fabrication of lightweight polymer composite materials
TRANSCRIPT
Fabrication of Lightweight Polymer Composite Materials
Using Electric Wire-Arc Spraying Process
by
Sudarshan Devaraj
A thesis submitted in partial fulfillment of the requirements for the degree of
Master of Applied Science
in
Mechanical and Industrial Engineering
Department of Mechanical and Industrial Engineering
University of Toronto
© Sudarshan Devaraj, 2020
ii
Fabrication of Lightweight Polymer Composite Materials Using Electric
Wire-Arc Spraying Process
Sudarshan Devaraj
Master of Applied Science
Department of Mechanical and Industrial Engineering, University of Toronto
2020
Abstract
Wire-arc spraying of zinc and aluminium onto polyethylene (PE) and polytetrafluoroethylene (PTFE)
was conducted. The impact of surface roughness and substrate temperature on the adhesion strength of
the coatings was investigated. Continuous coatings of zinc on PE and PTFE and aluminium on PTFE
were fabricated. It was not possible to form continuous aluminum coatings on PE. Individual metal
splats were observed using scanning electron microscope (SEM) to provide insights into the adhesion
mechanisms.
The challenge of metallizing PE was overcome by using porous PE as substrate materials. Copper,
aluminum and zinc were successfully deposited on the porous PE to form coatings with thickness of
about 400 µm. The coating surfaces and cross-sectional areas were characterized using a SEM, while
the coating adhesion strength and electrical resistivity was examined using pull testing and four-wire
sensing respectively. The use of polymer composite materials as lightweight heatsinks for LEDs was
then investigated.
iii
Acknowledgements
I would like to thank my supervisors, Dr. Sanjeev Chandra and Dr. André McDonald. Their guidance
and mentorship were critical to the success of my thesis. They provided me with a solid foundation
of mechanical and materials engineering with which I built my thesis and graduate studies upon.
I express my gratitude to all the supporting staff at the University of Toronto. Thank you to the staff
in the machine shop for helping me with all machining work related to the research. Thank you to
the engineering technologist, Sal Boccia at U of T’s Ontario Centre for the Characterization of
Advanced Materials (OCCAM) for providing characterization analysis and supporting my curiosity
with their new and innovative techniques to characterize materials.
To my lab mates and fellow graduate students at the Centre for Advanced Coating Technologies
(CACT), I appreciate the constant support in all my research endeavors during these two years. Thank
you also to those who collaborated on research with me and to those who helped proof-read my papers.
I wish you all the best of luck with the rest of your studies and in the future. I extend my thanks to Dr.
Larry Pershin for his guidance and insight throughout this study.
I would like to acknowledge the financial support for this work from Natural Science and Engineering
Research Council Green Surface Engineering for Advanced Manufacturing (Green-SEAM) Strategic
Network.
Finally, to my family and friends, this journey would not be the same without your help and support
throughout these last two years. I may not have said it then, but your kind words and encouragement
helped me through the long days of writing my thesis.
iv
Table of Contents
Abstract ............................................................................................................................. ...... ii
Acknowledgments ................................................................................................................... iii
List of Tables ........................................................................................................................... vii
List of Figures ......................................................................................................................... viii
Nomenclature/Notation ............................................................................................................ xii
Chapter 1 Introduction .......................................................................................................... 1
1.1. Motivation ................................................................................................................... 1
1.2. Literature Review ........................................................................................................ 3
1.3. Research Objectives .................................................................................................... 8
1.4. Thesis Organization ..................................................................................................... 9
Chapter 2 Experimental Method ……………………………….….………….................... 10
2.1 Wire-Arc Spraying …………………………………………………………………... 10
2.2 Experimental Assembly ……………………………………………………………... 11
2.3 Polymeric Substrate Surfaces ……………………………………………………….. 14
2.4 Coating Cross-section Characterization …………………………………………….. 15
2.5 Adhesion Strength Measurement ……………………………………………............ 16
2.6 Electrical Resistance Measurement ………………………………………………… 17
Chapter 3 Metallization of Porous Polyethylene ............................................................... 18
3.1 Introduction ………………………………………………………………………… 18
3.2 Results and Discussions ……………………………………………………………. 19
3.2.1 Particle Size Distribution ……………………………………………………. 19
3.2.2 Substrate Surface Temperature ……………………………………………… 20
v
3.2.3 Adhesion Strength ……………………………………………………………. 21
3.3 Conclusions …………………………………………………………………………. 31
Chapter 4 Metallization of Porous Polyethylene .............................................................. 33
4.1 Introduction ............................................................................................................... 33
4.2 Results and Discussions …………………………………………………………… 34
4.2.1 Spray Particle Size Analysis ………………………………………………… 34
4.2.2 Substrate Roughness ………………………………………………………… 34
4.2.3 Coatings Cross-section ……………………………………………………… 36
4.2.4 Single Splat Characterization ……………………………………………….. 39
4.2.5 Single Splats on Porous and Non-Porous Polyethylene …………………….. 45
4.2.6 Coatings Bond Strength …………………………………………………….. 48
4.2.7 Coatings Electrical Resistivity ……………………………………………… 49
4.3 Conclusions ……………………………………………………………………….. 52
Chapter 5 Fabrication and Characterization of Lightweight Heatsinks ……............... 54
5.1 Introduction ………………………………………………………………………. 54
5.2 Experimental Method ............................................................................................... 54
5.2.1 Experimental Assembly for Thermal Power Measurement of LED Strips … 54
5.2.2 Fabrication of Al foil Heatsinks ……………………………………………. 56
5.2.3 Fabrication of Polymer Composite Heatsink ………………………………. 57
5.2.4 Experimental Assembly for Temperature Distribution in Heatsinks ………. 57
5.3 Analytical Heat Conduction Model …………………………….............................. 58
5.4 Results and Discussions …………………………………………………………… 60
5.4.1 Heat Generation in LEDs …………………………………………………... 60
5.4.2 Surface Temperature Measurement and Prediction ………………………... 60
vi
5.4.3 Heatsink Fin Efficiency…………………………………………………….. 61
5.4.4 Performance Comparison of Bare Al and Polymer Composite Heatsink…... 63
5.4.5 Performance of Al Foil Heatsinks …………………………………………. 65
5.5 Conclusions ………………………………………………………………………. 68
Chapter 6 Conclusions ...................................................................................................... 69
Chapter 7 Recommendations for Future Work ………………………………………. 72
References ........................................................................................................................... 74
Appendix A - SEM Images of Coating Microstructure ...................................................... 79
Appendix B – MATLAB Code for Solidification Parameter Calculations ....................... 83
Appendix C – Splat Size and Circularity Data ................................................................... 85
vii
List of Tables
Table 2-1 Properties of spray materials………………………………………….................. 10
Table 2-2 Properties of polymer substrate materials .......................................................... 14
Table 4-1 Electrical resistivity of In-625 coatings of varying thickness …………………... 49
Table 5-1 Temperature and weight of lightweight heat sink models…………………...........64
viii
List of Figures
Figure 2-1 Schematic diagram of single splat experimental apparatus …………………… 12
Figure 2-2 Schematic diagram of substrate pre-heating experimental apparatus…………… 13
Figure 2-3 Experimental setup for four-wire sensing test .................................................... 17
Figure 3-1 SEM images of particles of (a) aluminum (mean particle diameter d50 = 78.6 µm)
and (b) zinc (d50 = 61.3 µm) captured by spraying into a water bath …………… 19
Figure 3-2 Adhesion strength of aluminum and zinc coatings on rough and smooth polymer
substrates. Aluminum coatings did not adhere to PE surfaces [48] …………….. 21
Figure 3-3 Cross sectional view of (a) zinc coating on smooth PTFE, (b) aluminum coating on
smooth PTFE, (c) zinc coating on rough PTFE and (d) aluminum coating on rough
PTFE substrates. Areas of mechanical interlocking are outlined in (c) [48] ……. 22
Figure 3-4 Single splats of (a) zinc and (b) aluminum sprayed on a smooth PTFE surface at
room temperature ……………………………………………………………….. 23
Figure 3-5 SEM image of a zinc splat trapped in the asperities of a rough PE substrate ….. 24
Figure 3-6 Cross sectional view of PE substrates on which aluminum was sprayed. The surfaces
were (a) smooth and (b) rough. The surfaces are shown inclined to the horizontal
so that both the coated surface and the cross-section through the substrate can be
seen ................................................................................................................... 25
Figure 3-7 Adhesion strength of aluminum and zinc coatings on polymer substrates that were
initially either at room temperature or preheated (initial temperature 95°C for
ix
PTFE, 55°C for PE) [48] ……………………………………. ............................ 27
Figure 3-8 Cross sectional view of preheated PE substrates on which zinc was sprayed. (a)
continuous coating formed after 7 passes of the spray torch and (b) individual
particles deposited after 1 pass of the spray torch .............................................. 28
Figure 3-9 Single splats of zinc on PE preheated to 55°C (a) viewed from above and (b) section
through substrate made using a focused ion beam (FIB ..................................... 29
Figure 3-10 Cross sectional view of preheated PE substrates on which aluminum was sprayed.
(a) after 1 pass and (b) after 7 passes of the spray torch ...................................... 30
Figure 4-1 SEM images of particles of (a) aluminum (mean particle diameter d50=78.6 µm), (b)
zinc (d50=61.3 µm) and (c) copper (d50=40.8 µm) captured by spraying into a water
bath .................................................................................................................... 34
Figure 4-2 SEM image of as-received porous polyethylene surface ..................................... 35
Figure 4-3 Cross sectional view of metallic coating at low and high magnification, (a) and (b)
zinc coating, (c) and (d) aluminum coating, (e) and (f) copper coating on porous
polyethylene substrates. Areas of mechanical interlocking are outlined in (e) and
(f). ...................................................................................................................... 37
Figure 4-4 SEM image of single splats at low and high magnification, (a) and (b) zinc, (c) and
(d) aluminum, (e) and (f) copper ........................................................................ 38
Figure 4-5 Solidification parameter values for Zn, Al and Cu on porous PE, assuming a thermal
contact resistance (Rc) value in the order of 10-5 m2-K/W……………………… 40
x
Figure 4-6 Average size of particles sprayed into a water bath and splats deposited on PE .... 41
Figure 4-7 Idealized splat formation on polymer substrate ................................................... 42
Figure 4-8 Circularity of Al, Zn and Cu splats ............................................................................ 43
Figure 4-9 Splats of (a) Al, (b) Zn and (c) Cu on non-porous polyethylene ............................ 44
Figure 4-10 Adhesion test sample (3 mm) .............................................................................. 45
Figure 4-11 Pull off pressures of zinc, aluminum and copper coated porous polyethylene
substrates (6 mm) ............................................................................................... 46
Figure 4-12 Electrical resistivity of the coatings .................................................................... 47
Figure 4-13 In-625 Coating, (a) top surface and (b) cross-section ........................................... 48
Figure 5-1 Experimental setup for heat generation estimation ............................................. 52
Figure 5-2 Lightweight heat sink models, (a) aluminum wrapped around polyethylene, (b)
aluminum foil in ABS frame and (c) copper foil in ABS frame ........................ 53
Figure 5-3 Cross-sectional view of Al-PE composite heat sink ……………………………. 54
Figure 5-4 (a) Bottom view of 46 W LED-heat sink assembly showing temperature measurement
locations, (b) Top view of heat sink .................................................................. 55
Figure 5-5 Analytical heat conduction model of the composite fin consisting of two domains. 56
Figure 5-6 Analytical temperature curves of the composite heatsink (Equation 5-7) fits the
experimental results ........................................................................................ 59
xi
Figure 5-7 Fin efficiency of the composite fin varies with coating thickness and fin length….60
Figure 5-8 LED strip mounted on polymer heat sink coated with (a) aluminum and (b)
copper…………………………………………………………………………... 61
Figure 5-9 Led strip mounted on bare aluminum plate (reference heatsink material) .......... 62
Figure 5-10 Performance of lightweight polymer composite heat sinks .............................. 63
Figure 5-11 LED strip mounted on 0.5 mm thick Al foil as heat sink, (a) flat and (b) U-
shaped………………………………………………………………………….. 65
Figure 5-12 Performance of heatsinks with different fin lengths .......................................... 66
xii
Nomenclature
All notation used in this study is outlined in this section. A list of acronyms and their meanings is
presented first, followed by all mathematical symbols. The acronyms and symbols are listed in the
order they first appear.
