications in the topical area of polymer thermal spray in intervals

of two years. Fifty five journal papers starting from 1996 have been published to date

using various thermal spray processes. The majority of these papers are cited within this

thesis at the appropriate time. The flame spray process has dominated, at around 62% of

the publications, with feedstocks of polymers, copolymers and polymer matrix

composites. The HVOF process contributes at around 22%, with 11% for plasma and

5% for cold spray. Around two thirds of the publications were related to polymers and

copolymers, while one third of the publications studied polymer matrix composites. The

area of examining polymer splats was limited within these research publications.

Fig. 2-4. Numbers of polymer thermal spray journal publications with respect to year. Note that data has been selected in two-year intervals.

2.2 Polymeric Thermal Spray Coating (PTSC)

Polymeric thermal spraying (PTS) represents an alternative method to processing

polymer powders to produce both polymeric coatings and free-standing forms. Thermal

spraying of polymers can be traced back to the 1940s when polyethylene (PE) was first

produced [2], and which was commercially practiced in the 1980s [34]. The

development was initially for infrastructure applications and later, many industries

including automotive, aircraft and petrochemical, began using thermally sprayed

polymer coatings to provide surface protection against humidity, corrosion, and

0

2

4

6

8

10

12

14

16

1996 1998 2000 2002 2004 2006 2008 2010 2012

N = 55

Year

Num

ber o

f pub

licat

ion



aggressive chemical environments. Since the mid-1990s, polymer spraying has gained

significant commercial attention and acceptance as a replacement for paint within the

United States of America [35]. However, thermal spraying of polymers is still in the

development stage and has not established major industrial applications. The market

size of polymeric thermal spray has not been documented in the open literature.

PTS coatings can be manufactured using high velocity oxy fuel (HVOF) [36-37],

plasma [38-39] and flame spray processes [40-41]. The temperature and velocity of the

spray process must be matched to the thermophysical characteristics of the feedstock.

Thus, materials that exhibit a high glass transition temperature (Tg) need to be sprayed

via plasma or HVOF methods to ensure adequate splat formation attributes.

PTS is gaining increased attention because of its ability to apply relatively thin

(50 µm) [42] and thick (5 mm) [43] coatings onto a wide variety of materials and

structures of any size [15] and may overcome the limitations of the other processes [31]

described in Table 2-2. The PTS process is capable of coating non-conductive

components (e.g., concrete, plastic, and fibreglass) [7] and can be used on heat-sensitive

materials (e.g., wooden products) [8]. Thermoplastic coatings could be applied in

applications where toughness, strength, and flexibility of coatings are required [9].

Thermoplastic materials can always be remelted and will fuse with new material in the

molten state to make it possible for coatings to be repaired in situ by remelting or

applying additional material to the damaged area [7].

Table 2-2. Coating technique limitations.

Coating technique

Limitations compared to thermal spray References

Painting Involves 50% - 80% VOCs by mass, limitation in film properties and thickness.

[22]

Fluidized bed Fixed in location, limitation of coating thickness of minimum 0.2 mm, difficulty in masking, and not applicable for heat sensitive substrates.

[44]

Electrostatic spray

High initial cost compared to TS, thickness limitation to 0.2 mm, and not applicable for non-conductive substrate

[1]

CVD Uses poisonous, volatile precursors that could be ignitable; and using chemical and high cost compounds

[31]

PVD Low deposition rate and not suitable for large objects such as structures.

[31]



Thermoset polymers are typically used in surface coating and painting for

applications such as corrosion protection and decoration [5-6]. The temperatures

associated with the flame spray process restrict the choice of polymer to thermoplastics

because thermoset polymers are likely to degrade under high temperature. Polymers

used in TS are engineering polymers that have thermal resistance up to approximately

100ºC. Other qualities of engineering polymers that make them suitable for thermal

spray are: (i) capable of being formed to precise and stable dimensions, (ii) exhibit

continuous high performance at temperatures exceeding 100ºC, and (iii) tensile

strengths of over 40 MPa. [45].

High performance polymers have the best thermal resistance of the polymer

family and can be used above 150ºC. Recent research has shown that a combination of

polymers can produce unique solutions for many types of industrial challenges [46] and

offer potential for general engineering applications [47]. The benefits of PTS could be

extended further by the incorporation of recycled, post consumer polymers into the

coating feedstock [4, 48-50]. Mixtures of thermoset and thermoplastic polymers have

been successfully deposited by thermal spray [11].

The PTS process uses a fuel gas to melt the polymeric powder. The molten or

semi-molten material is propelled towards a prepared substrate. The particles strike the

substrate, flatten, and form thin platelets (also called “splats”) that conform and adhere

to surface asperities (i.e., pits and grooves) of the rough surface and then to each other.

As the sprayed polymer particles impact upon the surface, they cool and build up splat

by splat into a lamellar structure that forms the polymeric coating. The formation of

splats is influenced by many parameters that must be optimised to form coatings (Fig. 2-

5). Coatings have microstructural characteristics that result from their flattening and

solidification.



Fig. 2-5. Fishnet diagram showing parameters involved in splat formation for thermal spray technology.



Wire arc thermal spray is not suitable for polymer processing because this process

is limited to conductive materials. The process economics, that is, factors such as the

cost of the equipment, feedstock materials and other consumables such as gases and grit

blast media, may also limit the viability of a particular process [51]. Flame spraying is

more preferable for deposition of coatings consisting of metals and polymers [52]. A

limited number of polymers have been sprayed by plasma and HVOF. A comparison to

flame spray was reported and detailed relative advantages and limitations [40, 43, 53].

The use of polymers in thermal spray requires close control of the molten particles.

Equipment manipulation is advantageous in spraying polymers. Several studies

were conducted using special tools. Consider, for example, a simple combustion torch,

which is well suited for low-melting-point polymers, that incorporates a large

processing window [35]. Polymer flame spray equipment has a lower flame temperature

or uses a gas shroud to cool the centre of the flame [54]. Special thermal spray

equipment designed exclusively for polyethylene-based polymers has been used [7].

Unique nozzles were employed to avoid thermal degradation while still melting the

plastic sufficiently [55]. Several examples of flame sprayed polymer were produced

with a suitable powder particle size range of 50 to 250 µm [56].

The most common polymers that have been thermal spray processed include

nylons [37]; polyetheretherketone (PEEK); ethylene-acrylic acid (EAA) copolymer

[57]; polymethylmethacrylate (PMMA); polytetrafluoroethylene (PTFE) and its

copolymers [58]; polyethylene (PE) [59]; polyphenylene sulfide (PPS) [14]; liquid

crystalline polymer (LCP) [55] and polypropylene (PP) [18]. Other thermally sprayed

polymers that have potential applications include ethylene vinyl acetate (EVA)

copolymer [60], and recycled feedstock polymers; including polyethylene terephthalate

(PET) [7, 48], PET–EMAA and polymeric blends [50].

2.2.1 Polymer Characteristics in Thermal Spray Coatings The characteristics of the polymer feedstock in thermal spray coatings depend on

the polymer used and the processing parameters used during the preparation, operation

and post-processing. A polymer such as PEEK generally shows good adhesion to a



metal substrate [41] and has become one of the most popular polymer materials used in

industry [14, 53], also having exceptionally high performance behaviour in an aqueous

medium [61]. The flame spray method was used to coat PEEK onto an aluminium

substrate. The coating exhibited excellent tribological performance with a relatively low

coefficient of friction and wear rate [62]. The wear rate and crystallinity decreased with

the increase of both the load and sliding velocity [40]. Applying PEEK on steel and

aluminium substrates with a preheating temperature of 200°C and spray distance of

100 mm showed that the adhesion increased with an increase in interfacial roughness

[63]. It is feasible to deposit dense PEEK coatings using a combined flame spray–laser

remelting process [41].

PET coatings have promising features for applications against corrosion [4, 64-

65]. Grafted low density polyethylene (LDPE) with maleic acid shows higher adhesion

strength than LDPE [22] and a higher corrosion resistance [66]. Ultra high molecular

weight polyethylene (UHMWPE) is used for wear resistance with a moderate

coefficient of friction against steel. UHMWPE generally provides a protective coating

on metals [47] and improves adhesion [67]. PMMA is transparent, with a similar

abrasive wear resistance on aluminium with potential as a protective coating [68].

Polymers normally used in thermal spray exhibit limitations that affect the coating

characteristics when additives are not considered. PTFE exhibits poor wear and abrasion

resistance leading to early failure [58]. It has shown leakage problems in seals [69] and

exhibits cold-flow phenomenon under load [70]. Polyphenylene sulfide is oxidized

during thermal spraying [71]. Nylon shows some limitations of high moisture pickup,

shrinkage and notch sensitivity [45]. Table 2-3 lists the most common polymers used in

thermal spray and their attributes.



Table 2-3. Major polymers used in thermal spraying with their related attribute. Polymers, References

Thermal spray related properties

Applications Industries Fields, examples

Processes

Nylon, Nylon 11, [36-37, 45]

Low friction coefficient, high tensile strength and toughness. Resistance to chemical corrosion, impact, abrasion, cavitation and various chemicals. Resistance to oil, greases, solvents, fatigue, and repeated impact,

Wear resistance , corrosion (abrasion) resistance

Industrial, machinery, water alliances, petrochemical, automotive

Valves, pipes, fittings, dishwasher racks, shopping carts, pipes, door rails, spline columns

Flame, HVOF, plasma

PEEK, [40, 63, 72]

Low friction coefficient of sliding parts, high tensile strength and toughness at elevated temperatures, high resistance to chemicals, solvents, acids, and alkalis

Wear, corrosion

Food processing, industrial

Magnetic field bearing and slider materials

HVOF

EMAA, [43, 49, 57, 73]

Low friction coefficient, good adhesion to the substrate, resistance to ultraviolet radiation, increase strains at fracture, increased tensile strengths, safe in contact with food, flexible coating

Wear, corrosion, food processing

Military, food industry

Lining of railcar, tank and hopper cars

Flame

PET, [4, 59, 65, 74]

Low friction coefficient, good adherence to the substrate, low permeability to gases and solvents

Wear, corrosion

Automobile, industrial

Fuel tanks, wall panels

Flame, HVOF, plasma

PTFE, [58, 70, 74-75]

Low friction coefficient, self-lubrication, high temperature stability and chemical resistance, temperature stability, good resistance to solvent

Wear, chemical resistance

Mechanical component

Biochemical equipment

Flame, plasma

PPS, [56]

Chemical resistance, electrical insulation, low dielectric loss factor

Dielectric Industrial, automotive

HVOF

PVDF, [76]

High impact resistance, toughness, non wetting surface, resist nuclear radiation, ozone and chemicals

Wear Industrial, food industry

Wall panels Plasma

PE, [47]

Insulation, high chemical strength

Dielectric, weathering

Industrial, automotive

Instrument panels

Flame, plasma

LCP, [2]

High temperature resistance, low gas and water permeability, high abrasion resistance

Abrasion wear

Aerospace Flame



2.2.2 Thermal Spray Variables of Polymers Thermal spraying of polymers is classified according to the spray method used

and materials by which the particle speed, flame temperature, and spray atmosphere are

decided. There are a large number of interrelated variables in the thermal spray process

that lead to significant changes in the performance and quality of the resulting

polymeric coating. The overall coating properties are a consequence of factors that

could be extrinsic variables (i.e., operating variables) that are controllable; while others

are intrinsic parameters, related to the process, substrate and feedstocks that are well-

defined; as well as some parameters that are still ambiguous.

The powder feedstock influences the mechanical properties of the final coating

[77]. The spray variables such as composition of working gas, fuel / O2 ratio, powder

feed rate, electric power input, and carrier gas flow rate affect the coating quality [78-

79]. Spray distance, torch traverse speed and the particle dwell time; factors that

influence the coating density, were shown to influence the quality of polymer

coatings [68, 80].

The low characteristic melting temperature and the irregular morphology of

particles with a large grain size present some difficulties for thermal spray [81]. The

particle size of the polymer influences the coating quality due to the small temperature

range between the melting and the degradation temperature of polymers that exhibit low

thermal conductivity. Thus, the typical thermal spray polymer particles are larger than

metal and ceramic particles.

The high temperature and the presence of oxygen are the two main reasons for

polymer deterioration, which occurs by chain scission, (i.e., the breaking of covalent

bonds) or by crosslinking, (i.e., a formation of bonds between chains). Hence,

degradation of particles (i.e., deterioration and reduction in properties) that might occur

in-flight or at the coating surface needs to be considered when selecting the spray

parameters for polymers. The higher the coating temperature, the more susceptible the

polymers will be to thermal oxidation.

The principle of surface coating is based on the formation of bonds between the

coating material and the coated surface; i.e., mechanical adhesion. The primary bonding



mechanism of thermal spray coatings is mechanical interlocking between the coating

material and asperities on the roughened substrate [12]. Grit blasting provides a further

increase in coating adhesion due to mechanical interlocking [82]. Surface pre-treatment

has been found to enhance the adhesion of polymer coatings [14]. A higher temperature

substrate allows splats to flow well and impeded the solidification of molten particles;

thus improving the adhesion between molten particles and the substrate, and producing

disc-shaped splats [14]. Sufficient substrate preheating can prevent crack nucleation,

and alleviate any mismatch that arises due to coefficient of thermal expansion effects

[27]. Substrate preheating also provides the required flow deformation during impact

due to an adequate cooling rate.

A heated droplet spreads on a substrate when the substrate surface energy is

higher than the droplet surface. This is influenced by both feedstocks and substrate

intrinsic and extrinsic parameters. The quality of droplet-substrate contact and the splat

or coating adhesion are controlled by wetting of the surface and penetration into the

substrate. Wetting influences the formation of uniform, adhered coatings on substrates.

Poor wetting between the polymer and substrate is the major factor for drawing the

liquid splat towards a spherical shape.

The liquid wettability is characterised by the contact angle (θ) as described in Fig.

2-6; i.e., the angle between the horizontal layer (the splat-substrate interface) and the

droplet interface. Droplet wetting can be achieved and therefore good adhesion becomes

possible when . Splat contact angle (CA) depends on the splat surface

tension; i.e., between the splat and substrate ( ), between the substrate and air ( ),

and between the splat and air ( ). Splat contact angle (θ) is calculated using equation

2-1.

Equation 2-1



Fig. 2-6. A schematic of the wetting contact angle of a droplet on a substrate with all surface tension forces acting on it.

2.2.3 Microstructure of Thermal Spray Polymer Coatings

The behaviour of a polymer feedstock during thermal spray processing is

different to that of metals and ceramics. The use of polymers in thermal spray is an

emerging technique. The polymer splat behaviour during melting and solidification is

not yet fully understood. Polymers are distinctive in their rheology; i.e., their flow

behaviour in the liquid state. The melting and solidification of a polymer splat is

complex and the non-Newtonian flow - where the melt viscosity is variable with

temperature - must be considered. The flow characteristics of liquid droplets during

thermal spraying are relevant and dimensionless quantities such as the Reynolds (Re),

Weber (We), Biot (Bi) and Peclet (Pe) numbers are defined to assist in mechanistic

understanding. These dimensionless quantities along with their definitions and values

for polymers and metals are shown in Table 2-4.

Table 2-4. Definitions of the dimensionless quantities acting on impacted particles of polymers and metal.

Quantity Definition Value Explanation of value

Reynolds number (Re)

Ratio of droplet inertia to viscous flow forces

Polymer = low Higher viscous flow forces

Metal = high Lower viscous flow forces

Weber number (We)

Ratio of droplet inertia to surface tension forces

Polymer = high Lower surface tension forces

Metal = low Higher surface tension forces

Biot

number (Bi)

Ratio of internal and external heat transfer resistances

Polymer = high Higher internal resistance

Metal = low Lower internal resistance

Peclet number (Pe)

Ratio of thermal energy convected to thermal energy conducted

Polymer = high Higher thermal energy convection than conduction

Metal = low Higher thermal energy conduction than convection



The differences in the solidification process of polymer particles lead to unique

splat behaviour with high temperature gradients between the surface and the core of a

splat. Visco-elasticity, slow partial crystallization, high shear / extension stresses and

high pressure have a profound effect on nucleation, growth and morphological

evolution in most polymer processing operations [83]. The properties of polymer

coatings depend on the molten state and velocity of the particle upon impact [24].

Establishment of the coating microstructure is determined by the behaviour of

individual splats during cooling and solidification.

The shear rate decreases polymer melt viscosity due to shear thinning of a

deforming droplet. Shear thinning causes an increase in the droplet spreading ratio and

generates splats with higher degrees of deformation than Newtonian droplets under the

same impact conditions. Shear thinning was demonstrated to have a major effect on the

flow dynamics of a spreading polymer droplet because it significantly reduces the

viscosity at the beginning of droplet deformation when the shear rate is the highest [11].

A PTSC is typically an inhomogeneous composite of splats that could be

spherical or elongated, and fully or semi- melted. Porosity, microcracks, inclusions and

oxides are also present. These microstructural constituents act to alter the coating

properties and ultimately affect the coating performance. Maximizing the performance

of a polymer coating requires realisation of processing-structure-property

interrelationships to ensure appropriate application. Figure 2-7 shows a schematic of

features that exist along a cross section through the thickness of a typical thermal spray

coating.

Spray parameter optimization is an essential step in microstructural evolution; as

is controlling the quality of the coating such as by using a shroud or a chamber to

control oxide formation. Greater thermal input is required to melt larger particles and

higher molecular weight organic molecules. A coarser particle size distribution or

molecular weight distribution may allow the formation of numerous heterogeneities

within the coating microstructure; thereby creating voids, a range of splat aspect ratios,

and degraded material [7].



Fig. 2-7. Schematic of general microstructural features of the major constituents recognised in thermal spray coatings.

Generally, higher particle velocities and lower jet temperatures result in denser

coatings, good adhesion to the substrate, and more uniform coating microstructures.

These can be better achieved by HVOF compared to conventional flame spray

processes. The lower jet temperatures of HVOF reduces vaporization of the polymer

while the higher velocity of the particles results in a reduced dwell time and provides

higher impact kinetic energy for enhanced splat behaviour. The high particle velocity

results in plastic deformation of particles upon impact at the substrate, even if the

particles are still relatively viscous. Thus, polymers can be processed below their

degradation temperatures.

.



2.2.4 Advantages of Polymer Use in Thermal Spray PTSC methods are attractive as a manufacturing process due to an ability to coat

large and complex shapes [11, 17-18] and a capability of producing relatively thin and

thick coatings. Widespread PTS applications include surface protection from wear [40,

68] and corrosion [4, 64, 66] in industries such as automotive [37, 65], aerospace [18],

marine [22], industrial [46, 62], food processing and mechanical systems [72].

The increasing use of polymers in coating industries is explained by the improved

quality of coatings, increased productivity, environmental compliance, and economic

benefits. Thermal spraying of polymers represents an environmentally attractive method

of processing polymer powders to produce polymeric barrier coatings, due to the lack of

hazardous volatile organic compounds (VOCs) that are associated with other

technologies such as solvent-based paints.

The potential of polymer thermal spray processes is vast. However, the

processing window for fully molten particles with no degradation is critical and

challenging to achieve due to the polymer characteristics; leading to narrow and limited

investigations for thermal spraying of polymers. Some examples of findings follow.

EMAA can be advantageous to be used in high humidity conditions and at temperatures

below freezing [43] and adheres exceptionally well to steel surfaces [84]. Also, thermal

sprayed EMAA and ETFE polymers are resistant to attack by hypersaline brine and

geothermal sludge [41]. Dense and homogenous coatings were produced of urethane;

polyester; and epoxy and hybrid polyester-epoxy resins for anti-corrosion applications

in hazardous waste storage containers [85]. Other advantages of using thermal spray

polymer coatings are listed in Table 2-5.



Table 2-5. Advantages of polymeric thermal spray coating

Polymeric thermal spray coating advantages

Reasons for advantage

Applications Wear resistance, corrosion protection, dielectric coating, impact resistance and weathering resistance.

One step process No need for post processes and coating acts as both the primer and the sealer.

Environmentally safe Solvent free and no VOCs.

Capability to coat complex shapes

Coating by particulate spray.

Capability of coating large objects

Movable technique (flame spray).

Versatility Non-conductive components and heat sensitive substrate materials.

Ease of coating maintenance

Thermoplastic coating capability for remelting or adding material.

Different environmental conditions

High humidity, high chemical resistance and below freezing temperatures environments.

Cost effective Coating extends substrate life at an efficient cost.

Wide thickness range Thickness depends on application ranging from coatings of 0.01 micrometres to thick coatings of 10 mm.

Low heat input No heat-affected zone and no residual stress.

Surface aesthetic improvement

Availability of different colours.

Particulate depositions Particles do not need to be fully molten to flatten out on the substrate surface.

2.2.5 Applications of Thermal Spray Polymer Coatings

Polymeric coatings have been used widely for improving the strength of glass

bodies, surface decoration, corrosion protection, and wear resistance. These applications

include the lining of vessels used in the chemical industry, coatings on light poles and

coatings on bridges for environmental protection, vehicles and infrastructure [49].

Polymer coatings can be used to improve the surface finish of metals for decorative or

protective reasons [6] and as sound absorbing media [86]. Coupled with metal or

ceramic powders, plastics work well for abrasion applications in the aviation industry,

as anti-slip surfaces for pedestrian use on steps, or in factories where walking surfaces



may become wet [57]. The intrinsic polymer properties of the coated surface include

high chemical and impact resistance, electrical insulation, low coefficient of friction,

ductility, and high toughness.

Polymer thermal spray coatings can be used for various applications including

corrosion protection, wear and abrasion resistance, solvent and chemical resistance,

weathering protection, antiskid protection, and non-stick and low friction surfaces.

Many thermal spray applications are well established and new ones are being developed

that allow pairing of an optimum base material and surface coating properties to obtain

characteristics required for special applications. The aircraft and aerospace industries

have provided a proving ground for testing and integrating coating concepts. This

technology has advanced to the point where it has increased the credibility and

reliability of coating applications. Polymer matrix composites (PMCs) are used

extensively in aircraft engines for many reasons including reduced weight, improved

strength, reduced part counts, and reduced manufacturing cost. PMC parts in current

aircraft engines include fan blades, outlet guide vanes, bypass ducts, nose cone spinners,

core engine fairings, variable vane rings, and cowl doors. These components are mainly

located in the ‘‘cold’’ sections (with operating temperatures below 320ºC, and 0.4 MPa

air pressure) of turbine engines [87].

Mechanical failures such as wear and corrosion can be avoided by applying a

thermal spray coating. The total cost of replacing parts because of wear is estimated to

be more than $US100 billion per year [88]. The direct cost of corrosion in the United

States was estimated to be $276 billion in a study released in 2002 by the U.S. Federal

Highway Administration [89]. Table 2-6 shows further significant polymeric thermal

spray applications.



Table 2-6. Polymeric thermal spray coating applications.

Polymers / polymer matrix

Industries Fields, examples References

Nylon-11 Appliances, petrochemical, automotive

Valves, pipes, door rails

[37]

PE Industrial, marine Gas pipe lines, steel and marine environment

[22]

PET Automotive Fuel tank, packaging [65] PEEK Automotive, food

processing, petrochemical industrial and medical

Bearing and sliding material, engine pistons, pumps and high pressure application, aqueous medium application, vessels in petrochemical

[40-41, 61-62, 72, 90]

EMAA Medical Painting, skin protection [49] PTFE Mechanical systems High performance

mechanical seals [75]

PVDF Pharmaceutical, aerospace, marine

Paper, pipe, valve, piling [76]

LCP Food processing and petrochemical

Vessels in petrochemical, breweries

[91]

Carbon / polyimide Aerospace [13] Al / PE Oil and petroleum Transformation steel duct [92] PTFE / Compound Mechanical system Radial shaft, hydraulic rod

seal [93]

WC-Co / polyimide

Industrial Turbine engine [13]

Polymer properties diversify the applications for functional and decorative

purposes to include chemical; oil and gas; paper; printing; steel; metal processing;

textile; synthetic fibres; glass; pumps; pneumatic and hydraulic systems; and the

nuclear, electronic and electrical industries [1]. Every polymer has a wide range of

applications depending on its particular characteristics, such as PEEK being suitable for

use in pump components and other high-pressure, wear sensitive situations [72]. Nylon-

11 is well suited for use as a protective coating in many industries such as water (e.g.,

valves, pipes, and fittings), domestic appliances (e.g., dishwasher racks and shopping

carts), petrochemical (e.g., pipes), and automotive (e.g., door rails and spline columns)

[37]. PET is largely used in the form of fibres, sheets, films and packing



applications [4]. Powder coatings that consist of polyesters and polyester-epoxy hybrids

are attractive and durable in both thin and thick film applications [64].

2.2.6 Polymer Matrix Composites in Thermal Spray

Polymers are widely used in many industries. However, polymers are susceptible

to damage by scratching and abrasive wear. The properties and performance of polymer

coatings are also frequently limited by high permeability to water and gas and poor

adhesion to metal substrates [81]. Reinforcing fillers are added and polymer matrix

composites (PMCs) are formed to improve the durability and chemical resistance

behaviour of polymers. PMCs were first developed during the 1940s for military and

aerospace applications [34] and today are becoming a favoured method to improve

mechanical strength and enhance the appearance of polymer components [94-95].

Compounds consisting of a polymer with metal or ceramic fillers are finding

applications in aerospace; food industries; medical treatment products; underwater and

transportation; and applications that cannot tolerate external lubrication or where there

are requirements for special properties [96-97].

Reinforcing fillers - such as ceramics (e.g.; alumina, silica, and silicon carbide),

organics (e.g.; thermosets and higher melting point thermoplastics) or metals (e.g.;

nickel-aluminium, and zinc) improve the wear and corrosion resistance of the so-formed

PMC [59, 98]. The costing of PTS materials can be optimized by incorporating recycled

polymers into the feedstock [49-50] without significantly influencing the material

properties of the composite coating system.

The application of polymer-composite coatings is vast with further potential.

Various coatings have been manufactured and include (i) post-consumer commingled

plastic coatings; (ii) abrasion-resistant fluoropolymer coatings for release applications

(iii) high friction, non skid coatings and (iv) magnetic and electrically conductive

coatings [2]. Common polymer matrices used in TS with their related characteristic

properties are shown in Table 2-7.



Table 2-7. Common polymer matrices used in thermal spray and their characteristics.

Polymer matrix Characteristics Processes References Silica / Nylon 11

Increased scratch resistance, wear resistance

Flame, HVOF and Plasma

[99-100]

SiO2 / PA1010

Increased crystallinity, reduce friction and wear


[101]

Carbon / polyimide

Excellent high-temperature mechanical properties, thermooxidative stability and good processability


[13]

Glass / polyamide

Increased hardness, reduce wear rate


[102]

CaCO3 / EMAA

Decreased tensile strain at fracture, increased modulus, peel strength

Plasma

[103]

Alumina / PET Increased wear resistance - [104] SiC, Graphite, Al2O3 / PEEK

Significantly increased wear resistance

Plasma

[62, 105]

Fe-B alloy / LDPE Increased wear resistance Flame

[59]

Maleic anhydride, Al / MDPE

Increased strength and rigidity, enhance coating adhesion to substrate

Flame

[59]

Fe-B alloy, Al / LDPE

Increased wear resistance - [92]

HA / EMAA Increased elastic modulus, Young's modulus and bioactivity

- [106]

Glass fibre, Bronze, MoS2-ZrO2, Al2O3, TiO2 / PTFE

Reduce wear rate and abrasiveness, increased wear resistance


[58, 69-70, 74, 98, 107]

Ceramic / epoxy Corrosion protection - [108] PC / SiO2

Increased scratch resistance and improve sliding wear

HVOF [78]

UHMWPE / EAA

Increased ear resistance, and make non stick and self lubricating

Plasma [76]

Al / MDPE

Corrosion protection, adequate adhesion

Flame

[59]



2.3 Flame Spray Process

The flame spray process was the first thermal spray technique practiced since the

early 1900s, when Dr. Schoop used a combustion flame as a heat source. Flame

spraying uses a fuel gas to generate heat, such as propane and hydrogen, with an

oxyacetylene torch being the most common configuration. The powder, wire, or rod

feedstock is fed axially through the rear of the nozzle into the flame at the nozzle exit.

Figure 2-8 shows a schematic of the flame combustion spray system. The feedstock is

melted and the particles / droplets accelerated toward the surface under the influence of

the expanding gas flow and any air jets that may be incorporated in the rig, Fig. 2-9.

Producing the desired thermal output is achieved by adjusting the fuel / oxygen

ratio and the gas flow rates to generate thermal environments up to 3,350ºC. The

temperature and characteristics of the flame, and therefore the final coating structure,

depend on the stoichiometric combustion temperature of the oxy-fuel mixture. A

combustion ratio of 2.5:1 acetylene to oxygen produces a neutral flame (i.e., one oxygen

molecule is consumed for each acetylene molecule during combustion) and a maximum

flame temperature occurs at an oxyacetylene ratio of approximately 2:1 [109].

Fig. 2-8. A schematic of the typical flame spray process.



Fig. 2-9. Schematic of a typical flame spray process illustrating the location of the external powder injection.

The flame spray process was the first thermal spray technique used to deposit

polymers [55]. Over the last 15 years HVOF [110] methods have also been employed.