Acronym Description
SEM Scanning Electron Microscope
CFRP Carbon Fiber Reinforced Polymer
PTFE Polytetrafluoroethylene or Teflon®
PE Polyethylene
ABS Acrylonitrile Butadiene Styrene
PVC Polyvinyl Chloride
LED Light Emitting Diode
ASTM American Society for Testing and
Materials
DAQ Data Acquisition Module
PEEK Polyether Ether Ketone
PEI Polyethyleneimine
GFRP Glass Fiber Reinforced Polymer
PU Polyurethane
PS Power Supply
PID Proportional-Integral-Derivative
FIB Focused Ion Beam
xiii
SP Solidification Parameter
PMC Polymer Matrix Composites
Al Aluminum
Zn Zinc
Cu Copper
In 625 Inconel 625
Symbol Description Units
Ra Surface Roughness Average [µm]
Tg Glass Transition Temperature 0C
Ŋfin Heatsink Fin Efficiency %
µ Dynamic Viscosity Pa. s
ν Kinematic Viscosity of Fluid m2/s
δ Characteristic Length of Fin m
β Coefficient of Volume Expansion K-1
ρ Density kg/m3
k1 Thermal Conductivity of Coating W/m-K
k2 Thermal Conductivity of Polymer W/m-K
m Fin Parameter
Ra Rayleigh Number
Pr Prandtl Number
L Fin Length m
H Thickness of Polymer m
xiv
t Thickness of Coating m
Nu Nusselt Number
A Surface Area of Splats m2
p Perimeter of Splats m
D Equivalent Diameter/Size of Splats m
Tair Temperature of Surrounding 0C
Tbase Temperature of Heatsink Fin Base 0C
h Heat Transfer Coefficient W/m2-K
1
Chapter 1
Introduction
1.1 Motivation
1.1.1 Lightweight Materials and Industrial Appeal
New lightweight composite materials, polymers and metals that have high strength-to-weight ratio
are emerging as key technologies for future automobiles and aircrafts. This is largely due to the ever-
increasing demand for fuel efficient vehicles and a rise in Government investments in renewable
energy projects. It has been found that a 10% reduction in vehicle weight can increase its fuel
economy by about 7% [1]. A major proportion of such lightweight materials used are either plastics
or plastic-based composites with enhanced mechanical strength and thermal resistance. According
to statistics from R. Geyer et al. [12] plastics have become the most used material in the world since
1976. Carbon Fiber Reinforced Polymer (CFRP) and Glass Fiber Reinforced Polymer (GFRP) are
the two most widely used polymers in automobile and aerospace industries due to their excellent
high strength-to-weight ratio. Currently, about 20% of the total weight of a typical commercial
passenger automobile comes from such polymeric materials. These composites are strong and light
but are expensive to produce and hard to machine. Although such high-performing composites are
receiving increasing interest, aluminum and its alloys still constitute a significant proportion of
aerospace structural weight [3]. The relatively high specific strength and stiffness, good ductility and
corrosion resistance, low price and excellent manufacturability and reliability make aluminum and
its alloys a popular choice of lightweight materials in many aerospace structural applications.
Therefore, research involving fabrication of lightweight polymer composites which are less
expensive and lighter than aluminum and which have desirable properties of both the metals and the
polymers is gaining a lot of attention [4]. The use of cheap and extremely lightweight plastic,
2
polypropylene (PP) [5] in the construction and automotive sectors is growing rapidly. Addition of
natural fibers (cellulose) in PP provided the desired mechanical strength. These polymer composites
are significantly less expensive than the glass fiber-reinforced plastics currently used. In addition,
the natural fillers used are advantageous for sustainable development as they are all derived from
renewable resources and are biodegradable. The tensile strength of the composite is a function of the
filler content in the polymer. It increases with increase in the percentage of natural fibers in the
plastic. Although such advancements have led to the development of novel lightweight composite
materials, they have complicated the manufacturing processes, which has in turn increased
production costs. There exists a trade-off between the performance and cost, which has motivated
researchers to identify inexpensive and efficient ways to manufacture lightweight materials with
industrial applications.
1.1.2 Challenges in Metallization of Polymer-Based Composites
People have employed different methods to manufacture polymer-based composites for applications
demanding lightweight properties. In this section, the challenges associated with some of the most
common techniques used in industries for polymer metallization will be discussed. The most preferred
processes to metallize plastics include vacuum metallization (physical vapor deposition and chemical
vapor deposition), plating (electroless and electroplating) and thermal spray coating processes.
Metalized plastic components that are coated with metals using the vacuum metallization process are
found in a range of applications, from automotive interior parts to certain types of foils [6]. In the
PVD process, atoms of solid materials (e.g., aluminum, copper or silver) are transferred into gas phase
by strafing with high-energy ions at vacuum. After this they condense on the desired polymer substrate
as a solid metal film. The automotive industry is using metallized plastics because of their stability
3
against corrosion, low density and processing ease. Such metallized polymers are widely used for
making reflectors in automotive, decorative surfaces, electromagnetic shield for polymer packages
and diffusion barriers in packaging foils [7]. However, electroplating most often yields non-uniform
coating deposition, requiring further processing to finish the coated surface. It is a comparatively
expensive, time-consuming process and care needs to be taken while disposing the used electroplating
chemicals that are toxic. Unlike the electroplating and vacuum metallization processes, thermal
spraying process has a high deposition rate and can create very thick metallic layers. Moreover, the
nature of the application process makes it possible for depositing metallic coatings to specific areas
of components, which is useful when working with intricate geometries. This process is relatively
simple, low cost and has a much higher deposition rate than other common techniques used to
metallize polymers.
1.2 Literature Review
1.2.1 Thermal Spray Metallization of Polymers
Polymeric materials are employed in many industrial applications since they are light, relatively
cheap, and can be easily processed; however, they have low thermal and electrical conductivity, which
limits their usefulness. The surface properties of polymers can be modified by applying coatings on
them, creating lightweight composite materials that can be very useful in specific applications [8, 9,
10]. Chen et al. [10] characterized the thermal performance of zinc sprayed Acrylonitrile Butadiene
Styrene (ABS) substrates as heatsinks for Light Emitting Diode (LED) cooling applications and found
that these composite materials could provide significant weight reduction in the heatsinks. Thermal
spray coating is a widely used method of applying metal coatings by directing a molten metal spray
onto a solid surface. One technology, wire-arc spraying, is a low-cost and well-developed way of
4
applying metallic coatings. In this process an electric arc is struck between the tips of two continuously
fed wires and a compressed air jet is used to strip off molten metal droplets and direct them onto a
substrate where they coalesce and freeze to form a solid layer [11]. Though this is a well-known
coating method little research has been done on the metallization of polymers using wire-arc systems.
Gonzalez et al. [12] summarized the current state of the art for thermal spray metallization of polymers
and found that the low melting point and soft nature of most thermoplastics made it difficult to apply
metal coatings on them using high-temperature thermal spray techniques. Thermoplastics have
comparatively low melting points [13] and can be severely damaged when molten metal particles and
hot gases impinge on them. Thermosetting plastics have higher heat-resistance than thermoplastics,
which reduces structural damage even at the elevated temperatures they experience when subjected
to a thermal spray.
Cold spraying, which has inherently low operating temperatures, has also been examined as a method
of coating thermoplastics [14,15]. Vucko et al. [14] embedded Cu particles into high-density
polyethylene (HDPE) (a thermoplastic) for anti-fouling applications using cold spraying but did not
report deposition of continuous copper coatings in their work. Ganesan et al. [15] investigated the
deposition efficiency of spherical and dendritic copper powder on both polyvinyl chloride (PVC) (a
thermoplastic) and brittle epoxy (a thermoset) substrates using the cold spray technique. They showed
that the particles could adhere to the PVC substrate due to its highly plastic nature and substrate
damage was less than that inflicted on epoxy substrates. Che et al. [16] cold sprayed Cu particles onto
CFRP (Carbon Fiber Reinforced Polymer) (thermosetting matrix). They observed substrate erosion
due to the large impact force of impinging metal particles when combined with the brittleness of the
substrate. However, good mechanical interlocking of Cu particles was achieved on the thermoplastic
5
polymers (Polyether Ether Ketone (PEEK), Polyethyleneimine (PEI) and ABS) used in their study.
This was a result of thermal softening exhibited by the thermoplastics at temperatures close to their
Tg, which promoted adhesion. Similarly, Rokni et al. [17] deposited dense Al coatings on PEEK, ABS
and PEI using a high-pressure cold spray process.
Some high-temperature thermal spray methods such as powder flame spray, electric wire-arc spray,
plasma spray have been used successfully to coat thermosetting plastics like polyurethane (PU) and
carbon fibre reinforced epoxy composites where the epoxy matrix was a thermoset [[18], [19], [21]].
Ashrafizadeh et al. [18] co-cured PU with Al-12Si particles on the surface before flame spraying with
Al-12Si. This approach minimized direct contact of impacting particles with PU substrates, avoiding
localized melting and decomposition of the low melting temperature polymer. However, Liu et al.
[22] deposited Zn and Al bond coat layer ( 50 µm), using both plasma and electric wire-arc spraying
process, on Graphite fiber-reinforced polyimide (thermoset) without any interlayer. Therefore, when
using high-temperature thermal spray techniques to deposit a metal coating, the need for an interlayer
depends on the heat sensitive nature of the polymers and the applications targeted. Polymer matrix
composites which have high temperature resistant reinforcements are generally sprayed without the
use of an interlayer [22, 23, 24]. Polymers with very low melting temperatures require an interlayer
to be first applied on the polymer substrate to protect it from damage by high-temperature gases and
molten particles impinging on it [18,25].
The effect of varying substrate temperature on adhesion of thermal spray coatings is well established
[26]. Pershin et al. [27] showed that coating adhesion strength of nickel coatings on steel could be
significantly improved by heating the substrate before coating. Che et al. [28] studied cold spraying
on thermoplastics and investigated the effect of varying carrier gas temperature during on particle
6
penetration. They observed that particles propelled by gas whose temperature was higher than the
glass transition temperature (Tg) of the polymer substrate penetrated deeper into it than those in gas at
a temperature below Tg. This resulted in development of thick metallic coatings of tin and copper onto
these thermoplastics. Similarly, Ganesan et al. [15] showed that the deposition efficiency of Cu
particles, deposited on PVC substrates by cold spraying method, increased with increase in the gas
temperature of the process up to the Tg value of PVC. The thermal softening behaviour exhibited by
these thermoplastics close to or higher than their Tg aids metal particle penetration into the softened
polymer substrates. This eventually leads to strong mechanical interlocking of the metallic particles
with the substrates upon cooling. Therefore, by controlling the gas temperature in the cold spray
process, thick metallic coatings could be achieved on thermoplastic substrates. The results from cold
spraying metals on polymers highlight the importance of polymer substrate temperature during
spraying. Anand [48] investigated the role of polymer substrate temperature on the coatings bond
strength by measuring the substrate temperature throughout the spray period. The results from his
work have been used for reference in the current study. The bond strength of the Al and Zn coatings
on PTFE and PE were found to improve by pre-heating the polymer substrates. Additional
experiments were carried out in the present study to support these results. It is also important to know
the actual substrate temperature during spraying because of the significant changes in polymer
properties that can occur within a small temperature range [29,30].
Grit blasting metallic substrates prior to thermal spraying to increase their roughness is the most
common way of achieving strong mechanical interlocking of coating particles with the substrate.
However, the process of sandblasting polymer substrates can produce significant damage as
demonstrated by Ganesan et al. [20] who concluded that sandblasting thermoset CFRP structures
caused localized destruction of the polymer surface. Sandblasting polymer matrix composites which
7
have brittle fiber reinforcements is not a viable option. Therefore, other surface preparation methods
like co-curing of metallic mesh and powder into the CFRP substrates have been used to enhance the
bond strength of the coatings [31]. Thermoplastics, by contrast, have excellent impact resistance,
which makes them suitable for mechanical roughening processes. The difficulty in achieving a
uniform surface roughness on the polymers by grit blasting has limited research in this direction [20].
1.2.2 Applications of Lightweight Polymer Composite Materials
The substitution of metallic materials with polymers in heat transfer applications allows reduction of
weight and cost of components. They are also much more resistant to corrosion and fouling. Although
polymers are already used in heat exchangers, their low thermal conductivity is still considered the
main limitation for electronics cooling applications. Recent advances in manufacturing polymeric
matrix composites together with modern processing techniques have made it possible to overcome the
problem of reduced conductivity. Marchetto et al. [32] added highly conductive reinforcements and
fabricated thin walls, reducing the overall thermal resistance of heat sinks fabricated with polymeric
materials. Development in the field of materials science has paved the way to overcome some typical
complications associated with using polymers in heat exchangers. Adding suitable fillers that possess
high thermal conductivity can increase the overall thermal conductivity of the composite by forming
a percolating network for thermal transport (Hussain et al., 2017). Such advancements expanded the
applications of plastic materials to the fields of solar water heating [33], automotive radiators [34],
water desalination [35], condensing boilers [36] and electronics cooling [37]. Polymers also have
potential applications in power electronics. They can be used as lightweight heatsinks or heat
exchangers to cool electronics and prevent them from exceeding temperatures above their operating
range. However, for such applications, the thermally non-conductive polymer surface needs to be
8
engineered to make it thermally conductive. Chen et al. [10] fabricated a lightweight heatsink
consisting of Acrylonitrile Butadiene Styrene (possesses similar moulding temperature as
polyethylene) core and a thin zinc coating layer on the polymer surface which was deposited using
wire-arc thermal spray process. Here, the zinc coating provided the desired heat conduction path and
the polymer acted as a lightweight structural support material. Metals such as aluminium, copper and
silicon are usually used in thermal applications because of their high thermal conductivity and
reasonably low coefficient of thermal expansion.
1.3 Research Objectives
The present work aims to fabricate lightweight polymer composites for electronic cooling applications
using a thermal spray process. Polytetrafluoroethylene (PTFE), polyethylene (PE) and porous
polyethylene were chosen as the substrate materials to be coated. An electric wire-arc spray coating
system was used to deposit zinc, copper and aluminum onto these polymers. There were four principle
objectives for this work:
i. Analyze individual metal splats on thermoplastics to provide insight into the interaction
of molten metal particles with the polymer substrates.
ii.Overcome the challenge of metallizing non-porous polyethylene substrates by using
porous polyethylene as the substrate material to be coated.
iii. Evaluate the coating bond strength and electrical resistivity of Zn, Al and Cu coatings
deposited on porous PE.
iv. Fabricate lightweight heatsinks for LED cooling applications using polymer
composite materials.
9
1.4 Thesis Organization
The present thesis document has several chapters with the following structure: Chapter 2 reports
the experimental methods used for characterizing the coatings formed on PTFE, non-porous PE
and porous PE. The experimental setup that was employed for conducting the single splat tests and
the procedure for characterizing the splat morphology are included in this chapter. Chapter 3
focuses on the single splat analysis of Al and Zn on PTFE and PE. In Chapter 4 of this thesis
document, thermal spray metallization of porous polyethylene with Zn, Al and Cu was studied.