The flame spray method is the more desirable thermal spray process due to its lower

capital cost, simplicity, and ability to be transferred to field applications; which is

important for large applications such as construction projects, bridges and ships. The

attributes of the flame spray coating generally are that it is (i) relatively low cost, (ii)

low energy is required to melt the particle, (iii) the velocity and temperature of the jet

are low, and (iv) high heat transmission occurs from particles to the substrate.

Deposition capability is excellent and is ideal for large spray coverage areas. The flame

technique is responsible for about 25% of the turnover in the thermal spray

business [24]. On the downside, the coatings tend to be high in oxides and are highest in

void content.

Flame spraying involves relatively low particle velocities and temperatures,

which limit the range of polymers that may be processed; while other limitations

constrain the use of plasma and HVOF techniques. The higher particle velocity

associated with HVOF produces greater deformation and denser coatings; but polymers

are exposed to a higher temperature. Although the shorter dwell time due to the high gas

velocity is considered an advantage to avoid decomposition, polymer materials can be

partially degraded or oxidized in-flight during the HVOF process [37]. The use of

plasma spraying is limited in polymer applications due to the low temperature of

polymer thermal degradation and the low powder feed rate. The use of cold spray

methods to deposit polymers has not been documented. Cold spray methods are not

currently suitable for the formation of polymer coatings because the low density of



polymer particles implies an insufficient mass to penetrate the bow-shock pressure zone

that forms where the gas stream impinges on the substrate [3].

The flame spray process was the first thermal spray technique used to deposit

polymers [55] and it remains the preferred option for their deposition [52]. In addition

to varying the process parameters that affect the polymer coating quality, process design

and equipment modification have also been considered to allow polymer particles to be

totally melted in the flame without degradation. Instances of this include: (i) a nozzle

designed to avoid the thermal degradation, (ii) a modified combustion torch suited for

low melting point polymers with a large processing window [35], (iii) using a shroud to

cool the centre of the flame [54], and (iv) using cold air to decrease the flame

temperature [55]. Examples of flame sprayed polymer coatings are found on pipes;

valves; fittings; ladder racking; bridges; bridge and municipal handrails; river tugboat

hulls and barges; wastewater clarifiers; wastewater surge tanks; lighting poles; railroad

cars; and exhaust systems for submarines [55].

2.4 Single Splats in Thermal Spray

Deposited particles are the building blocks of coatings and the first deposited

layer plays a significant role in the coating / substrate adhesion. A deposited particle is

referred to as a ‘splat’; a generic term that encompasses distinct morphologies. The splat

formation and flattening that occur when a sprayed droplet impacts onto a substrate is a

key unit process during the production of thermal spray coatings. Splat formation

depends strongly on the temperature, velocity, morphology and chemistry of the

impacting particles [10].

Splat flattening and solidification are controlled by the internal mechanical and

thermal forces attained from the energy source used in the process. The particle kinetic

energy is transformed into the work of viscous deformation and surface energy [111].

The flow mechanism of liquid droplets during the thermal spray process is controlled by

forces captured in the dimensionless Reynolds (Re), Weber (We) and Peclet (Pe)

numbers. Reynolds number is the ratio of inertial to viscous flow forces, and the Weber

number is the ratio of inertial to surface tension forces. The high viscosity of the

polymer results in thicker splats.



Mechanical and thermal constraints influence splat flattening and solidification.

The quality of interfacial contact is a function of the particle impact force, which varies

non-uniformly along the contact surface. The contact quality depends on the droplet

wetting on the substrate and desorption of adsorbates and condensates at the surface of

the underlying layer. In addition, the contact between the piled-up splats is controlled by

the relief of the quenching stress induced by the thermal contraction of splats upon

cooling [111]. Splat formation, topology and microstructure have an important

fundamental implication concerning the structure-property relationships of coatings,

since the design of these intrinsic building units also controls the composite nature of

the coating. Figure 2-10 shows the splat formation of a droplet before and after impact

onto a substrate.

The induced material solidification depends on splat thickness, thermal

diffusivities of both the sprayed material and the underlying solid layer and the quality

of contact between the substrate and the flattening particle [111]. Research into the area

of splat formation is a difficult area from an experimental view point. This gap in the

field of study arises because: (i) the first layer of splats may contribute less than 1 / 50th

of the total coating thickness and, therefore, the relevance of examining such splats may

not be considered important towards understanding the coating as a whole, (ii) the ideal

splat morphology is not easily observed within the body of a coating; therefore, single

splat studies might be considered of no consequence to understanding the coating

behaviour, and (iii) the experimental methods and analytical techniques to understand

splats have not been standardised and, thus, it is difficult to calibrate the outcomes of

such research efforts.

Fig. 2-10. A two-step mechanism of single splat formation in thermal spray deposition.



2.4.1 Splat Morphologies in Thermal Spray Understanding the fundamental building blocks or the unit process of a thermal

spray coating is essential to comprehend the physical characteristics of the ensemble of

splats that constitute a coating. The implication is that single splats need to be examined

from the viewpoint of how their morphology influences intrinsic characteristics. This

forensic process then allows coatings to be designed from a materials engineering

perspective rather than from an experimental trial and error approach.

Splat morphology, and consequently the microstructure of the coating, is

controlled and determined by a number of factors including the spray process, spray

conditions, feedstock properties, and substrate conditions. Furthermore, it depends on

transient properties and the physical setup. When molten or semi-molten droplets

contact a cold substrate or a previously deposited layer that is also cold due to rapid

solidification, then conductive heat transfer from splat to substrate occurs. The

flattening and spreading of droplets stop when the kinetic energy of the droplet is

dissipated; however, solidification of the splat can occur before spreading is complete.

Depending on the kinetic energy, the heat transfer and solidification conditions, planar

or cellular microstructures may be achieved.

Spreading of a droplet commences when the droplet touches the substrate and

continues until it is fully deformed into a static microstructural feature. Solidification of

the lower part of the splat during flattening leads to a decrease in the splat thickness.

Reducing the spray angle and contact wetting angle also yield a decreased splat

thickness [112]. As impact velocity increases, droplet spreading time and splat thickness

both decrease [113]. It has been observed that a higher solid volume fraction leads to an

increase in the splat thickness [114].

Experimental [115-116], analytical [117-119] and computer modelling [11, 120-

121] studies have been conducted to determine the splat formation, morphology,

microstructure and splat bonding with the substrate and any prior-formed splats. Splat

classifications and coating features are correlated to the spray parameters [79]. The

controls of the numerous parameters that govern splat formation are typically related to

the impact, spreading, and solidification of droplets directed at the substrate or

previously deposited layers [79].



The two extreme cases of thermal spray splat morphology are (i) extensively

splashed and highly fragmented or (ii) disc-like splats. However, there are other splat

shapes that lie between these two extremes. The splat morphology in thermal spray

technology can be described in a variety of terms; including disc-like, pancake-like,

flower-like, mushroom-like and fingered splats. The disc-like splat shape with a

uniform solidification rate is the more desirable for coating formation and represents

better contact with the underlying surface. The accumulation of disk-like splats is

accompanied by lower porosity and oxidation than other splat shapes; thus resulting in

better adhesion and cohesion of coatings [122].

Different splat types may be produced by altering the spray parameters. The

geometry of polymer splats includes (i) disc, (ii) “fried-egg” and (iii) splashed or

fingered shapes [14]. The nature of the splat-substrate interface is significant in

determining the coating properties. Substrate parameters such as temperature and

roughness, and the interfacial interactions at impact, will also influence splat formation.

2.4.2 Splat Splashing in Thermal Spray

The occurrence of splashing during polymeric droplet deposition is an undesired

outcome for the coating. Splat solidification has a significant influence on splash

behaviour [119], with splashing observed during rapid formation of the initial solidified

layer just after impact on the substrate [123]. Thus, the liquid portion of the impacting

polymer particle is not constrained by surface tension forces. Conversely, at higher

substrate temperatures the particle remains liquid at the bottom and this allows

formation of a pancake shape [17].

Splashing can occur in a thermal spray process (i) upon impact (so-called “impact

splashing”) which begins immediately at impact and (ii) during the spreading and

flattening process (so-called ‘spreading splashing’) while solidification is starting [115].

Both splashing modes are related to the Sommerfeld parameter (K) of the particle at

impact; which is defined in equation 2-2 [124]:

Equation 2-2



Where We and Re are dimensionless Weber and Reynolds numbers, respectively

that were discussed in Section 2.2.3. If K < 3 then particle rebound occurs. A value of

3 < K < 58 results in deposition and K > 58 induces splashing.

The splashing phenomena can be minimized if the droplet does not commence

solidification until the completion of spreading [119]. The necessary condition is

achieved by (i) raising the deposition surface temperature, (ii) increasing the thermal

contact resistance at the droplet-substrate interface, or (iii) using a substrate with lower

thermal diffusivity [125].

Several hypotheses and models have been proposed concerning splashing

behaviour [126-129]. For a substrate held below the transition temperature,

solidification starts at the contact points of the interface. For a preheated substrate, the

solidification occurs once the whole flattening process is complete. It was suggested

that rapid solidification on cold substrates and splashing produced by the Rayleigh–

Taylor instability causes the splat periphery to become unstable and a splash

morphology forms [130-131]. The Rayleigh–Taylor instability is the theory developed

to predict the number of fingers around the periphery of a splash. It states that the more

rapidly a droplet flattens, then the more complex its shape such that a splash occurs.

Simulation studies proposed that the splashing of molten droplets occurs due to

local solidification [132-133]. Other authors [124, 134] suggested that the liquid particle

flattens to a maximum area dictated by its initial kinetic energy until this is completely

dissipated, at which point the surface tension causes surface shrinkage and jetting out of

liquid from the upper part.

The quality and the behaviour of the splat is a function of the particle impact

force at the splat-substrate interface. These thermo-physical conditions vary non-

uniformly over the contact surface during impact. The contact quality also depends on

the droplet wetting of the substrate and desorption of adsorbates and condensates at the

surface of the underlying layer [111]. These mechanisms have been observed to occur

on polished substrates where the average surface roughness is several orders of

magnitude smaller than the splat height.



The majority of splats involve some degree of splashing; i.e., they are not disc-

like. The degree of splashing is a splat metric that needs to be quantified ideally as a

volume fraction; however, it is not well documented in the field of thermal spray

technology and particularly in experimental studies with polymers. Although the

majority of particles splash during coating formation, a quantification of the degree of

splashing and splashing mechanism has remained elusive despite recent advanced

simulation models of coating formation [135].

2.4.3 Splats of Thermal Spray Polymeric Coatings

Research concerning splat formation will ultimately reveal fundamental

relationships between spray parameters and splat characteristics. However, the spray

parameters required to obtain the optimum polymer splat have not been determined.

This lack of advancement arises due to (i) the large number of factors involved in the

formation of splats (Fig. 2-4), (ii) the unique processing characteristics for each

polymer, (iii) the low processing temperature that the particles must reach to be fully

spread, and (iv) the large difference between thermal conductivity of polymer feedstock

powder and the substrate, which results in thermal gradients within the polymer particle.

The splat shape is used as a measure of compatibility and adhesion with the substrate.

The splat morphology is a reflection of the degree of polymer particle melting

achieved during processing of the feedstock. Polymer splat morphologies may be

classified as (i) disc-like (also described as a pancake), (ii) shaped in the form of a fried-

egg that is raised in the centre, (iii) exhibiting a reasonably coherent but highly

deformed splash pattern, (iv) a fragmented and quite discontinuous splat, or (v) a

fingered morphology that has also been described as flower-shaped [11-12, 18].

Coatings formed of disc or “fried-egg” splats tend to exhibit high adhesion and cohesion

and low coating porosity, while coatings comprised of splashed splats demonstrate poor

adhesion and cohesion with high porosity [18].

Several studies have been conducted on PP coating microstructures, as well as PP

mechanical and physical properties [59, 63]; however, there is little documentation

about PP single splat formation. Research concerning polymer splat formation and its

analyses is limited and focused on HVOF spraying of Nylon-11, with the effects of a



roughened surface on the final splat morphology studied both experimentally and

numerically [12]. Roughness promotes splat instability, and results in jetting and / or

molten particle break-up and formation of radial fingers. Roughness lowers the

spreading ratio. The PEEK splat was studied using HVAF spraying on substrates held at

323ºC, showing the effect of substrate pretreatment on the circularity and area of single

splats, and also on the number of splats deposited on the substrates [72]. PEEK was also

deposited using plasma thermal spray [14].

Many factors influence splat formation; including particle velocity, particle dwell

time [136], torch traverse speed [80], feedstock size, substrate temperature, substrate

surface, powder feed rate, gas flow rate [43], spray method, and many other process

variables [78]. All of these factors need to be tuned towards achieving a designed

microstructure.

2.5 Influence of Stand-off Distance on Splat

Formation

Stand-off distance (SOD), also called the spray distance, is a spray parameter that

typically represents the distance between the exit of the torch nozzle and the substrate

being coated. Changes in SOD significantly influence the in-flight particle and the final

coating because it controls the particle temperature and velocity, which influences the

splats and the overall coating adhesion and cohesion. The maximum and minimum SOD

is conditional upon the type of thermal spray process and equipment used, as well as the

materials employed as the substrate and feedstock.

The temperature and velocity of particles are directly proportional to the length of

time that the particles are exposed to heating and acceleration and thus influence the

degree of particle melting. The velocity at which the particles impact on the substrate is

also determined by how far the particle must travel until impacting the target. Particle

kinetic energy and velocity decrease as particles move away from the energy source due

to the frictional forces from air molecules. Adhesion is directly proportional to the

particle kinetic energy, which varies with particle velocity. Cooler and slower particles

result in weaker adhesion and cohesion and hence an overall unfavourable quality of



coating. With a shorter SOD, the substrate experiences more exposure to heat from the

flame and, thus, is maintained at a higher temperature.

Conversely, a greater SOD suggests that the substrate experiences less of the

heating effects from the flame, and then particles will exchange heat with the relatively

colder substrate and solidify more quickly. The excessive heat exposure associated with

longer dwell time contributes to heat transfer to the local atmospheric environment and

may lead to in-flight particle degradation and particle cooling prior to impact. The flame

of thermal spray processes exits rapidly from a small opening in the spray torch. The

temperature and velocity of the flame are maximized at the torch head, and gradually

decrease with increasing distance. Decreasing the spray distance was found to increase

the hardness of the coatings [137] and the deposition efficiency was shown to depend

on the stand-off distance [138]. The maximum efficient SOD increases with increasing

particle size or weight. This effect arises because the increase in the particle mass

transfers to a decrease in the particle velocity and hence prolongation of its residence

time at elevated temperatures [139].

Stand–off distance is one of many parameters examined in thermal spray

technology to find the optimum SOD and improve the process efficiency. SOD has been

substantiated with evidence to affect coating properties and qualities in a number of

ways, by a number of researchers and for different materials. The best mechanical

properties were found to result at a spray distance of 10 cm for hydroxyapatite.

Increasing the spray distance led to undesirable coating properties, including increasing

porosity and unmelted particles with non-uniform deposition [140]. Increasing the SOD

from 10 cm to 15 cm for plasma sprayed Al2O3 resulted in a decrease of the bonding

ratio [141]. Spray distances of 75-125 mm are necessary to produce adherent coatings of

yttria-stabilized zirconia using HVOF [142].

Several researchers have examined the effect of SOD on splat formation and

coating properties. An increase in SOD produced splashing behaviour. Increasing the

spray distance from 20 cm to 50 cm decreased the coating thickness and residual

stresses of polyurethane resin with Inconel 625 by plasma spraying. The content of both

oxides and porosity were higher [143]. Schadler et al. [144] reported that the optimum

spray distance for producing a desirable splat was 200 mm for Nylon 11 particles



thermally sprayed using HVOF, when compared to results at other SODs between

150 mm and 250 mm. At 200 mm, the particles deformed significantly upon impact,

indicating good melting, while at 175 mm particles were not found to have sufficient

time to melt or deform perfectly, and at 225 mm poor splat deformation was also

indicated.

T. Palathai et al. studied the effect of SOD on adhesion index, micro-hardness and

the overall coating roughness of PEEK sprayed onto aluminium and low carbon steel

substrates, using the flame spray process. The optimum roughness, micro-hardness

properties and the highest adhesion index were obtained at a spray distance of 100 mm,

as shown in Fig. 2-11 [63].

Fig. 2-11. Effect of spray distance of PEEK coatings onto aluminium and low carbon steel substrate on (a) top-surface roughness, (b) adhesion index and (c) hardness [63].



2.6 Influence of Substrate Factors on Splat

Formation

Adhesion of the coating to the substrate is of great importance in thermal spray

technology. Chemical bonding and mechanical interlocking are the two prime bonding

mechanisms proposed for thermal spray coatings. Substrate preparation and conditions

are significant considerations for splat-to-substrate or coating-to-substrate interfacial

bonding. Good mechanical bonding can be achieved when splats penetrate into the

asperities of a roughened surface. Flattening and solidification of splats are greatly

influenced by the substrate parameters including the substrate temperature. When the

substrate temperature is low, the solidification would start at a few uneven contact

points at the interface, while solidification would occur after the flattening process for a

substrate preheated over the transition temperature.

2.6.1 Substrate Surface Temperature Substrate surface treatment affects the morphology, wetting, flattening and

solidification of splats for ceramics [122, 145-146], metals [147-148] and polymers [27,

72]. Preheating the substrate is one factor that influences the coating temperature and is

important for low temperature materials such as polymers. The primary purpose of

increasing the substrate temperature is to remove water, but this also decreases the rate

of solidification. Such a physical behaviour produces highly flattened disc-shaped splats

that increase the physical contact between the splat and substrate [149]. Substrate

heating also permits grain growth perpendicular to the substrate throughout splats;

thereby potentially enhancing the integrity of the lamellae [150]. The temperature must

be sufficiently high to thermally activate the required interfacial bonding mechanism

and low enough to avoid degradation of the coating material.

Higher substrate temperatures allow the coating and substrate to cool down

together and, thus, minimise the mismatch in temperature and stress. Moreover, the

higher the particle (and / or substrate) temperature, the higher the adhesion and the

lower the coating porosity; hence achieving a greater degree of flattening [151].

However, if the substrate temperature is too high, porosity increases inside the coating



and also the substrate may degrade the polymer particles. At lower temperatures, the

formation of splashes is more probable than that of the disc-like splats [3]. C. Li et al.

[152] reported that when a preheated substrate surface temperature reached 150–350ºC,

the splashing would be suppressed and subsequently a regular disc-shape splat would be

formed regardless of the droplet materials and substrate materials.

Preheating the substrate affects the splat morphologies of polymer deposits by

means of post-deposition flow – activated primarily by surface tension after initial fried-

egg splats are fully melted by the preheated substrate. Splats were seen to exhibit a

flattened hemispherical shape when the substrate was preheated to a temperature around

the melting temperature of Nylon-11(~190°C) (Fig. 2-12) [11].

The surface temperature of the substrate also demonstrates an effect on splat area

(Fig. 2-13) for plasma spraying of PEEK [14]. The increase in splat area was thought to

be due to surface adsorbates being driven off at the higher substrate temperatures and,

therefore, enabling spreading of the splats on impact. A smaller temperature difference

between the splat and the substrate would lower the driving force of the splat cooling

such that cooling and solidification would not affect the splat spreading.

Fig. 2-12. Nylon-11 splats deposited onto a glass slide substrate (a) at room temperature and (b) preheated to 190°C [11].



Fig. 2-13. PEEK splats deposited by plasma spray onto a polished substrate at (a) 23°C and (b) 323°C [14].

Y. Bao et al. [47] realised that flame sprayed UHMWPE splats deposited onto a

glass slide flowed extensively to form disc-shaped splats. It was also noticed that there

was scatter in the degree of particle flow, which related to the different heating (and

subsequent melting) of in-flight particles (Fig. 2-14a). Heating splats to 200°C for 5

minutes resulted in overheated particles exhibiting less flow, although they were more

uniformly heated (Fig. 2-14b).

It has been observed that splat solidification occurs after the complete flattening

process [147], and splats become circular with no occurrence of break-up, when the

substrate temperature is raised beyond a “transition” temperature [153]. The transition

temperature is the substrate temperature above which the splat morphology changes

from splash to disc-like [154]. Increasing the substrate temperature leads to reduced

splashing [124, 147, 155], with the small amount of splashing observed on a room

temperature substrate eliminated on a preheated substrate.

The transition temperature depends on surface material properties and increases

with increasing substrate thermal conductivity [123]. Transition temperatures that have

been examined are between 350 - 610ºC for metal substrates and between 300 - 350ºC

for ceramics with lower thermal conductivity. Y. Tanaka et al. [156] reported that a

higher substrate thermal conductivity led to an increased transition temperature.

However, the ideal substrate transition temperature for polymer deposition has not been

determined, but is expected to be low because polymers exhibit a lower thermal

conductivity compared to other materials. A substrate with greater thermal conductivity

will enhance nucleation of crystals to give rise to a finer microstructure [79].



Fig. 2-14. (a) Morphologies of UHMWPE deposited by flame spray onto a glass slide and (b) the same splat after being oven-heated for 5 minutes at 200°C [47].

The relationship between the transition temperature and the thermal conductivity

of feedstock materials deposited on a steel substrate collected from various publications

[127, 157-160] is shown in Fig. 2-15. The graph shows that an increase in the thermal

conductivity of the feedstock leads to an increase in the transition temperature for a steel

substrate. E. Petrovicova et al. [18] recommended a preheat temperature of 80-90°C

using the flame spray process for spraying PP, PEEK. Polyphenylene sulphide (PPS)

was successfully sprayed at substrate temperatures of 127, 350 and 280°C.

Fig. 2-15. The relationship between substrate transition temperature and the thermal conductivity of feedstock materials deposited onto a steel substrate.



2.6.2 Surface Topography The surface morphology of the substrate is an important variable in thermal

spraying that influences droplet spreading and adhesion, and subsequently the overall

coating. The scale of the coating / substrate interface roughness is a typically significant

fraction of the coating thickness. Since coating reliability is contingent primarily upon

the mechanical bonding between the coating material and substrate, it is critical that the

substrate be properly prepared to ensure maximum bond strength. Substrate

preparations include cleaning, drying and roughening by using the grit blast process.

Grit blasting is a common practice for surface preparation, in which abrasive

media pressurised with compressed air is accelerated towards the target. Variables

affecting the final finish include grit size, grit morphology, grit hardness, air pressure,

distance to substrate and angle of impingement. A surface profile of 40–50 µm

roughness (Ra) is recommended for improved adhesion [18]. Cleaning of the substrate

following grit blasting is essential to remove residual sand and dust and it is important

that the prepared surface is coated soon after preparation to prevent surface oxidation or

contamination.

The theory of mechanical interlocking proposes that coating adhesion results

from penetration of the coating material into surface irregularities. The literature on

mechanical adhesion is concerned with the significance of interlocking in explaining

surface adhesion. Authors report that the degree of splashing decreases with increasing

surface roughness height, and it is believed that roughening of the surface is simply

increasing the surface area for more molecular bonding interactions [10, 161]. The

coating surface is classified as a low roughness substrate when Ra < 0.2 µm, or

otherwise a high roughness substrate when Ra > 0.2 µm [10]. It was shown

experimentally that thermal spray coatings have significantly higher bond strengths on

roughened surfaces than on smooth surfaces; promoting surface roughening as a method

to improve adhesion [162].

Only several authors have studied the effects of surface roughness either

numerically [163-165] or experimentally [12, 72, 166]. Although thought to be the main

bonding mechanism in thermal spray technology, there have not been many attempts to

study the effect of surface roughness on droplet impact and solidification. It was



reported that that the majority (~98 %) of published papers discussing splat formation

processes addressed droplet impact onto smooth surfaces, and very few were related to

rough substrates (with Ra > 0.2 μm) [10].

The process of impacting, flattening and solidification of particles, including

partially-molten particles, is complex; especially when droplets impact onto an irregular

surface. The bonding mechanism is still unclear regarding the unmelted or partially

melted feedstock during the coating process. Thermal spray techniques are typically

based on the materials being applied in a plastic state; hence, the bond is not due to

fusion between the coating and the substrate and is not chemical in nature. However,

chemical bonding mechanisms have been reported as playing a role in thermal spray

coating processes [162, 167-168].

2.7 Substrate Material: Mild Steel

Mild steel contains from 0.2% to 2.1% carbon as well as chromium, manganese,

tungsten and vanadium as alloying agents designed to impart specific properties. Mild

steel is the least expensive, most useful and most versatile steel that can serve in many

situations requiring a bulk amount of steel. Furthermore it is a readily available raw

material with low manufacturing costs. Mild steel is used in a wide range of

applications including manufacturing industries, construction, infrastructure,

automotive, and marine. Although mild steel exhibits favourable mechanical properties

and is quite cost-effective, it has low corrosion resistance and is not suitable for long-

term service in many aggressive environments.

Mild steel manufacturing technologies require industrial processes such as acid

cleaning and pickling. As a result of the composition and production, mild steel is

susceptible to corrosion when exposed to conditions within the process. The majority of

mild steel applications involve contact with aqueous media such as minerals, oils, and

chemicals, or entail weather exposure or use in water cooling systems. Therefore,

isolating the metal from corrosive agents is the routine approach for preventing

corrosion. Conventional protection methods are mostly applied to steel surfaces where

aqueous contact is necessary and corrosion occurs.



Techniques that have been used to protect metal surfaces against corrosion

include painting [169-170], sol-gel processes [171-172], electrochemically synthesized

polymer coatings [173-174], and thermal spray coatings [175-176]. Thermal spray

coatings with desirable adhesion and mechanical properties provide a cost-effective way

to protect and prolong the lifespan of not only steel components but many other

substrates used in engineering applications [79]. A polymer coating is often used to

protect metallic materials from the environment, by any one of the following

mechanisms: (i) a decrease of anodic or cathodic reaction, (ii) the introduction of a high

electrical resistance into the circuit of the corrosion cell, and (iii) the creation of a

physical barrier to oxygen, moisture and electrolytes [177].

Different types of thermoplastic polymers have been used to protect steel against

corrosion and wear. Branco et al. [4] used poly-ethylene-terephthalate (PET) powder

obtained from post-consumer beverage bottles that was applied to steel using a flame

spray process. The same process was employed by Duarte et al. [65] to produce a

coating for a fuel tank application. Bao et al. [47] used UHMWPE powder to coat steel

using combustion flame spraying that gave rise to improved adhesion and ductility of

the coating.

2.8 Feedstock Material: Polypropylene

Among the thermal spray polymer candidate materials, PP exhibits a low density,

provides high resistance to chemical attack, and is economically attractive for high

volume application. Further, PP demonstrates high tensile strength, compressive

strength and impact resistance as well as superior dielectric properties and a high

strength-to-weight ratio [5]. Therefore PP has found wide use in domestic appliances, as

well as in the automotive, medical and packaging industries [5]. It is cautioned that

these material characteristics of the bulk polymer can be used only as a guide for the

thermal spray coating that evolves from the polymer feedstock, since the chemistry of

the polymer and the microstructure of the coating are different.

Polypropylene is a viscoelastic material and consequently its mechanical

properties depend on time, temperature and stress. It is a semi-crystalline polymer;



therefore the degree of crystallinity also affects the mechanical properties [178]. The

microstructure of the PP splats will influence the bulk thermal, mechanical and

electrical properties of the coating and there is an associated relationship with the spray

parameters. The cooling rate of the sprayed PP influences the degree of crystallinity

[179]. Crystallisation is generally favoured by slower cooling rates from the melt, while

very rapid cooling, achieved by spraying onto an air-cooled substrate, can suppress

crystallization.

The longer chain lengths of PP result in a higher molecular weight. The weight-

average molecular weight of polypropylene generally ranges from 220,000 to

700,000 g / mol, with melt flow indices from less than 0.3 g / 10 min. to over

1,000 g / 10 min. [178]. Toughness is directly related to molecular weight: the higher

molecular weights provide greater toughness. As a result, higher molecular weight

polypropylenes have greater impact resistance and elongation [5].

2.8.1 Polypropylene Morphology

Polypropylene is a semi-crystalline polymer with varying degrees of crystallinity

and different types of crystal structures in the presence of various additives. Generally,

polymer crystallisation occurs when melted material solidifies or when the solvent is

evaporated. As heat is removed from the melt during processing, molecules begin to

lose the ability to move freely, and the molten PP becomes more viscous [178]. At the

crystallisation temperature, molecules begin to arrange themselves into crystals and

ordered crystalline regions are formed, along with disordered amorphous regions.

Crystal growth may be spontaneous (when the molecular structure is favourable to a

highly ordered structure) or may be induced by the presence of a foreign particle (such

as a nucleating agent or the mould surface).

The crystallisation rate depends on the nucleation rate and the rate of crystal

growth. As the polymer melt is cooled down, the nucleation rate increases, while the

rate of crystal growth decreases. The optimum crystallisation rate usually occurs at a

temperature equal to about 0.8 Tm [178]. The crystalline structure of PP disappears at

the melting point, Tm (171°C) when the material undergoes a phase change from solid



to liquid. Crystallisation is generally favoured by slower cooling from the melt, and the

degree of crystallisation can be controlled by the rate of melt quenching and subsequent

annealing. Crystallisation can be suppressed by very rapid cooling. Polymers, including

PP, generally melt over a narrow temperature range rather than at a distinct point. At the

melting point, physical properties of the material change abruptly as the material

becomes a viscous liquid. The high melting point of PP provides resistance to softening

at elevated temperatures of over 107°C and over 121°C for short periods of time [5].

2.8.2 Thermal Properties of Polypropylene

Synthetic polymers including PP undergo degradation of their physical and

mechanical properties at elevated temperatures in the presence of oxygen. When PP is

heated, the molecular chain length is reduced as a function of time by the thermal

cleavage of intra-chain covalent bonds. This chain scission at elevated temperatures

results in a reduction in the polymer molecular weight and mechanical properties over

time. In addition to thermal degradation, the chains can be broken mechanically by

shear stresses induced during processing, and also by exposure to ultraviolet radiation.

The auto-oxidation degradation of PP at elevated temperatures follows a process

involving several steps, including initiation, propagation, chain branching, and

termination [136].

Polypropylene can be protected from thermal oxidative degradation by

incorporating stabilisers into the polymer. Stabilisers are used to keep the polymer

chains and the original molecular structure intact, thus enabling properties such as

strength, stiffness and toughness to be retained over a longer time. Unmodified PP

ignites at a temperature of about 360°C and burns. Flame retardant grades are available

with increased fire resistance. The improvement is produced by flame retardant

additives that have the dual effect of increasing resistance to ignition and reducing the

spread of a flame. However, there is a severe downside to the use of flame retardants.

High concentrations are needed to produce fire resistance and the resulting compound is

generally unsuitable for food applications. Retardants substantially impair the

mechanical properties of PP and generally impose a limit on the melt processing

temperature which should not exceed 230°C.



2.8.3 Polypropylene Applications Polypropylene has a good balance of properties that can be customized to an

extensive range of manufacturing and fabrication methods and applications, along with

the low cost, that make PP useful in many different industries. The use of PP is

increasing in applications including automobile components; packaging films and

containers; medical devices; textiles; and large and small appliances. The use of PP in

automotive applications provides a good price to performance ratio, weight reduction,

recyclability, and improved acoustic damping compared to other polymers. The

advantages of plastics as packaging materials include easy processing, low cost, and

lower weight due to a lower density than other packaging materials. Plastic packages

can be optimized for a particular application, and they are more impact resistant than

glass.

The properties of PP that make it useful for medical applications include an

excellent resistance to solvents and autoclave heat; good tensile strength and stiffness; a

long flex life; a high heat distortion temperature; low density; and low moisture vapor

transmission rate. Medical applications of PP include medical devices, drug delivery

systems, non-woven fabrics, packaging for medical devices, solutions, and drugs.

Beneficial properties for appliances include abrasion resistance when used in food

industries. In addition, high-flow grades allow easier colouring and processing of more

unique shapes than possible for metals. Automotive applications include exteriors,

interiors, and under-hood applications. Other PP uses include small kitchen appliances

and large appliances such as washing machines.

The good insulating properties of PP fibres are used in carpeting, automobile

interiors, apparel, geo textiles, and in non-woven applications where a soft textile feel is

of benefit and where there is a need for fibres that are insensitive to moisture and dirt.

Polypropylene can be found in recreational items and in the building and construction

industry; used in walls and partitions; as insulation for power cables and telephone

wires; and in pipes. Pipe applications include under-floor heating; hot and cold water;

sanitary engineering; and pipe fittings.

Chapter 3: Experimental Procedures


Chapter 3

3 Experimental Procedures

This chapter describes the experimental procedure used throughout this research

along with testing and characterisation techniques implemented to examine and analyse

splats. Note that all other chapters containing results and discussion have their own

specific experimental procedures when necessary.

3.1 Powder Processing

Commercially available Moplen EP203N, a polypropylene manufactured by

Basell Australia Pty. Ltd., was used as feedstock. Table 3-1 shows the characteristic

properties of this PP [178]. The feedstock material, received in granules of a millimetre

in size, was ground into powder. An OMNI mixer homogenizing system, Lomb

Scientific, Australia, was employed in the grinding process where the blade rotated at

around 10,000 RPM. The system was cooled using liquid nitrogen to prevent the

polymer powder from melting and agglomerating. The powder morphology produced

from the grinding machine was irregular due to the high speed used in grinding. These

particle shape irregularities reduced the powder flowability and increased

agglomeration. The powder was, therefore, stored in a vacuum oven at 80ºC for three

days before spraying. This procedure prevented any environmental reactions that may

result from weathering reactions and, thus, improved the powder flow properties.

Table 3-1. Thermo-physical properties of polypropylene used in this study.

Melting temperature (°C) 160-170 Glass transition temperature (°C) -10 Melt density (g / cm3) 0.85 Water absorption (%) 0.02 Coefficient of linear thermal expansion (mm / °C) Thermal conductivity at 20°C (W / m.°C) 0.19 Dynamic coefficient of friction on steel 0.23



The powder was fed into the spray gun type “6P-II thermoSpray®” torch from

Sulzer Metco via a powder feeder type Metco 3MP (USA). The substrates used in this

study were standard microscope glass slides and mild steel grade 250. The substrates

were mounted on a fixture that was specially designed for this study. The fixture was

designed to hold all substrates at the specified SOD so that there was consistency of the

sample procurement procedure in manually spraying one pass to cover all substrates.

Figure 3-1 shows the substrate fixture with a magnified view one substrate holder that

exhibited deposited splats.

The selection and variation of SODs were based on the published literature. The

maximum and the minimum SODs of polymer flame sprayed with external powder feed

were selected considering the flame temperature and the melting or softening

temperature of the polymer used. The 50 mm SOD variation was chosen because it was

the most common variation used throughout the literature. This distance was also

reasonable since results could be discriminated by a change in outputs. One hundred

particles or splats were used throughout this thesis to be examined for each processing

parameter. The selection of a one hundred sample size provided the required confidence

level to represent the population measured within the deviation from the mean. Refer to

Appendix A that details the procedure implemented for this process.

The morphology and size of powder particles were characterised using a ZEISS

Supra 40 VP field emission Scanning Electron Microscope (SEM). Refer to Section

3.4.1 for more details. Samples were gold coated with a DYNAVAC (CS 300)

deposition system prior to the SEM analysis. An Ambios Technology XP-2 surface

profilometry technique (refer to Section 3.4.2) was used to inspect the cross section of

splats and to examine the splat thickness. The splat profiles and thicknesses were

analysed with a VECCO WYKO NT1100 non-contact optical surface profilometer and

associated software (Vision V3.60). Refer to Section 3.4.3 for more details.



Fig. 3-1. Substrate mounting fixture used in this study designed to perform one spray run at multiple stand-off distances.

The geometrical merits of the particles and splats were measured via imageJ

software. Using computer image software is a practical method for accurate results and

calculations as well as simplicity of data analysis along with the ability for further data

manipulations. Particle and splat size are widely used terminology in the area of thermal

spray technology. They refer to the diameter of an approximated circle of the calculated

geometry. However, using the actual area of the geometry allows an exact and accurate

reading. The nominated new terms are particle and splat projected area. Appendix B

presents graphs using the projected areas of particles and splats.

3.2 Powder Injection Port

The PP splat samples were deposited using a Sulzer Metco, 6P-II flame spray

torch with an acetylene / oxygen gas mixture as the heat source and air as the carrier

gas. The spray parameters differed for each experiment and a spray table will be

documented when required. An external powder feed port with an internal diameter (ID)

of 2 mm (Fig. 3-2) was used in this study to avoid the high core temperature of the

torch. Figure 3-3 shows the spray torch used in this study with the location of the

injection powder feeder. Powder was fed normal to the direction of the flame and the

substrates were held at room temperature (~23°C) with 50 mm increments in SOD. It is

cautioned that the SOD is a different geometrical dimension that the distance between

the powder injection port and the substrate.



Fig. 3-2. Schematic of the external powder feed port used in this study, dimensions in millimetres.

The exit point of the powder from the feeder was termed the “powder injection

point”. This location was generally 50 mm distant from the exit of the flame spray torch

to prevent premature melting of the PP within the powder delivery tube. This geometry

was atypical from usual thermal spray practice and was implemented in this case

because (i) the thermal spray torch used was not designed to spray the chosen PP

feedstock, and (ii) the powder feeder used was not designed to be used with this

feedstock. The powder port exit was located 50 mm from the axial axis of the torch.

This was an optimised distance that was a compromise between (i) being close enough

to the flame plume to permit PP entry into the flame and (ii) sufficiently distant to

prevent melting of the PP in the metallic powder feed tube.

Fig. 3-3. Photos of the spray torch and the location of the external powder port used in this study.



3.3 Characterisation Techniques

The physical, chemical and mechanical properties of particles, splats and

substrates require characterisation tools to examine the quality of the thermal sprayed

particles deposited onto the substrate. Various techniques have been used extensively

for polymer thermal spray splats and coating characterisation. The commonly used

techniques in this study are optical microscopy (OM), Scanning Electron Microscopy

(SEM), Atomic Force Microscope (AFM), Raman spectroscopy (RM) and Focused Ion

Beam microscopy (FIB).

3.3.1 Scanning Electron Microscopy (SEM)

SEM was the primary characterisation technique used to classify and characterise

the splat shapes. SEM provides images of surfaces at high magnification. Besides

morphological imaging, SEM is used to detect signals that can provide compositional

information; i.e., characteristic X-rays as well as backscattered electrons and Auger

electrons. SEM has a large depth of focus that allows viewing of three dimensional

aspects of samples at high resolution. The use of SEM was ideal for viewing individual

particles, splat morphologies and splat-substrate cross-sections typically at

magnification ranges between 10 and 200,000. All characterisations using SEM were

conducted with a ZEISS Supra 40 VP field emission scanning electron microscope.

The physical size of spray samples must be appropriate to fit inside the SEM

specimen vacuum chamber; these being rigidly mounted on a specimen holder. The

surface of the specimen must be electrically conductive. Non-conductive specimens

such as polymers will charge when scanned by the electron beam and cause scanning

faults and other image artefacts. Electrically conducting materials, commonly gold or

platinum, are, therefore, used to coat the polymer specimen surface. The metal

deposition is performed either by a low vacuum sputter coating process or by high

vacuum evaporation. All PP samples were gold coated with a DYNAVAC (CS 300)

deposition system prior to the SEM analysis, Fig. 3-4. Examples of the instrument

outputs with regard to the testing and characterisation techniques are provided in the

figures in this section.

http://en.wikipedia.org/wiki/Sputter_deposition



Fig. 3-4. The ZEISS Supra 40 VP field emission scanning electron microscope used for imaging in this study, and an image of a typical splat obtained using this system.

3.3.2 Energy Dispersive X-ray Spectroscopy (EDS)

Energy dispersive X-ray spectroscopy (EDS) is an analytical technique that

allows determination of the elemental composition of a specimen. It is sometimes also

referred to as EDX analysis. Analysis using EDS spectroscopy relies on analysing X-

rays emitted by the matter in response to it being bombarded with charged particles. The

capabilities of EDS characterization are based on the principle that each element has a

unique atomic structure. The energy of the emitted X-rays can be measured by an

energy-dispersive spectrometer. The EDS analysis system used in this study worked as

an integrated feature of the SEM.

The electron beam bombarded the sample and collides with electrons from the

sample atoms; knocking some of these from the regular orbital arrangement in the

process. During the transferring process, the atom of every element releases X-rays with

a characteristic amount of energy. The identity of the atom from which the X-ray was

emitted can be established by measuring the amount of energy present in the X-rays

released by the specimen. An EDS spectrum displays peaks corresponding to the levels

of energy for which the most X-rays had been received. Each of these peaks

corresponds to a single element. The higher a peak in a spectrum, the more concentrated

the element is in the specimen.

http://en.wikipedia.org/wiki/Spectroscopy

http://en.wikipedia.org/wiki/X-ray

http://en.wikipedia.org/wiki/X-ray

http://en.wikipedia.org/wiki/Atom

http://www.siliconfareast.com/SEMTEM.htm



3.3.3 Two-dimensional Profilometry Two-dimensional profilometry allows measurement of a contour of a solid

object by moving the sample under a stylus to characterize the surface roughness and

morphology. The technique works by actual contact between a diamond stylus tip and

the object. The stylus moves and encounters the surface features by vertical motion,

forming signals that are then converted into two-dimensional data. The sample surface

can be read in terms of the converted data to represent the height, which could be a

surface roughness or a splat thickness. A two-dimensional system – Ambios XP-2

Profilometer – was used for exploring surface roughness and morphology, and to inspect

the cross-section of splats and reveal the splat thickness.

This technique can be operated in either the instrument coordinate system or in an

arbitrary coordinate system defined relative to the specimen. The software used to

process the data from a scan was also used to compare data and process the curves. The

facility for exploring polymer splat thickness was an important feature in using the

profilometer stylus. It was necessary during the experiment to locate the stylus tip on

the middle of the splat to ensure that a reading of the highest thickness was obtained.

Figure 3-5 shows the technique with the output data displayed by the associated

software. Note that the unit displays dimensions in Angstroms (Aº).

Fig. 3-5. Photograph of the two-dimensional Ambios XP-2 surface profilometry technique used for surface profile measurements, and an image of the splat thickness profile obtained using the system.



3.3.4 Three-dimensional Non-contact Profilometry Three-dimensional (3D) surface profilometry is one testing and characterisation

technique that examines surface topography. It is a non-destructive technique with the

ability to perform surface roughness measurements, thickness measurements and to

examine splat topology. Further, using the accompanying software, additional analyses

are obtainable. The Veeco WYKO NT1100 used in this study is an optical non-

destructive profilometry tool providing three-dimensional surface profile measurements

without contacting the specimen, Fig. 3-6.

The working mode used is Vertical Shift Interference (VSI), which is based on

white light and an interferometer. This is an optical device that splits a beam of light

exiting a single source into two separate beams and then recombines them. The

interferometer employs a system where one beam is reflected from the object under test

and the other beam is reflected from a reference mirror. The beams are recombined to

create bright and dark bands called “fringes”, representing the topography of the object,

where the best-contrast fringe occurs at optimal focus. Interference fringes are present

only over a very shallow depth for each focus position. The system scans through the

focus range as the camera captures frames of interference data at evenly spaced

intervals and forwards the frame to a computer for processing using interferometric

phase mapping programs. As the system scans downward, an interference signal for

each point on the surface is recorded and finally the vertical position corresponding to

the peak of the interference signal is extracted for each point on the surface. The

maximum measurable topography is 1 mm and there are some limitations to measuring

slopes, depending on the optical numerical aperture and surface roughness.

The system is accompanied with software (WYKO Vision) that allows further

manipulation and analysis of images. It includes the availability of an X-tilt and a Y-tilt

with 2 encoders; which are adjusted to capture the necessary fringes.



Fig. 3-6. Photograph of the non-contact optical Veeco WYKO NT1100 three-dimensional profilometry technique used for surface profile measurements, and an image of a splat thickness profile obtained using the system.

3.3.5 Focused Ion Beam (FIB)

Focused ion beams have been used since the 1960’s to investigate the chemical

and isotopic composition of materials because of its sputtering capability. FIB is also

used as a micro-machining tool to modify or machine materials at the micro- or nano-

scales [180]. FIB has been extensively used to image and analyse materials to

investigate the morphology of splats, splat cross-sections and splat-substrate interfaces.

Further, FIB milling technology allows effective preparation of TEM lamellae with

predefined thickness, while the process can be simultaneously monitored by SEM. In

polymer materials science, the use of FIB for the preparation of polymeric samples is

delicate, because it can induce amorphization of the sample surface, scission and / or

cross-linking of polymer chains, shrinkage of the chains, and modification of the

surface chemistry [181].

A source of charged particles (ions) is located at the top of the electro-optical

columns of the FIB. A series of electrodes, electron lenses and mechanical beam

current-limiting apertures extract the charged particles and focus them into a beam with

the desired characteristics, either a "large" beam of high current or a "small" beam of

low current. The position of the beam over the sample is controlled by the deflection

plates that provide the scanning feature. The amplitude of a secondary signal, generated

by the beam-sample interaction is displayed synchronously with the beam position to

form an image of the scanned area. A beam of high energy electrons can yield



information on topography, morphology, chemical composition, and crystallographic

information.

Using the FIB technique in polymer milling is a delicate task due to the

requirement of a high energy ion beam for the milling process that may sputter surface

atoms away and affect the material’s structure or topology. To reduce this drawback,

several stages of rough and fine milling are executed. A thin film of approximately

1 μm platinum may be deposited on the top of the area to be milled for surface

protection.

FIB systems can be operated at low-beam currents for imaging or high-beam

currents for site-specific sputtering or milling. To study the splat-substrate interface and

the interface between overlapped splats, splat cross-sections were produced using the

FIB milling technique. Quanta 3D FEG high-resolution, low vacuum SEM / FIB was

used for splat sectioning. All of the FIB work was carried out using gallium ions

accelerated through a 30 kV ion beam, Fig. 3-7. The initial rough cutting was performed

at 15 nA and the surface clean-up cutting using 5 nA.

Fig. 3-7. Photograph of the Quanta 3D FEG high-resolution, low vacuum SEM / FIB used for splat sectioning, and an image of a splat cut obtained using the system.

3.3.6 Raman Spectroscopy

Raman spectroscopy can be described as the inelastic interaction of laser light

with a material. When the light strikes an object, some of the light is transmitted while

the rest is either absorbed or scattered. A large portion of the scattered light has the



same wavelength as the incident light but a small amount of the light is shifted in

wavelength as a result of the molecular vibrations and rotation of molecules in the

sample. The spectrum of these wavelengths is called the Raman spectra when the

scattering is not elastic. If the scattering is elastic, the process is termed Rayleigh

scattering. In Raman spectroscopy, no probe physically touches the material; the laser

light is the only contact with the specimen, making it a physically non-destructive

testing and characterisation technique.

When the scattered photons are shifted to lower frequency or higher final energy

than the exciting photon, then the shift in scatter observed is termed as “Raman Stokes”

scattering. If this interaction shifts the frequency of the scattered photon to a higher

frequency, corresponding to a lower final energy, then it is a Coherent Raman Anti-

Stokes (CARS) scattering. Various materials including liquids, solids and gases can be

evaluated using Raman spectroscopy because of the distinct spectra that certain

materials produce due to their structural arrangement. The composition of unknown

substances can be determined using Raman spectroscopy and a qualitative analysis of

materials performed.

Raman spectroscopy is a practical tool in polymer science and analysis because of

the non-destructive nature of the test method, and the fact that sample preparation is

quite minimal. Useful information on composition, chemistry, and structure may be

obtained from the spectra that are generated from specimens. Furthermore, it reveals

information on crystallinity and the average chain orientation. Analyses were conducted

with an excitation wavelength of 514 nm at a spectrum resolution of 1 cm-1 over the

range of 200 to 3,200 cm-1 with three accumulations of data acquisition and 10% power

for 10 seconds with cosmic ray removal. Micro-Raman scattering experiments were

performed using an inVia Raman spectrometer (Renishaw plc, UK), Fig. 3-8. The

spectra were collected from the original PP feedstock as well as from splats that had

experienced different heating conditions. The collected data was modified using the

accompanying software to remove the cosmic ray peaks and for further data

manipulation.



Fig. 3-8. Photograph of the Renishaw inVia Raman spectrometry system and an example of a spectrum obtained using the system.

3.3.7 Image Processing

Particle and splat size measurements can be used to optimise the thermal spray

process parameters. Computer imaging techniques present an accurate method for

measuring the particle size and provide an opportunity for gathering metrics on particle

and splat shapes. Measurements of shape variations are necessary when particles are

small and densely packed. Image analysis generates statistically relevant data with no

subjective bias: that is, data is generated from each individual particle and splat – the

individual microstructural constituents of a coating – rather than the coating as a whole.

Image analysis software can display different image formats, and allow the user

to edit, analyse, process and save results based on the selected format type and output

settings, which can then be converted to an MS Excel file for further manipulations. For

a selected enclosed shape with set scale, imageJ calculates and displays area statistics

including the shape area, standard deviation, perimeter, diameter, circularity and feret

diameter. The program imageJ can create density histograms and line profile plots and

supports standard image processing functions such as contrast manipulation,

sharpening, smoothing, edge detection and median filtering. With the use of optical

microscopy, hundreds of powder particle and splat images were recorded. The above-

described techniques were employed to evaluate the splats using the imaging analysis

software.

Chapter 4: Influence of Stand-off Distance on Polypropylene Splats Deposited onto

a Flat Surface


Chapter 4

4 Influence of Stand-off Distance on

Polypropylene Splats Deposited onto a Flat

Surface

4.1 Introduction

The stand-off distance represents the length dimension between the exit of the

torch nozzle and the substrate surface upon which the particles impact. SOD depends on

the type and source of spray process equipment, structural application required and the

feedstock materials used. A moderate SOD produces more oxidization and more

porosity; hence, deposition efficiency decreases and a coating with less cohesion and

adhesion are created. The larger the SOD; then the lower the kinetic energy, lower

velocity and the greater degree of cooling that the thermal spray particles experience

prior to impact. The cooler and slower impacted droplets have less coalesce to the

substrate and to each other; thus resulting in a less adherent coating.

Particle temperature and velocity depend on the SOD, both being highest with a

shorter SOD and lowest with a longer SOD. The particle melting stage is linked to the

particle dwell time; the longer SOD will provide more time for the particles to melt.

Droplets may solidify before they reach the substrate if the SOD is much longer;

whereas a much shorter SOD may not provide sufficient time for particles to melt. This

behaviour depends on the type of thermal spray equipment, the coating and substrate

materials. However, it was reported that plasma sprayed coatings at a shorter SOD

exhibited better spreading and cohesion with greater cracking activity, reflecting the

presence of a larger amount of residual stresses [182]. The maximum and minimum

SODs need to be optimised with regard to the morphology of splat formation and

improved deposition efficiency.


a Flat Surface


In this chapter, the influence of SOD on PP splat morphology was examined to

establish a fundamental understanding of the manner in which processing variables

influence microstructural characteristics. It should be noted that the SOD considered in

this study is a geometrical distance between the powder injection port and the substrate.

This is a different definition from the conventional description of SOD that is measured

from the head of the torch to the substrate, Fig. 2-9. A disc-shaped splat is more

desirable to manufacture a dense coating with minimum porosity and enhanced

mechanical properties. The prime operational parameter of SOD was varied to produce

splats that exhibited minimum splash; keeping in mind the overall intent of this body of

work is to understand the basic building blocks of polymer coatings that are formed via

the flame spray process.

4.2 Methodology

The location of the powder injection point was either 50 mm or 100 mm distant

from the exit of the flame spray torch, as represented by X in Fig. 3-3 (Section 3.2).

Two values for X were tested to establish the influence of changing this distance on

splat formation. The selected distances were not close to the flame plume and at an

adequate distance to prevent PP degradation. The spray parameters used in this study

are summarised in Table 4-1.

Table 4-1. Flame spraying parameters used in this study.

Torch Sulzer Metco (Thermospray) torch 6P-II

Stand-off distance (mm) 150, 200, 250, 300, and 350

Oxygen pressure (kPa) 200

Acetylene pressure (kPa) 103

Substrate temperature (°C) RT

Spray angle (deg.) 90

Particle size range (µm) 110-220

Oxygen-to-acetylene flow ratio 1.66


a Flat Surface


4.3 Results and Discussion

4.3.1 Particle Size and Morphology Analysis The feedstock after grinding, Fig. 4-1, depicts an irregular and angular

morphology with small flakes adhering to the larger particles, which manifested

themselves by impeding the flow characteristics into the flame. Equivalent diameters of

an elliptical shape were used to characterise the particle size. The particle size

distribution (PSD) of 100 particles evaluated from the SEM observations is shown in

Fig. 4-2. The mean PP particle size was estimated to be 150 µm within a range of 120 to

220 µm. The aspect ratio of the particles ranged from 1 to 4 with an average of 1.7.

Fig. 4-1. SEM images showing the irregular morphology PP feedstock powder: (a) at low magnification (200x) and (b) at high magnification (600x).

Fig. 4-2. Particle size distribution of polypropylene feedstock powder.

0

10

20

30

110 130 150 170 190 210

N=100

Particle size (µm)

Freq

uenc

y (%

)


a Flat Surface


The particle shapes are not spherical and needed to be approximated to the shape

of an ellipse. Figure 4-3 shows the frequency of the aspect ratio of the long to short axis

of an ellipse. The peak of the curve indicates that the maximum particle frequency was

at an aspect ratio of about 1.5, and the preferred dispersion of particles to the right hand

side indicates the presence of many longer axis particles with a maximum of around 3.5.

Particle shapes tended to be more elliptical than spherical.

Fig. 4-3. Relationship between aspect ratio and frequency of PP particles for the approximated elliptical shape.

4.3.2 Splat Analysis Many of the PP splats sprayed at short distances (150 mm) exhibited a “fried-egg”

shape; i.e., a large, nearly-hemispherical, unmelted or still-viscous core in the centre and

a fully melted wide thin rim, also termed as the leading edge. The leading edge was

generated due to a higher flow velocity of the molten material than the original particle

impact velocity. The splat leading edge influenced the adjacent splat to coalesce and

increased cohesion, as shown in Fig. 4-4. The higher flow velocity arose due to

squeezing of the melt between the relatively viscous particle core and the rigid

substrate. The high internal thermal resistance (Biot number > 5) of polymer particles,

which can be contrasted to a Biot number < 0.1 for metal particles, impedes heat

transfer to the polymer particle interior [12]. The “fried-egg” splat shape indicated a

large radial difference in the flow properties of the molten surface of the PP droplets as

0

2

4

6

8

10

12

0.5 1 1.5 2 2.5 3 3.5 4

Freq

uenc

y (%

)

Aspect Ratio

N =100


a Flat Surface


well as a large gradient temperature profile. When the particle diameter was sufficiently

large (> 100 µm), the surface temperature could exceed the polymer degradation

temperature while the core temperature remains almost unchanged [37].

Fig. 4-4. Microscopic image of coalesced splats.

The SEM image and profilometry analyses of PP splat morphology indicated that

the contact angle of the splat core was much larger than the contact angle of the outer

edge. The profile of splats at a 150 mm SOD, measured with an Ambios Technology

XP-2 surface profilometer, is shown in Fig. 4-5. The splats were characterised as dense

with no pores or cracks. They were partially melted or not melted in the centre. The

splat size to thickness ratio, also called the flattening ratio in this work, was about 5;

which is small compared with metals or ceramics that exhibit diameter to thickness

ratios of 20 or larger. That is, these polymer feedstocks did not deform as much as

metals or ceramics under the given spray conditions.


a Flat Surface


Fig. 4-5. (a) SEM image of a PP “fried-egg” splat on a glass substrate at room temperature; (b) Profile of splat cross section along A–B with splat core and rim contact angles. (Note that the X and Y axes are different scales).

The spray distance between the substrate and torch nozzle affects the splat

morphology and the substrate-coating adhesion, since it influences the extent of

polymer melting. Undesired splat morphology arises due to partial or no melting, or as a

result of overheating of polymer particles. Most of the splats sprayed at SODs greater

than 250 mm exhibited a well spread morphology with larger flattening ratios

represented by a wider diameter of very thin splats. Altering the SOD is a practical

method of controlling the amount of particle energy at impact with the substrate [183]

and, thus, greatly influences the material characteristics and engineering properties of

the coating. The results of PP splat profiles and thicknesses obtained from the WYKO

optical surface profilometry agree with the results using XP-2 surface profilometry.

SEM observations, Fig. 4-6, indicated that splat morphologies varied with SOD,

attributed to the differences in particle heat exposure and acceleration. A longer SOD

allowed the feedstock to experience a longer dwell time in the flame and provided

exposure to higher temperatures and created a more intense particle impact onto the

substrate. Excessive heat would lead to in-flight particle degradation. On the other hand,

the longer particle dwell time contributed to heat transfer to the local atmospheric

environment and droplet cooling prior to impact. The number of droplets that impacted

orthogonally decreases as the SOD increased because trajectory deviation was

manifested as the particle travels through longer distances.


a Flat Surface


Fig. 4-6. SEM images of single polypropylene splats on a glass substrate at different stand-off distances at room temperature: (a and b) a “fried-egg” splat with few splashes, (c and d) a “fried-egg” splat with little splashing, and (e and f) an overheated fully spread splat.

The thickness to diameter ratio was obtained for different SODs using the WYKO

optical surface profilometry and VISION software. A three dimensional view with a 1:1

scale of thickness:diameter for a few splats sprayed at about 150 mm SOD indicated a

larger splat size with the maximum observed thickness (Fig. 4-7). Another measurement


a Flat Surface


of thickness and diameter was obtained for a 200 mm SOD and indicated a larger splat

diameter of about 250 μm with maximum thickness of 14 μm, Fig. 4-8a. Furthermore, a

SOD of 350 mm or more led to a larger splat size and smaller splat thickness, Fig. 4-8b,

and a longer dwell time in the flame led to a thinner and larger PP splat. On the other

hand, unmelted or partially melted splats occurred at small SODs.

Fig. 4-7. 3D view showing thickness and diameter of PP splat scanned by WYKO surface profilometer at 150 mm stand-off distance.

Fig. 4-8. 3D view showing thickness and diameter of PP splat scanned by WYKO surface profilometer: (a) splat at 200 mm stand-off distance and (b) splat at 350 mm stand-off distance.


a Flat Surface


The splat thickness measurement was performed using 2D profilometry. The

results agreed with the outcomes of statistical calculations for PP splat profiles and

thicknesses (Section 8.3.10). Figure 4-9a depicts the 2D profilometry data, across the

point on the splat indicated by the inset, as a trace line of the profilometer. An

alternative technique was also used to measure the splat thickness, which involved

cutting the glass slide using a diamond dicing saw close to the edge of the deposited

splat, and viewing under the SEM. This SEM side view image is shown in Fig. 4-9b,

and reveals a splat thickness of around 20 µm, which has been validated through

calculations on splats with the same topology.