Discussion about the impact the microstructure of the fabricated coatings, coatings adhesion
strength and single splat characterization has been included in this chapter. Furthermore, the
deposition of Inconel 625 on polyethylene and its electrical resistivity was discussed. Chapter 5
presents the details related to the fabrication of lightweight polymer composite heatsinks for LED
cooling applications. An analytical heat conduction model that was developed for determining the
fin efficiency and distribution of temperature in the composite heatsinks was also included in this
chapter. Evaluation of the accuracy of the analytical results by comparing them with experimental
data was added to this chapter. Chapter 6 summarizes the conclusions from this thesis. Finally,
Chapter 7 provides the suggestions for future work as for extension and modification of this
research work.
10
Chapter 2
EXPERIMENTAL METHOD
2.1 Wire-Arc Spraying
A high-density wire-arc spray coating system (Thermion, Silverdale, Washington, USA, P/N: 57456)
was used to coat the polymer substrates. Commercially available pure Al (Oerlikon Metco, Westbury
NY, USA, DSMTS-0003.10) and Zn (Oerlikon Metco, Westbury NY, USA, DSMTS-0010.6), copper
(Sulzer Metro Inc., NY, United States, DSMTS-0009.7) and Inconel 625 (Sulzer Metro Inc., NY,
United States, DSMTS-0052.7) wires were used in this study. Some important thermophysical
properties of the coating metals are listed in Table 2-1. The mean size of the Al, Zn and Cu spray
particles was determined by spraying the metal particles into a water bath and then allowing them to
dry under natural convection. Particle size distribution of both Al and Zn particles was obtained using
a Malvern Mastersizer X laser analyser (Malvern Instruments Ltd., Malvern, Worcestershire, UK,
P/N: 2000).
Table 2-1 Properties of spray materials [38,39,40]
Metal Density
(kg/m3)
Tm (oC) Surface Tension at
Melting Point (Nm-1)
Specific Heat
(J/kg-K)
Aluminum 2700 660 0.9 1180
Zinc 7200 420 0.65 427
Copper 8960 1085 1.5 385
Inconel 625 8440 1350 1.8 410
11
All spray parameters for Al and Zn were kept constant except the input arc voltage, which was 28 V
for Zn and 32 V for Al, Cu and Inconel 625. Dry air was chosen as the atomizing gas and the gas
pressure was fixed at 690 kPa (100 psig). A nozzle stand-off distance of 152 mm (6 in) was used. All
samples used in this work were cooled using compressed air during spraying. The pressure of the
cooling air was 690 kPa (100 psig). The robotic arm holding the spray nozzle was programmed to
move at a speed of 1000 mm/min in a serpentine pattern consisting of parallel passes spaced 5 mm
apart. Identical process parameters (e.g., gas pressure, stand-off distance) were used for single splat
studies.
2.2 Experimental Assembly
2.2.1 Setup for Single Splat Study
A schematic of the experimental apparatus for single splat study is shown in Fig. 2-1. The polymer
samples were mounted to a steel support and positioned in such a way that the wire arc spray cone
core impacts approximately the centre of the sample during spraying. A protective barrier (mask) with
an orifice 1 mm in width between the spray and the sample prevented coating deposition on the
polymer substrate. The spray of droplets was passed through the orifice before they impacted on the
substrate so that after a single pass of the spray torch over the polymer surface, individual metal splats
could be obtained. To deposit a complete coating on the substrate, the protective barrier was removed
from the experimental assembly.
12
2.2.2 Setup for Substrate Pre-heating
The experimental facility used for preheating the polymers (PTFE and PE) is shown in Fig. 2-2. Some
of the polymer samples were heated prior to depositing individual metal splats to investigate the effect
of polymer substrate temperature on coating bond strength. The polymer samples were clamped to a
heating block whose temperature was monitored with a Proportional-Integral-Derivative (PID)
controller (Omega, Laval, Canada, P/N: C9000A). This block was machined to hold three cartridge
heaters (OMEGALUX™ CS, Omega, Laval, Canada, 350 W, P/N: CSS-10150) to heat the system to
the desired temperature. Power to the heaters were controlled by a variable voltage power supply.
Figure 2-1 Schematic diagram of single splat experimental apparatus
Steel Support
1 mm
Spray directing
air jet Feed Wire
Atomizing air
jet
Feed Wire Air gap
Wire guide and
current pickup
Spray
stream
Mask
Sample
13
Figure 2-2 Schematic diagram of substrate pre-heating experimental apparatus
The polymer samples were heated using a 320 W electrical resistance heater (KHA-808/5, Omega,
St-Eustache, Quebec, Canada,) while measuring and controlling the sample surface temperature using
K-type thermocouples and a PID controller, respectively. Heating experiments were only carried out
on smooth polymer samples and rough polymers were not considered for this study. The surface
temperature of the polymer samples was recorded using five type-K thermocouples (FF-K-20-100,
Omega), which were connected to a data acquisition module (DAQ-2408, Omega). The
thermocouples were attached to the substrates through holes drilled through the heater block, and the
thermocouple junctions were placed flush in contact with the substrate surface.
2.3 Polymeric Substrate Surfaces
14
2.3.1 PTFE and PE
Polytetrafluoroethylene (PTFE , McMaster-Carr, Sante Fe Springs, California, USA, P/N: 8545K24)
and Ultra-high-molecular-weight polyethylene (PE, McMaster-Carr, Sante Fe Springs, California,
USA, P/N: 8752K111), were used in this study to investigate the effect of substrate roughness and
substrate temperature on the coatings adhesion strength. Table 2-2 lists their thermophysical
properties. Polymer sheets as received from the manufacturer are labelled smooth in the entire study.
Table 2-2 Properties of polymer substrate materials [[13], [30], [41], [42], [49]]
Density
(kg/m3)
Tm (0C) Glass-
transition
temp. (oC)
Thermal
Diffusivity
(mm2/s)
Specific
Heat
(J/kg-K)
Impact
Strength
(J/m)
Elastic
Storage
Modulus at
1 Hz
(MPa)
PTFE 2200 320-330 115-125 0.12 970 186.8 1377
PE 900 125-135 ~75 0.27 1900 896.8 1938
For the single splat study, sample dimensions of 25 mm x 25 mm x 3 mm were used. Samples were
cleaned using isopropyl alcohol (99%, Commercial Alcohols, Brampton, ON, Canada, P/N: 028668)
that was spread over the entire surface and allowed to dry to removes any contaminants [43].
The polymer sheets obtained from the manufacturer had roughness (Ra) values of approximately 0.20
± 0.05 µm, which are called “smooth” in this work. In order to study the effect of polymer surface
roughness on the adhesion strength of the coatings, the polymer surfaces were roughened by grit
blasting with #20 aluminum oxide (3418K46, McMaster-Carr, Grand Haven, Michigan, USA) using
15
a constant air pressure of 690 kPa (100 psig) and a nozzle-substrate stand-off distance of
approximately 100 mm (4 in). The surface was roughening by grit blasting with #20 aluminum oxide
(3418K46, McMaster-Carr, Grand Haven, Michigan, USA) using a constant air pressure of 690 kPa
(100 psig) and a nozzle-substrate stand-off distance of approximately 100 mm (4 in) was used. The
surface roughness of all the samples was measured before deposition using a skid-reference
profilometer (Precision Devices Inc., Michigan, USA, P/N: PDA-400ao) with at least 10
measurements taken for each sample. The average surface roughness (Ra) value after grit blasting was
1.60 ± 0.05 µm and these are labelled “rough” substrates.
2.3.2 Porous PE
All porous PE samples (Scientific Commodities, Inc., Arizona, United States, P/N: BB2062-35) used
in this work were manufactured by free sintering of ultra-high molecular weight polyethylene
powders. Porosity of the samples stated by the manufacturer was approximately 35-40% (void volume
percentage). The porous substrates were hydrophilic, and their mean pore size was about 70 µm. In
preparation for spraying of thick coatings (400 ± 20 µm), all samples were cut to 50 mm x 50 mm,
with a thickness of 3 mm. 6 mm thick samples were also used for conducting coating adhesion strength
tests. It should be noted that no surface preparation techniques were employed, and samples as
received from the manufacturer were used to deposit metals in this study.
2.4 Coating Cross-section Characterization
Coated polymer samples were mounted in epoxy resin, cut and polished. These samples were
examined using a scanning electron microscope (SEM) (Hitachi Tabletop, country, P/N: TM 3000) at
16
low voltage mode (5 V) to avoid specimen charging of the non-conducting polymer samples. To
observe the cross-section of single metal particles deposited on polymers the samples were cooled
with liquid nitrogen so that they became brittle and were easily broken by instigating a crack in the
sample [44]. A focused ion beam (FIB) was used (NB 5000 Dual Beam) to take sections through
particles embedded in the polymer. Parameters such as splat equivalent diameter, perimeter,
circularity and degree of splashing were calculated using image processing software, ImageJ-NIH.
2.5 Adhesion Strength
Pull tests (PosiTest™ AT-M Manual Tester, DeFelsko, St. Catherines, Ontario, Canada, P/N:
ATM20) on the continuous coating samples with dimensions of 50 mm x 50 mm was conducted to
determine the metal-polymer coating bond strength as per ASTM standard D4541. The bond strength
of the coatings deposited on the 3 mm and 6 mm thick porous polyethylene samples were compared.
A total of 5 samples of each Zn, Al and Cu coatings were tested. A standard ATM20 Al pull stub of
20 mm diameter was bonded to the coated samples using an epoxy adhesive (Devcon No.19770
‘plastic steel’ two-part epoxy, Aurora, Ohio, USA, P/N: 19770). Once the epoxy hardened, the pull
test was performed. The pull-off pressure range used for the samples was 0.5 - 20 MPa. The results
of bond strength of the coatings on PTEF and non-porous PE reported by Anand [48] were used in
this study for reference.
2.6 Electrical Resistivity
The electrical resistance of the deposited coatings was measured using a four-wire sensing method
with a 5.5-digit precision multimeter (Fluke 8808A, Fluke Electronics Canada LP, Mississauga, ON,
17
Canada). Four holes were drilled in the samples to insert aluminum bolts as shown in Fig. 2-3. The
aluminum bolts act as electrical ends. The wires connecting the multimeter were attached to the
aluminum bolts using alligator clips.
Figure 2-3. Experimental schematic of four-
wire sensing setup
Polymer
Coating
Al bolts as electrical
leads
18
Chapter 3
Single Splat Study of Zn and Al on PTFE and PE
3.1 Introduction
The following study investigates the effect of substrate temperature and surface roughness on coating
adhesion when using a high temperature thermal spray technique, wire-arc spraying to deposit
aluminum and zinc onto thermoplastic polymers, namely, polyethylene and polytetrafluoroethylene.
While previously published studies investigated the cold spray deposition behaviour of metals on
polymers by single particle impact experiments, in the present study, the interaction between the
molten metal particles and the polymers was explicitly studied. Single splats of metal on polymer
substrates were examined using SEM images. The single splat morphological characterization done
in this work will provide greater understanding of the process of impingement, penetration and
embedment of molten metal particles into the polymer substrates and provide a significant
contribution in defining why metals coat polymers. The adhesion strength results and cross-sectional
images of the coatings that are used in this study for reference are part of the work conducted by
Anand in his thesis [48].
The work presented in this chapter has been published in a journal, Surface and Coatings
Technology, under the title, ‘Thermal spray deposition of aluminum and zinc coatings on
thermoplastics.’
19
3.2 RESULTS AND DISCUSSIONS
3.2.1 Particle Size Distribution of Aluminum and Zinc
Figure 3-1 shows SEM images of aluminum and zinc spray particles that were captured by spraying
into a water bath. Most of the particles were irregular in shape and using a particle size analyzer the
mean particle diameters (d50) of aluminum and zinc were measured to be 78.6 um and 61.3 um,
respectively, which are typical for wire-arc spraying [45,46,47]. Zinc has a higher density and lower
surface tension than aluminum (see Table 3-1) and would be expected to fragment into smaller
particles.
Figure 3-1 SEM images of particles of (a) aluminum (mean particle diameter d50 = 78.6 µm) and (b)
zinc (d50 = 61.3 µm) captured by spraying into a water bath
3.2.2 Substrate Surface Temperature
(a) (b)
20
Individual metal splats were deposited on polymer substrates that were preheated before the start of
spraying. The goal was to raise the peak substrate temperature during spraying so that it approached
the glass transition temperature of the polymer, making it soft enough for impacting particles to
penetrate it. However, care had to be taken not to overheat the substrate and damage it.
The PTFE samples were heated at a rate of 5 °C/min and maintained at an elevated temperature of
about 95 °C for 5 minutes before aluminum was sprayed on it. When spraying zinc splats on PE the
substrates were preheated to 55°C (20 °C lower than the maximum operating temperature) before
spraying.
3.2.3 Adhesion Strength
Samples at Room Temperature
The adhesion strength of the coatings and the SEM cross-sections shown in Fig. 3-2 and Fig. 3-3
respectively were reported by Anand [48] in his thesis. It was found that zinc coatings adhered to both
PTFE and PE substrates, while aluminum adhered only to PTFE. No continuous coating of aluminum
could be applied on PE. Grit blasting the surface to increase roughness to 1.6 µm prior to coating
doubled the adhesion strength. Therefore, increased surface roughness promotes mechanical
interlocking of the coating material with asperities in the surface and therefore increases adhesion
strength.