Fig. 4-9. 2D Profile of splat cross section (along X-Y shown in inset) using data obtained from XP-2 surface profilometry; note that the X and Y axes are different scales, and (b) SEM image of the splat from side view.

4.3.3 Statistical Analysis of Splat Morphologies Statistical analyses were carried out using optical microscopy to determine the

splat shape frequency in terms of three categories: (i) few splashes (Fig. 4-6a and b), (ii)

little splashing (Fig. 4-6c and d), and (iii) overheated splats (Fig. 4-6e and f). The

maximum number of uniform disc-like splats was formed at 150 mm and 200 mm, Fig.

4-10a. The number fraction of disc-like splats decreased as the SOD increased beyond

200 mm. The incidence of splashed splats was greater at longer SODs up to 300 mm,

beyond which degraded and carbonized splats were observed. Similar behaviour was

observed at SODs of 200 mm to 250 mm, Fig. 4-10b, when the distance between the


a Flat Surface


powder injection port and the torch head increased from 50 mm to 100 mm. A 100 mm

distance resulted in more splashing and overheated splats than the 50 mm distance (i.e.,

the distance between the powder injection port and the torch head) that was employed

for all other experiments.

Fig. 4-10. Frequency of splat shape formations for different spray distances and powder injection distances with N = 100 for each measurement.

The splat formation depended on the position of particle injection and the flame

temperature, which gradually decreased from the powder port at the head of the torch.

Splat formation and splat structure were related to the distance between the particle

injection and the substrate because this geometrical parameter was related to the degree

of particle melting. SEM observations demonstrated that an increase in SOD from

150 mm to 350 mm could increase the degree of melting of the particle, Fig. 4-11. A

SOD maximum of 350 millimetres was used in these experiments and was the largest

SOD before the PP fully degraded. Partially melted particles were more likely to form

disc-shaped splats due to their unmelted particle centre; as well as the fact that any

melted portion of the particle ‘skin’ around the solid centre would contribute to a thin,

fine splat thickness around the splat periphery.


a Flat Surface


Fig. 4-11. Relationship between the degree of particle melting and stand-off distance of flame sprayed PP on a glass substrate at room temperature. Data obtained from SEM observations.

Three zones, representing a transition from low to high stand-off distances, could

be recognized on the basis of SOD. The zones were (i) a partially melted splat zone with

no or little splashing and less flow deformation due to insufficient heating, (ii) a fully

melted splat zone with or without splashing and full flow deformation due to sufficient

heating, and (iii) a splat evaporation zone and a highly deformed splat with full flow

deformation due to particle overheating. Table 4-2 describes the characteristics of splats

formed in each of these zones. Figure 4-12 shows the location of the three splat zones

identified with respect to SOD.


a Flat Surface


Table 4-2. Description of the splat formation zones identified in the thermal spraying of polymers.

Zone 1

(Partially melted splat zone)

Zone 2

(Fully melted splat zone)

Zone 3

(Splat evaporation zone)

Less flow deformation

Little or no splashes

Insufficient heating

Increased viscosity

Preservation of a rounded splat

Full flow deformation

Little splashing

Sufficient heating

Lower viscosity

Highly deformed splat

Overheating

Full flow deformation

Evaporated splat

Highly deformed splat

Fig. 4-12. Schematic of the three splat zones identified with respect to SOD.

4.3.4 Spreading Factor The splat spreading factor was defined as the ratio of equivalent splat size to

droplet diameter. Consider, for example, a SOD of 150 mm that gives rise to a

maximum fraction of disc-like splats: the mean splat size was 206 μm and the mean

droplet diameter was 159 μm. Thus, the splat spread factor was 1.3. The extent of

droplet spread was determined by measuring the splat size via optical microscopy and

image analysis, Fig. 4-13.

The experimental value for the maximum spread factor of splat average diameter

to particle average diameter was 1.3, which represented a thicker splat. The spread

factor was larger for metals [132, 158] and ceramics [146]. The explanation for this


a Flat Surface


observation was based on two facts: (i) the internal thermal resistance between the

polymer and the substrate was greater than that for other materials, and (ii) the PP

particle size used in this study was significantly larger than those of typical metal- and

ceramic-based feedstocks. Appendix C demonstrates a method of finding the

transformation function for the probability density function of splats.

Fig. 4-13. Particle size (diameter) and splat size distributions of polypropylene.

4.3.5 Molecular Structure (Micro–Raman Spectroscopy) The Stokes Raman spectra of the PP flame sprayed fully melted splats (near the

edge and at the centre) are shown in Fig. 4-14. The vibrational behaviour of helical

molecules of PP is spatially confined and so-called regularity bands occur at 809, 841,

900, 973, 998 and 1220 cm-1. These bands depend on the conformation of sections of

the chain that have 31 helical structures [184]. Nielsen et al. [185] reported that the

Raman lines of PP peaked at about 809 and 841 cm–1. These Raman lines could be

assigned to the vibrations of helical molecules localized in the amorphous regions and

the vibrations of molecules in the crystalline phase, respectively. Both of these peaks

were identified for the PP splat in both edge and centre locations. As the degree of

crystallinity increased, the intensity of the band at 809 cm-1 increased in relation to that

at 841 cm-1 [186].


a Flat Surface


Fig. 4-14. Raman spectroscopy spectra for flame sprayed PP splats and bulk PP.

The random orientation of the PP chains was reflected in the low peak ratio of the

Raman bands at 809 and 841 cm-1. The Raman data showed no major changes between

the centre and the edge. Moreover, both centre and edge spectra corresponded to the

spectrum of the bulk PP, which exhibited the typical Raman bands of oriented PP-

chains [184-186]. This data confirms that the fully melted PP splats had not undergone

any thermal structural degradation during the thermal spray process.

4.3.6 Crystallization of Polypropylene Splats in Thermal

Spray Polypropylene is a semi-crystalline thermoplastic that softens when heated and

hardens when cooled. Polypropylene consists of crystalline regions or spherulites

contained in an amorphous matrix. These two constituents behave differently at

different temperatures. At the glass transition temperature (Tg), the amorphous regions

soften to change from a brittle or glassy state to a rubbery or ductile state. At a

temperature higher than Tg and lower than the melting temperature, the crystallites or

spherulites begin to lose cohesion. At around 160 to 170°C (Tm), the crystalline regions

start to melt. Increasing temperature above Tm does not change the polymer state;

however, it will reduce viscosity. This higher temperature will deteriorate the

1300

1800

2300

2800

200 400 600 800 1000 1200 1400

Raman shift (cm-1)

Splat centre

Splat edge

Powder feedstocks

809

841 97

3

900 12

20

998


a Flat Surface


polypropylene structures. Figure 4-15 shows the behaviour of a PP splat during melting

and solidification.

Cooling the melted polymer will reform crystallites at the crystallization

temperature, which is lower than the crystalline melting point. Crystallization

temperature depends on the cooling rate and on the polymer composition (such as the

presence of fillers). For most plastics, the lower end of the recommended range of

processing temperatures will be well above the crystalline melting point, generally by

30 to 50°C. The crystallization temperature of polypropylene is in the range of 100 to

130°C [5].

Fig. 4-15. Description of a polypropylene splat behaviour during melting and solidification.

The final splat morphology and crystallization were influenced by the cooling

rate, which is rapid for thermal spray processes. The crystallization of the structure

starts in the colder portion of the solidified melt; this is represented by the thinner part

of the splat (splat rim). While these parts fill with spherulites, the high temperature

region in the thicker portion of the splat (splat core) remains amorphous, owing to the

weak nucleation and slow growth of spherulites [187]. The spherulites nucleated in the

colder region grow in the direction of increasing temperature. Therefore, the larger-

sized crystals are found toward the splat core where the cooling rate is lower [188].

Figure 4-16 shows spherical structures that radiate outward from a central nucleus, with


a Flat Surface


lamellae that grow in a radial direction forming spherulites. These spherulites were

observed on splats deposited at a 250 mm SOD where they were fully melted.

Fig. 4-16. SEM images of single polypropylene splats onto a glass substrate at 250 mm: (a) the whole splat with different degree of PP crystallizations indicated by dotted circles, (b) magnification of crystallization nuclei and spherulites, and (c) splat edge zoom depicting a region of two joining spherulites.

SEM images showed spherulites of different sizes formed throughout the PP

splats, Fig. 4-17. The mechanical properties of PP depend strongly on the shapes and

sizes of the spherulite crystals. Increasing crystallinity increases stiffness, yield stress,

and flexural strength; however, it decreases toughness and impact strength. The secant

flexural modulus at 1% displacement can range from 2,067 – 2,412 MPa for ultra-high

crystallinity PP; however, for general-purpose PP it decreases to 1,378–1,654 MPa

[178].

The SEM splat images showed a structural change in semi-crystalline polymers in

different regions of the splat. Figure 4-17b showed spherulites stacked in irregular

arrangements. The spherulite size tended to decrease toward the middle of the splat

where the thickness was higher and cooling rate lower. This indicated that there were

less amorphous materials in this region of the splat. However, the spherulite size


a Flat Surface


increased on the splat edge where the thickness was lower and cooling rate higher, as

shown in Fig. 4-17c. This result agreed to a great extent with most literature, which also

suggested that adding nucleating agents to PP will usually increase the spherulite size.

Fig. 4-17. SEM images of single polypropylene splats on a glass substrate at SOD of 250 mm: (a) the whole splat with different-sized bough-like spherulites, (b) the spherulites’ shape and size at the splat centre and (c) the spherulites’ shape and size at the splat edge.

The degree of crystallinity was reported to influence deterioration of the

mechanical properties of PP [189]. Low crystalline PP was reported to retain its

physical properties for a longer time than high crystalline PP, despite the fact that the

formation of photoproducts is quantitatively and qualitatively similar. The

microhardness increased [190] with the volume fraction of spherulites and can be

defined by equation 4-1:

Equation 4-1

The terms and are the microhardness values of spherulites and

interspherulitic ("amorphous") material, respectively, and is the volume fraction of

spherulites. No microhardness assessments could be conducted in this study; however, a

shear force applied to drag or pull off a splat enabled a qualitative test of the

coating / substrate bonding strengths.


a Flat Surface


Fig. 4-18 shows an adhered splat that failed to overcome such an applied force.

The splat behaviour was attributed to a lack of adhesion at the thicker central region of

the splat, along with less mechanical interlocking between the polymer and the smooth

surface of the glass slide. The adhesion at the thinner part of the splat was greater.

Figure 4-18b illustrates a transition with regard to adhesion mechanisms. The resistance

to fracture in the area between the thick and thin regions of the splat was not sufficient

to withstand the imposed stresses and resulted in the fracture shown, Fig. 4-18c. The

more crystalline structures at the splat rim allowed more coalescence between the splat

and substrate. Figure 4-18d depicts a larger spherulite structure with good adhesion to

the substrate and an area with smaller spherulites closer to the area of failure.

Fig. 4-18. SEM images of bonding failure for a PP splat deposited onto a glass substrate: (a) full splat, (b) magnification of areas with differing adhesion, (c) higher magnification of splat edge ripped from the substrate, and (d) area of PP crystallization with higher splat adhesion.


a Flat Surface


4.4 Conclusions

Flame spraying is a practical method of producing polypropylene coatings.

Feedstock properties and thermal spray parameters play an important role in splat

morphology. The results of this study show that optimal equipment configurations that

produce desirable disc-shaped splats are achieved when either (i) a stand-off distance

(SOD) of 150 to 200 mm is employed when the powder port is 50 mm from the torch

exit, or (ii) a stand-off distance of 200 to 250 mm is employed when the powder port is

10 mm from the torch exit. Increasing the SOD to 300 mm produces splats that exhibit

significant splash patterns, which suggest non-ideal building blocks for a coating.

Overheated fully spread, degraded and carbonized splats were observed at SODs

greater than 300 mm. Stand-off distance also influenced the degree of particle melting.

The processing window for fully molten particles with no degradation was critical and

challenging to achieve due to the polymer characteristics. That is, if the centre of a

particle was fully molten, then it was likely that the particle periphery would degrade

due to excessive heat. Increasing the stand-off distance also decreased the particle

velocity at impact. Splat size and thickness are factors influenced by particle dwell time

in the flame, which is a function of particle SOD. The longer particle dwell time

produced larger diameter PP splats that are thin. However, a shorter SOD yielded

unmelted or partially molten splats. Desirable morphological features for polymer splat

formation could only be achieved over a limited range of stand-off distances.

SEM images revealed that flame spraying could melt only the outer shell of

particles with diameters more than 100 μm, while the core of such particles stayed

unmelted. The splat size to thickness ratio of polypropylene was determined to be about

5.

Since the degree of porosity, density and coating adhesion depends highly on the

individual splat morphology, a 150-200 mm stand-off position was recommended to

make a dense coating with acceptable mechanical properties.

Raman measurements on a PP fully melted splat showed that the PP chains were

unchanged from the original PP grains. There are no major changes between the centre

and the edge of splats throughout the entire Raman shift.


a Flat Surface


A polypropylene spherulite structure resulted when a larger SOD of 250 mm was

employed. Splats were fully melted under these conditions. Spherulites radiated

outward from a central nucleus. The spherulite size tended to decrease toward the

middle of a splat where the thickness was higher and cooling rate lower. However, the

spherulite size increased on the splat edge where the thickness is lower and cooling rate

greater.


a Rough Surface


Chapter 5

5 Influence of Stand-off Distance on

Polypropylene Splats Deposited onto a Rough

Surface

5.1 Introduction

The variables investigated for their effect on splat formation in this section were

substrate material and substrate surface roughness. A mild steel substrate was the

practical material of focus. Mild steel, with a higher thermal conductivity than a glass

substrate, has different effects on the PP splat flattening and solidification and

consequently the shape and the thickness of deposited splats. Using a mild steel

substrate will exaggerate the differences in thermal conductivity between coating and

substrate materials, creating high thermal gradients through the polymer splat.

The surface roughness of the mild steel substrate also influences the dynamics of

droplet spread. Increasing surface roughness promotes splashing of impacting molten

droplets, in turn increasing the roughness of coating layers formed by solidified splats.

Therefore, the initial roughness of the surface may impact the porosity of thermal spray

coatings.

Interlocking of particles to the surface enables coating adhesion that generally

improves with a more rough substrate surface; a direct consequence of the larger area of

contact between the two materials. The bonded area increases as the asperities on the

substrate increase and the droplet is flattened by plastic deformation. Therefore, coating

interfacial toughness increases with increasing roughness. This leads to less crack

propagation than would be seen in a smooth splat-substrate interface and most coating

failures originate at this interface. Mean adhesion values can be higher for coatings that

have a higher surface roughness [191].


a Rough Surface


Flattening and cooling begin immediately on droplet impact against the substrate.

The lower part of the splat attached to the substrate imparts heat to the substrate. This

heat removal depends on the initial temperatures and thermophysical properties of the

substrate and the splat, and also on the contact thermal resistance at the substrate-splat

interface. The solidification front moves from the interface to the interior of the splat.

In this chapter, the influence of SOD on PP splat morphology deposited onto a

rough surface of a mild steel substrate was examined to establish a relationship between

the processing variables and the splat microstructural characteristics.

5.2 Methodology

Mild steel substrates were grit blasted using aluminium oxide provided by EMAS,

Abrasive Salto Ltd., Carborundum, Aloxite, Brazil, type EC31 and grit size 60. The

surface profile roughness of the substrate was obtained and analysed with a VECCO

WYKO NT1100 non-contact optical surface profilometry with (Vision V3.60) software.

The spray parameters used in this study are shown in Table 5-1.

Table 5-1. Spraying parameters used in this study.


Stand-off distance (mm) 150, 200, 250, 300, and 350




Particle size range (µm) 110-220




a Rough Surface



5.3.1 Mild Steel Surface Roughening The ability to deposit coating materials onto a wide range of substrates with high

quality bonding is desirable. The surface the mild steel after being grit blasted is shown

in Fig. 5-1 and the roughness is illustrated by the profile scan shown in Fig. 5-2. The

average surface roughness exceeded Ra 3.2 µm. This value agreed with the typical

average surface roughness (Ra) of a grit blasted surface used in thermal spray of around

2 to 15 μm, which needs to be considered relative to a mean droplet size of 50 μm [12].

Fig. 5-1. Mild steel grit blasted substrate.

Increasing the substrate roughness for deposition of smaller particles might result

in a symmetrical splat due to higher peak to valley distances. The magnitude of surface

roughness is related to the fluid perturbation at the moment of droplet impact. The X-

Profile (Fig. 5-2c) and Y-Profile (Fig. 5-2d) reveal the uniformity of the surface

roughness of the grit blasted substrates. The result of the three-dimensional image

developed for the roughened surface of the mild steel substrate as a function of grit-

blasting using 3D profilometry showed a uniform mild steel surface roughness. Figure

5-3 shows a 3D image of the rough substrate generated by (Vision V3.60) software of

the non-contact optical surface profilometry surface image.


a Rough Surface


Fig. 5-2. Surface roughness of mild steel substrates obtained by optical surface profilometry: (a) contour plot of the mild steel substrate surface with lines X and Y indicating the locations of profile measurements; (b) SEM image of the rough surface of mild steel substrate; (c and d) plot of depth vs. horizontal distance of X and Y profiles, respectively.


a Rough Surface


Fig. 5-3. Three-dimensional image of mild steel surface roughness using 3D profilometry of the mild steel roughened surface (SEM shown in inset figure).

5.3.2 Splat Formation on a Rough Surface When a molten polymer droplet impacts on a rough surface, the impact pressure

propels the molten droplet to completely conform to the substrate topology. The force

inside the molten droplet is an important factor that may not be sufficient to overcome

surface tension and force liquid into small cavities. As the splat rate of solidification is

high, the kinetic energy of the splat leading edge enables it to flow over peaks, then

downwards into grooves; as shown in Fig. 5-4 and indicated by A. On the other hand, if

the material flows with insufficient forces, the splat leading edge may encounter a peak

too high to allow movement upwards and over the substrate asperities (peaks), as shown

in Fig. 5-4 and indicated by B. Thus, material might solidify or change direction,

depending on the substrate and spray direction.

The SEM observations of multiple splats indicated that the substrate roughness

promoted splat instability derived from the radial material jetting away. Thus, irregular

splat shapes could be created by splat peripheries filling up adjacent valleys through the

rough substrate asperities. At droplet impact, a splat rim at a higher velocity than the

original impact velocity could be generated due to flow between the relatively viscous

particle core and the rigid substrate materials. This high shear flow leads to low

viscosity due to shear thinning. The low viscosity and high shear rate at the rim and

interaction with the substrate roughness contributed to the formation of jets and fingers,

as shown in Fig. 5-5.


a Rough Surface


Fig. 5-4. Schematic of mechanical adhesion of a splat onto substrate asperities with indication of a lower surface peak at A and higher surface peak at B.

Fig. 5-5. SEM images of splat formation of PP on a rough surface of mild steel, where (a) is a full splat and (b and c) are two areas of the substrate where PP flow is downwards into grooves.

The surface roughness affected the spread of the PP droplet and consequently the

splat topologies. Impact of molten droplets against a rough surface promoted splashing;

thereby increasing the roughness of coating layers for the new arriving droplets.

Significant changes in the splat formation and topologies of PP deposited on the mild

steel rough surface were observed by varying the SOD from 150 to 350 mm. Splats

sprayed at low SOD exhibited “fried-egg” shapes with a large core in the centre and a

thin rim. This morphology resulted from particle heat exposure due to a larger particle


a Rough Surface


dwell time. The relationship between the splat properties and SOD showed agreement

with results obtained for splats deposited onto a flat glass substrate surface.

The SEM observations indicated that there was a tendency for splats deposited on

a rough mild steel substrate to be more disc-like, with elimination of the thin rim in

multiple locations. Thus, for splats deposited onto rough mild steel substrates, the splat

size decreased compared to splats deposited on a flat surface. On the other hand, the

overall splat diameter could be larger when taking into consideration the extent to which

the polymer splat edge spread along the tiny grooves of the substrate.

The other observation refers to splashing that could be considered

disadvantageous to the coating properties. Splashing of PP on a rough surface was

governed by a mechanism that strives to fill the groves and valleys of the rough

substrate. Filling-in of the substrate surface increases splat-to-substrate contact and,

thus, enhances bonding and intensifies an interlocking mechanism, as indicated earlier.

The splat morphologies of PP deposited on the mild steel roughened surface with SODs

of 100, 150, 200, 250, 300 and 350 mm are shown in Fig. 5-6a, b, c, d, e and f,

respectively.

The mechanism of particle flattening and solidification was based on a balance

between the particle initial kinetic energy, viscous energy, and surface tension energy.

The particle kinetic energy is transformed into work of viscous deformation and

dissipates quickly during particle impact resulting in droplet spreading and flattening.

Kinetic energy dissipation is caused by the droplet viscosity and the surface tension.

Generally, the higher polymer viscosity acts against the droplet splashing, so that

polymer splats produce less splashes than other materials. Thus, a higher kinetic energy

is required for sufficient polymer particle deformation and flow at impact. Figure 5-7

shows a schematic of a droplet deposited onto a rough surface with splats formed at

high and low SOD, (a) and (b), respectively. Semi-molten particles necessitate even

higher energy. The high kinetic energy is an artefact of thermal spray technology that

makes it unique to manufacture coatings with particles that need not be in the fully

molten condition.


a Rough Surface


Fig. 5-6. SEM images of polypropylene single splats deposited onto a mild steel rough substrate at room temperature with different stand-off distances: (a and b) splats at 100-150 mm, (c and d) at 200-250 mm, and (e and f) at 250-300 mm.


a Rough Surface


Fig. 5-7. Schematic of a droplet deposited onto a rough surface with SEM images showing the effect of stand-off distance: (a) polyproplene splat deposited at low SOD, and (b) polyproplene splat deposited at high SOD.

Altering the SOD caused the most significant change in the splat morphology,

including the degree of splashing, compared to other spray parameters. However, the

general splat morphologies of polypropylene were consistent; they did not demonstrate

significant differences when deposited onto different substrate materials at the same

SOD. Figure 5-8 shows two splats deposited at the same SOD onto a flat glass substrate

and a rough mild steel substrate. The differences in thermal conductivity of the

substrates affected the degree of splashing, as a consequence of faster solidification. The

mild steel substrate, with a higher thermal conductivity, exhibited a higher level of

splashing than the glass substrate. The substrate thermal conductivity exerted the


a Rough Surface


greatest influence on the quench rate of the particles and greatly influenced the particle-

to-substrate bond [192].

Fig. 5-8. SEM images of splats at a SOD of 150 mm deposited onto (a) a flat glass substrate and (b) a rough mild steel substrate.

5.3.3 Splat Cracking and Delamination The splat-substrate interface region is the focus of most discussions on thermal

spray technology because this region represents the contact between two materials with

different properties. Inadequate adhesion between the splats and substrate can occur due

to different chemical and thermal characteristics of the two materials. Cracks and

delamination are examples of adhesive failure. A crack can be defined as a line of

fracture without complete separation, and delamination is the separation of splats or the

coating either within itself or from the substrate. The SEM observations of splats

indicated that the increase in substrate roughness promoted splat instability. The result

was splat radial jetting and break-up in terms of cracking and delamination, Fig. 5-9.

The impact force was greater at the splat centre and gradually decreased at the

splat periphery where thickness was low. The splat centre was typically attached to the

underlying substrate topology. During splat solidification, the splat edge kinetic energy

was insufficient to overcome the capillary pressure required to force liquid flow into

substrate asperities. Splat surface tension pulls the splat edge inwards and overcomes

the peripheral bonding. The splat edge bonding may be sufficient to adhere well to the

substrate, otherwise it delaminates. The high velocity of the splat leading edge and

interaction with the substrate roughness features can create surface valleys with trapped


a Rough Surface


air or other gases that are unable to escape prior to solidifcation. Thus, stress is

produced at the splat edge, leading to cracking (Fig. 5-9).

Fig. 5-9. Schematic of splat deposited on a rough surface with an air bubble trapped in substrate valleys.

The peaks on the rough substrate surface correspond to potential crack initiation

sites and the valleys correspond to potential crack arrest locations. Bubbles are

prevented from escaping the fast flow behaviour of droplets prior to solidification of the

splat and may leave a hole or crack as shown in Fig. 5-10. The thin layer of material at

the splat edge resulted in the coating experiencing this type of failure.

Fig. 5-10. SEM images of a polypropylene splat deposited onto a rough surface of mild steel, showing a trapped gas bubble leaking out of the splat leaving a hole or crack; (a) the full splat and (b) a higher magnification image of the crack.

Cracks are initiated in the tensile region of splats, presumably originating from

relaxation of strained grains and resulting in an increase in local stress. After the initial

cracks were created, crack propagation progressed throughout the loading direction and


a Rough Surface


stressed area of the splat, along the splat periphery, or pullout at the splat-substrate

interface. The rapid solidification of the splat top surface compared to the substrate

interface encouraged its compression and lifting up of the splat edges, causing splat

delamination. Under particular conditions, the high stress of the splat edge could be

interpreted in terms of delamination when splat-substrate bonding was insufficient, or in

terms of cracking when bonding was adequate. Figure 5-11 shows a splat suffering from

both cracking and delamination when deposited onto a substrate preheated to 70°C. The

region of both failures is shown in 5-11b at higher magnification, while 5-11c

confirmed that splat delamination occurs after solidification, with an indication of some

splat material still attached to the substrate.


a Rough Surface


Fig. 5-11. SEM images of a polypropylene splat deposited onto a substrate at SOD of 150 mm showing bonding failure: (a) full splat, (b) higher magnification of splat crack and delamination, and (c) higher magnification of the delaminated splat edge.


a Rough Surface


5.3.4 Energy Dispersive X-ray Analysis The chemical composition of a polypropylene splat deposited on a mild steel

substrate was confirmed from energy dispersive X-ray spectroscopy (EDX) analysis by

examining multiple points across the splat centre from edge to edge, as indicated by AB

in Fig. 5-12. The EDX spectrum of eight points examined across line AB shown in Fig.

5-13 indicated that the dominant composition for all points along the line to be carbon

(C), reflecting the main constituent of the PP grains. No significant changes in the

carbon content were distinguished along the tested AB line. The weight percentage of

carbon ranged between 80% - 95% of the total weight, which fluctuated around the 88%

carbon content for the original feedstocks, Fig.5-13. Note that hydrogen (H), the other

main constituent of polypropylene, is too light to be detected by EDX analysis.

Fig. 5-12. SEM image of a PP splat deposited onto a rough mild steel surface at room temperature at SOD of 100 mm, showing the points examined by EDX across line AB.


a Rough Surface


Fig. 5-13. EDX spectra results of eight points indicated by Fig. 5-12 and PP powder.


a Rough Surface


Other elements detected in the EDX spectra were (i) a small amount of oxygen

on the splat surface after the flame deposition and (ii) a significant amount of gold on

the surface. The detection of gold arose from the gold coating of the sample that is

necessary for conductivity purposes and an essential experimental protocol for imaging

polymer particles with the characterisation techniques that were employed. A similar

argument, focussed on some contamination, might also explain a slight amount of iron

and chromium being introduced at some points across the line. The EDX spectra were

obtained for four splats deposited at SODs of 100, 150, 200 and 250 mm at eight points

across the line AB, as shown in Fig. 5-14. The results showed that carbon dominated the

compositions for all splats deposited at different SODs, fluctuating around 75% to 95%.

Figure 5-15 shows the carbon weight percentage of eight points across the splat surfaces

at different SODs.

The results did not indicate a consistent relationship between the SOD increase

and the level of decarbonization. No change in the carbon content indicated that there

was no carbon decrease and, furthermore, no chemical structure changes in the polymer

composition. It is recommended that additional testing with more samples and

employing EDX of splats is required to further substantiate studies of this nature.


a Rough Surface


Fig. 5-14. SEM images of PP single splats on a glass substrate at SODs of 100 mm, 150 mm, 200 mm and 250 mm; a, b, c and d respectively showing the eight points of the EDX reading across the splat diameters (between A and B).

Fig. 5-15. Chart depicting the weight percentage of carbon in the PP feedstocks and eight points along the splat diameters (between A and B as shown in Fig. 5-13) for PP particles deposited onto a mild steel substrate at four SODs.


a Rough Surface


5.4 Conclusions

Feedstock and thermal spray parameters play an important role in splat

morphology. This study showed that the general splat morphologies of polypropylene,

using the same SOD and other spray parameters, were consistent between observations

using a glass substrate and when particles were deposited onto a rough surface of mild

steel. A stand-off distance of 150-200 mm is recommended to achieve a dense coating

having the desired mechanical properties. Increasing the SOD to 300 mm produced

splats that exhibited significant splash patterns. Overheated fully spread splats with

significant loss of the particle material, suggested to be due to degradation and

evaporation, were obtained at the larger SOD of 300 mm.

The EDX analysis results showed that no significant changes in the carbon

content could be distinguished along the tested points throughout the cross-section of

splats. The weight percentage of carbon ranged between 75% - 95% of the total weight

for all SODs.

Chapter 6: Influence of Substrate Temperature on Polypropylene Splats Deposited

onto a Flat Surface


Chapter 6

6 Effect of Substrate Temperature on

Polypropylene Splats Deposited onto a Flat

Surface

6.1 Introduction

The morphology, wetting, flattening and solidification of splats for all feedstock

materials including ceramics, metals and polymers are a consequence of substrate

surface pre-treatment as well as the spray parameters. Substrate preheating is an

important factor that influences the splat behaviour during impact and deformation.

Substrate preheating reduces the temperature differential between the feedstock and

substrate along with eliminating residual moisture and preventing condensation from

forming on the surface. Preheating the substrate surface above a critical temperature,

known as the transition temperature, Tt, results in disc-shaped splats. This results in

improved droplet wettability on the substrate above this critical temperature due to

vaporization of adsorbates and condensates on the substrate.

Experimental studies have shown that the cooling rate of a particle impacting on a

surface preheated above Tt was up to two orders of magnitude higher than it was on a

cold surface. A high cooling rate corresponded to low thermal contact resistance at the

interface splat-substrate [193]. This transfer rate of thermal energy from polymer splats

to the substrate was controlled by the polymer thermal conductivity. If the substrate was

cold, then the splat would complete spreading before splat solidification could begin.

Thus, such droplets solidify before fully spreading. They were observed to coalesce

with each other on the substrate surface.


onto a Flat Surface


In this chapter, the influence of the flat mild steel substrate temperature on PP

splats was studied. Studies on splat morphology and the nature of the splat-substrate

interface were performed to comprehend the manner in which processing variables of

the substrate condition influenced microstructural characteristics of splats.

6.2 Methodology

6.2.1 Substrate and Particle Preparation The mild steel substrates were prepared using Struers products as follows: (i)

grinding with 320 grit silicon carbide paper for 1 minute and water as the lubricant; (ii)

step 1 polishing with MD Largo for 3 minutes and DiaPro Allegro / Largo as a

lubricant; (iii) step 2 polishing with MD MOL for 5 minutes and DiaPro Mol as a

lubricant; and (iv) step 3 polishing with MD Chem for 1 minute and OP-U as a

lubricant. The surface profiles of the polished substrate were measured using a Taylor

Hobson precision, Surtronic 25 roughness tester with TalyProfile Silver software. The

average surface roughness (Ra) was 0.036 µm. The spray parameters used in this study

are listed in Table 6-1. The PP particles size distribution used ranged between 70-

120 µm (Fig. 6-1).

Table 6-1. Flame spray parameters used in this study.


Stand-off distance (mm) 150



Substrate temperature (ºC) Room temperature, 70, 120, 170


Particle size (µm) 70-120



onto a Flat Surface


Fig. 6-1. Particle size distribution of polypropylene powder feedstock.

6.2.2 Microstructural Analysis

The morphology and size of powder particles were characterized using a Zeiss

Supra 40 VP field emission scanning electron microscopy (SEM). Samples were gold

coated with a Dynavac (CS 300) deposition system prior to the SEM analysis. A Quanta

3D FEG high-resolution, low vacuum SEM / FIB was used for splat sectioning using a

30 kV ion beam. The rough cutting was performed at 15 nA whereas the surface clean-

up employed a 5 nA cut.

Raman spectroscopy is an effective method for the structural analysis of PP that

allows characterization of the molecular state at the various stages of its processing

schedule. Micro-Raman scattering experiments were performed using an inVia Raman

spectrometer (Renishaw plc, UK). Vibrational modes of the micro-Raman effect arise

from inelastic photon scattering [194]. Analyses were conducted with an excitation

wavelength of 532 nm at a spectral resolution of 1 cm-1 within the range of 200 to

3,200 cm-1. The spectra were collected from the original PP feedstock as well as from

splats that experienced different heating conditions.

0

10

20

30

40

70 80 90 100 110 120

N=100

Freq

uenc

y (%

)

Particle size (µm)


onto a Flat Surface



6.3.1 Splat Characterization SEM observations of splats indicated that the PP splat morphologies varied with

substrate temperature. Splat size and morphology generally increase with increasing

substrate temperature, Fig. 6-2. Increases in splat size and degree of circularity arose

due to the increased time available for the splats to spread and solidify after impact

since the temperature difference between the splat and the substrate was decreased.

Disc-like splats, distinguished by a “fried-egg” morphology with little splashing

evident, were seen in the SEM images. The “fried-egg” phenomenon was exhibited by

polymer splats on substrates held at room temperature due to the large radial difference

in the flow properties of the molten PP droplets. The lower viscosity of the fully molten

rim layer contrasted with the highly viscous core that did not spread upon impact [11,

72].

Fig. 6-2. Particle and splat size distribution of polypropylene splats deposited onto a mild steel substrate preheated to various substrate temperatures.

Particles deposited onto the preheated substrates showed flattened

hemispherical-shaped splats that were partially melted, forming “fried-egg” splats as

well as other splat morphologies with a greater degree of melting and post-deposition

material flow, as shown in Fig. 6-3 (a-d). These particles experienced a longer cooling


onto a Flat Surface


time and reflected solidification conditions that were dictated by the substrate

temperature and relative surface tensions. Heating the substrate close to the PP powder

melting point range of 160-170ºC produced fully molten splats with a distinctive tri-

cusp on the centre of the splat surface. This could be attributed to splat shrinkage during

solidification, see Fig. 6-3dc.

Fig. 6-3. SEM images of flame sprayed polypropylene particles deposited onto a polished mild steel substrate at different preheating temperatures; showing the effect of substrate temperature on the splat morphology, size, surface texture and adhesion to the substrate. Figs. a, b, c and d are a top view of full single splats at different preheating temperatures (i.e., room temperature, 70, 120 and 170ºC, respectively). Figs ac, bc, cc and dc show the surface texture at the splat centre for the same splats. Figs ae, be, ce and de show the splat edge in contact with the polished substrates at higher magnification.

Tri-cusp


onto a Flat Surface


The SEM images showed patterns with different scales of undulating surface

texture in the splat surface mound; i.e., the thicker portion of the splat, Fig.6-3 (ac, bc, cc

and dc). Surface ripple patterns increased with increasing substrate temperature. It is

proposed that there are two forces acting during droplet flattening and solidification;

i.e., (i) the effect of the substrate temperature that influences polymer flow and (ii) the

splat surface tension that impedes material flow. The PP flow increased as the substrate

temperature increased and was reflected in a decrease in splat thickness and an increase

in splat size.

The size scale of the undulating surface features varied with respect to the

substrate preheat from a slightly wavy surface texture at a lower preheat temperature

(Fig.6-3a) to a coarse texture at the higher temperature (Fig. 6-3d). Another reason for

rippling in the thickest part of the splat relates to inhomogeneous cooling where the

outer surface solidified first. Then, as the interior cooled and shrinks at rates depending

on localized cooling conditions, the now solid surface buckles. There was more

shrinkage for high temperature substrates and therefore more buckling. This effect could

not be observed in the thinner periphery where the shrinkage was restricted by the low

thermal expansion of the mild steel.

No significant change in the degree of splat splashing was observed from

changing the substrate temperature. In comparison, a previous study where the same

material was deposited onto a flat surface, demonstrated that stand-off distance had a

large influence on the degree of splashing.

The transition temperature is defined as the substrate temperature where the splat

morphology changes from a splash to a disc morphology [154] and this temperature

increased with increasing substrate thermal conductivity. Several studies have been

carried out on preheated metal substrates to determine the transition temperature. There

is no record of the optimum metal temperature for the splat transition of polymer

coatings. Solidification occurs after the completion of the flattening process when the

substrate is preheated above the transition temperature [147]. The limited degree of

splashing observed on the room temperature substrate was eliminated on preheated


onto a Flat Surface


substrates and agreed with reports that increasing the substrate temperature leads to

reduced splashing for different materials [124, 155].

Good adhesion between the substrate and splats is required and the adhesion

mechanism depends on the surface characteristics of the abutting materials.

Metallurgical bonding is not applicable in the present case since the polymer is coating

a metal substrate. Other physical mechanisms of adhesion can occur and mechanical

interlocking under the combined factors of heat and kinetic energy from the impinging

particle will promote intimate contact between the particle and substrate. Bonding by

mechanical interdigitation of the molten or semi-molten PP with the substrate is the

most likely adhesion mechanism; especially with a roughened substrate when the high

force of an impacting droplet forces liquid into surface crevices that then freeze and

form interlocking connections.

The crystalline regions of polypropylene exhibited a greater shrinkage than the

surrounding amorphous regions. Shrinkage of PP is related to the degree of crystallinity

and hence to the cooling rate. A greater degree of crystallinity leads to a higher rate of

shrinkage and also to a greater differential between shrinkage in the flow direction and

shrinkage measured transverse to the flow [188]. The differential effect is another

consequence of the visco-elastic property of long-chain molecules. During flow, the

molecules are aligned to a limited extent in the flow direction and are extended to a

degree that is proportional to shear rate. On cooling, a partial recovery of this extension

gives rise to a higher shrinkage.

Figure 6-4 summarises the splat and substrate forces that influence the flattening

and solidification of a splat. The figure shows the flame sprayed PP splats that evolve

when the substrate is held at room temperature and as the substrate preheat temperature

increases. Table 6-3 specifies the characteristics of splats that are impacted onto

substrates under different preheating temperatures.


onto a Flat Surface


Table 6-2. Characteristics of splats impacted onto a substrate held at different temperatures.

Fig. 6-4. Schematic of a droplet deposited onto a smooth surface showing the forces acting on droplet flattening and the effect of substrate temperature, with SEM images of a polyproplene splat deposited onto a mild steel substrate held at (a) a low temperature of 70ºC and (b) a high temperature of 170ºC.

Characteristics of PP splat impacted onto a substrate held at low

temperature (RT)

Characteristics of PP splat impacted onto a preheated substrate (170ºC)

Thicker splat core and thinner centre Smaller splat size Smaller size of rippling surface texture Inhomogeneous cooling Fast solidification rate Less shrinkage

Uniform splat thickness Larger splat size Larger size of rippling surface texture Homogeneous cooling Slow solidification rate More shrinkage


onto a Flat Surface


6.3.2 Splat-Substrate Interface Disc-like splats exhibited good adhesion while splashed splats demonstrated

poor adhesion. An adhesion defect described as splat delamination may occur at the

interface between a splat and substrate due to poor adhesion. Good bonding was

expected to occur over the region of the highest pressure within a droplet. The

combination of low contact pressure and high surface tension at the droplet edge leads

to poor contact, which induces splat curling, Fig. 6-3ae. If the next deposited droplet

cannot fill in the gap caused by the previous curled-up splat, then porosity is created: a

microstructural artefact that is undesirable for the majority of applications.

The current study demonstrated that delamination occurred as a result of splat

edge lifting at the periphery of disc-shaped splats for substrates held at room

temperature. The poor mechanical bonding to the substrate observed was attributed to

(i) the surface tension at the periphery of the molten splat during spreading [150], (ii)

the large mismatch in the coefficients of thermal expansion between the polymer and

substrate causing shrinkage in the splats on cooling, and (iii) the relief of stress

associated with the bonded portion of splats that did not contract to the same degree as

other segments of the splat during solidification [195].

FIB images for cross-sections of splats deposited onto substrates at room

temperature and 120ºC are shown in Fig. 6-5a and b. Preheating the substrate vaporized

any trapped water vapour or organic residue and allowed the molten particle to flow into

the substrate asperities for maximum adhesion [2]. There was no observation of

substrate melting or intermixing between the splat and the substrate for both conditions.

This behaviour indicated no diffusion or metallurgical bonding for the polymer coating

on the metal substrate.


onto a Flat Surface


Fig. 6-5. SEM images of polypropylene splats deposited onto polished mild steel substrates at different preheat temperatures: (a) a splat deposited at room temperature and the side view at higher magnification after FIB milling showing delamination across the splat edge in contact with the substrate; (b)a splat deposited at 120ºC and the side view indicating good adhesion between the splat edge and the preheated mild steel substrate.

The degree of particle deformation reflected the thermal input into the particle

and substrate. Raising the substrate temperature enabled the formation of well-melted

splats that could coalesce. Figure 6-6 illustrates the differences in splat morphologies

due to the changes in substrate temperature. The splats were still rounded and partially

deformed; indicative of high viscosity with little subsequent deformation after impact

due to the relatively high cooling rate on the 70ºC substrate (Fig. 6-6a). The highly

deformed disc-like splats signified particle melting and viscous flow deformation after

impact due to a lower cooling rate at the higher substrate temperature of 170ºC, Fig. 6-

6b.


onto a Flat Surface


Fig. 6-6. SEM images of PP splats deposited onto mild steel substrates showing the degree of melting deformation at two different conditions where the substrate was preheated to (a) 70ºC and (b) 120ºC.

The temperature of the deposited polypropylene splat depended on the substrate

preheat temperature that, therefore, had a large influence on the coating temperature.

Mechanical bonding of coatings composed of single particles could be achieved in

thermal spraying without intensive preheating of the substrate; a slight increase in the

substrate temperature enabled a significant change in splat formation. Splats formed on

substrates preheated at temperatures above RT exhibited a high level of splat-to-splat

coalescence and formation of fully molten particles, even at lower stand-off distances.

Figure 6-7 shows the effect of substrate preheat temperature on PP splat formation for a

large number of splats obtained from a single pass of the flame spray torch.

The SEM images in Fig. 6-7 show that preheating a mild steel substrate

generally changed the formation and structure of deposited splats. As indicated earlier,

the splat formation and structure varied because the splat solidification rate changed.

Polypropylene splat morphology in the first layer sprayed onto a mild steel substrate

surface affected considerably the adhesive strength of the coating. Furthermore, the

topology of the first splat pattern enabled the formation of the next splat layers as well

as contributing to the incorporation of porosity into the coating microstructure.


onto a Flat Surface


Fig. 6-7. SEM images of many polypropylene splats deposited onto a mild steel substrate at preheat temperatures of (a) 70°C, (b) 120°C and (c) 170°C.


onto a Flat Surface


Generally, successful application of thermal spray coatings depends on the

quality of adhesion between the coating and the substrate. The splat formation is based

on particle kinetic energy and its ability to deform plastically on impact. Splat adhesion

is directly proportional to the kinetic energy of the spray particles; this kinetic energy

varies with the particle velocity. The higher particle velocity of thermal spray processes

explains the mechanism of coating formation by particles that need not be fully molten

to spread out on the substrate surface. An example of such a mechanism would be cold

spray technology where particle adhesion and cohesion depend strongly on the kinetic

energy of the particles.

When a molten droplet impacts onto a substrate surface, the droplet kinetic

energy and momentum induces a lateral flattening of the droplet liquid along the

substrate surface. The contact pressure dissipates quickly with droplet flattening and

decreases during droplet spreading; particularly at the splat edge. Thus, solidification

will start in an area where the impact pressure is not too low and where the flattening

droplet is thinner. Since the maximum force occurs at the centre of the impacting

droplet, the collision force; i.e., the force per unit area, is focused at the centre of the

splat during the flattening process.

The high central collision force promotes a firm contact and indicates good

adhesion between the splat and the substrate. The stagnation pressure and low shear

flow rate at the splat centre direct the droplet flow towards the substrate surface;

encouraging the fluid to flow into surface asperities. Figure 6-8a depicts an FIB image

of the transverse section of a splat. Figure 6-8b shows no porosity or delamination

throughout the splat-substrate interface except at the splat edge. This morphological

feature indicated that the comparatively low kinetic energy of flame spray is sufficient

to produce a good splat-to-substrate adhesion.


onto a Flat Surface


Fig. 6-8. FIB images of a PP splat deposited on mild steel: (a) the full splat and (b) a high magnification depicts a good adhesion with no defects at the splat-substrate interface.

6.3.3 Structural Characterisation

Vibrational analysis of semi-crystalline PP by Raman spectroscopy is well

adapted for characterisation of overall chain structure regularity since the Raman

scattering spectra are sensitive to both inter- and intra-molecular interactions. Analysis

of the Raman spectra obtained from the deposited PP splats onto different preheated

substrates exhibited vibrational peaks in the region of 200 to 1,600 cm-1. Sensitivity to

the level of crystallinity in PP occurs predominantly in the 750 to 880 cm-1 region; and

the so-called regularity bands occur at 809, 841, 900, 973, 998 and 1,220 cm-1 [196].

Previous studies have shown that the Raman lines peaking at about 841 and

809 cm–1 can be assigned to the vibrations of helical molecules localized in the non-

crystalline regions and the vibrations of molecules in the crystalline phase, respectively

[185, 197]. Arruebarrena de Báez et al. [186] demonstrated that as the degree of

crystallinity increases, then the intensity of the band at 809 cm-1 rises relative to that at

841 cm-1. In the present study, the Raman spectra generally showed no major changes

or shifts of the principal peaks between the feedstock and the PP particles deposited on

substrates at various temperatures.

Table 6-3 shows the average ratio of heights for the 809 cm-1 and 841 cm-1 peaks

for four splats on each substrate; while representative Raman spectra are shown in Fig.

6-9. The ratio of peak heights was consistently lower for the splats than the feedstock,


onto a Flat Surface


which suggested that partial melting in the flame tended to reduce the level of particle

crystallinity. The increasing trend in the ratio suggested that the degree of crystallinity

increased with higher substrate temperatures, which was consistent with a slower

solidification rate. The temperature of the substrate relative to the glass transition

temperature of the deposited polymer is an important parameter that influences the level

of crystallinity. Polypropylene will be heated when the substrate temperature is well

above its glass transition temperature, thereby promoting crystallisation.

Table 6-3. Average ratio of Raman spectra peak intensities at 809 cm-1 and 841 cm-1 for polypropylene under different substrate conditions, each value was based on measurements of four different splats, with an average standard deviation of ±0.06.

Raw

material

Substrate at

RT

Substrate at

70ºC

Substrate at

120ºC

Substrate at

170ºC

1.25 1.15 1.17 1.31 1.37


onto a Flat Surface


Fig. 6-9. Raman spectra for PP powder and flame sprayed PP on mild steel substrates preheated to various temperatures with the two significant PP peaks highlighted.

6.4 Conclusions

Polypropylene splats were formed using the flame spray process by impacting

the feedstock particles onto polished, preheated mild steel substrates at room

temperature, 70, 120, and 170°C. The PP splat morphologies were affected by the

substrate temperature as was the degree of particle melting. Raising the substrate


onto a Flat Surface


temperature from RT to 170ºC provided splats that were well melted and displayed

good adhesion characteristics to the substrate. The higher substrate temperatures

promoted coalescence between splats as well as keeping splash morphologies to a

minimum.

The diameter of disc-like splats was increased with increasing substrate

temperature. As a result the flattening ratio was increased. The splats deposited onto the

RT substrate exhibited a “fried-egg” shape with a large central viscous core and a fully

melted wide rim with a thin edge. At higher substrate temperatures, the splat shape

changed to a thicker splat edge and there was more flow of the centre material towards

the perimeter to form a more uniform splat with reduced thickness differences.

Delamination features were observed along the edge of the splat-substrate interface for

the RT substrate with no observation of substrate melting or interfacial atomic-scale

bonding.

The Raman spectra exhibited no major changes or shifting of the principal peaks

for feedstock and the PP particles deposited at different substrate temperatures. These

results confirmed that the material did not suffer any thermal degradation. However, the

increase in the intensity of the 809 cm-1 band relative to the 841 cm-1 peak suggested

that the degree of crystallinity increased with increased substrate temperature from RT

to 170ºC.


onto a Rough Surface


Chapter 7

7 Effect of Substrate Temperature on the Splat

Formation of Flame Sprayed Polypropylene onto a

Rough Surface

7.1 Introduction

Satisfactory mechanical performance of thermal spray coatings requires good

adhesion between the substrate and splats. The adhesion mechanism depends on the

surface characteristics of the abutting materials. According to the American Society for

Testing and Materials (ASTM) definition (D907-70), adhesion is “the state in which

two surfaces are held together by interfacial forces which may consist of valence forces

or interlocking forces, or both” [15]. Coating adhesion is usually considered as being

the result of a combination of mechanisms based on diffusion, mechanical interlocking,

electrostatic attraction, physical adsorption, and chemical bonding. Mechanical forces

originate from coatings wedging or keying into their substrates and interacting with

surface asperities. Chemical forces originate from chemical reactions between the

coating and the substrate. Physical forces originate from covalent bonding and van der

Waals interactions.

Metallurgical bonding is not relevant in the present study since the polymer is

coating a metal substrate. There was no possibility of the molten polymer causing

substrate melting at contact due to the high melting temperatures of metals. Other

mechanisms of adhesion can occur through mechanical interlocking under the combined

factors of heat and kinetic energy from the impinging particle. This promotes intimate

contact between the particle and substrate, particularly for roughened surfaces. Bonding




by mechanical interdigitation of the molten or semi-molten PP with the substrate was

the most likely adhesion mechanism in the present study, with the high force of

impacting droplets forcing liquid into surface crevices of the roughened substrate that

then freeze; forming an interlocking connection.

Improvement in the coating adhesion can be achieved through substrate pre-

heating. Pre-heating the substrate drives off condensates and adsorbates on the surface

as well as improving the flow of the splats on the substrate by reducing the cooling rate.

This process lowers the surface wettability or the contact angle and results in the

polymer material being able to flow into the pores and other morphological surface

features; thereby increasing the substrate-splat contact area and, thus, the adhesion. The

substrate surface pre-treatment (heating and / or roughening) has a significant effect on

the adhesion of polymer coatings [18, 43].

In this chapter, the influence of the substrate preheating temperature on the splat

morphology of PP powder particles deposited onto a rough substrate of mild steel was

examined. The study was conducted to understand the relationship between the

processing variables and splat characteristics.

7.2 Methodology

The substrates were roughened by grit blasting in a similar method to that

described in Chapter 5 (Section 5.2.1). Particle characterisation and other spray

parameters are analogous to the aforementioned description in Chapter 6 (Section

6.2.1). However, the focus of this chapter is the influence of the substrate preheating

temperature on the splat morphology.


The influence of surface roughness has not been completely described within the

literature, since most of the prior work addressed droplet impact onto smooth surfaces.

The majority of thermal spray publications involving rough substrates have employed

simulation modelling. However, irregular surface roughness is still a challenge for




realistic modelling of splat predictions, and the number of factors that contribute to splat

formation increases with a rough surface.

The difference in the splat shapes for particles deposited onto smooth and rough

surfaces are generally manifested at the moment of impact. At the first point of contact

between droplet and substrate, the kinetic energy begins to dissipate and the splat shape

starts to form. While there is a single point of contact between the droplet and a smooth

substrate, it is possible for more than one surface peak to contact a droplet impacting

onto a rough substrate, Fig. 7-1. This may induce break up and non-symmetric splats.

Fig. 7-1. Schematic of droplets before and at impact onto smooth and rough substrates.

In this investigation the substrates were grit blasted so that adhesion could be

qualitatively observed. The surface of the rough substrates is illustrated, using ImageJ

software, by a three-dimensional image of a mild steel rough surface, Fig. 7-2. The

average surface roughness Ra was confirmed to be around 3 µm.




Fig. 7-2. Three-dimensional surface roughness image of mild steel substrates used in this study. Obtained by ImageJ software. The inset figure shows the SEM image of this surface.

The SEM splat observations indicated that the PP splat morphologies varied with

preheat temperature, Fig. 7-2. Splat areas generally increased with increasing substrate

temperature following the same tendency demonstrated in Fig. 6-2 and for the same

reasons. Disc-like splats, partially or fully melted, forming a “fried-egg” morphology

with little splashing were shown on the SEM images, with a fully molten rim layer and

highly viscous core. The “fried-egg” phenomenon has been discussed previously.

Heating the substrate close to the PP powder melting point range of 160-170ºC

produced fully molten splats. A distinct cusp in the vicinity of the splat centre was

evident on some thicker splats, Fig. 7-3.

The high force of an impacting droplet forces liquid into surface crevices of the

roughened substrate, which then freezes and forms an interlocking bond. The pressure

and surface tension at the droplet edge varies and leads to features around the edges,

Fig. 7-4.




Fig. 7-3. SEM images of flame sprayed polypropylene particles impacted onto a grit blasted mild steel substrate. Figures a, b, c and d are the top view of full single splats at preheating temperatures of room temperature, 70, 120 and 170ºC, respectively.

Fig. 7-4. SEM images of flame sprayed polypropylene particles impacted onto a grit blasted mild steel substrate. Figures a, b, c, and d show the splat edge in contact with the rough substrate at preheating temperatures of room temperature, 70, 120 and 170ºC, respectively.




Coating formation of substrates held at room temperature involves rapid splat

solidification due to the large difference in temperature between the droplet and

substrate. The next droplets overlap on previously solidified splats and do not fuse

sufficiently to fill the surface cavities located at the splat edges; thus inducing shadow

porosity within the coating at the splat boundaries. The observation of the first wave of

PP splats deposited onto a substrate preheated to 120°C (Fig. 7-5a and Fig. 7-5c)

indicated that splat-to-splat reciprocal interactions occurred before the start of

solidification; that is when splats were in the liquid state. Along with high substrate

wetting, these interactions allowed splat capillaries to flow and induce splat fusion.

Figures 7-5b and 7-5d show PP splats in a single layer coating with good splat

interactions, induced by post-deposition flow due to a higher substrate temperature of

170ºC.

Fig. 7-5. SEM images of PP splats impacted on a grit blasted mild steel substrate at two preheat temperatures; (a) 120ºC, (b) 170ºC, (c) higher magnification of image “a” and (d) higher magnification of image “b”.




The SEM images (Fig. 7-6) of several splats in one coating layer of PP,

deposited onto a mild steel substrate preheated close to the PP melting point range,

indicated that fully molten splats were produced, with many spherulites forming a

crystalline structure. Melting of PP as a semi-crystalline polymer requires an additional

heat for melting of the crystalline structure; known as the latent heat of fusion. [188].

The latent heat of fusion release increases the temperature and slows crystal growth in

the thicker region of the splat due to insufficient heat removal, as this is restricted by the

low thermal conductivity of PP. Polypropylene crystallisation proceeds largely under

splat cooling and solidification well below its melting point.

The crystallisation growth of thermal sprayed PP dominated the splat and

coating structure. The porosity level of one coating layer was noticeably increased as

the coating structure became more crystalline. The porosity level depended primarily on

the semicrystallinity degree of PP and the growth rate of crystals induced by the

substrate temperature. The spherulites of adjacent splats appeared to be well coalesced

with each other, as shown in Fig. 7-7. In this instant, crystallisation then occurred

during splat cooling and solidification when splats were in contact during the melting

process.




Fig. 7-6. Polypropylene splats deposited onto a mild steel substrate preheated at 170°C, showing a crystalline structure and induced porosity; (a) layer of splats and (b) higher magnification indicating a porous structure.

Fig. 7-7. Spherulite adhesion between splats of particles deposited onto a mild steel substrate preheated at 170 ° C; (a) two splats in close contact and (b) three splats in close contact.




7.4 Conclusions

Polypropylene splats were formed using the flame spray process by impacting

feedstock particles onto rough, preheated mild steel substrates at room temperature, 70,

120, and 170°C. The effect of preheat temperature was noticeable, influencing the splat

morphologies and degree of particle melting. Improved adhesion characteristics of the

polypropylene particle to the substrate were produced when the substrate temperature

was raised from RT to 170ºC. The higher substrate temperatures allowed coalescence

between splats. Disc-like splats, partially or fully melted, forming a “fried-egg”

morphology with little splashing, were exhibited on the SEM images with a fully

molten rim layer and highly viscous core.

Heating the substrate close to the PP powder melting point range of 160-170ºC

produced fully molten splats with a greater proportion of crystalline structures and

increased porosity. A distinct cusp on the splat centre on some thicker splats was

evident. Splat-to-splat reciprocal interactions increased as the preheating temperature

increased. Spherulites of adjacent splats were well coalesced.

Chapter 8: Splat Taxonomy of Thermally Sprayed Polypropylene


Chapter 8

8 Splat Taxonomy of Thermally Sprayed

Polypropylene

8.1 Introduction

The morphology of a powder particle and the splat that evolves during the

thermal spray process are key to determining the physical attributes of a coating. The

splat shape is an important feature that requires accurate characterisation, preferably by

means of a quantified metric, so that the rapid solidification process of thermal spray

deposition can be understood.

This study reports on a PP powder that was sprayed onto glass substrates at

room temperature using the flame spray process, at various stand-off distances (SODs).

Several statistical concepts were employed using image analysis techniques

complemented by optical microscopy, scanning electron microscopy, and 2D

profilometry: methods that measured the splat metrics of formation, including an

estimate of splash area. Measurement of equivalent diameter, degree of splashing,

spreading factor, deposition efficiency and circularity have quantified splats so that

thermal spray parameterisation can be related to the process efficiency. The results of

this study indicated that increasing the SOD from 100 to 250 mm reduced the splat

thickness and deposition efficiency; while other metrics (listed above) were not

routinely correlated, though an increase in splash area was demonstrated. Splat

circularity was steady around 0.9, indicating that the deposited particles exhibited a

circular shape at all SODs.

This chapter consolidates information related to splat geometry for a specific

polymer feedstock. Several geometric quantities were investigated with the aid of



imaging software to determine the splash behaviour of polymer particles sprayed at

different SODs. Splat profiles were evaluated to determine the area of splashing fraction

for different SODs and establish a relationship between the SOD and the splash

occurrence. Statistical tools were implemented to study the effect of the SOD on the

geometric characteristics of polymer particles after their impingement against a glass

substrate.

8.2 Experimental Procedures

Standard microscope glass slides were held at room temperature (~23°C) and

used as substrates. Substrates were cleaned with acetone and dried before spraying.