Figure 3-3 shows SEM cross-sections of zinc and aluminum coatings on both smooth (Figs. 3-3a and
3-3b) and rough (Figs. 3-3c and 3-3d) PTFE substrates. In the rough substrates (Figs. 3-3c and
21
Figure 3-2 Adhesion strength of aluminum and zinc coatings on rough and smooth polymer
substrates. Aluminum coatings did not adhere to PE surfaces [48]
Fig. 3-3d) there is evidence of penetration of metal into crevices in the substrates, which would
enhance coating adhesion (see the areas outlined in Fig. 3-3d). The extent of penetration in the smooth
substrates (Fig. 3-3a and Fig. 3-3b) was comparatively less which resulted in coatings with lower
adhesion strength.
The size and shape of individual splats can affect adhesion strength [26, 27]. Figure 3-4 shows
individual splats of aluminum and zinc splats after being sprayed on smooth PTFE at room
22
temperature. 20-25 splats were chosen from each sample to calculate the equivalent diameter and
circularity of the splats. Separation lines were drawn manually around each splat to calculate the
equivalent diameter (D) and circularity of splats (C), respectively, as
Figure 3-3 Cross sectional view of (a) zinc coating on smooth PTFE, (b) aluminum coating on
smooth PTFE, (c) zinc coating on rough PTFE and (d) aluminum coating on rough PTFE substrates.
Areas of mechanical interlocking are outlined in (c) [48]
(a) (b)
(c) (d)
23
2A
D
= and (3-1)
2
4 AC
P
=
, (3-2)
where P and A are the perimeter and surface area of the splat, respectively [50]. The size of irregularly
shaped splats was expressed in terms of D, which was defined as the diameter of a circle with the
same area as the selected splat. Here the measured area corresponds to the area of a 2D image of the
splat shape. Splat circularity lies between 0 and 1. A circularity of 1 corresponds to a perfectly circular
splat, whereas splashed, fragmented and irregularly shaped splats typically have circularity values
closer to zero. Molten zinc particles spread out significantly more than those of aluminum: the average
equivalent diameter of zinc splats was measured to be 100 ± 10 µm, whereas that of aluminum splats
was 60 ± 10 µm, even though zinc droplets in the spray were smaller (see Fig. 3-1). Aluminum has a
surface tension value approximately 30% higher than zinc (see Table 2-1) which would restrict
spreading during impact. The circularity values of aluminum splats (0.41) were calculated to be higher
Figure 3-4 Single splats of (a) zinc and (b) aluminum sprayed on a smooth PTFE surface at room
temperature.
(a) (b)
24
than those of zinc (0.28), by about 45%, which also promotes adhesion strength [51]. As a result,
aluminum coatings on PTFE substrates had higher adhesion strength than zinc coatings. All SEM
images of single splats of metals on polymers at different surface temperatures and roughness values
can be found in Appendix A.
Increasing surface roughness makes it easier for metal splats to adhere to surface asperities. Figure 3-
5 shows zinc splats trapped in surface cavities on a PE substrate that were created by grit-blasting.
Strong interlocking between the metal and rough polymer increases adhesion strength of the coating.
Similarly, Liu et al. [22, 52] observed excellent mechanical interlocking between grit blasted polymer
matrix composites and zinc bond coating layer. Here, corundum powder was used for roughening the
polymer substrate and it was found that by optimizing the grit blasting parameters, coatings with high
bond strength could be achieved without damaging the substrate. Therefore, by increasing the
roughness of polymer substrates the bond strength of the coatings could be enhanced significantly.
Figure 3-5 SEM image of a zinc splat trapped in the asperities of a rough PE substrate
25
Aluminum was found not to adhere to PE substrates irrespective of the surface roughness. Figure 3-6
show cross-sections of smooth (Fig. 3-6a) and rough (Fig. 3-6b) PE substrates respectively after seven
passes of the spray torch. Aluminum particles adhered to both surfaces but did not form a continuous
coating. On the roughened PE (Fig. 3-6b) substrates, the aluminum splats were observed to adhere to
the asperities on the surface. Several more passes of the spray torch over the rough surface produced
a thin, porous coating with thickness less than 50 µm and delamination at the corners of the samples.
. A similar type of coating delamination was observed by Lie et al. [22] while arc spraying Cu onto
carbon fiber reinforced thermosetting polyimide. This occurred due to the high melting point of Cu
and low temperature resistance of polyimide.
Figure 3-6 Cross sectional view of PE substrates on which aluminum was sprayed. The surfaces
were (a) smooth and (b) rough. The surfaces are shown inclined to the horizontal so that both the
coated surface and the cross-section through the substrate can be seen
(a) (b)
26
The failure of aluminum particles, unlike those of zinc, to penetrate PE during impact may be due to
the high impact strength of PE (almost an order of magnitude greater than that of PTFE, see Table 2-
1) and the low density of aluminum (one-third that of zinc, see Table 2-1). As a result, cold spraying
of Al on thermoplastics is generally carried out under high pressures and temperatures to provide
greater impact force [54]. However, in the current study, melting of the substrates of the substrates
was observed when the stand-off distance was reduced to increase particle velocity. This was due to
the increase in substrate temperature at low stand-off distances. Additionally, aluminum also has a
higher melting point than zinc, which may have caused local melting and decomposition of PE under
impacting particles, preventing adhesion. The temperature of the polymer substrate during Al particle
impact was also found to be consistently above 80 0C (maximum operating temperature of PE)
throughout the spraying period [48]. This could have also resulted in poor adhesion of aluminum onto
PE substrates.
Samples at Elevated Temperatures
Preheating smooth polymer substrates before the start of spraying was found to enhance adhesion
strength by roughly the same amount as was achieved by grit blasting them to make them rougher.
Figure 3-7 shows the adhesion strength of aluminum and zinc coatings on smooth polymer substrates
measured by Anand [48], both for those sprayed at room temperature and those sprayed on preheated
substrates (95°C for PTFE and 55°C for PE).
Pre-heating the substrates makes them softer and allows impacting molten metal droplets to penetrate
them before solidifying. The adhesion strength of zinc was approximately the same on both PTFE
27
and PE substrates at room temperature. Preheating the substrates increased adhesion strength
significantly since it promoted penetration of metal particles into the polymer. Figure 3-8a shows a
cross-section through a 260 ± 20 µm thick zinc coating formed after 7 passes of the spray torch over
a pre- heated PE surface showing good bonding at the metal-polymer interface. A very similar bonding
line as shown in Fig. 3-8a was observed when Che et al. [28] cold sprayed Cu on PEEK with a gas
temperature above the Tg of PEEK. Here, the samples were not pre-heated but the high gas temperature
(425 0C) combined with the extremely high velocity of Cu particles aided in achieving excellent
Figure 3-7 Adhesion strength of aluminum and zinc coatings on polymer substrates that were
initially either at room temperature or preheated (initial temperature 95°C for PTFE, 55°C for
PE) [48].
28
Figure 3-8 Cross sectional view of preheated PE substrates on which zinc was sprayed. (a)
continuous coating formed after 7 passes of the spray torch and (b) individual particles deposited after
1 pass of the spray torch
adhesion with the substrates through particle embedment. However, in the current study, owing to the
comparatively low particle velocity and the molten state of the spray particles in the wire-arc spray
process [46,47], the substrates had to be pre-heated to promote such particle penetration into the
substrate. Figure 3-8b shows a cross-section through the interface after a single pass of the torch,
showing deep penetration of metal particles up to a maximum depth of 20 µm. King et al. [53]
previously observed copper particles penetrating to a depth of up 50 µm when a cold spray was used
to deposit them on HDPE surfaces. A portion of the zinc particles in Fig. 3-8b remained protruding
above the surface so that they could serve as anchors for subsequent splats that coalesced with them.
When Ganesan et al. [8] cold sprayed copper particles on PVC, they observed similar kind of particle
(a) (b)
29
embedment phenomenon through FIB bisection of a copper particle on the substrate. It was found that
such bonding mechanisms helped in developing thick copper coatings on PVC.
Figure 3-9 shown splats of zinc after impact on PE surfaces that were preheated to 55°C before
spraying. Figure 3-9a shows a view from above, in which a splat appears to be partially buried in the
polymer, showing how preheating promoted penetration of the metal into the substrate and good
adhesion. Figure 3-9b shows splats that were sectioned using a focused ion beam to create a trench in
the substrate. Splats can be seen both above and below the surface of the PE substrate, showing the
deep penetration of the splats. This kind of deep penetration provided excellent interlocking with the
substrate. As a result, these initial set of splats served as an interlayer between PE and the subsequent
molten Zn particles impacting the substrate, promoting overall adhesion of the coatings. Some of the
Figure 3-9 Single splats of zinc on PE preheated to 55°C (a) viewed from above and (b) section
through substrate made using a focused ion beam (FIB)
(a) (b)
30
particles appear to be almost completely inside the substrate. These particles could have travelled with
a comparatively higher momentum, resulting in embedment of the whole particle in the substrate [30].
The adhesion strength of aluminum on PTFE increases with substrate temperature, but it still did not
adhere to PE. Figure 3-10 shows cross-sectional view of pre-heated PE substrates sprayed with
aluminum after one (Fig. 3-10a) and seven (Fig. 3-10b) passes of the spray torch. The splats are
flattened out and fragmented but show no evidence of particles piercing the substrate. Since there was
no initial penetration there was no way for mechanical interlocking of the coating with the substrate
to occur. Aluminum particles could also have caused local melting and decomposition of polymer
substrates during impact due to
Figure 3-10 Cross sectional view of preheated PE substrates on which aluminum was sprayed. (a)
after 1 pass and (b) after 7 passes of the spray torch
(a) (b)
31
the very high particle temperatures created by the exothermic oxidation of aluminum in air.
Ashrafizadeh et al. [12] observed such decomposition at locations that were directly exposed to the
high temperature flame while depositing Al-12Si on polyurethane substrates using flame spraying.
3.3 Conclusions
Coatings of aluminum and zinc were applied on two thermoplastic materials, PE and PTFE, using
electric wire-arc spraying. Single splats of metal were observed after impact to understand the
interaction of molten metal droplets (Al and Zn) with polymer substrates. Substrate roughness was
increased using grit blasting. The temperature of the polymer substrates was monitored during
spraying to understand how it affects the adhesion of metal particles.
Zinc coatings, about 260 µm thick, formed on both polymer substrates irrespective of their surface
roughness. Increasing surface roughness significantly enhanced adhesion strength since it promoted
mechanical interlocking of the metal with surface asperities. Increasing the initial substrate
temperature so that the maximum substrate temperature during spraying reached the glass transition
temperature of the polymer also enhanced adhesion strength. Individual zinc splats were observed to
be buried deep inside heated polymer substrates.
Aluminum adhered only to the PTFE but not PE. Aluminum particles have low density and therefore
may not have enough momentum to penetrate PE, which has a much higher impact strength than
PTFE. It is also possible that impacting droplets of aluminum, which have a higher melting point than
zinc, caused localized melting of the substrate upon impact and therefore could not flow into surface
32
cavities and freeze, which would have led to mechanical bonding.
The results from this work provide insight into the interaction of molten metal particles with
thermoplastic polymer substrates. It therefore provides a step-change in understanding why thermal
sprayed metals coat polymer surfaces. Knowledge of this metal-polymer interaction mechanism
eventually aids in selecting appropriate polymer surface modifications techniques to achieve a strong
bond between coatings and the substrate using high-temperature thermal spray coating processes.
33
Chapter 4
Metallization of Porous Polyethylene
4.1 Introduction
In this study dense coatings of zinc, aluminum and copper were deposited on polyethylene, a low
melting temperature and lightweight polymer, using a high temperature thermal spray technique,
wire-arc spraying. This was made possible by using porous polyethylene as the substrate to be
coated. While previously published studies modified the surface of the polymer substrates to
achieve good adhesion with the thermally sprayed metallic particles, in the present work, no prior
surface preparation techniques were employed. The adhesion strength of metal coating on porous
surfaces was measured by conducting standard pull tests and the coating-porous polymer interface
was examined using SEM images. Individual metal splats on the polymeric surface were also
analyzed using the SEM to provide insight into the interaction of molten metal particles with the
porous polymer during impact. A thin layer of Inconel 625 was also deposited to study the
feasibility of spraying a metal with a very high electrical resistance on the porous polymer. The
electrical resistivity of all the metallic coatings were measured using a four-point probe method.
Finally, the splat morphologies and adhesion strength of the coatings on porous and non-porous
polyethylene was compared. The results obtained in the previous chapter was used for these
discussions.
34
4.2 Results and Discussions
4.2.1 Size Distribution of Spray Particles
The SEM images (Fig. 4-3) revealed the size and shape of the different spray particles. Al and Zn
particles were observed to have a more irregular form compared to the Cu particles, which were
approximately spheroidal. The mean particle sizes (d50) of Zn, Al and Cu were measured to be
61.3 µm, 78.6 µm and 40.8 µm, respectively using a particle size analyzer. The difference between
the mean particle size and the mean pore size of the substrate was less than 10 µm in the case of
Al and Zn. Cu particles on the other hand were almost about 30 µm smaller in diameter than the
pores on the polymer surface.
4.2.2 Substrate Roughness
The polymers received from the manufacturer had mean surface roughness (Ra) values of
approximately 3.4 ± 0.3 μm (n = 10). Sintering of the polymer powders during manufacture
(b) (a)
35
resulted in an uneven surface topography which can be seen in the SEM image of the bare
polyethylene surface (Fig. 4-4). It should be noted here that the surface roughness of porous
Figure 4-4 SEM image of as-received porous polyethylene surface
(c)
Figure 4-3 SEM images of particles of (a) aluminum (mean particle diameter d50=78.6 µm), (b) zinc
(d50=61.3 µm) and (c) copper (d50=40.8 µm) captured by spraying into a water bath
(a) (b)
36
polyethylene was significantly higher than that of the non-porous polyethylene substrates (Ra = 1.5
± 0.05 μm for n = 10), which were roughened by grit blasting [48].