Table 8-1 shows the spray parameters used in this section.

Table 8-1. Flame spraying parameters used in this study.


Stand-off distance (mm) 100, 150, 200 and 250




Powder size (μm) 70-120



Different shape factors were assessed. A measure of the data variability is

represented by the standard deviation of the mean that is included on the graphs. A two-

step technique was employed to evaluate the splat shape. Using image analysis

software, the splat periphery was manually traced and an enclosed shape inside the

splat, without the splash features, measured as shown in Fig. 8-1a. The second step was

to measure the peripheral projection of all splat material around the core of the flattened

particle and, hence, obtain the area of the splat that included splashed material, Fig. 8-

1b. The equivalent diameter (ED) was used to define particle size and offers a practical

technique for measuring the size of irregular splat shapes. ED was defined as the



diameter of a circle with the same area as the selected feature and can be calculated after

measuring the area of a 2D image of the particle shape.

Fig. 8-1. SEM images of the same splat showing the two different measurement techniques using image analysis: a) measuring splat particle size without splash features, and b) measuring actual splat area with splash features.

Degree of splashing (DS) was defined as the peripheral projection of material at

impact. Spreading factor (SF) was defined as the ratio of maximum spreading diameter

to the initial droplet diameter. Splat perimeter is a measure of the outside boundary of

each splat. The splat areas, circularity and the length of the periphery of all splats were

determined using a software imaging technique. The ED was then calculated using the

circularity approximation relation of equation 8-1.

Equation 8-1

The splat DS was calculated according to the method defined by Montavon et al.

[198] by applying equation 8-2, and the circularity calculated using equation 8-3.

Equation 8-2

Equation 8-3

Here, A is the area of the splat (µm2) and P is the splat perimeter (µm). One

hundred splats were analysed for each processing parameter. A splat counting method

was used to evaluate deposition efficiency (DE) in terms of SOD. The method was

conducted using scanning electron microscopy, optical microscopy and image analysis.



The related deposition efficiency was calculated for two powder sizes with mean

particle sizes of 75 µm and 125 µm. The counting of splats was constructed using a

125 mm2 area of the glass slide substrate. Splat perimeter is a measure of the outside

boundary of each splat.

Another metric considered during this study was the feret diameter. Feret

diameter (Df) is defined as the perpendicular distance between parallel tangents

touching opposite sides of a profile [199]; and is also known as the “maximum calliper

distance”. The definition also used minimum feret diameter (D f,min. ), maximum feret

diameter (D f,max. ) and mean feret diameters. The feret diameter (Df) was obtained at

90° to the direction of the maximum feret diameters (Fig. 8-2). The minimum feret

diameter is often used as the diameter equivalent to a sieve analysis. The splat inter-

relationship of Df and other splat metrics such as splat area and splat perimeter enabled

a full statistical understanding of the splat formation. Figure 8-3 is a schematic of these

size and shape factors.

Fig. 8-2. Feret diameter definitions.



Fig. 8-3. Schematic of the top view of particle and splat shape factors.


8.3.1 Particle Size Analysis

The particle size distribution and particle morphology are related to the

characteristics of the deposit. It is expected that a larger particle size will give rise to

larger splats and splashes. However, if the velocity and temperature are high then this

effect may be disguised due to severe fragmentation on impact. The particle

morphology will influence its flow within the flame spray process envelope. Thus, a

spherical particle is more likely to exhibit ideal behaviour where the flow path is

reasonably linear. On the other hand, granular and angular feedstocks will tend to

tumble and twist in-flight; thereby giving rise to non-linear flow and probably a slightly

more divergent spray angle. The heat transfer to symmetrical particles will also be non-

uniform compared to spherical feedstock and the outcome would be revealed as angular

features within the coating.



The characteristics of the feedstock evolve from their manufacturing process.

Grinding processes, as in the case of this study, will lead to rough, angular

morphologies that are likely to lead to more variability in the microstructures of the

deposits. The major attribute of such feedstocks is their availability and lower cost. In

many instances such feedstocks are quite acceptable for thermal spray processes. The

PP feedstock after grinding is shown in Fig. 8-4, which depicts an irregular and angular

morphology with small flakes adhering to the larger particles. The larger particles

manifested themselves by impeding the flow characteristics into the flame. The particle

size distribution of PP powder, Fig. 8-5, was determined by counting and measuring

particles under optical microscopy and image analysis.

Fig. 8-4. SEM image of polypropylene showing the particle size used in this study.

Fig. 8-5. Particle size distribution of polypropylene feedstock.

0

10

20

30

40

70 80 90 100 110 120

N=100

Freq

uenc

y (%

)

Particle size (µm)



8.3.2 Splat Morphology Figure 8-6 demonstrates the different levels of splashing seen in the splat

formations of PP sprayed onto a glass substrate, collected at various SODs. Generally,

the degree of polymer splashing on the flat substrate at room temperature was minimal,

but more significant at the larger SOD. SEM images of PP splats sprayed at the lowest

spray distance of 100 mm exhibited the smallest degree of splashing, with a large,

symmetrical and nearly-hemispherical unmelted core, and a fully melted wide thin rim

around it, forming a “fried-egg” shaped splat. No voids or cracks were distinguishable,

with good splat edge adhesion to the substrate. Increasing the SOD led to a proportional

increase in the splat areas and splash size.

Fig. 8-6. SEM images of PP splats deposited onto a glass slide, with SODs of 100 mm, 150 mm, 200 mm and 250 mm for a, b, c and d respectively.

The splat areas were generally proportional to the preheating of the substrate;

i.e., increasing splat areas were observed when the substrate temperature was increased

since there was a greater thermal flux generated by the torch. Disc-like splats, exhibiting



“fried-egg” characteristics with little or no splashing, were found under SEM

observation. The “fried-egg” phenomenon that primarily occurs for polymer splats is

believed to be caused by the large radial difference in the flow properties of the molten

PP droplet surface. There was a low viscosity fully melted rim layer, with the unmelted

or still highly viscous core that did not spread upon impact; an observation documented

by others for polymer splats [11-12, 72].

8.3.3 Splat Equivalent Diameters

One-hundred splats were assessed with regard to their equivalent diameters and

ranked within size ranges, without splashes, for each SOD, Fig. 8-7. A wide spread of

splat sizes was demonstrated for all SODs. The lower and upper limits of the diameter

sizes were established; these ranges increase with the SOD. A maximum splat size of

220 µm was found for an SOD of 250 mm.

Figure 8-8 illustrates the second step in the analysis, where the splash portions of

the splat are included. A similar trend was observed, with a maximum splat size of

250 µm obtained for the 250 mm SOD. Figure 8-9 shows the average splat sizes for

each SOD and each analysis stage. Table 8-2 provides details of lower and upper limits

for all splat EDs and SODs.

Fig. 8-7. Comparison of particle and splat size distributions analysed without splashes, for different SODs.



Fig. 8-8. Comparison of particle and splat size distributions, analysed with splash features included, for different SODs.

Fig. 8-9. Comparison between the average splat particle size with and without splash features for different SODs.

Table 8-2. Equivalent diameter of splats* with and without splash features for different stand-off distances.

Equivalent dia. of splats without splashes

Equivalent dia. of splats with splashes

Stand-off distance (mm)

Lower dia. (µm)

Upper dia. (µm)

Average dia. (µm)

Lower dia. (µm)

Upper dia. (µm)

Average dia. (µm)

100 68.9 158.6 116.3 73.9 179.1 122.2

150 75.0 181.7 120.3 81.6 204.0 127.2

200 85.9 197.6 133.6 90.3 197.5 153.4

250 90.9 216.4 142.4 117.7 264.3 180.7 * 100 splats were considered for each measurement at each stand-off distance. Measurements were made twice, first without splash features, and second with splash features included.



8.3.4 Degree of Splashing (DS) The degree of splashing, DS, could offer an alternate method for identifying and

optimising the SOD to produce circular splats. As splashing decreases so does DS, with

a DS of 1 indicating an ideal circular splat. The DS of flattened particles was used to

compare splat morphologies obtained under different spray conditions. The DS

presented in Fig. 8-10 shows that the PP splats were not circular, with values ranging

from 1.6 for the smallest SOD of 100 mm to 2.5 for an SOD of 250 mm, which

demonstrates the dependence of DS on SOD. The values obtained here were lower than

those achieved by other researchers when measuring DS as a function of preheat

temperature [148]. This behaviour arises because the polymer splat thickness was

typically much greater than those of metal and ceramic splats. The splash features of

polymeric material were produced as the first layer of polymer contacts the substrate,

while the remaining bulk of the splat would solidify before reaching the splat edge. The

low thermal conductivity of polymers controls the interplay between spreading and

freezing of splats.

Fig. 8-10. Relationship between the degree of splashing and SOD.

The anticipated particle behaviour during impact on a low temperature substrate,

which gives rise to a rapid solidification rate, was the formation of a solid ring that

restricted materials from spreading freely. Due to the high inertial energy of the droplet,

the upper portion of the liquid droplet escaped beyond the solid ring, resulting in the

formation of fingers around the splat periphery. Generally, splat morphology was

dictated by the heat transfer between the splat and the substrate. Droplets with a low



melting point, such as polymers, were observed to solidify quickly during impact and,

therefore, trigger freezing-induced splashing [125]. Thus, the more rapid the heat was

conducted from the splat to the substrate, then the higher the solidification rate within

the droplet.

The rate of solidification depends on many factors, including thermal contact

resistance at the splat-substrate interface, substrate temperature, substrate material, and

droplet melting temperature [193]. A simulation [157] has shown that a molten tin

droplet impacting on a substrate at low temperature begins to freeze first around its

periphery, thereby obstructing liquid flow, and, when the solid rim becomes sufficiently

thick, this flow instability triggers splashing.

8.3.5 Circularity Circularity is a measure of the compactness of a shape. It is a value that allows

estimation of splat formation and how far the splat shape is from the ideal disc shape, as

well as examination of how the DS develops during impact at different SODs. A

circularity of 1 equates to a perfect circle and the closer to zero then the more likely that

splats are splashed, fingered or fragmented. All results of the circularity analysis

indicated some DS for both calculations of splats; i.e., with and without splash features

as depicted in Fig. 8-1. However, the lowest SOD of 100 mm presented higher

circularity values than the largest SOD of 250 mm. It is proposed that there was

insufficient heating so that the partially melted splats did not contain adequate molten

material to cause significant splashing. This postulate follows the trend that the

occurrence of splashing is more likely due to the jetting of liquid from the upper part of

the flattening particle [124, 134].

Figure 8-11 shows the relationship between splat circularities and SOD, and

demonstrated that the shape of splats with splash features was less circular than splats

without splashes. No significant difference in circularity between SODs was evident for

splats without splash features with all values around 0.9, indicating that these splat

shapes were regular and almost circular. A slight decrease in circularity from 0.66 to

0.50 was observed for splats analysed with splash features.



Fig. 8-11. Relationship between average splats circularity and SOD, for analysis with and without splash features.

8.3.6 Feret Diameter The effect of SOD variation on splat feret diameters followed the same curvature

tendency as for splat ED; i.e., increasing the SOD yielded an increase in the feret

diameters for both measurements of splats with and without splashes, across the 100

particles that were measured. Increasing the SOD from 100 to 150 mm produced a

slight increase in the Df, while a larger difference was observed for a SOD change from

150 to 200 mm, and greater again for the next increment in SOD to 250 mm. The

measurement of Df, therefore, demonstrated the tendency for splats to splash.

Measuring the splats without splash features is associated with small and steady

Df increases, resulting in a nearly linear relationship. Figure 8-12 shows the relationship

between the splat feret diameter and SOD. The ferret diameter was larger than the splat

diameter for all measured splats with and without splash features, with the same

curvature tendencies.

It can be concluded that a larger feret diameter was associated with a larger

difference between the maximum and minimum feret diameters. This indicated that the

splats were not circular, but more of an elliptical shape. A small difference in the splat

equivalent and feret diameters indicated that splats were more circular, as shown in

Table 8-3. These results were due to the variation and irregularities in the powder



particle shape with aspect ratios shown in Fig. 4-3. Another possibility was that the

spray direction, as the spray torch was traversed from one side to the other, tended to

form an asymmetric splat that was larger in the spray direction than the orthogonal

direction. Figure 8-13 indicates how the definition of feret diameter was applied to a

thermal spray splat.

Fig. 8-12. Comparison of the feret diameter for splats with and without splash features at all SODs of 100 splats.

Fig. 8-13. Feret diameter definition applied to a thermal spray splat.

N = 100 for each measurement



Table 8-3. Comparison of the splat equivalent and feret diameter of 100 splats at each SOD with and without splash features.

SOD (mm)

Average splat diameter with

splash features (µm)

Average splat diameter

without splash features (µm)

Average splat feret diameter

with splash features (µm)

Average splat feret diameter without splash features (µm)

100 122.8 116.3 153.8 128.2

150 127.5 120.4 165.9 132

200 153.8 133.7 201.9 151

250 180.7 142.4 266.1 155.3

8.3.7 Splat Perimeter

The perimeter of one hundred splats was assessed under both classifications –

with splash features and without splash features for each SOD, Fig. 8-14. The smallest

splat perimeters were obtained for SOD of 100 and 150 mm, for results both with and

without splash features. Increasing the SOD produced a small increase in the splat

perimeters calculated for splats without splash features, with the smaller standard

deviations shown as error bars. Assessing splats with splash features shows increasing

SOD from 150 to 200 and then to 250 mm also incurs significant increases in the splat

perimeter with larger error bars. Thus, as the SOD increased, there was a larger

variation and inconsistency in splat perimeter as a result of larger splash features. The

results revealed that there was no big difference in splat perimeter when splash features

were not considered; i.e., splash features played a major role in increasing the splat

perimeter, especially for larger SODs of 200 and 250 mm.

Fig. 8-14. Comparison of the splat perimeter with and without splash features at all SODs averaged over 100 splats.



8.3.8 Spreading Factor The rate of droplet spreading was quantified by measuring the splat diameter at

successive stages during droplet deformation; from the initial droplet diameter to the

maximum splat diameter. A nearly linear relationship exists between the spreading

factor and the SOD across the range used in this study, Fig. 8-15, with the smallest SOD

of 100 mm having the lowest spreading factor value. The splat with splash feature

analysis of 100 particles did not show a significant increase in the droplet spreading on

increasing SOD from 100 to 150 mm. However, a large increase in the splash area

occurred when increasing the SOD from 200 to 250 mm. Calculation of the splash areas

for 100 splats can be achieved by subtraction of the sum of the areas of 100 splats

without splash features from the sum of the area of 100 splats with splash features. The

splash area of 100 splats reached 1 mm2 at the 250 mm SOD, while it is less than

0.1 mm2 for the SOD of 100 mm, as shown in Fig. 8-16.

Fig. 8-15. Relationship between the splat spreading factor and SOD.

Fig. 8-16. Comparison of the spreading factors for splats with and without splash features at all SODs, including calculation of the splash area for 100 splats.




8.3.9 Deposition Efficiency (DE) DE can be determined by calculating the number of splats with consideration of

the different splat size. The significant effect of the SOD on the number of splats

deposited on the surface was evident from an analysis of a 5 mm2 area of the sprayed

glass slide, Fig. 8-17. The number of splats decreased proportionally to the increase in

SOD distance between 100 mm and 200 mm, while there was a larger drop in the

number of splats seen when the SOD reached 250 mm. This decline in the number of

drops impacted orthogonally could be explained by deviations in the particle trajectory

over longer distances, with polymer particle decomposition due to the excessive heat

that the in-flight particle experiences because of the increased dwell time.

Fig. 8-17. Relationship between the number of splats per unit area (5 mm2) and SOD.

8.3.10 Splat Thickness

Splat thickness can be calculated from the flattening ratio by applying a

geometric equation, defined by Madejski’s relationship [200], which links the diameter

of an impacting particle with the diameter and thickness of splats. The effect of surface

tension was considered as negligible towards the end of the spreading process. It was

assumed that a solidified splat was a thin cylindrical disc with a volume equal to that of

the initially spherical droplet [135]. Splat thicknesses (h) were calculated according to

equation 8-4.

Equation 8-4



The terms “d” and “D” are the diameters of the in-flight particle and splat,

respectively. The variation of the splat thickness with the SOD was investigated. As

Fig. 8-18 illustrates, thicker splats were observed at shorter SODs. It was noted that the

average splat thickness ranged from 15 to 35 µm, equivalent to 1 / 5 to 1 / 8 of the

droplet diameter. The results confirmed the visual observations and diagnostic data that

showed the highest spreading factor occurs for the highest SOD; explained by particle

dwell time and a higher particle temperature and velocity. The splat thickness trend was

opposite to that observed for the diameters because the volume of the particle and splat

were identical except in instances when the particle was overheated or evaporated at

longer SODs. This relationship; i.e., smaller diameters with increased thickness, is

illustrated in Fig. 8-19. The splat thickness measurement was performed using 2D

profilometry. The results agreed with the outcomes of statistical calculations for PP

splat profiles and thicknesses as stated earlier (refer to Section 4.3.2)

Fig. 8-18. Relationship between splat thickness and SOD.

Fig. 8-19. Relationship between splat size and splat thickness for different SOD.




8.3.11 Effect of Particle Size on Deposition Efficiency Several factors influenced the deposition efficiency. DE measured in this study

was not the ultimate quantity, which is normally calculated in terms of powder weight

input (used in the powder feeder) and the powder weight output (in terms of coating).

DE in this work was characterised by the “relative DE” which measures the number of

splats and calculates the relative volumes at each specific spray distance.

Different splats diameters were calculated to determine the DE, in order to

evaluate the relationship between SOD and the number of splats. The analysis was

conducted for a 125 mm2 area of the sprayed glass slide. Generally, for the two particle

sizes, the number of splats decreased proportionally to an increase in SOD between 100

and 200 mm. Increasing SOD to more than 200 mm enabled a larger drop in splat

numbers for both particle sizes. Figure 8-20 illustrates the variations in the number of

splats for the two mean particle sizes as a function of SOD. The reason for this decline

in the number of splats is explained in Section 8.3.9. A slight difference was noted in

the number of splats counted between both sizes: more splats were present for the larger

particle size than the smaller particle size.

Fig. 8-20. Comparison of number of splats for two mean particle sizes at different SOD of 100 splats.

The number of splats for different particle sizes, and thus the relative deposition

efficiency was greater for larger particle sizes as shown in Table 8-4. The relative

deposition efficiency decreased as SOD increased. The relative DE was greater and



nearly the same for SODs of 100 and 150 mm, but smaller and almost similar for 200

and 250 mm. The particle momentum dissipates and the velocity decreases as SOD

increases. Table 8-4 documents the number of splats for the two mean particle sizes.

The relative deposition efficiency of the smaller to larger feedstock particle size was

calculated using the volume of the spherical particle shape and the mean particle size.

Table 8-4. Number of splats for different particle sizes and the relative deposition efficiency between them.

SOD (mm)

Number of splats obtained using larger feedstocks with mean particle size =125 µm

Number of splats obtained using smaller feedstocks with mean particle size =75 µm

Relative deposition efficiency of smaller to larger particle size

100 427 400 20%

150 357 340 20%

200 270 220 17%

250 63 50 17%

300 24 20 18%

The acceleration of each particle depends on its size and momentum during the

thermal spray process. The larger particles have higher momentum and higher velocity

than the smaller particle size. Thus, the particle size is proportionally related to the final

particle velocity on impact, and a large powder particle size distribution leads to a large

velocity distribution. This is not the case for heavier powders such as metals or

ceramics, where the smaller the particle then the higher the velocity.

As particle size increases, the number of splats increased, owing to a decrease in

velocity. At different SODs it is necessary for particles to achieve a velocity higher than

the critical velocity to improve the deposition efficiency. The velocity determines

whether deposition of the particle or erosion of the substrate occurs on impact of a spray

particle [201]. Another possibility why larger particles have lower deposition efficiency

could relate to particle flow rate. Polymer powders have an atomic structure that enables

electrostatic attraction to be induced between the polymer particles and the powder

carrier tube. Thus, larger particles will have a greater surface area and give rise to a



greater electrostatic force of attraction within the powder delivery tube than small

particles. The net result will be that large particles will flow at a lower rate than small

particles for the same carrier gas flow rates.

Smaller particles have less momentum to enter the flame. Powder particles were

injected orthogonally to the flame centreline at its nozzle exit. The high velocity of the

flame prevents polymer particles entering the flame owing to the low density of the

polymer. Small particles might be driven to the outer regions of the flame downstream

when they interact with the flame, Fig. 8-21. This condition is more relevant to the

external powder feeding system used in this current study.

Fig. 8-21. Schematic of small and large polymer particles fed externally into the flame. The inset figure shows a small particle ‘bouncing off’ the flame.

The number of splats deposited for each particle size could be a function of

deposition efficiencies relative to the SODs. The number of splats collected at each

SOD relative to the maximum splat number collected enables a relative deposition

efficiency calculation. Considering the highest deposition efficiency for a SOD of

100 mm, the number of splats was 400 and 427 for mean particle sizes of 75 µm and

125 µm, respectively. Table 8-5 shows the results, with no significant difference evident

between the two particle sizes. The effect of increasing the SOD, for example from 100

to 150 mm, will reduce the efficiency to around 15%, while increasing SOD to 300 mm

reduces the efficiency by 95%.



Table 8-5. Relative deposition efficiency (achieved at lower SOD of 100 mm) for two different mean particle sizes.

SOD (mm)

Deposition efficiency using feedstocks with mean particle size =125 µm relative to SOD =100 mm

Deposition efficiency using feedstocks with mean particle size =75 µm relative to SOD =100 mm

100 -- -- 150 84% 85% 200 63% 55% 250 15% 12% 300 5% 5%

8.4 Conclusions

The shape of feedstock and impacted particles was measured by employing

software tools in conjunction with statistical analysis. The complexities of the particle

shape were analysed using splat metrics; e.g., particle area; splat area and diameter;

degree of splashing; deposition efficiency; circularity; spread factor; feret diameter and

splat thickness. The results indicate that as the stand-off distance was increased from

100 to 250 mm; then the splat size, degree of splashing, splat perimeter, feret diameter,

and spread factor increased, while the deposition efficiency and splat thickness

decreased. Splat circularity was steady around 0.9 indicating that the splats analysed

were close to circular for all stand-off distances and for splats calculated without splash

features. Splat circularity for the actual size with splash features was lower. Table 8-6

summarises the results obtained for all splat metrics for the actual size of splats (with

splash features).

Table 8-6. Results summary of all splat metrics relative to stand-off distance increase.

SOD (mm) 100 150 200 250 Increasing SOD

Splat size (µm) 122.8 127.5 153.8 180.7 Increases

Degree of splashing 1.6 1.7 2.3 2.5 Increases

Feret diameter (µm) 153.8 165.9 201.9 266.1 Increases

Splat perimeter (µm) 508.7 533.2 756.1 904.7 Increases

Spread factor 1.92 2.18 2.43 2.64 Increases

Splat count or (DE) 427 357 270 63 Decreases

Splat thickness (µm) 35.9 33.3 22.9 16.6 Decreases

Circularity 0.66 0.6 0.5 0.5 Decreases

Chapter 9: Conclusions, Recommendations and Future work


Chapter 9

9 Conclusions, Recommendations and Future

Work

9.1 Conclusions

The flame spray process was a practical method of producing polypropylene

splats using oxy-acetylene fuel. The process was implemented under different spraying

and substrate conditions. Splats were produced on the flat surfaces of glass slides and

roughened surfaces of mild steel substrates by spraying at different stand-off distances.

Splats were also produced on flat and rough surfaces of mild steel substrates preheated

to different temperatures. The feedstock properties and thermal spray parameters were

demonstrated to have an important role in splat morphology. “Fried-egg” shaped splats

with a large viscous central core and a fully melted wide rim with a thin edge were a

frequent observation for the polypropylene thermally sprayed splat.

The present research has substantiated a relationship between the spray

parameters and the microstructural characteristics of splats that significantly affect the

overall coating. Numerous intrinsic and extrinsic thermal spray parameters (Fig. 2-5)

enable alteration of the splat microstructure and topology, and thus, the coating

properties. Stand-off distance, substrate topology and substrate temperature have been

specified as extrinsic conditions that affect the splat formation. Experiments were

conducted and analysed to demonstrate the effect of the various parameters on splat

formation. The influence of parameter variation on splat formation was examined.



9.2 Recommendations

Based on the knowledge obtained from the literature and on the experience

acquired through conducting these experiments, the characteristics of thermal spray

coatings manufactured using polymer feedstocks can be improved in several ways.

These are listed below.

1. Development of a special polymer grade, manufactured specifically for thermal

spray use, would enable advancement in thermal spray polymer applications.

There is the potential for using recycled polymers. Such standardization in

polymer feedstock would reduce the current trial and error process that is

employed for selection of an appropriate polymer chemistry.

2. Future improvements in polymer manufacturing technology should consider

reducing polymer particle size variations, which affect the melting points of the

polymer feedstocks that influence the coating microstructure and characteristics.

3. Particle flowability is an important factor to be considered in a polymer coating:

the more spherical the powder morphology then the better the powder flow

characteristics.

4. Robotic manipulation of thermal spray equipment to suit specific polymer

applications would be highly beneficial.

5. A torch designed exclusively for spraying the relatively low cost and widely

available polymers such as polyethylene or polypropylene would promote

applications for many coating industries. Examples include torch design in the

combustion processes, especially the flame process, which needs to be well suited

for low melting point polymers to avoid thermal degradation while melting the

polymer sufficiently. Using a shroud to cool the centre of the flame or feeding the

polymer powder externally could also reduce the effect of the high flame core

temperature.

6. Understanding the polymer particle behaviour, including the rheology during

melting and solidification, will enable more complete understanding of the

requirements and the specifications that need to be modified to achieve high-

quality polymer-based coatings.



9.3 Future work

The ultimate aim of this splat research and analysis was to create an alternative

coating technology using polymeric thermal spray that can serve many industries.

Understanding the basic building blocks of a coating (the splat) is the initial stage in

building a relationship between the processing conditions and the microstructural

coating properties.

The results and conclusions of this work have contributed to the area of polymer

science and thermal spray technology. The technological parameters must be optimised;

however, they are not sufficient enough to produce the promise of a coating that can replace

paint, for example.

Additional future research and development are essential. These recommendations

include optimisation of important variables as follows:

Feedstock particle size and particle size distribution.

Gas mixture and flow rate.

Post processing treatments.

Delivery rate of the feedstock.

Carrier gas type and flow rate.

Coating onto a bond layer.

Polypropylene composites.

The next step would be to produce an industrial-sized coating using the

optimised parameters, and then conduct tests based on the proposed applications. This

examination would include:

Different weathering conditions.

Scratch tests.

Corrosion resistance tests.

Adhesion and peel-off tests.

Measurement of crystallinity and other classical materials science studies as

documented in this thesis.



The aim and focus of this follow up, industrial-based research would be to link the

structure of many thousands of splats to the properties and performance of the

consolidated thermal sprayed polymeric coatings.

Chapter 10: References


Chapter 10

10 References

1. J. J. Licari and L. A. Hughes, Handbook of Polymer Coatings for Electronics - Chemistry, Technology and Applications (2nd Edition). 1990, William Andrew Publishing / Noyes: Park Ridge, New York, USA.

2. J. A. Brogan, Thermal-spraying of polymers and polymer blends, MRS Bulletin, 2000, 25(7): p. 48-53.

3. J. R. Davis, Handbook of Thermal Spray Technology, 2004, USA: ASM International: Materials Park, OH, USA.

4. J. R. T. Branco, S. R. V. Campos, L. T. Duarte, and V. De Freitas Cunha Lins, Combustion-thermal-sprayed recycled poly(ethylene terephthalate), Journal of Applied Polymer Science, 2004, 92(5): p. 3159-3166.

5. H. G. Karian, Handbook of Polypropylene and Polypropylene Composites 2003, Michigan, U.S.A.: Marcel Dekker Inc

6. A. A. Tracton, Coatings Materials and Surface Coatings 2006: CRC Press 532.

7. J. Brogan, C. C. Berndt, S. Sampath, H. Herman, and S. A. Drozdz, Development of Recycled Polymer Blends for Thermal Spray Applications, A.C.o. Engineers, Editor. 1997, (CPAR) PROGRAM: US.

8. V. Landry, B. Riedl, and P. Blanchet, Alumina and zirconia acrylate nanocomposites coatings for wood flooring: Photocalorimetric characterization, Progress in Organic Coatings, 2008, 61(1): p. 76-82.

9. C. S. Jeffcoate, T. L. Wocken, and G. P. Bierwagen, Electrochemical assessment of spray-applied thermoplastic coating barrier properties, Journal of Materials Engineering and Performance, 1997, 6(4): p. 417-420.

10. P. Fauchais, M. Fukumoto, A. Vardelle, and M. Vardelle, Knowledge concerning splat formation: An invited review, Journal of Thermal Spray Technology, 2004, 13(3): p. 337-360.

11. M. Ivosevic, R. A. Cairncross, and R. Knight, 3D predictions of thermally sprayed polymer splats: Modeling particle acceleration, heating and deformation on impact with a flat substrate, International Journal of Heat and Mass Transfer, 2006, 49(19-20): p. 3285-3297.



12. M. Ivosevic, V. Gupta, J. A. Baldoni, R. A. Cairncross, T. E. Twardowski, and R. Knight, Effect of substrate roughness on splatting behavior of HVOF sprayed polymer particles: Modeling and experiments, Journal of Thermal Spray Technology, 2006, 15(4): p. 725-730.