4.2.3 Coating Cross-Section Characterization
The interface between the metallic coating and polymer substrate is shown in Fig. 4-5. An average
coating thickness of about 400 ± 20 µm was achieved after approximately 9 to 10 passes of the
spray torch over the polymer surface, with the three metals used in this study. Zn, Al and Cu
coatings were all observed to penetrate deep into the pores present on the surface. Penetration
depth of the coatings into the pores was even over 50 µm (from the surface) in certain regions,
thus providing excellent mechanical anchorage. The smaller particle size of Cu (40 µm) enabled
such deep penetration into the pores which was not observed in Al and Zn coatings. The porous
structures provided a strong mechanical bond between the metals and the polyethylene substrates,
thereby eliminating any requirement for substrate preparation prior to thermal spraying.
4.2.4 Single Splat Characterization
The surface topography (SEM image) of the porous substrates with individual metal splats at
different magnification is shown in Fig. 4-6. Analysis of these images provided a fundamental
knowledge on the behavior of three different molten metal particles, having different melting
temperatures, on porous polyethylene during impact. A majority of the Cu spray droplets with
mean size of about 40 µm were observed to penetrate the surface pores (order of 70 µm) and adhere
to the polymer.
37
Figure 4-5 Cross sectional view of metallic coating at low and high magnification, (a) and (b)
zinc coating, (c) and (d) aluminum coating, (e) and (f) copper coating on porous polyethylene
substrates. Areas of mechanical interlocking are outlined in (e) and (f).
(a) (b)
(c) (d)
(e) (f)
Mounting Resin
Porous Polyethylene
Mounting Resin
Porous Polyethylene
Mounting Resin
Porous Polyethylene
Zn
Al
Cu
>50 µm
38
(c) (d)
(e) (f)
(a) (b)
Figure 4-6 SEM image of single splats at low and high magnification, (a) and (b) zinc, (c) and (d)
aluminum, (e) and (f) copper
39
On the other hand, the spray particles of Zn and Al with mean size of about 60 µm and mean size
of about 40 µm were observed to penetrate the surface pores (70 µm) and adhere to the 80 µm
respectively, were found to be distributed on both the porous and non-porous regions on the
substrate surface. Cu droplets penetrated the pores more frequently due to their comparatively
smaller size. Additionally, Al splats stuck to the crevices on the polymer surface more often than
Zn and Cu (highlighted in Figs. 4-6c and 4-6d). These kind of adhesion mechanisms enhance the
coating bond strength by providing strong mechanical interlocking which usually is achieved by
sandblasting the coating substrates prior to thermal spraying [55]. Splats that were significantly
smaller than the initial spray droplet size exhibited splat fragmentation (indicated by arrows in
Figs. 4-6b and Figs. 4-6f). Splat-breakup was a common phenomenon observed in all the three
cases. This was possibly due to the high surface roughness of the substrate that promoted splashing.
McDonald et al. [56] showed that increasing the surface roughness of the coating substrate
promoted splat splashing, restricted splat spreading and thus resulted in splats with skewed
morphologies. Zinc and aluminum splats that landed directly on the non-porous locations on the
polymer surface was observed to spread with fingers radiating out from their periphery. Splat
spreading in the case of Cu was found to be restricted largely due to the presence of Cu particles
inside the pores, where there was comparatively less room for spreading.
To further understand splat morphology, the solidification parameter (SP) for the three metallic
splats were determined. Dhiman et al. [57] defined SP as the ratio of the thickness of the solid
layer formed on the splat before it attained maximum spread (t) to the final thickness of the splat
(H). They also carried out a 1D heat conduction model to determine the SP. Figure 4-7 shows a
spray droplet with diameter D and velocity Vi impacting a polymer substrate and spreading into a
40
splat of uniform thickness H. It was found that, if the solid layer grew by a significant amount
during spreading (SP∼0.1 to 0.3), it would restrain the splat from spreading and becoming thin
enough to rupture, and therefore would produce a disk-shaped splat. On the other hand, if the
solidification was very rapid (SP>>0.3), they observed a solid ring that form around the edges of
the spreading droplet. This obstructed the outward flowing liquid and destabilized it, resulting in
a splat with fingers radiating out from its periphery. In the present work, an identical model was
used to calculate the SP. The value of this parameter depends on the thermal contact resistance
(Rc) between the splats and the substrate. In this study, the SP was determined assuming Rc is of
the order 10-5 m2-K/W. This was chosen based on Rc values calculated in previous studies
involving splat cooling rate determination in thermal spray processes [58], [59]. In these studies,
Substrate
Solid layer of thickness ‘t’
Spray
droplet
Vi
H
D
t
Figure 4-7 Idealized splat formation on substrate
41
the thermal contact resistance values were found to depend on the substrate temperature. The
resistance values that corresponded to non-heated substrates in their study were used in the current
study for calculating SP. The SP values for the three metallic splats are shown in Fig. 4-8. Zinc
and aluminum splats were observed to have similar SP while SP for Cu was found to be
significantly higher in comparison.
Therefore, based on the definition of the solidification parameter and the SP values obtained for
the three metals in the present work, it can be concluded that Al and Zn splats had enough time to
flatten out and spread (shown in Figs. 4-6b and 4-6d) before solidifying completely. On the other
hand, SP values for Cu splats suggested that they experienced rapid solidification during impact.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
Zinc Aluminum Copper
Soli
dif
icati
on
Para
met
er
Figure 4-8 Solidification parameter values for Zn, Al and Cu on porous PE, assuming a
thermal contact resistance (Rc) value in the order of 10-5 m2-K/W
42
As a result, the initial spray droplets of Cu, which penetrated the pores, froze much faster than the
other two metals. Splat spreading was highly restricted in this case, and the splats assumed the
shape of the pores during solidification. The MATLAB code that contains all equations used in
the calculation of solidification parameters can be found in Appendix B.
4.2.5 Single Splats on Porous and Non-Porous Polyethylene
The average size of metallic splats on porous and non-porous polyethylene substrates is shown in
Fig. 4-9. The average diameter of the all the three metallic splats was found to be higher in the case
of non-porous polyethylene. The comparatively smaller splat size observed in porous polyethylene
0
10
20
30
40
50
60
70
80
90
100
Zn Al Cu
Part
icle
Siz
e (µ
m)
Initial Particle Size
Non porous PE
Porous PE
Figure 4-9 Average size of particles sprayed into a water bath and splats deposited on PE
43
could be due to their uneven surface morphology created by the presence of pores on the surface.
The surface roughness Ra of the porous polyethylene substrates was about 3 µm higher than that
found on non-porous polyethylene. The increased roughness caused more splat fragmentation
which resulted in the smaller size of splats in porous substrates.
The circularity of splats sprayed on to both porous and non-porous polyethylene substrates are
shown in Fig. 4-10. The splat circularity of copper was found to be significantly higher than both
aluminum and zinc. This can also be observed in the SEM images of the single splats on
polyethylene shown in Fig. 4-11. This could be due to the higher solidification parameter value of
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Zn Al Cu
Sp
lat
Cir
cula
rity
Non porous PE
Porous PE
Figure 4-10 Circularity of Al, Zn and Cu splats
44
Cu. The solidification of copper could be rapid enough to prevent it from spreading out unlike
aluminum and zinc. The other reason for this kind of behavior could be the vaporization of the
polyethylene substrate due to the higher melting point of copper. This could have caused the splats
to disintegrate during solidification leaving behind only the central solidified core [26]. Splat
surface area and perimeter values of copper splats measured using ImageJ software can be found
(a) (b)
Figure 4-11 Splats of (a) Al, (b) Zn and (c) Cu on non-porous polyethylene
(c)
45
in Appendix C.
4.2.6 Adhesion Strength
All pull tests conducted on the 3 mm thick polymer samples resulted in the fracture of the porous
polyethene substrates before coating detachment. Therefore, the adhesion strength of the coatings
was higher than the yield strength of the polymer substrates. The samples consistently fractured
during the tests at a pull off pressure of about 3.2 ± 0.4 MPa (n = 5), regardless of the metal used
to coat the polymer. This type of fracture lead to the formation of a through hole in the sample as
shown in Fig. 4-12. Since the rigidity of the substrates affects the pull strength results, pull tests
were also conducted on thicker polyethylene substrates (6 mm) coated with zinc, aluminum and
copper. The pull tests conducted on the 6 mm samples also resulted in fracture of the polyethylene
substrates. However, the fracture did not lead to complete removal of the polyethylene material
unlike the 3 mm sample. The pull off stress values of the three metallic coatings on the 6 mm
samples are shown in Fig. 4-13. It should also be noted that
Fractured pull
test region
Al coated PE
adhered to pull stub
Figure 4-12 Adhesion test sample (3 mm)
46
all pull stress values measured were on the coated porous polyethylene substrates were
significantly higher than the adhesion strength of the Zn coating on non-porous polyethylene (0.7
± 0.05 MPa for n = 5) [48].
4.2.7 Coating Electrical Resistivity
Electrical resistivity (ρ) is a material property of the coating and is not a function of the physical
dimensions of the coating. The electrical resistivity of the coatings was calculated according to
the following equation,
0
0.5
1
1.5
2
2.5
3
3.5
4
Pu
ll o
ff P
ress
ure
(M
Pa)
Zn
Al
Cu
Figure 4-13 Pull off stresses of zinc, aluminum and copper coated porous polyethylene
substrates (6 mm)
47
Rwt
l = (4-1)
where R, w, l, and t are the electrical resistance, width, length, and thickness of the coating, s,
respectively.
Figure 4-14 summarizes the electrical resistivity of zinc, aluminum and copper coatings deposited
on the porous polyethylene substrates. The electrical resistivity of the coatings was measured at
two different coating thicknesses. To compare the electrical resistivity of the coatings with bulk
metals, metallic foils of similar thickness were chosen, and four-point tests were conducted on
them too. The measured values were in good agreement with available literature [60]. The
electrical resistivity of the metallic coatings was affected by the coatings’ thickness. With increase
0.00E+00
2.00E-08
4.00E-08
6.00E-08
8.00E-08
1.00E-07
1.20E-07
1.40E-07
1.60E-07
1.80E-07
Zn Al Cu
Ele
ctr
ical R
esis
tivit
y (Ω
m)
CP Metal
300 um
500 um
Figure 4-14 Electrical resistivity of the coatings (n = 4)
48
in coating thickness from 300 µm to 500 µm, the coatings electrical resistivity was observed to
decrease significantly for all the three cases in accordance to Eq. 4-1. This could be due to higher
compaction in thick coatings (500 µm) which can reduce the number of pores that may be present
within the coating layer. The bulk metal electrical resistivity was however found to be an order of
magnitude lower than that of the thermally sprayed coatings. This could largely be due to the
oxidation of metallic particles in air during thermal spray and the associated porosity in the
coatings [46,47]. The aluminum and copper coatings had comparatively higher electrical resistivity
owing to their greater affinity towards oxygen than zinc [61].
Figure 4-15 shows the SEM images of the cross-section and surface topography of Inconel 625
coating deposited on the porous polymer. The pores on the polymer surface aided in achieving
excellent mechanical interlocking of In-625 with the substrate. As a result, In-625 coatings of
thickness up to 200 µm could be deposited. However, the coatings were porous which affected the
electrical resistance of the coatings.
Mounting Resin
Porous Polyethylene
Inconel 100 µm
Pores
Figure 4-15 In-625 Coating, (a) top surface and (b) cross-section
(b) (a)
49
The electrical resistivity of the coatings was observed to decrease with increase in the coating
thickness from about 100 µm to 200 µm. This could be due to the difference in the coating
densities. A thin Inconel-625 sheet (0.5 mm) was used a reference material to compare the
electrical resistance of the coatings with the bulk. The electrical resistivity of the 200 µm thick
coating was found to be an order of magnitude less than the bulk compound. This could be
attributed to the oxidation and porosity of the coatings deposited by thermal spray.
In 625 Bulk 100 ± 15 µm 200 ± 20 µm
Resistivity
(Ohm-m)
1.29E-06 2.40E-04 3.70E-05
4.3 Conclusions
Dense metallic coatings were deposited on porous polyethylene with 70 µm average pore size,
using an electric wire-arc spray system. Zn, Al and Cu coatings were all observed to penetrate
deep into the pores, creating excellent mechanical bond with the substrates. The bond strength of
all three metallic coatings was found to be higher than the yield strength of the porous polymer
which fractured before the coating detached during pull tests. The coating adhesion strength of all
the three metals on the porous polyethylene samples was found to be significantly higher than the
coatings deposited on non-porous polyethylene substrates. Cu spray droplets, which were much
smaller than pores on the substrate surface, were mostly observed to penetrate pores. They
Table 4-1 Electrical resistivity of In-625 coatings of varying thickness
50
experienced rapid solidification on impact, which restricted splat spreading. Zn and Al splats were
observed to spread to a comparatively greater extent and adhere to both porous and non-porous
regions. Therefore, by increasing the void volume or the size of the pores, an even lighter metal-
polymer composite material can be fabricated which has tremendous potential to develop cooling
solutions for consumer electronics.
51
Chapter 5
Fabrication and Characterization of Lightweight Heat Sinks
5.1 Introduction
In this chapter, the fabrication of lightweight heat sink materials Light Emitting Diode (LED)
cooling applications will be discussed. The heat generated by the LEDs during operation was first
calculated and then the heatsink design and materials were selected accordingly. In this study, the
lightweight heat sinks were designed using polymer composite materials and aluminum foils. Thin
aluminum foils with thickness in the range of 200-500 µm were tested. The composite heatsinks
were made by depositing thin aluminum and copper layer on porous PE using wire-arc spray
process. The thin thermally conductive aluminum/copper layer enhances heat conduction. Polymer
composite heatsinks were mounted on the LEDs to keep their temperatures below operational
range. The cooling performance of the different heatsink models were then characterized. The
weight and performance of these composite heatsinks were then compared with a bare aluminum
heatsink material (reference), which is widely used to cool LEDs.