13. M. Ivosevic, R. Knight, S. R. Kalidindi, G. R. Palmese, A. Tsurikov, and J. R. Sutter, Optimal substrate preheating model for thermal spray deposition of thermosets onto polymer matrix composites, in International thermal spray conference and exposition. 2003, Glenn Research Center: Orlando, Florida.

14. B. P. Withy, The effect of substrate parameters on the morphology of thermal sprayed PEEK splats, in Engineering. 2008 University of Auckland: New zealand.

15. R. F. Bunshah, Handbook of hard coatings, Deposition Technologies, Properties and Applications, 2001, Norwich, New York, U.S.A.: Noyes Publications / William Andrew Publishing, LLC.

16. M. Sexsmith and T. Troczynski, Peel adhesion test for thermal spray coatings, Journal of Thermal Spray Technology, 1994, 3(4): p. 404-411.

17. L. Pawlowski, The Science and Engineering of Thermal Spray Coatings, Second Edition ed, ed. L. John Wiley & Sons, 2008 England: John Wiley & Sons, Ltd. , 647.

18. E. Petrovicova and L. S. Schadler, Thermal spraying of polymers, International Materials Reviews, 2002, 47(4): p. 169-190.

19. N. O. Eddy, P. Ekwumemgbo, and S. A. Odoemelam, Inhibition of the corrosion of mild steel in H2SO4 by 5- amino-1-cyclopropyl-7-[(3R, 5S) 3, 5-dimethylpiperazin- 1-YL] -6, 8-difluoro-4-oxo-uinoline-3-carboxylic acid (ACPDQC), International Journal of Physical Sciences, 2008, 3(11): p. 275-280.

20. A. Ramniceanu, Investigation of parameters governing the corrosion protection efficacy of fusion bonded epoxy coatings, in Civil and Environmental Engineering. 2007, Virginia Polytechnic Institute and State University: Virginia, USA. p. 129.

21. J. Matejicek, P. Chraska, and J. Linke, Thermal spray coatings for fusion applications - Review, Journal of Thermal Spray Technology, 2007, 16(1): p. 64-83.

22. S. K. Singh, S. P. Tambe, V. S. Raja, and D. Kumar, Thermally sprayable polyethylene coatings for marine environment, Progress in Organic Coatings, 2007, 60(3): p. 186-193.

23. JA Brogan, CC Berndt, A Claudon, and C. Coddet, The mechanical properties of combustion-sprayed polymers and blends, Thermal spray practical solution for engineering problems 1996, Cincinnati, OH, USA, Pub. ASM International, Materials Park, OH, USA.



24. P. Fauchais, A. Vardelle, and B. Dussoubs, Quo Vadis thermal spraying?, Journal of Thermal Spray Technology, 2001, 10(1): p. 44-66.

25. P. Fauchais and M. Vardelle, Plasma spraying: Present and future, Pure and Applied Chemistry, 1994, 66(6): p. 1247-1258.

26. P. Fauchais, G. Montavon, M. Vardelle, and J. Cedelle, Developments in direct current plasma spraying, Surface and Coatings Technology, 2006, 201(5): p. 1908-1921.

27. M. Ivosevic, Splat Formation during Thermal Spraying of Polymer Particles: mathematical modeling and experimental analysis. 2006, Drexel University: Philadelphia, USA.

28. C. C. Berndt and G. Montavon, Thermal spray: Preserving 100 years of technology, Journal of Thermal Spray Technology, 2006, 15(1): p. 5-8.

29. M. U. Schoop, Device for coating articles with glass enamel, quartz, and metals - JTST historical patent #19, Journal of Thermal Spray Technology, 2001, 10(2): p. 244-245.

30. W. J. Trompetter, Splat-substrate interactions in high velocity thermal spray coatings. 2007, University of Auckland: Auckland. p. 234.

31. V. Gupta, Thermal Spraying of Polymer-Ceramic Composite Coatings with Multiple Size Scales of Reinforcements, in Master of science thesis. 2006, Drexel University: Philadelphia, USA.

32. S. T. Sidhu, S. Prakash, and R. D. Agrawal, Studies on the properties of high velocity oxy fuel thermal spray coatings for higher temperature applications. A review, Fiziko-Khimicheskaya Mekhanika Materialov, 2005, 41(6): p. 80-95.

33. J. R. Davis, Surface Engineering for Corrosion and Wear Resistance. 2001, Woodhead Publishing. pp. 1-10.

34. S. M. Vyawahare, Protective thermal spray coating for polymer matrix composites, in Departement of Mechanical Engineering. 2006, Wichita State University: India. p. 99.

35. M. L. Berndt and C. C. Berndt, Thermal Spray Coatings, Corrosion: Fundamentals, Testing, and Protection, in Advancing the Science & Applying the Technology M. B. and M. C., Editors. 2003 ASM International, Materials Park, OH, USA: Orlando, Florida, USA. p. 803-813.

36. L. Jackson, M. Ivosevic, R. Knight, and R. A. Cairncross, Sliding wear properties of HVOF thermally sprayed nylon-11 and nylon-11 / ceramic composites on steel, Journal of Thermal Spray Technology, 2007, 16(5-6): p. 927-932.



37. M. Ivosevic, R. A. Cairncross, and R. Knight, Melting and degradation of nylon-11 particles during HVOF combustion spraying, Journal of Applied Polymer Science, 2007, 105(2): p. 827-837.

38. K. Y. Tufa and F. Gitzhofer, DC plasma sprayed polymer composite coatings for abrasion resistant protective surfaces, Meeting the Challenges of the 21st Century, 1998, Nice, France, Pub. ASM International, Materials Park, OH, USA.

39. J. Wu, P. R. Munroe, B. Withy, and M. M. Hyland, Study of the splat-substrate interface for a PEEK coating plasma-sprayed onto aluminum substrates, Journal of Thermal Spray Technology, 2010, 19(1-2): p. 42-48.

40. J. Li, H. Liao, and C. Coddet, Friction and wear behavior of flame-sprayed PEEK coatings, Wear, 2002, 252(9-10): p. 824-831.

41. G. Zhang, H. Liao, M. Cherigui, J. Paulo Davim, and C. Coddet, Effect of crystalline structure on the hardness and interfacial adherence of flame sprayed poly(ether-ether-ketone) coatings, European Polymer Journal, 2007, 43(3): p. 1077-1082.

42. C. Mateus, S. Costil, and C. Coddet, Ceramic / fluoropolymer composite coatings by thermal spraying: effects of parameters, Thermal spray: Advancing the science and appling the technology, 2003, Orlando, Florida, USA, Pub. ASM International, Materials Park, Ohio, USA.

43. J. A. Brogan and C. C. Berndt, The coalescence of combustion-sprayed ethylene-methacrylic acid copolymer, Journal of Materials Science, 1997, 32(8): p. 2099-2106.

44. J. J. Licari, Coating Materials for Electronic Applications - Polymers, Processes, Reliability, Testing. 2003, William Andrew Publishing / Noyes.

45. D. D. B. Jung, Characterisation of Poly (ethylene naphthalate) based polymer blends, in Department of Mechanical Engineering. 2002, The University of Auckland: Auckland, New Zealand. p. 212.

46. J. McCook and B. Hyllberg, Industrial Applications for Polymer Coatings with Release Properties in Combination with Thermal Spray Coatings, Expanding Thermal Spray Performance to New Markets and Applications, 2009, Las Vegas, Nevada, USA, Pub. ASM International, Materials Park, OH, USA.

47. Y. Bao, T. Zhang, and D. T. Gawne, Influence of composition and process parameters on the thermal spray deposition of UHMWPE coatings, Journal of Materials Science, 2005, 40(1): p. 77-85.

48. R.S. Lima, A.S. Takimi, M.D. Lima, and C. P. Bergmann, Production of recycled polyethylene terephtalate coatings by HVOF and plasma spray, Thermal spray; A united forum for scientific and technological advances, 1997: p. 215-221.



49. V. F. C. Lins, J. R. T. Branco, F. R. C. Diniz, J. C. Brogan, and C. C. Berndt, Erosion behavior of thermal sprayed, recycled polymer and ethylene-methacrylic acid composite coatings, Wear, 2007, 262(3-4): p. 274-281.

50. V. F. C. Lins, F. M. Cury, and R. Moreira, Infrared and ultraviolet-visible spectroscopy study of the degradation of polyester and polyester / ethylene methyl acrylate copolymer blend coatings on steel, Journal of Applied Polymer Science, 2008, 109(4): p. 2103-2112.

51. C. C. Berndt and C. K. Lin, Measurement of adhesion for thermally sprayed materials, Journal of Adhesion Science and Technology, 1993, 7(12 pt 2): p. 1235-1264.

52. Y. Borisov, I. Sviridova, V. Korzhik, and N. Fialko, Analysis of deposition process for composite polymer thermally sprayed coatings., Thermal spray: Advancing the science and appling the technology, Ed. by C. Moreau and B. Marple, Published by ASM International , Materials Park, Ohio, USA, 2003: p. 1651-1657.

53. G. Zhang, H. Liao, H. Yu, V. Ji, W. Huang, S. G. Mhaisalkar, and C. Coddet, Correlation of crystallization behavior and mechanical properties of thermal sprayed PEEK coating, Surface and Coatings Technology, 2006, 200(24): p. 6690-6695.

54. E. M. Leivo, T. H. Kellberg, M. J. Kolari, P. M. J. Vuoristo, and T. A. MÃntylÃ, Structure and Thermal Conductivity of Flame Sprayed Polyamide 11 Coatings Filled with Al2O3, AlN and BN Ceramics, New Surfaces for a New Millennium, 2001, Singapore, Pub. ASM International, Materials Park, OH, USA.

55. M. Ivosevic, S. L. Coguill, and S. L. Galbraith, Polymer thermal spraying: A novel coating process, Expanding Thermal Spray Performance to New Markets and Applications, 2009, Las Vegas, NV, USA, Pub. DVS-ASM.

56. R. Winkler, F. Bultmann, S. Hartmann, and A. Jerz, Thermal spraying of polymers: spraying processes, material and new trends, Thermal spray: Advancing the science and appling the technology, 2003, Materials Park, Ohio, USA, Pub. ASM International

57. F. Y. Yan, K. A. Gross, G. P. Simon, and C. C. Berndt, Peel-strength behavior of bilayer thermal-sprayed polymer coatings, Journal of Applied Polymer Science, 2003, 88(1): p. 214-226.

58. M. Yamane, T. A. Stolarski, and S. Tobe, Influence of counter material on friction and wear performance of PTFE-metal binary coatings, Tribology International, 2008, 41(4): p. 269-281.



59. G. V. M. Anjos, T. H. D. Sydenstricker, R. S. C.Paredes, and S. C. Amico, A Preliminary Investigation into Thermal Spray and Other Metal / Polymer Deposition Processes and their Potential Use in the Oil Industry, Engenharia Térmica, No 4, 2003 p. 3-8.

60. S. P. Tambe, S. K. Singh, M. Patri, and D. Kumar, Ethylene vinyl acetate and ethylene vinyl alcohol copolymer for thermal spray coating application, Progress in Organic Coatings, 2008, 62(4): p. 382-386.

61. B. Normand, H. Takenouti, M. Keddam, H. Liao, G. Monteil, and C. Coddet, Electrochemical impedance spectroscopy and dielectric properties of polymer: Application to PEEK thermally sprayed coating, Electrochimica Acta, 2004, 49(17-18): p. 2981-2986.

62. G. Zhang, W. Y. Li, M. Cherigui, C. Zhang, H. Liao, J. M. Bordes, and C. Coddet, Structures and tribological performances of PEEK (poly-ether-ether-ketone)-based coatings designed for tribological application, Progress in Organic Coatings, 2007, 60(1): p. 39-44.

63. T. Palathai, J. Tharajak, and N. Sombatsompop, Hardness, adhesion index and microstructure of PEEK coating on Al or Fe substrate by LVOF flame spray, Materials Science and Engineering: A, 2008, 485(1-2): p. 66-73.

64. M. Öksüz and H. Yildirim, Structure and properties of flame-sprayed poly(ethylene-co-vinyl alcohol) copolymer coatings, Journal of Applied Polymer Science, 2004, 94(4): p. 1357-1364.

65. L. T. Duarte, E. M. Paulae Silva, J. R. T. Branco, and V. F. C. Lins, Production and characterization of thermally sprayed polyethylene terephthalate coatings, Surface and Coatings Technology, 2004, 182(2-3): p. 261-267.

66. S. K. Singh, S. P. Tambe, G. Gunasekaran, V. S. Raja, and D. Kumar, Electrochemical impedance study of thermally sprayable polyethylene coatings, Corrosion Science, 2009, 51(3): p. 595-601.

67. Y. Borisov, V. Korzhik, I. Sviridova, and A. Skorokhod, Influence of a filler composition and conditions of flame spraying on a structure and mechanical properties of composite polymer coatings, United forum for scientific and technological advances, 1997: p. 239-242.

68. T. Zhang, D. T. Gawne, and Y. Bao, The influence of process parameters on the degradation of thermally sprayed polymer coatings, Surface and Coatings Technology, 1997, 96(2-3): p. 337-344.

69. J. Khedkar, I. Negulescu, and E. I. Meletis, Sliding wear behavior of PTFE composites, Wear, 2002, 252(5-6): p. 361-369.

70. F. Li, F. Y. Yan, L. G. Yu, and W. M. Liu, The tribological behaviors of copper-coated graphite filled PTFE composites, Wear, 2000, 237(1): p. 33-38.



71. R. Kawase and A. Nakano, Production of heat and corrosion-resistant plastic coatings, Thermal spray practical solution for engineering problems 1996, Material Park, Ohio-USA, Pub. ASM International, Materials Park, OH, USA.

72. B. P. Withy, M. M. Hyland, and B. J. James, The Effect of Surface Chemistry and Morphology on the Properties of HVAF PEEK Single Splats, Journal of Thermal Spray Technology, 2008, 17(5-6): p. 631-636.

73. D. Horton and P. J. Loustaunau, Advancements in Application of Thermoplastic Powder Coatings for Railcar Linings, Corrosion, 674, 1996.

74. S. E. McElwain, T. A. Blanchet, L. S. Schadler, and W. G. Sawyer, Effect of particle size on the wear resistance of alumina-filled PTFE micro- and nanocomposites, Tribology Transactions, 2008, 51(3): p. 247-253.

75. H. Unal, U. Sen, and A. Mimaroglu, An approach to friction and wear properties of polytetrafluoroethylene composite, Materials and Design, 2006, 27(8): p. 694-699.

76. G. K. Sweet, Applying thermoplastic / thermoset polymer with a modified plasma system, National Thermal Spray Conference 1993, Anaheim, CA, Pub. ASM International, Materials Park, OH, USA.

77. E. Rajamaki, M. Leino, P. Vuoristo, P. Jarvela, and T. Mantyla, Effect of Powder Properties such as Particle Size, Density and Melt Flow Rate on the Properties of Flame Sprayed PE Coating, Surface Engineering via Applied Research, 2000, Montreal, Canada, Pub. ASM International, Materials Park, OH, USA.

78. R. Knight, X. Fang, and T. E. Twardowski, Structure and Properties of HVOF Sprayed Amorphous Polymer Matrix Nanocomposite Coatings, Proceedings of the International Thermal Spray Conference, 2001, Singapore, Pub. ASM International, Materials Park, OH, USA.

79. S. Saber-Samandari and C. C. Berndt, IFTHSE Global 21: Heat treatment and surface engineering in the twenty-first century: Part 10 - Thermal spray coatings: A technology review, International Heat Treatment and Surface Engineering, 2010, 4(1): p. 7-13.

80. Y. Bao, D. T. Gawne, and T. Zhang, Effect of thermal spray parameters on the adhesion of polymer-based coating on steel Thermal Spray: A United forum for Scientific and Technological Advances, 1997, Indianapolis, IN, USA, Pub. ASM International, Materials Park, OH, USA.

81. C. Mateus, S. Costil, R. Bolot, and C. Coddet, Ceramic / fluoropolymer composite coatings by thermal spraying - A modification of surface properties, Surface and Coatings Technology, 2005, 191(1): p. 108-118.



82. D. J. Varacalle Jr, D. P. Guillen, D. M. Deason, W. Rhodaberger, and E. Sampson, Effect of grit-blasting on substrate roughness and coating adhesion, Journal of Thermal Spray Technology, 2006, 15(3): p. 348-355.

83. T. Huang and M. R. Kamal, Morphological modeling of polymer solidification, Polymer Engineering and Science, 2000, 40(8): p. 1796-1808.

84. K. A. Gross, F. Y. Yan, C. C. Berndt, and G. P. Simon, Repair of multi-layered, polymer-based thermal spray coatings, Thermal Spray connects: Explore its surfacing potential, 2005, Basel, Switzerland, Pub. DVS-ASM.

85. J. D.J. Varacalle, KW. Couch, V.S. Budinger, Studies of the flame spraying of polymers, Thermal spray: practical solutions for engineering problens, Published by ASM International Materials Park, Ohio, USA, 1996: p. 251-255.

86. Y. Bao and D. T. Gawne, Particle size selection for plasma deposition of low temperature materials, Thermal Spray Science & Technology, 1995, Houston, TX, USA, Pub. ASM International, Materials Park, OH, USA.

87. M. Ivosevic, R. Knight, S. R. Kalidindi, G. R. Palmese, and J. K. Sutter, Solid particle erosion resistance of thermally sprayed functionally graded coatings for polymer matrix composites, Surface and Coatings Technology, 2006, 200(16-17): p. 5145-5151.

88. S. Kalpakjian, Manufacturing Engineering and Technology, Vol. Third Ed, 1995, USA: Addison Wesley.

89. D. J. Flynn, Nalco Water Handbook (3rd Edition). 2009, McGraw-Hill.

90. G. Zhang, H. Liao, H. Li, C. Mateus, J. M. Bordes, and C. Coddet, On dry sliding friction and wear behaviour of PEEK and PEEK / SiC-composite coatings, Wear, 2006, 260(6): p. 594-600.

91. R. Winkler, F. Buitmann, S. Harmann, and A. Jerz, Thermal spraying of polymers: spraying processes, material and new trends, thermal spray 2003; Advancing the science & Appling the Technology, Published by ASM International Materials Park, Ohio, USA, 2003.

92. Gustavo Vicente Maximo Anjos, Thais H. D. Sydenstricker, Ramón S. Cortés Paredes, and S. C. Amico, Characterization of Al / Polymer Composite Coating for Mechanical and Adhesion Properties on Steel Advancing the Science & Applying the Technology 2003, Orlando, Florida, USA, Pub. ASM International, Materials Park, OH, USA.

93. W. Haas, U. Putzier, H. Haiser, and J. Putzier, Wear resistant coatings for radial shaft sealing systems, Meeting the Challenges of the 21st Century, 1998, Nice, France, Pub. ASM International, Materials Park, OH, USA.

94. S. Saber-Samandari and A. A. Khatibi, The effect of interphase on the elastic modulus of polymer based nanocomposites. 2006. p. 199-204.



95. S. Saber-Samandari, A. A. Khatibi, and D. Basic, An experimental study on clay / epoxy nanocomposites produced in a centrifuge, Composites Part B: Engineering, 2007, 38(1): p. 102-107.

96. S. Saber-Samandari and A. Afaghi-Khatibi, Evaluation of elastic modulus of polymer matrix nanocomposites, Polymer Composites, 2007, 28(3): p. 405-411.

97. Y. Dong, Multi-Scale Effects on Deformation Mechanisms of Polymer Nanocomposites: Experimental Characterisation and Numerical Study, in Department of Mechanical Engineering 2008, The University of Auckland: New Zealand. p. 271.

98. D. L. Burris and W. G. Sawyer, Improved wear resistance in alumina-PTFE nanocomposites with irregular shaped nanoparticles, Wear, 2006, 260(7-8): p. 915-918.

99. E. Petrovicova, R. Knight, L. S. Schadler, and T. E. Twardowski, Nylon 11 / silica nanocomposite coatings applied by the HVOF process. II. Mechanical and barrier properties, Journal of Applied Polymer Science, 2000, 78(13): p. 2272-2289.

100. T. Twardowski, M. Reilly, and R. Knight, Properties of HVOF Sprayed Multi-Scale Polymer / Silica Nanocomposite Coatings, Proceedings of the International Thermal Spray Conference, 2001, Singapore, Pub. ASM International , Materials Park, Ohio, USA.

101. Y. Li, Y. Ma, B. Xie, S. Cao, and Z. Wu, Dry friction and wear behavior of flame-sprayed polyamide1010 / n-SiO2 composite coatings, Wear, 2007, 262(9-10): p. 1232-1238.

102. D. T. G. Y. Bao, Plasma spraying of glass- polyamide composite coatings, Precedings of the national thermal spray conference, Anaheim, CA, 1993: p. 417-422.

103. F. Y. Yan, K. A. Gross, G. P. Simon, and C. C. Berndt, Mechanical and erosion properties of CaCO3-EMAA thermal sprayed coatings, Polymer Engineering and Science, 2004, 44(8): p. 1448-1459.

104. P. Bhimaraj, H. Yang, R. W. Siegel, and L. S. Schadler, Crystal nucleation and growth in polyethylene terephthalate / alumina- nanoparticle composites, Journal of Applied Polymer Science, 2007, 106(6): p. 4233-4240.

105. R. Kawase and Y. Tanaka, Sprayed Composite Coating of High Performance Thermoplastics, Proceedings of the International Thermal Spray Conference, 2001, Singapore, Pub. ASM International, Materials Park, OH, USA.

106. L. Sun, C. C. Berndt, and K. A. Gross, Characterization and mechanical properties of flame sprayed hydroxyapatite / polymer composite Ccatings, Proceedings of the International Thermal Spray Conference, 2001, Singapore, Pub. ASM International, Materials Park, OH, USA.



107. R. Gadow and D. Scherer, Composite coatings with dry lubrication ability on light metal substrates, Surface and Coatings Technology, 2002, 151-152: p. 471-477.

108. M. L. Allan, C. C. Berndt, and D. Otterson, Polymer Coatings for Corrosion Protection in Biochemical Treatment of Geothermal Residues, Geothermal Resources Council Transactions, 1998, 22.

109. D. J. Varacalle, Jr, D. P. Zeek, K. W. Couch, D. M. Benson, and S. M. Kirk, Flame spraying of polymers. thermal spray, Proceedings of the International Thermal Spray Conference, 1997, Indianapolis, IN, USA, Pub. ASM International , Materials Park, Ohio, USA.

110. H. Liao, E. Beche, F. Berger, and C. Coddet, On the microstructures of thermally sprayed "PEEK" polymer, Meeting the Challenges of the 21st Century, 1998, Nice, France, Pub. ASM International, Materials Park, Ohio, USA.

111. A. Fridman and Y. I. Cho, Transport Phenomena in Plasma-Advances in Heat Transfer, Volume 40, first ed, 2007, Amsterdam: Elsevier.

112. V. V. Sobolev and J. M. Guilemany, Flattening of droplets and formation of splats in thermal spraying: A review of recent work - Part 2, Journal of Thermal Spray Technology, 1999, 8(2): p. 301-304.

113. V. V. Sobolev, J. M. Guilemany, and A. J. Martín, Flattening of composite powder particles during thermal spraying, Journal of Thermal Spray Technology, 1997, 6(3): p. 353-360.

114. V. V. Sobolev and J. M. Guilemany, Formation of splats during thermal spraying of composite powder particles, Materials Letters, 2000, 42(1): p. 46-51.

115. M. Pasandideh-Fard, V. Pershin, S. Chandra, and J. Mostaghimi, Splat shapes in a thermal spray coating process: Simulations and experiments, Journal of Thermal Spray Technology, 2002, 11(2): p. 206-217.

116. C.-J. Li, H.-L. Liao, P. Gougeon, G. Montavon, and C. Coddet, Experimental determination of the relationship between flattening degree and Reynolds number for spray molten droplets, Surface and Coatings Technology, 2005, 191(2-3): p. 375-383.

117. H. Fukanuma, R. Huang, Y. Tanaka, and Y. Uesugi, Mathematical modeling and numerical simulation of splat cooling in plasma spray coatings, Journal of Thermal Spray Technology, 2009, 18(5-6): p. 965-974.

118. C. J. Li and J. L. Li, Transient contact pressure during flattening of thermal spray droplet and its effect on splat formation, Journal of Thermal Spray Technology, 2004, 13(2): p. 229-238.



119. J. Mostaghimi, M. Pasandideh-Fard, and S. Chandra, Dynamics of Splat Formation in Plasma Spray Coating Process, Plasma Chemistry and Plasma Processing, 2002, 22(1): p. 59-84.

120. M. Pasandideh-Fard, S. Chandra, and J. Mostaghimi, A three-dimensional model of droplet impact and solidification, International Journal of Heat and Mass Transfer, 2002, 45(11): p. 2229-2242.

121. H. B. Parizi, L. Rosenzweig, J. Mostaghimi, S. Chandra, T. Coyle, H. Salimi, L. Pershin, A. McDonald, and C. Moreau, Numerical simulation of droplet impact on patterned surfaces, Journal of Thermal Spray Technology, 2007, 16(5-6): p. 713-721.

122. L. Bianchi, A. C. Leger, M. Vardelle, A. Vardelle, and P. Fauchais, Splat formation and cooling of plasma-sprayed zirconia, Thin Solid Films, 1997, 305(1-2): p. 35-47.

123. M. Fukumoto and Y. Huang, Flattening mechanism in thermal sprayed nickel particle impinging on flat substrate surface, Journal of Thermal Spray Technology, 1999, 8(3): p. 427-432.

124. A. A. Syed, A. Denoirjean, B. Hannoyer, P. Fauchais, P. Denoirjean, A. A. Khan, and J. C. Labbe, Influence of substrate surface conditions on the plasma sprayed ceramic and metallic particles flattening, Surface and Coatings Technology, 2005, 200(7): p. 2317-2331.

125. R. Dhiman and S. Chandra, Freezing-induced splashing during impact of molten metal droplets with high Weber numbers, International Journal of Heat and Mass Transfer, 2005, 48(25-26): p. 5625-5638.

126. M. Bussmann, J. Mostaghimi, and S. Chandra, On a three-dimensional volume tracking model of droplet impact, Physics of Fluids, 1999, 11(6): p. 1406-1417.

127. M. Fukumoto, T. Yamaguchi, M. Yamada, and T. Yasui, Splash splat to disk splat transition behavior in plasma-sprayed metallic materials, Journal of Thermal Spray Technology, 2007, 16(5-6): p. 905-912.

128. M. Mani, S. Mandre, and M. P. Brenner, Events before droplet splashing on a solid surface, Journal of Fluid Mechanics, 2010, 647: p. 163-185.

129. A. P. Newbery and P. S. Grant, Droplet splashing during arc spraying of steel and the effect on deposit microstructure, Journal of Thermal Spray Technology, 2000, 9(2): p. 250-258.

130. M. Fukumoto, E. Nishioka, and T. Matsubara, Flattening and solidification behavior of a metal droplet on a flat substrate surface held at various temperatures, Surface and Coatings Technology, 1999, 120-121: p. 131-137.



131. S. Shakeri and S. Chandra, Splashing of molten tin droplets on a rough steel surface, International Journal of Heat and Mass Transfer, 2002, 45(23): p. 4561-4575.

132. M. Pasandideh-Fard, R. Bhola, S. Chandra, and J. Mostaghimi, Deposition of tin droplets on a steel plate : Simulations and experiments, International Journal of Heat and Mass Transfer, 1998, 41(19): p. 2929-2945.

133. S. D. Aziz and S. Chandra, Impact, recoil and splashing of molten metal droplets, International Journal of Heat and Mass Transfer, 2000, 43(16): p. 2841-2857.

134. P. Gougeon and C. Moreau, Simultaneous independent measurement of splat diameter and cooling time during impact on a substrate of plasma-sprayed molybdenum particles, Journal of Thermal Spray Technology, 2001, 10(1): p. 76-82.

135. J. Mostaghimi and S. Chandra, Heat transfer in plasma spray coating processes, Advances in Heat Transfer, 2007, 40: p. 143-204.

136. T. E. Twardowski, X. H. Fang, and R. Knight, Structure and Properties of Thermally Sprayed Amorphous Polymer Coatings, Proceedings of the International Thermal Spray Conference, 2000, Montreal, Canada, Pub. ASM International , Materials Park, Ohio, USA.

137. E. P. Song, J. Ahn, S. Lee, and N. J. Kim, Effects of critical plasma spray parameter and spray distance on wear resistance of Al2O3-8 wt.%TiO2 coatings plasma-sprayed with nanopowders, Surface and Coatings Technology, 2008, 202(15): p. 3625-3632.

138. J. Pattison, S. Celotto, A. Khan, and W. O'Neill, Standoff distance and bow shock phenomena in the Cold Spray process, Surface and Coatings Technology, 2008, 202(8): p. 1443-1454.

139. V. V. Sobolev and J. M. Guilemany, Dynamic processes during high velocity oxyfuel spraying, International Materials Reviews, 1996, 41(1): p. 13-32.

140. S. W. K. Kweh, K. A. Khor, and P. Cheang, Plasma-sprayed hydroxyapatite (HA) coatings with flame-spheroidized feedstock: Microstructure and mechanical properties, Biomaterials, 2000, 21(12): p. 1223-1234.