5.2 Experimental Method
5.2.1 Experimental Assembly for Thermal Power Measurement of LED
Strips
LEDs are not 100 % efficient at converting input power to light. Some of the energy is converted
into heat and must be transferred to the ambient air. All LED strips used in this study were rated
at 46 W. The total heat generated by the 46 W LED strips was measured using the experimental
52
setup shown in Fig. 5-1. A 180 mm long bare LED board was mounted on an Aluminum block
(heatsink) using a thermally conductive tape (Heat-Transfer Mounting Tape, Polyester Plastic,
McMaster-Carr, Sante Fe Springs, California, USA, P/N: 1761N11). The bottom surface and sides
were then thermally insulated using a high-temperature insulation material (Pyrogel® XTE, Aspen
Aerogels, Northborough, Massachusetts, USA) as shown in the figure. The LED was turned on and
allowed to reach steady state conditions. This took approximately 50 minutes. The steady-state
temperature was noted. In order to determine the thermal output of the LED, a strip heater (pseudo-
LED) was substituted in place of the LED. The heater power was varied with a variable transformer
until the heatsink reached the same temperature as the real LED. These experiments were then
repeated 5 times to ensure consistency.
Heat dissipation direction
LED strip Contact
material
DAQ Computer
Al Block (heat sink)
Thermal insulation
Thermocouple wire
Power Supply
Figure 5-1 Experimental setup for heat generation estimation
53
5.2.2 Fabrication of Al Foil Heatsinks
The surface area of the heatsinks is the most important parameter that affects its thermal
performance. Typically, the heatsinks with large surface area dissipate more heat compared to
heatsinks with comparatively lower surface area [62]. The thickness of the heatsinks can
therefore be minimized without significantly affecting its performance while designing
lightweight heatsinks. In this work, lightweight heatsinks were fabricated by using thin
aluminum/copper foils with different thicknesses.
Three different heatsink models were designed using the foils. In the first design, an aluminum
foil (200 µm) was wrapped around porous polyethylene to enhance heat conduction. The second
and third heatsink design comprised of a thick Al foil (0.5 mm) fixed into a 3D printed ABS
frame as shown in Fig. 5-2.
Figure 5-2 Lightweight heat sink models, (a) aluminum wrapped around polyethylene, (b) aluminum
foil in ABS frame and (c) copper foil in ABS frame
(a) (b) (c)
54
5.2.3 Fabrication of Polymer Composite Heatsink
The composite heat sink comprised of a porous polyethylene core of thickness 3 mm which acted
as the support structure. A thin (400 ± 20 µm) thermally conductive metallic coating of aluminum
and copper were deposited on both sides of the polymer using an electric wire-arc spraying process
(shown in Fig. 5-3). The metallic coatings provided the thermal conduction path and the polymeric
base helps in achieving weight reduction in the heatsink.
5.2.4 Experimental Assembly for Temperature Distribution Measurement in
Heatsinks
The 46 W LED fixtures were suspended on an aluminum profile with the lights facing down as
shown in Fig. 5-4. Thermocouple wires were attached to the heat sinks using thermal paste (Heat
Sink Compound, McMaster-Carr, Sante Fe Springs, California, USA, P/N: 76645A14) as shown
below in order to acquire local temperature measurements. The LEDs were turned on and the
temperature at different locations were recorded using a DAQ until steady-state conditions were
reached. All temperature measurements were recorded in units of ⁰ C.
Porous Polyethylene
Al coating layer
Fig.5-3 Cross-sectional view of Al-PE composite heat sink
Al coating layer
55
5.3 Analytical Heat Conduction Model
The cross section of the composite heatsink is shown in Fig. 5-5. The composite fin was
constructed as a sandwich structure with a polymer core with thickness H and Al coating layer
with thickness t, which is in perfect thermal contact with the polymer.
IR
Camera Computer
Heat Sink
46 W LED strip
Al Profile support
structure
DAQ Thermocouple wire
Suspended LED strip
T1
T2
T3
T4
T5
T7
T6
Figure 5-4 (a) Bottom view of 46 W LED-heat sink assembly showing temperature measurement locations, (b)
Top view of heat sink
(a) (b)
10 in
3 in 1 in
7 in
56
The thermal conductivity of the polymer core and the metallic coating layers is k1 and k2,
respectively; at X = 0, the temperature is uniform (T1 = T2 = Tbase) while the outer surface X = L
is insulated The natural convection heat transfer coefficient h is estimated using the experimental
correlation [64] for an upward-facing horizontal plate,
0.25 4 7Nu = 0.54Ra , 10 < Ra < 10 (5-1)
Nu is the Nusselt number and Ra is the Rayleigh number defined as
( ) 3
2
base airg T TRa Pr
−= (5-2)
h
Nuk
= (5-3)
where Pr is Prandtl number, δ the characteristic length of the heatsink, β the coefficient of volume
expansion, ν the kinematic viscosity of the fluid, and k the thermal conductivity of the heatsink.
Thermal properties were estimated at an average temperature, base airT + T
2.
Heat conduction in the two-dimensional polymer composite heatsink is governed by the following
partial differential equations,
2 2
1 2
2 20
d T d T
dx dy+ = (5-4)
Polymer (1)
Y
X
H
t
h, Tair
= 22 0C
Tbase
Figure 5-5 Analytical heat conduction model of the composite fin consisting of two domains
Al coating layer (2)
L
57
2
2
dT= 0
dx (5-5)
The following boundary conditions were applied to equations 5-4 and 5-5.
At 1 2 0, basex T T T= = = 1 2 0, basex T T T= = =
At x = L, 1 2dT dT= 0, = 0
dx dy (adiabatic fin tip)
At 1 21 2 1 H, x 2 H, x
dT dTy = H, T = T ; k = k
dy|
y|
d
At ( ) 22 x, H+t 2 air| ( )
dTy = H +t , - k = h T -T
dy
Now, performing an energy balance gives,
2
2
2
d θ- m θ = 0,
dx (5-6)
where airθ = T -T and ‘m’ is the fin parameter
Thus, the fin equation is,
base air2 air
T -TT (x)= coshm(L - x) +T
coshmL (5-7)
where 1 2
hm =
k H +k t (5-8)
This model was then verified with the experimental results obtained. This can be seen below in
Fig. 5-7.
58
5.4 Results and Discussions
5.4.1 Heat Generation in LEDs
The thermal power dissipated by the 46 W LED fixtures was 29.3 ± 0.6 W. An average of 35% of
energy consumed was converted to light and the rest 65% was dissipated as heat. This amount of
heat generated is typical of all LEDs [63].
5.4.2 Surface Temperature Measurement and Prediction
Figure 5-6 shows both measured and calculated temperature distribution on the top surface of the
heatsink at steady state conditions. The experimental surface temperature profile is determined by
allowing the LED strip-heatsink assembly to reach steady state conditions and then recording the
temperature distribution using the IR camera and local thermocouple temperature measurements.
The temperature is maximum closest to LED strip where heat is generated and decreases radially.
The calculated heatsink surface temperature distribution is a function of the base temperature, as
seen in Eq 5-7, and is independent of the power input. The predictions agree well with the
experimental measurements. At regions far away from the LED strip, the measured temperature is
slightly higher (< 2 0C) than the prediction, but the difference is of the same magnitude as the
associated experimental uncertainty of the temperature measurement.
59
5.4.3 Heatsink Fin Efficiency
The fin efficiency (ŋfin) is an important heatsink design parameter which helps in optimizing the
size of the fins. It is defined as the ratio between the actual heat transfer rate from the fin to the
ideal heat transfer rate from the fin if the entire fin were at the base temperature. The fin efficiency
for a straight rectangular fin which was used in this study is given by,
tanhfin
mL
mL = , (5-9)
where m is the fin parameter and L is the fin length.
The effect of coating thickness and fin length on the fin efficiency is shown in the Fig. 5-7. The fin
efficiency was observed to increase with increase in the coating thickness. This is because the fin
efficiency is a function of the fin parameter m which decreases as the coating thickness increases.
20
25
30
35
40
45
50
0 0.008 0.016 0.024 0.032
Tem
per
atu
re (
0C
)
Distance from Heatsink Base (m)
Experimental Data
Analytical Data
Figure 5-6 Analytical temperature curves of the composite heatsink (Equation 5-7) fits the
experimental results
60
The fin efficiency was found to increase with decrease in the fin length. Heatsinks with 0.5 in fins
on both sides of the LED strip had the highest fin efficiency assuming the same coating thickness.
This is because the temperature drops along the fin exponentially and reaches the environment
temperature at some length. Therefore, increasing fin length beyond this point does not help in
heat transfer and only adds weight and cost. The fin model also predicted that the improvement in
fin efficiency would be insignificant beyond a coating thickness of about 1 mm.
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 0.2 0.4 0.6 0.8 1
Fin
Eff
icie
ncy
Coating Thickness (mm)
Existing FinLength_1in
FinLength_0.5in
Fin Length_2in
Figure 5-7 Fin efficiency of the composite fin varies with coating thickness and fin
length
Fin
Length_1in
61
5.4.4 Performance Comparison of Bare Aluminum and Polymer Composite
Heatsink
The polymer composite heatsinks used in this study are shown in Figure 5-8. A thin (400 µm)
thermally conductive aluminum/copper layer present on porous polyethylene aided in enhancing
heat conduction. A bare, commercially pure aluminum plate (see Fig. 5-9) with dimensions of
180 mm x 45 mm x 3 mm was used as the reference heatsink material for performance
comparison purposes. From the analytical model, the dimensions of the polymer composite
heatsink that would yield similar temperature values as the bare aluminum heatsink was
identified. The composite heatsinks used in this study had dimensions of 250 mm x 80 mm x 3
mm.
To evaluate the performance of the polymer composite heatsinks, the base temperature of the
Figure 5-8 LED strip mounted on polymer heat sink coated with (a) aluminum and (b) copper
(a) (b)
62
heatsinks at steady state conditions was noted. The base temperature values of the heatsinks and
their weight is shown in Fig. 5-10. Both the composite heatsinks performed better than the bare
aluminum heatsinks for the dimensions considered in this study. By replacing the bare aluminum
heatsink with the aluminum coated polyethylene material, a weight reduction of about 65%
could be achieved for similar thermal performance. Whereas by using copper coated
polyethylene heatsinks, a significant reduction in base temperature ( 3 0C) was observed. This
was due to the comparatively higher thermal conductivity of copper compared to aluminum.
However, the high density of copper resulted in much lesser weight reduction ( 25%) in the
heatsink as can be seen in Fig. 5-10.
Figure 5-9 LED strip mounted on a bare aluminum plate (reference
heatsink material)
63
5.4.5 Performance of Aluminum Foil Heatsink
The base temperature and the weight of the three lightweight heatsink models along with the
reference heatsink material are listed in Table 5-1. The base temperatures of the lightweight
heatsink were predicted from the analytical heat conduction model. The heatsink dimensions
(250 mm x 80 mm x 0.5 mm) were chosen based on the predicted temperature values. The model
predicted similar thermal performance in the lightweight heatsink models when they were of the
dimensions mentioned above.
It can be seen from the table below that by wrapping aluminum foil around polyethylene, weight
reduction of about 33% was achieved in the heatsinks. By removing the polymer core significant
0
50
100
150
200
250
300
41
42
43
44
45
46
47
48
49
Bare Al heatsink
used by Zortech
Al coated
polymeric
heatsink
Cu coated
polymeric
heatsink
Weig
ht
of
heats
ink
(g)
Base
Tem
peratu
re (
0C
)
Base temperature at steady state
(deg C)
Weight of heatsink (g)
Figure 5-10 Performance of lightweight polymer composite heat sinks
64
weight reduction can be achieved. This was observed in the second heatsink design where a
polymer frame was used to hold the aluminum, thereby eliminating the use of a polymer core
for structural support. With this design, weight reduction of about 70% was achieved in the
heatsinks for similar thermal performance as the bare aluminum heatsink. When the aluminum
foil was replaced with a copper foil, the base temperature of the heatsink reduced by about 4 0C
due to the higher thermal conductivity of copper.
The performance of the aluminum foil was further characterized by changing the length of the
heatsink fins. The aluminum foil heat sink models that were used in this study are shown in Fig. 5-
11. The heatsink base temperature and their weight for different fin lengths are shown in Fig. 5-
12. It can be observed that the temperature of the heatsink base reduces with increase in fin length.
This reduction was however found to be very meagre beyond a fin length of about 2 inches.
Bare Aluminum
Heatsink (3 mm)
– Reference
Material
Al foil (200 µm)
wrapped around
polyethylene
Al foil (500 µm)
fixed in ABS
frame
Cu foil (200 µm)
fixed in ABS
frame
Base
Temperature
(0C)
48 ± 1 48 ± 1 47 ± 1 44 ± 1
Weight of
Heatsink (g)
270 177 83 85
Table 5-1 Temperature and weight of lightweight heat sink models
65
72% reduction in weight of the heatsinks was achieved with the 2 in fin design for similar thermal
performance as the reference aluminum heatsink considered in this study. There existed a trade-
off between thermal performance and weight of the heatsinks. The results shown in figure 5-10
aided in optimization of these two parameters. Significant reduction in the heatsink base
temperature (about 3 0C) was observed by increasing fin length from 1 to 1.5 in. The increase in
heatsink weight due to added material was however still much lower than the reference heatsink
material. It should also be noted that the thermal performance of the heatsink does not improve
substantially by increasing the fin length from 1.5 in to 2 in. As a result, it only adds more weight
for minimum improvement in its performance. Therefore, the heatsink with 1.5 in was found to
both economically feasible and thermally efficient. Figure 5-11b shows a compact model of the
Figure 5-11 LED strip mounted on 0.5 mm thick Al foil as heat sink, (a) flat and (b) U-shaped
(a) (b)
66
heatsink with the same weight as the flat aluminum foil heatsink.
5.5 Conclusions
Lightweight heatsink materials for LED cooling applications were designed using aluminum foils
and metal-polymer composite materials. Polymer composites were fabricated by spraying
aluminum onto porous polyethylene substrates, where the thin aluminum coating layer provided
the heat conduction path. 45% weight reduction in the heatsinks was achieved with composite
materials for similar thermal performance as the reference aluminum heatsink material considered.
An analytical heat conduction model of the heatsink fin was developed and was found to be in
good agreement with the experimental measurements of heatsink temperatures at different fin
Figure 5-12 Performance of heatsinks with different fin lengths
200
400
600
800
1000
1200
1400
47
47.5
48
48.5
49
49.5
50
50.5
51
51.5
52
Zortech Al (1 in) Al (1.5 in) Al (2 in)
Wei
gh
t (g
)
Base
Tem
p.
(0C
)
Heatsinks with various fin lengths
Juction Temp
Weight
Base Temp.
Reference
Al Heatsink
67
lengths. The fin efficiency of the composite heatsinks was observed to increase with increase in
the aluminum coating thickness thereby improving the overall performance of the heatsinks. The
model also predicted that the improvement in fin efficiency would be insignificant beyond a
coating thickness of about 1 mm.
Al foils were also used as lightweight heatsinks for LEDs. The thickness of these foils that would
yield highest heat transfer was found using the analytical heat conduction model. By using Al foil
as the heatsink, about 70% weight reduction was achieved. Heatsinks with different fin lengths
were tested. The heatsink with fin length of about 1.5 in was found to have the highest fin
efficiency.
Chapter 6 Conclusions
This study investigated the thermal spray metallization of zinc, copper and aluminum on different
thermoplastic substrates. An electric wire-arc spray coating system was used to deposit these
metals on polytetrafluoroethylene (PTFE), polyethylene (PE) and porous polyethylene. The
thermoplastic substrate was sandblasted and heated at different temperatures before spraying to
understand the effect of polymer surface temperature and surface roughness on the adhesion
properties of the coatings. Polymer metallization was studied by analyzing the splat morphology,
conducting adhesion strength tests and evaluating the electrical conductivity of the coatings. The
results from these studies provide insight into the interaction of molten metal particles with
thermoplastic polymer substrates and an understanding of why thermal sprayed metals adhere to
polymer surfaces. Finally, the application of these lightweight composite materials as heatsinks
68
for cooling LEDs was discussed. The weight reduction and performance of these composite
heatsinks were compared to the conventional aluminum heatsinks. Several specific conclusions
were drawn from this study and are as follows:
1. Wire-arc spraying of zinc and aluminum on PTFE and PE was studied. Zinc coatings,
about 260 µm thick, formed on both PTFE and PE. Aluminum adhered only to the
PTFE but not PE substrates irrespective of their surface roughness. Substrate pre-
heating did not help adhesion. Single splats of these metals were observed under the
SEM to provide insight into the coating adhesion mechanisms. Increasing surface
roughness significantly enhanced adhesion strength since it promoted mechanical
interlocking of the metal with surface asperities. Increasing the initial substrate
temperature so that the maximum substrate temperature during spraying reached the
glass transition temperature of the polymer also enhanced adhesion strength. Individual
zinc splats were observed to be buried deep inside heated polymer substrates.
Aluminum particles have low density and therefore may not have enough momentum
to penetrate PE, which has a much higher impact strength than PTFE. It is also possible
that impacting droplets of aluminum, which have a higher melting point than zinc,
caused localized melting of the substrate upon impact and therefore could not flow into
surface cavities and freeze, which would have led to mechanical bonding.
2. Zinc, aluminum and copper coatings of thickness about 400 µm were deposited on
porous polyethylene with 70 µm average pore size, using an electric wire-arc spray
system. The metallic particles were all observed to penetrate deep into the pores,
69
creating excellent mechanical bond with the substrates. The bond strength of all three
metallic coatings was found to be higher than the yield strength of the porous polymer
which fractured before the coating detached during pull tests. The coating adhesion
strength of all the three metals on the porous polyethylene samples was found to be
significantly higher than the coatings deposited on non-porous polyethylene substrates.
A thin layer (100 µm) of Inconel 625 was deposited on the porous polyethylene and
four-point probe testing was conducted to measure its electrical resistivity. The
electrical resistivity values of all the metallic coatings considered in this study were
found be an order or two of magnitude higher than the bulk metals due to the oxidation
of metal particles during thermal spraying.
3. Lightweight heatsink materials for LED cooling applications were designed using
aluminum foils and metal-polymer composite materials. Polymer composites were
fabricated by spraying aluminum onto porous polyethylene substrates, where the thin
aluminum coating layer provided the heat conduction path. 45% weight reduction in
the heatsinks was achieved with composite materials for similar thermal performance
as the reference aluminum heatsink material considered. An analytical heat conduction
model of the heatsink fin was developed and was found to be in good agreement with
the experimental measurements of heatsink temperatures at different fin lengths. The
fin efficiency of the composite heatsinks could also be increased by increasing the
aluminum coating thickness thereby improving the overall performance of the
heatsinks. By using Al foil as the LED heatsink, about 70% weight reduction was
achieved.
70
Chapter 7 Recommendations for Future Work
Recommendations for work that could be completed in the future are as follows:
1. Investigate the feasibility of using the Inconel 625 coated porous polyethylene
composite material for heating applications. The possibility of coating a thin layer of
Inconel 625 on porous polyethylene was discussed in this work. However, further
analysis into the heating performance of these composites by conducting Joule heating
tests could open up interesting heating applications.
2. Examine other porous polymer substrate materials for thermal spray metallization. In
this work, the challenge of metallizing polyethylene was overcome by using porous
polyethylene as the coating substrates. Similarly, polymers which pose difficulties in
metallization could be made porous (especially at the surface level) and thermally
sprayed to investigate the effect of these pores on the coating deposition.
3. Polymer substrate cooling before and during the thermal spray process could be
considered especially for low melting thermoplastics like polyethylene. A possible test
of polyethylene cooled to sub zero temperatures prior to copper and aluminum
deposition could help in the formation of thick coatings on polyethylene. The localized
melting of polyethylene could be prevented by following this approach.
71
4. Further investigate the feasibility of using polymer composite materials as heatsinks
for electronic cooling applications. The use of aluminum-porous polyethylene
composites materials as lightweight heatsinks for LED cooling applications was
investigated in this work. However, these composite materials also have potential in
cooling other power electronics. Their cooling performance could be improved by
passing coolant into the pores which can act as micro channels enhancing heat transfer.
72
References
[1] R.K. Lattanzio, J.E. McCarthy, Tier 3 motor vehicle emission and fuel standards, in: U.S.
EU Mot. Veh. Stand. Elem. Considerations Trade Issues, 2014.
[2] R. Geyer, J.R. Jambeck, K.L. Law, Production, use, and fate of all plastics ever made, Sci.
Adv. 3 (2017). https://doi.org/10.1126/sciadv.1700782.
[3] S. Zhang, D. Zhao, Aerospace materials handbook, 2016, pp. 281-359.
https://doi.org/10.1201/b13044.
[4] T.K. Das, P. Ghosh, N.C. Das, Preparation, development, outcomes, and application
versatility of carbon fiber-based polymer composites: a review, Adv. Compos. Hybrid
Mater. 2 (2019) 214–233. https://doi.org/10.1007/s42114-018-0072-z.
[5] L. Mohammed, M.N.M. Ansari, G. Pua, M. Jawaid, M.S. Islam, A Review on Natural Fiber
Reinforced Polymer Composite and Its Applications, Int. J. Polym. Sci. (2015).
https://doi.org/10.1155/2015/243947.
[6] J. Delmonte, Plastics Coated Metals and Metal Coated Plastics, in: Met. Compos., 1990: pp.
102–134. https://doi.org/10.1007/978-1-4684-1446-2_5.
[7] B. Mahltig, Y. Kyosev, Inorganic and composite fibers: Production, properties, and
applications, 2018, pp. 243-276. https://doi.org/10.1016/C2016-0-04634-X.
[8] D. Tejero-Martin, M. Rezvani Rad, A. McDonald, T. Hussain, Beyond Traditional
Coatings: A Review on Thermal-Sprayed Functional and Smart Coatings, J. Therm. Spray
Technol. 28 (2019) 598–644. https://doi.org/10.1007/s11666-019-00857-1.
[9] C. Feng, S. Ygeswaran, M. Gibbons, S. Chandra, Analytical heat conduction model of
annular composite fins, in: CSME International Congress 2018 (Toronto, ON), 2018.
https://doi.org/10.25071/10315/35434.
[10] C. Feng, M. Gibbons, S. Chandra, Fabrication of Composite Heat Sinks Consisting of a
Thin Metallic Skin and a Polymer Core Using Wire-Arc Spraying, J. Therm. Spray Technol.
28 (2019) 974–985. https://doi.org/10.1007/s11666-019-00864-2.
[11] H.H. Talib RJ, Saad S, Toff MRM, Thermal spray coating technology, Solid State Sci
Technol. 11 (2003) 109–117.
73
[12] R. Gonzalez, H. Ashrafizadeh, A. Lopera, P. Mertiny, A. McDonald, A Review of Thermal
Spray Metallization of Polymer-Based Structures, J. Therm. Spray Technol. 25 (2016) 897–
919. https://doi.org/10.1007/s11666-016-0415-7.
[13] P. Gramann, A. Rios, B. Davis, Failure of thermoset versus thermoplastic materials, in: SPE
Reg. Top. Conf. - A World Thermosets - Growth Through Appl. Dev., 2005.
[14] M.J. Vucko, P.C. King, A.J. Poole, C. Carl, M.Z. Jahedi, R. de Nys, Cold spray metal
embedment: an innovative antifouling technology, Biofouling. 28 (2012) 239–248.
https://doi.org/10.1080/08927014.2012.670849.
[15] A. Ganesan, M. Yamada, M. Fukumoto, Cold spray coating deposition mechanism on the
thermoplastic and thermosetting polymer substrates, J. Therm. Spray Technol. 22 (2013)
1275–1282. https://doi.org/10.1007/s11666-013-9984-x.
[16] H. Che, P. Vo, S. Yue, Investigation of cold Spray on polymers by single particle impact
experiments, J. Therm. Spray Technol. 28 (2019) 135–143. https://doi.org/10.1007/s11666-
018-0801-4.
[17] M.R. Rokni, P. Feng, C.A. Widener, S.R. Nutt, Depositing Al-Based Metallic Coatings
onto Polymer Substrates by Cold Spray, J. Therm. Spray Technol. 28 (2019) 1699–1708.
https://doi.org/10.1007/s11666-019-00911-y.
[18] H. Ashrafizadeh, P. Mertiny, A. McDonald, Determination of temperature distribution
within polyurethane substrates during deposition of flame-sprayed aluminum-12silicon
coatings using Green’s function modeling and experiments, Surf. Coatings Technol. 259
(2014) 625–636. https://doi.org/10.1016/j.surfcoat.2014.10.020.
[19] H. Boyer, A. McDonald, and P. Mertiny, Flame spray deposition of electrically conductive
traces on polymer substrates for system integrated composite structures, in: Composites
2012 (Las Vegas, NV), 2012, 1-6.
[20] A. Ganesan, M. Yamada, M. Fukumoto, The effect of CFRP surface treatment on the splat
morphology and coating adhesion strength, J. Therm. Spray Technol. 23 (2014) 236–244.
https://doi.org/10.1007/s11666-013-0003-z.
[21] R. Gonzalez, A. McDonald, P. Mertiny, Effect of flame-sprayed Al-12Si coatings on the
failure behaviour of pressurized fibre-reinforced composite tubes, Polym. Test. 32 (2013)
1522–1528. https://doi.org/10.1016/j.polymertesting.2013.10.002.
74
[22] A. Liu, M. Guo, J. Gao, M. Zhao, Influence of bond coat on shear adhesion strength of
erosion and thermal resistant coating for carbon fiber reinforced thermosetting polyimide,
Surf. Coatings Technol. 201 (2006) 2696–2700.
https://doi.org/10.1016/j.surfcoat.2006.05.012.
[23] G.H. Sun, X.D. He, J.X. Jiang, Y. Sun, Parametric study of Al and Al2O3 ceramic coatings
deposited by air plasma spray onto polymer substrate, Appl. Surf. Sci. 257 (2011) 7864–
7870. https://doi.org/10.1016/j.apsusc.2011.04.057.
[24] A. Wypych, P. Siwak, D. Andrzejewski, J. Jakubowicz, Titanium plasma-sprayed coatings
on polymers for hard tissue applications, Materials (Basel). 11 (2018) 2536.
https://doi.org/10.3390/ma11122536.
[25] H. Ashrafizadeh, A. McDonald, P. Mertiny, Deposition of Electrically Conductive Coatings
on Castable Polyurethane Elastomers by the Flame Spraying Process, J. Therm. Spray
Technol. 25 (2016) 419–430. https://doi.org/10.1007/s11666-015-0376-2.
[26] S. Chandra, P. Fauchais, Formation of solid splats during thermal spray deposition, J.
Therm. Spray Technol. 18 (2009) 148–180. https://doi.org/10.1007/s11666-009-9294-5.
[27] V. Pershin, M. Lufitha, S. Chandra, J. Mostaghimi, Effect of substrate temperature on
adhesion strength of plasma-sprayed nickel coatings, J. Therm. Spray Technol. 12 (2003)
370–376. https://doi.org/10.1361/105996303770348249.