141. C. J. Li and A. Ohmori, Relationships between the microstructure and properties of thermally sprayed deposits, Journal of Thermal Spray Technology, 2002, 11(3): p. 365-374.

142. T. A. Dobbins, R. Knight, and M. J. Mayo, HVOF thermal spray deposited Y2O3-stabilized ZrO2 coatings for thermal barrier applications, Journal of Thermal Spray Technology, 2003, 12(2): p. 214-225.



143. B. Henry, F. Borit, V. Guipont, M. Jeandin, D. Pachoutinsky, J. Lu, and E. Welvaert, Residual Stresses in Inconel 625 Coating of Organic-Based Composites Obtained by Plasma Spraying with Atmosphere and Temperature Control (ATC), Advancing the Science & Applying the Technology 2003, Orlando, Florida, USA, Pub. ASM International, Materials Park, OH, USA.

144. L. S. Schadler, K. O. Laul, R. W. Smith, and E. Petrovicova, Microstructure and mechanical properties of thermally sprayed silica / nylon nanocomposites, Journal of Thermal Spray Technology, 1997, 6(4): p. 475-485.

145. S. Saber-Samandari and K. A. Gross, The use of thermal printing to control the properties of calcium phosphate deposits, Biomaterials, 2010, 31(25): p. 6386-6393.

146. S. Saber-Samandari, C. C. Berndt, and K. A. Gross, Selection of the implant and coating materials for optimized performance by means of nanoindentation, Acta Biomaterialia, 2011, 7(2): p. 874-881.

147. H. Li, S. Costil, H. L. Liao, C. J. Li, M. Planche, and C. Coddet, Effects of surface conditions on the flattening behavior of plasma sprayed Cu splats, Surface and Coatings Technology, 2006, 200(18-19): p. 5435-5446.

148. S. Costil, H. Liao, A. Gammoudi, and C. Coddet, Influence of surface laser cleaning combined with substrate preheating on the splat morphology, Journal of Thermal Spray Technology, 2005, 14(1): p. 31-38.

149. W. Trompetter, M. Hyland, D. McGrouther, P. Munroe, and A. Markwitz, Effect of substrate hardness on splat morphology in high-velocity thermal spray coatings, Journal of Thermal Spray Technology, 2006, 15(4): p. 663-669.

150. M. Friis, C. Persson, and J. Wigren, Influence of particle in-flight characteristics on the microstructure of atmospheric plasma sprayed yttria stabilized ZrO2, Surface and Coatings Technology, 2001, 141(2-3): p. 115-127.

151. A. Papyrin, V. Kosarev, S. Klinkov, A. Alkhimov, and V. M. Fomin, Cold Spray Technology. 2006, Elsevier Science: Burlington.

152. C. J. Li and J. L. Li, Evaporated-gas-induced splashing model for splat formation during plasma spraying, Surface and Coatings Technology, 2004, 184(1): p. 13-23.

153. J. Mostaghimi and S. Chandra, Splat formation in plasma-spray coating process, Pure and Applied Chemistry, 2002, 74(3): p. 441-445.

154. M. Fukumoto, E. Nishioka, and T. Matsubara, Effect of interface wetting on flattening of freely fallen metal droplet onto flat substrate surface, Journal of Thermal Spray Technology, 2002, 11(1): p. 69-74.



155. C. J. Li, C. X. Li, G. J. Yang, and Y. Y. Wang, Examination of substrate surface melting-induced splashing during splat formation in plasma spraying, Journal of Thermal Spray Technology, 2006, 15(4): p. 717-724.

156. Y. Tanaka and M. Fukumoto, Investigation of dominating factors on flattening behavior of plasma sprayed ceramic particles, Surface and Coatings Technology, 1999, 120-121: p. 124-130.

157. N. Z. Mehdizadeh, M. Raessi, S. Chandra, and J. Mostaghimi, Effect of substrate temperature on splashing molten tin droplets, Journal of Heat Transfer, 2004, 126(3): p. 445-452.

158. A. Abedini, A. Pourmousa, S. Chandra, and J. Mostaghimi, Effect of substrate temperature on the properties of coatings and splats deposited by wire arc spraying, Surface and Coatings Technology, 2006, 201(6): p. 3350-3358.

159. X. Jiang, J. Matejicek, and S. Sampath, Substrate temperature effects on the splat formation, microstructure development and properties of plasma sprayed coatings: Part II: case study for molybdenum, Materials Science and Engineering A, 1999, 272(1): p. 189-198.

160. S. Sampath, X. Y. Jiang, J. Matejicek, A. C. Leger, and A. Vardelle, Substrate temperature effects on splat formation, microstructure development and properties of plasma sprayed coatings Part I: Case study for partially stabilized zirconia, Materials Science and Engineering A, 1999, 272(1): p. 181-188.

161. F. Awaja, M. Gilbert, G. Kelly, B. Fox, and P. J. Pigram, Adhesion of polymers, Progress in Polymer Science (Oxford), 2009, 34(9): p. 948-968.

162. W. J. Trompetter, M. Hyland, P. Munroe, and A. Markwitz, Evidence of mechanical interlocking of NiCr particles thermally sprayed onto Al substrates, Journal of Thermal Spray Technology, 2005, 14(4): p. 524-529.

163. A. M. Ahmed and R. H. Rangel, Metal droplet deposition on non-flat surfaces: effect of substrate morphology, International Journal of Heat and Mass Transfer, 2002, 45(5): p. 1077-1091.

164. H. Liu, E. J. Lavernia, and R. H. Rangel, Modeling of molten droplet impingement on a non-flat surface, Acta Metallurgica et Materialia, 1995, 43(5): p. 2053-2072.

165. M. Raessi, J. Mostaghimi, and M. Bussmann, Effect of surface roughness on splat shapes in the plasma spray coating process, Thin Solid Films, 2006, 506-507: p. 133-135.

166. K. Patel, C. S. Doyle, D. Yonekura, and B. J. James, Effect of surface roughness parameters on thermally sprayed PEEK coatings, Surface and Coatings Technology, 2010, 204(21-22): p. 3567-3572.



167. H. D. Steffens, B. Wielage, and J. Drozak, Interface phenomena and bonding mechanism of thermally-sprayed metal and ceramic composites, Surface and Coatings Technology, 1991, 45(1-3): p. 299-308.

168. H. Li, K. A. Khor, and P. Cheang, Impact formation and microstructure characterization of thermal sprayed hydroxyapatite / titania composite coatings, Biomaterials, 2003, 24(6): p. 949-957.

169. E. Almeida, D. Santos, F. Fragata, D. de la Fuente, and M. Morcillo, Anticorrosive painting for a wide spectrum of marine atmospheres: Environmental-friendly versus traditional paint systems, Progress in Organic Coatings, 2006, 57(1): p. 11-22.

170. A. Sakhri, F. X. Perrin, E. Aragon, S. Lamouric, and A. Benaboura, Chlorinated rubber paints for corrosion prevention of mild steel: A comparison between zinc phosphate and polyaniline pigments, Corrosion Science, 2010, 52(3): p. 901-909.

171. D. Wang and G. P. Bierwagen, Sol-gel coatings on metals for corrosion protection, Progress in Organic Coatings, 2009, 64(4): p. 327-338.

172. H. Li, K. Liang, L. Mei, S. Gu, and S. Wang, Oxidation protection of mild steel by zirconia sol-gel coatings, Materials Letters, 2001, 51(4): p. 320-324.

173. T. Tüken, B. YazIcI, and M. Erbil, A new multilayer coating for mild steel protection, Progress in Organic Coatings, 2004, 50(2): p. 115-122.

174. T. Tüken, B. Yazici, and M. Erbil, The use of polythiophene for mild steel protection, Progress in Organic Coatings, 2004, 51(3): p. 205-212.

175. Y. Wang, W. Tian, T. Zhang, and Y. Yang, Microstructure, spallation and corrosion of plasma sprayed Al2O3-13%TiO2 coatings, Corrosion Science, 2009, 51(12): p. 2924-2931.

176. P. K. Aw, A. L. K. Tan, T. P. Tan, and J. Qiu, Corrosion resistance of tungsten carbide based cermet coatings deposited by High Velocity Oxy-Fuel spray process, Thin Solid Films, 2008, 516(16): p. 5710-5715.

177. J. J. Suay, M. T. Rodríguez, K. A. Razzaq, J. J. Carpio, and J. J. Saura, The evaluation of anticorrosive automotive epoxy coatings by means of electrochemical impedance spectroscopy, Progress in Organic Coatings, 2003, 46(2): p. 121-129.

178. C. Maier and T. Calafut, Polypropylene The Definitive User’s Guide and Databook, Plastics Design Library, 1998, Norwich, NY, USA: Plastics Design Library a division of William Andrew Inc.

179. J. Martin, P. Bourson, A. Dahoun, and J. M. Hiver, The ß-spherulite morphology of isotactic polypropylene investigated by Raman spectroscopy, Applied Spectroscopy, 2009, 63(12): p. 1377-1381.



180. V. Lešer, M. Milani, F. Tatti, Ž. Pipan Tkalec, J. Štrus, and D. Drobne, Focused ion beam (FIB) / scanning electron microscopy (SEM) in tissue structural research, Protoplasma, 2010, 246(1): p. 41-48.

181. A. Friedmann, A. Hoess, A. Cismak, and A. Heilmann, Investigation of cell-substrate interactions by focused ion beam preparation and scanning electron microscopy, Acta Biomaterialia, In Press, Corrected Proof.

182. 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, Materials Science and Engineering A, 2000, 284(1-2): p. 29-40.

183. R. Chow, T. A. Decker, R. V. Gansert, D. Gansert, and D. Lee, Properties of aluminum deposited by a HVOF process, Journal of Thermal Spray Technology, 2003, 12(2): p. 208-213.

184. A. Brookes, J. M. Dyke, P. J. Hendra, and S. Meehan, The FT-Raman spectroscopic study of polymers at temperatures in excess of 200°C, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 1997, 53(13): p. 2313-2321.

185. A. S. Nielsen, D. N. Batchelder, and R. Pyrz, Estimation of crystallinity of isotactic polypropylene using Raman spectroscopy, Polymer, 2002, 43(9): p. 2671-2676.

186. M. Arruebarrena de Báez, P. J. Hendra, and M. Judkins, The Raman spectra of oriented isotactic polypropylene, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 1995, 51(12): p. 2117-2124.

187. A. Pawlak and E. Piorkowska, Crystallization of isotactic polypropylene in a temperature gradient, Colloid and Polymer Science, 2001, 279(10): p. 939-946.

188. J. Karger-Kocsis, Polypropylene - An A-Z Reference. 1999, Springer - Verlag.

189. E. L. Bedia, M. A. Paglicawan, C. V. Bernas, S. T. Bernardo, M. Tosaka, and S. Kohjiya, Natural weathering of polypropylene in a tropical zone, Journal of Applied Polymer Science, 2003, 87(6): p. 931-938.

190. S. Ayrault, A. Chateauminois, J. P. Soulier, D. Tréheux, and A. B. Vannes, Deposition of a ceramic coating on a thermoplastic polymer by atmospheric plasma and laser cladding, Surface and Coatings Technology, 1996, 79(1-3): p. 119-130.

191. C. R. C. Lima and J. M. Guilemany, Adhesion improvements of Thermal Barrier Coatings with HVOF thermally sprayed bond coats, Surface and Coatings Technology, 2007, 201(8): p. 4694-4701.



192. C. C. Berndt, The adhesion of flame and plasma sprayed coatings, in Department of Materials Engineering. 1980, Monash University: Victoria, Australia.

193. S. Chandra and P. Fauchais, Formation of solid splats during thermal spray deposition, Journal of Thermal Spray Technology, 2009, 18(2): p. 148-180.

194. T. Sundell, H. Fagerholm, and H. Crozier, Isotacticity determination of polypropylene using FT-Raman spectroscopy, Polymer, 1996, 37(15): p. 3227-3231.

195. M. Xue, S. Chandra, and J. Mostaghimi, Investigation of splat curling up in thermal spray coatings, Journal of Thermal Spray Technology, 2006, 15(4): p. 531-536.

196. C. Minogianni, K. G. Gatos, and C. Galiotis, Estimation of crystallinity in isotropic isotactic polypropylene with Raman spectroscopy, Applied Spectroscopy, 2005, 59(9): p. 1141-1147.

197. Y. V. Shemouratov, K. A. Prokhorov, E. A. Sagitova, G. Y. Nikolaeva, P. P. Pashinin, Y. A. Lebedev, and E. M. Antipov, Raman study of polyethylene-polypropylene blends, Laser Physics, 2009, 19(12): p. 2179-2183.

198. G. Montavon, S. Sampath, C. C. Berndt, H. Herman, and C. Coddet, Effects of the spray angle on splat morphology during thermal spraying, Surface and Coatings Technology, 1997, 91(1-2): p. 107-115.

199. W. H. Walton, Feret's statistical diameter as a measure of particle size [2], Nature, 1948, 162(4113): p. 329-330.

200. J. Madejski, Solidification of droplets on a cold surface, International Journal of Heat and Mass Transfer, 1976, 19(9): p. 1009-1013.

201. C. J. Li, W. Y. Li, Y. Y. Wang, G. J. Yang, and H. Fukanuma, A theoretical model for prediction of deposition efficiency in cold spraying, Thin Solid Films, 2005, 489(1-2): p. 79-85.

Appendix A


Appendix A

SAMPLE-SIZE DETERMINATION

A statistical analysis method was implemented to determine the minimum sample

size needed to estimate the measured metrics. The sample size should be sufficient to

represent the populations of particles and splats analysed in this study. A satisfactory

estimation of assumption should be met to accept or reject the results based on required

confidence limits and the maximum accepted error. For the powder particle, the sample

preparation method used consisted of a few stages. First the initial powder was sieved,

therefore the maximum and minimum particle sizes are known. Powder particles were

then affixed on double-sided sticky carbon tape to be imaged by SEM. Finally, image

analysis software was employed to image the particle shape and apply the statistical

tools to answer the question of sample size requirements.

The influence of particle size on the mean particle size was evaluated using a

randomly generated data subset of specific particle sizes with Excel. One hundred

particle size measurements were calculated using ImageJ and considered to be the main

data in the simulation. The mean, standard deviation and coefficient of variation for

each data subset generated from the main data with a specific size number were

calculated. The sample size calculated must be equal to or greater than the sample size

needed to obtain the predetermined tolerances for decision errors; the appropriate

sample size required to generate a sufficiently precise estimate of the true mean.

The mean ( ) used in mathematics and statistics is also called the "arithmetic

mean" or average and defined as the sum of all the observations ( divided by the total

number of observations ( . The mean is used to find the central tendency of

observations and can be calculated by equation A-1. The standard deviation (SD)

measures the variability or diversity of individual observation from the mean and can be

calculated by equation A-2. The lower the standard deviation, the closer the data points

to the mean, whereas a high standard deviation indicates that the data are spread out

over a large range from the mean. SD is the square root of variance. The variance is

http://en.wikipedia.org/wiki/Mean

Appendix A


defined as the sum of the squared distances of each term in the distribution from the

mean (μ), divided by the number of terms in the distribution ( ) and can be calculated

by equation A-3.

Equation A-1

Equation A-2

Equation A-3

Where σ is the standard deviation, σ2 is the variance; μ is the mean, is the ith

measurement in the raw data set and n is the number of measurements in the data set.

The activity aimed to determine the minimum sample size required to give a good

estimate of the average particle size for powders and splats used in this study. The

confidence interval for the estimate of the average of a population (μ) is calculated by

using equation A-4:

Equation A-4

Where is the critical value of the Student's t-distribution for the number of

degrees of freedom, n is the sample size and s is the standard deviation calculated on the

studied sample. The equation written above represents the (1 – α)*100% confidence

interval for μ where X and s are the mean and the standard deviation, respectively, of the

random sample of the powders for a normal population with unknown variance σ2.

Using 95% confidence intervals (α = 0.05) the range on either side of the average

value is the mean of the sample size minus / plus the width (i.e., the uncertainty term).

This term consists of the sample size and its standard deviation in relation to the level of

confidence required (Equations A-5 and A-6). The smaller the uncertainty term, then the

better will be the estimate and precision of the population mean. The uncertainty term

can be reduced if the sample size n is large and the sample standard deviation s is small.

Appendix A


Equation A-5

Equation A-6

The results can be presented in the form of percentages compared to the

calculated average of the sample size. The coefficient of variation (CV) represents the

ratio of the standard deviation to the mean, and it is a useful statistic for comparing the

degree of variation from one data series to another. The higher the CV, then the greater

is the dispersion in the variable. If the sample size is increased and the standard

deviation is small, the uncertainty in the result can be lowered, and the estimate and the

precision of the population average will be better.

A relatively large sample size would be required to reduce the uncertainty level.

Figure A-1 provides the results of the calculation of the sample size required to achieve

the required confidence levels (99%, 95% and 90%) measured within the deviation from

the mean value that is shown in Table A-1, for the values of (upper critical values of

Student's t distribution with a specific number of degrees of freedom) shown in Table

A-2.

Appendix A


Table A-1. Different margin of errors calculated for different confidence levels for all sample sizes.

Sample size

Error of

mean (µm)

Standard deviation

Margin of error for 90%

confidence (µm)


confidence (µm)


confidence (µm)

5 11.11 24.84 18.33 21.77 28.43

10 5.88 18.61 9.71 11.53 15.06

15 4.45 17.23 7.34 8.72 11.39

20 3.08 13.80 5.09 6.05 7.90

25 3.27 16.33 5.39 6.40 8.36

30 3.11 17.03 5.13 6.10 7.96

35 2.88 17.06 4.76 5.65 7.38

40 2.68 16.98 4.43 5.26 6.87

45 2.40 16.09 3.96 4.70 6.14

50 2.39 16.91 3.95 4.69 6.12

55 2.25 16.70 3.72 4.41 5.77

60 2.10 16.24 3.46 4.11 5.37

65 2.13 17.15 3.51 4.17 5.45

70 1.95 16.33 3.22 3.83 5.00

75 1.88 16.27 3.10 3.68 4.81

80 1.93 17.24 3.18 3.78 4.93

85 1.71 15.80 2.83 3.36 4.39

90 1.62 15.38 2.67 3.18 4.15

95 1.81 17.63 2.98 3.55 4.63

100 1.64 16.40 2.71 3.22 4.20

Table A-2. Upper critical values of Student t distribution with specific degrees of freedom. 99% 95% 90%

2.56 1.96 1.65

Appendix A


Fig. A-1. Graph of results of calculations of the sample size required to achieve the required confidence level measured within the deviation from the mean.

The sample size of 100 particles yielded 2.7%, 3.2% and 4.2% margins of error

for confidence levels of 99%, 95% and 90% respectively. Further analysis was

performed using bootstraps in Excel to generate a random number for particle diameter

for each set of data ranging between 5 and 100, with an increment of 5 each time. The

analysis was repeated 20 times to generate the following Table A-3. The relationship

between the confidence interval and the sample size is shown in Fig. A-2.

Fig. A-2. Graph showing relationship between the confidence interval and the sample size.

Appendix A


The bootstrapping technique was used for estimation of the sample size.

“Bootstrap” is a Monte Carlo method to assess or estimate statistics from samples using

a computer-based method such as Excel. A special command was used to enter the 100

particle size diameters considered, called the ‘original sample’. The trial could be

repeated as many times as required, however is longer when the next trial is executed,

so the data needs to be stored and then analysed later. Twenty trials of sample sets

containing some of the 100 samples in increments of 5 were attempted first, the results

are shown in Table A-3. Each of the sample values were selected at random using a

random number generator executed using F9.

Table A-3.Twenty trials of 100 samples with increments of 5. n i=1 i=2 i=3 i=4 i=5 i=6 i=7 i=8 i=9 i=10 5 124.3 123.7 133.7 140.5 148.1 147.4 133.6 138.2 130.5 130.2 10 126.5 139.7 136.6 127.7 133.5 129.2 133.2 129.1 132.7 129.9 15 133.1 129.8 136.2 134.9 132.0 130.1 131.2 128.3 127.6 129.0 20 134.1 135.2 125.8 129.2 132.4 131.6 129.2 125.9 134.3 133.7 25 131.1 133.5 130.5 134.1 131.9 128.1 130.2 129.5 126.4 124.8 30 129.1 127.0 129.6 133.3 131.5 129.6 131.5 135.8 133.0 127.3 35 132.3 130.4 127.1 129.5 131.6 129.9 130.8 130.2 132.8 125.3 40 129.3 129.0 130.0 131.5 127.1 128.3 128.2 129.6 132.3 129.0 45 133.2 127.9 131.2 132.1 126.8 131.8 134.1 131.0 128.2 127.7 50 125.9 129.8 129.4 128.7 130.9 130.4 130.3 132.4 133.4 132.8 55 129.4 130.3 128.4 130.6 129.9 132.2 132.7 131.1 127.4 130.5 60 130.2 128.5 132.7 133.2 130.6 132.5 134.1 131.5 130.3 132.2 65 129.6 126.0 134.5 130.5 128.7 134.8 132.3 128.5 133.6 129.5 70 132.4 129.1 127.9 131.1 133.9 131.3 131.3 129.9 131.7 127.0 75 128.0 131.9 131.4 133.6 129.8 130.0 128.9 130.1 128.5 133.0 80 132.5 131.6 127.8 128.2 125.7 127.9 128.5 133.8 132.5 130.2 85 130.9 131.0 129.5 134.6 128.4 133.6 132.2 129.5 130.6 127.5 90 131.0 132.2 128.0 132.4 130.0 129.6 128.1 130.3 129.3 130.4 95 133.4 132.6 129.7 132.4 128.7 131.0 130.6 133.9 131.2 129.6 100 128.6 130.4 134.5 131.5 131.6 129.1 132.3 129.0 129.8 127.7

Appendix A


Table A-3 continued. n i=11 i=12 i=13 i=14 i=15 i=16 i=17 i=18 i=19 i=20 5 128.2 116.1 138.6 120.8 137.5 124.9 139.3 133.8 116.9 126.4 10 130.3 141.3 134.0 123.5 124.2 126.7 133.9 126.2 142.7 125.1 15 131.0 126.4 125.8 126.1 133.8 133.4 131.9 131.4 131.3 135.4 20 130.6 129.8 133.4 133.8 125.6 130.6 127.4 125.1 135.9 130.5 25 127.8 129.9 129.4 134.0 128.0 129.7 127.6 133.8 135.4 134.1 30 130.4 132.4 125.1 130.8 130.1 135.2 127.7 127.4 132.3 131.2 35 128.1 130.0 129.2 130.6 130.7 127.2 128.5 131.5 128.3 125.9 40 126.8 135.2 131.5 131.2 129.6 128.9 130.1 127.1 131.7 129.5 45 131.7 131.0 131.0 131.7 128.7 134.4 131.0 132.1 125.1 130.4 50 130.3 129.9 128.8 132.5 128.8 130.5 131.4 133.9 130.0 128.3 55 133.2 130.7 131.4 132.2 129.2 129.8 127.9 132.7 129.1 128.9 60 133.1 125.8 133.3 132.3 132.1 129.3 130.5 130.3 131.2 129.5 65 128.6 129.6 127.9 130.2 130.9 134.4 130.2 129.9 130.6 132.0 70 128.0 129.6 131.1 128.9 131.9 131.1 132.4 127.0 129.7 132.4 75 129.2 129.0 133.9 130.8 132.0 132.5 129.1 130.4 128.0 131.7 80 129.2 132.5 131.0 132.0 130.9 128.3 131.4 133.1 129.6 128.1 85 130.2 132.9 129.7 131.0 132.1 129.8 129.4 126.4 130.3 130.1 90 129.5 130.7 130.0 128.5 128.7 127.7 132.3 130.9 129.5 130.3 95 128.7 133.6 128.5 131.2 129.7 128.7 130.0 128.9 125.7 133.0 100 128.2 131.4 128.3 129.3 126.2 129.9 128.1 128.3 131.0 130.0

Table A-3 continued. µ Σ CV CI 5 131.9 7.9 0.060 59.0 10 130.6 5.7 0.043 41.3 15 130.5 3.9 0.030 33.7 20 130.7 3.7 0.029 29.2 25 130.3 3.2 0.025 26.1 30 129.9 2.8 0.022 23.7 35 129.9 2.2 0.017 22.0 40 130.5 2.0 0.016 20.6 45 130.3 2.1 0.016 19.4 50 130.4 2.1 0.016 18.4 55 129.8 2.2 0.017 17.5 60 130.6 2.0 0.016 16.9 65 130.5 2.2 0.017 16.2 70 130.0 1.7 0.013 15.5 75 130.4 1.6 0.012 15.1 80 130.3 1.8 0.014 14.6 85 130.4 1.9 0.014 14.1 90 130.0 1.5 0.012 13.7 95 130.2 1.9 0.015 13.4 100 130.0 1.5 0.012 13.0

Appendix A


For a 20 trial number generator, the confidence interval (Equation A-7) reaches

60 %, while for a 40 trial number generator, the confidence interval reaches 40 %, as

shown in Fig. A-4. The samples mean and the deviations from the mean of all sample

sizes are shown in Figure A-5 and A-6, respectively.

Equation A-7

Fig. A-3 Graph showing the relationship between the confidence interval and the sample size for 40 trials.

Fig. A-4 Graph showing relationship between the confidence interval and the sample size for 60 trials.

Appendix A


Fig. A-5. Graph showing relationship between the samples’ mean and the sample size.

Fig. A-6. Graph showing relationship between the deviation of the means and the sample size for both upper and lower critical values of Student's t distribution with different confidence intervals.

Appendix B


Appendix B Appendix B presents graphs using the projected areas of particles and splats. Refer to

Section 3.1 on “Powder Processing”.

PARTICLE SHAPE READING VIA COMPUTER IMAGES

Fig. B-1. Comparison of particle and splat projected area analysed without splashes, for different SODs.

Fig. B-2. Comparison of particle and splat projected area analysed with splashes, for different SODs.

Appendix C


Appendix C Appendix C demonstrates a method of finding the transformation function for

the probability density function of splats. This Appendix is linked to Section 4.3.4 on

“Spreading Factor”.

TRANSFORMATION FUNCTION OF SPREADING FACTOR

Fig. C-1. Probability density function of the real data of particles and splat. The probability density function of the particle size is measured to be:

f(x) = -0.0037x2 + 1.2247x - 89.11

Fig. C-2. Plot of input and output data to calculate the transfer function.

Appendix C


Calculations of the transfer function were employed using the original data that

is labeled as “1” for particles. The transferred function for the splat data is labeled as

“2”. The X axis is the “particle size” and the Y axis is the “frequency”. The real data

points on Fig. C-2 are listed below.

X1= 110 X2 = 230 X3 = 110 X4 = 300 Xm1 = 160 Xm2 = 210 Ym1 = 16 Ym2 = 17 If: and where and where: A and D are Y-affecters; and B and C are X-affecters as follows:

A: dilate graph vertically or flip over X axis

B: dilate graph horizontally or flip over Y axis

C: shift graph left or right (horizontal shift)

D: shift graph up or down (vertical shift)

Combining all of these conditions and variables we have, for an unchanged

transformation function, the values for A, B, C and D can be expressed in the form:

The values of A, B, C and D can be found as follows:

1. The splat graph is wider than the particle graph (wider X-axis). Thus the particle

graph needs to be stretched. Then, B is < 1 and equal to 160 / 210= 0.762.

Appendix C


2. The splat graph is higher than the particle graph (wider Y-axis). Thus the

particle graph needs to be stretched. Then, A is > 1 and equal to 17 / 16=

1.0625.

3. The centre of the splat graph is shifted to the right of the particle graph on

the X-axis. Thus the particle graph needs to be shifted. Then, C is > 0 and

equal to 210-160=50.

4. Since there is no shifting or moving in the Y-axis, then D= 0

5. Substituting the value of B in the transformation function gives:

6. Substituting the value of A in the transformation function gives:

7. Substituting the values of C and D in the transformation function gives:

8. The result of the transformation function is:

9. Using Excel to generate the probability density function for the splat size gives:

This result can be compared with the probability density function of the splat

size generated using Excel. This provides an approximate curve fit of the splat size.

Both the derived equation and Excel-generated equation follow the same trends; thereby

indicating that the method is mathematically rigorous.

y = -0.0017x2 + 0.6926x - 54.808

List of Publications


List of Publications Alamara K., Saber-Samandari S., Stoddart P. and Berndt C. C., “Effect of

substrate temperature on the splat formation of flame sprayed polypropylene”,

Journal of Surface and Coatings Technology, 206, (2011), 1180–1187

Alamara K., Saber-Samandari S. and Berndt C. C., “Splat taxonomy of

polymeric thermal spray coating”, Journal of Surface and Coatings Technology,

205, (2011), 5028–5034

Alamara K., Saber-Samandari S. and Berndt C. C., “Splat formation of

polypropylene flame sprayed onto a flat surface”, Journal of Surface and

Coatings Technology, 205, (2010), 2518-2514

Conferences

Alamara K., Saber-Samandari S. and Berndt C. C., “Effect of substrate

roughness on splat formation of thermally sprayed polymer” in Thermal Spray

2011: Proceedings of the International Thermal Spray Conference, Hamburg,

Germany, Pub. (DVS-ASM), pp 685-689.

Alamara K., and Berndt C. C., “Effect of substrate chemistry on splat

formation of thermally sprayed polypropylene”, Pac Rim 9, 10-14 July 2011,

Cairns, Australia

Documents

Optimization of polypropylene splats using the flame spray