[28] H. Che, X. Chu, P. Vo, S. Yue, Metallization of Various Polymers by Cold Spray, J. Therm.
Spray Technol. 27 (2018) 169–178. https://doi.org/10.1007/s11666-017-0663-1.
[29] M. Iijima, N. Kohda, K. Kawaguchi, T. Muguruma, M. Ohta, A. Naganishi, T. Murakami,
I. Mizoguchi, Effects of temperature changes and stress loading on the mechanical and
shape memory properties of thermoplastic materials with different glass transition behaviors
and crystal structures, Eur. J. Orthod. 37 (2015) 665–670.
https://doi.org/10.1093/ejo/cjv013.
[30] M. Biron, 4 - Detailed Accounts of Thermoplastic Resins BT - Thermoplastics and
Thermoplastic Composites, second ed., in: Plast. Des. Libr., 2013.
https://doi.org/10.1016/B978-1-4557-7898-0.00004-4.
[31] A. Rezzoug, S. Abdi, A. Kaci, M. Yandouzi, Thermal spray metallisation of carbon fibre
reinforced polymer composites: Effect of top surface modification on coating adhesion and
mechanical properties, Surf. Coatings Technol. 333 (2018) 13–23.
75
https://doi.org/10.1016/j.surfcoat.2017.10.066.
[32] D. Borba Marchetto, D. Carneiro Moreira, G. Ribatski, A Review on Polymer Heat Sinks
for Electronic Cooling Applications, in: 2019.
https://doi.org/10.26678/abcm.encit2018.cit18-0394.
[33] C. Wu, S.C. Mantell, J. Davidson, Polymers for solar domestic hot water: Long-term
performance of PB and nylon 6,6 tubing in hot water, J. Sol. Energy Eng. Trans. ASME.
126 (2004) 581–586. https://doi.org/10.1115/1.1638786.
[34] I. Krásný, I. Astrouski, M. Raudenský, Polymeric hollow fiber heat exchanger as an
automotive radiator, Appl. Therm. Eng. 108 (2016) 798–803.
https://doi.org/10.1016/j.applthermaleng.2016.07.181.
[35] L. Song, B. Li, D. Zarkadas, S. Christian, K.K. Sirkar, Polymeric hollow-fiber heat
exchangers for thermal desalination processes, Ind. Eng. Chem. Res. 49 (2010) 11961–
11977. https://doi.org/10.1021/ie100375b.
[36] R. Trojanowski, T. Butcher, M. Worek, G. Wei, Polymer heat exchanger design for
condensing boiler applications, Appl. Therm. Eng. 10 (2016) 150–158.
https://doi.org/10.1016/j.applthermaleng.2016.03.004.
[37] H. Lee, Y. Jeong, J. Shin, J. Baek, M. Kang, K. Chun, Package embedded heat exchanger
for stacked multi-chip module, Sensors Actuators, A Phys. 114 (2004) 204–211.
https://doi.org/10.1016/j.sna.2003.12.026.
[38] D. V. Atterton, T.P. Hoar, Surface tension of liquid metals, Nature. 167 (1951) 602.
https://doi.org/10.1038/167602a0.
[39] L. Battezzati, A.L. Greer, The viscosity of liquid metals and alloys, Acta Metall. 37 (1989)
1791–1802. https://doi.org/10.1016/0001-6160(89)90064-3.
[40] E.D. Eastman, A.M. Williams, T.F. Young, The specific heats of magnesium, calcium, zinc,
aluminum and silver at high temperatures, J. Am. Chem. Soc. 46 (1924) 1178–1183.
https://doi.org/10.1021/ja01670a010.
[41] J.A. Jansen, Failure Analysis and Prevention, in: W.T. Becker, R.J. Shipley, (Eds.), Charact.
Plast. Fail. Anal. ASM Handb., ASM International, 2002, pp. 437–459.
https://doi.org/10.31399/asm.hb.v11.9781627081801.
76
[42] C. Vasile, M. Pascu, Practical Guide to Polyethylene, Rapra Technology Ltd., Shrewsbury,
UK, 2005, pp. 31-85.
[43] P.M. Martin, Handbook of Deposition Technologies for Films and Coatings, third ed.,
William Andrew Publishing, 2010, pp. 93-132.
[44] W. Brostow, B.P. Gorman, O. Olea-Mejia, Focused ion beam milling and scanning electron
microscopy characterization of polymer + metal hybrids, Mater. Lett. 61 (2007) 1333–1336.
https://doi.org/10.1016/j.matlet.2006.07.026.
[45] A. Pourmousa, J. Mostaghimi, A. Abedini, S. Chandra, Particle size distribution in a wire-
arc spraying system, J. Therm. Spray Technol. 14 (2005) 502–510.
https://doi.org/10.1361/105996305X76522.
[46] D.L. Hale, W.D. Swank, D.C. Haggard, In-Flight particle measurements of twin wire
electric arc sprayed aluminum, J. Therm. Spray Technol. 7 (1998) 58–63.
https://doi.org/10.1361/105996398770351043.
[47] G.D. Lunn, M.A. Riley, D.G. McCartney, A study of wire breakup and in-flight particle
behavior during wire flame spraying of aluminum, J. Therm. Spray Technol. 26 (2017)
1947–1958. https://doi.org/10.1007/s11666-017-0639-1.
[48] B. Anand, Thermal Spray Deposition of Metals on Polymer Substrates, M.A.Sc. Thesis,
University of Toronto, 2019.
[49] I.C. Mcneill, Thermal Degradation, in: Compr. Polym. Sci. Suppl., Elsevier Science, 1996,
pp. 451–500. https://doi.org/10.1016/b978-0-08-096701-1.00195-6.
[50] K. Alamara, S. Saber-Samandari, C.C. Berndt, Splat taxonomy of polymeric thermal spray
coating, Surf. Coatings Technol. 205 (2011) 5028–5034.
https://doi.org/10.1016/j.surfcoat.2011.05.002.
[51] M. Fukumoto, T. Yamaguchi, M. Yamada, T. Yasui, Splash splat to disk splat transition
behavior in plasma-sprayed metallic materials, J. Therm. Spray Technol. 16 (2007) 905–
912. https://doi.org/10.1007/s11666-007-9083-y.
[52] A. Liu, M. Guo, M. Zhao, H. Ma, S. Hu, Arc sprayed erosion-resistant coating for carbon
fiber reinforced polymer matrix composite substrates, Surf. Coatings Technol. 200 (2006)
77
3073–3077. https://doi.org/10.1016/j.surfcoat.2005.01.042.
[53] P.C. King, A.J. Poole, S. Horne, R. de Nys, S. Gulizia, M.Z. Jahedi, Embedment of copper
particles into polymers by cold spray, Surf. Coatings Technol. 216 (2013) 60–67.
https://doi.org/10.1016/j.surfcoat.2012.11.023.
[54] X.L. Zhou, A.F. Chen, J.C. Liu, X.K. Wu, J.S. Zhang, Preparation of metallic coatings on
polymer matrix composites by cold spray, Surf. Coatings Technol. 2016 (2011) 132–136.
https://doi.org/10.1016/j.surfcoat.2011.07.005.
[55] S.K. Asl, M.H. Sohi, Effect of grit-blasting parameters on the surface roughness and
adhesion strength of sprayed coating, Surf. Interface Anal. 42 (2010) 551–554.
https://doi.org/10.1002/sia.3184.
[56] A. McDonald, S. Chandra, C. Moreau, Photographing impact of plasma-sprayed particles
on rough substrates, J. Mater. Sci. 43 (2008) 4631–4643. https://doi.org/10.1007/s10853-
008-2669-z.
[57] R. Dhiman, A.G. McDonald, S. Chandra, Predicting splat morphology in a thermal spray
process, Surf. Coatings Technol. 201 (2007) 7789–8801.
https://doi.org/10.1016/j.surfcoat.2007.03.010.
[58] Y. Heichal, S. Chandra, Predicting thermal contact resistance between molten metal
droplets and a solid surface, J. Heat Transfer. 127 (2005) 1269–1275.
https://doi.org/10.1115/1.2039114.
[59] A. McDonald, C. Moreau, S. Chandra, Thermal contact resistance between plasma-sprayed
particles and flat surfaces, Int. J. Heat Mass Transf. 50 (2007) 1737–1749.
https://doi.org/10.1016/j.ijheatmasstransfer.2006.10.022.
[60] K. Bobzin, M. Öte, M.A. Knoch, X. Liao, C. Hopmann, P. Ochotta, Electrical Resistivity
of Wire Arc Sprayed Zn and Cu Coatings for In-Mold-Metal-Spraying, in: IOP Conf. Ser.
Mater. Sci. Eng., 2018. https://doi.org/10.1088/1757-899X/373/1/012011.
[61] V.A. Greenhut, J.D. Idol, R.L. Lehman, D.J. Strange, S.H. Kosmatka, B.C. Goswami, W.
Wang, R.A. Ridilla, M.B. Buczek, W.F. Fischer, Materials, in: CRC Handb. Mech. Eng.
Second Ed., 2004. https://doi.org/10.1115/1.4008051.
[62] V.M. Hameed, M.A. Khaleel, A study on the geometry and shape effects on different
78
aluminum fin types of a vertical cylindrical heat sink, Heat Mass Transf. Und
Stoffuebertragung. 56 (2020) 1317–1328. https://doi.org/10.1007/s00231-019-02750-7.
[63] X. Luo, R. Hu, S. Liu, K. Wang, Heat and fluid flow in high-power LED packaging and
applications, Prog. Energy Combust. Sci. 56 (2016) 1–32.
https://doi.org/10.1016/j.pecs.2016.05.003.
[64] D.D. Incropera.F.P, Fundamentals of heat and mass transfer 6th edition, 2015.
https://doi.org/10.1007/978-3-319-15793-1_19.
79
Appendix A
SEM Images of Coating Microstructures
SEM images of bare polymer surface and metallic splats that were not shown in the earlier
discussions are shown below. Zinc splats were observed to spread more on smooth PE surfaces
than rough substrates. The surface asperities on rough samples restricted splat spreading.
Figure A.1 Zinc splats on polyethylene
Smooth Rough
PE
Zn-PE
80
Zinc splats on heated PTFE surfaces were observed to be smaller in size than splats on substrates
at room temperature as shown below. This was due to the penetration of zinc particles into the
thermally softened substrates which restricts splats from spreading.
Figure A.2 Zn deposited on PTFE
Smooth Rough
PTFE
Zn-
PTFE
Zn-PTFE
(95 0C)
81
Aluminum splats on polyethylene substrates experienced significant fragmentation on impact.
This could be attributed to the formation of vapor film on the polymer surface due to the high
temperature aluminum particles.
Figure A.3 Al deposited on PE
Smooth Rough
Al-PE
(55 0C)
Al-PE
82
The aluminum splats on heated PTFE surfaces were found to push the substrate material down on
impact due to thermal softening of the substrate. This kind of behavior can be found in the images
below.
Figure A.4 (a) PTFE surface heated to 95 0C, (b) Al splats on heated PTFE, (c) Al splats
on rough PTFE and (d) Al splats on rough PTFE heated to 95 0C
(a) (b)
(c) (d)
83
Appendix B
MATLAB Code
Evaluation of Solidification Parameter of Zinc Deposited on Porous
Polyethylene
%% Created by Sudarshan Devaraj
%% Date: 2019-09-23
close all;
clear all;
clc;
Tm = 420;
Ts = 23;
cd = 427;
hf = 1.12*(10^(5));
pd = 7200;
v = 300;
d = 60*(10^(-6));
s = 0.7;
r = 1*(10^(-7));
kd = 60;
u = 0.0018;
cs = 1900;
ps = 500;
84
ks = 0.5;
Ste = (cd*(Tm-Ts))/(hf);
We = (pd*(v^2)*d)/s;
Bi = d/(r*kd);
Re = ((pd*v*d)/u)^(0.5);
Pe = (v*d*pd*cd)/(kd);
A = ((8*3.14*pd*cd*kd)/(3*Pe*ps*cs*ks))^(0.5);
Num = (16/3)*(Ste/Pe*A)*(1-(log(1+(Bi*A))/(Bi*A)));
Den = (Num/4)+(4/We)+(8/(3*Re));
Solidification Parameter = Num/Den
85
Appendix C
Splat Size and Circularity Measurement Data for Copper Deposited on Porous
Polyethylene
Splat
Count
Splat
Surface
Area (A)
Splat
Perimeter
(P)
Splat
Diameter
(4A/P)0.5 Splat Circularity (4πA/P2)
1 149.982 64.333 64.333 0.455156943
2 1020.674 244.084 244.084 0.2151781
3 104.855 45.456 45.456 0.637376946
4 74.327 38.544 38.544 0.628380469
5 907.855 205.193 205.193 0.270820195
6 151.309 61.074 61.074 0.509497047
7 53.091 26.743 26.743 0.932374305
8 65.036 32.026 32.026 0.796412491
9 74.327 49.389 49.389 0.382715259
10 191.127 57.536 57.536 0.725157513
11 327.837 114.281 114.281 0.315282265
12 69.018 29.443 29.443 0.99997212
13 74.327 34.89 34.89 0.766892172
14 134.055 73.549 73.549 0.311256989
15 292 101.363 101.363 0.35695509
16 58.4 30.397 30.397 0.793854735
17 104.855 45.736 45.736 0.629596675
18 91.582 40.848 40.848 0.689379188
19 75.655 41.243 41.243 0.558633246
20 67.691 39.218 39.218 0.552776644
21 285.364 102.992 102.992 0.337895058
22 58.4 36.914 36.914 0.538294907
23 50.436 24.835 24.835 1.0270745
24 82.291 64.053 64.053 0.251920221
25 135.382 63.378 63.378 0.423324598
26 108.836 64.053 64.053 0.333183327
27 119.455 72.595 72.595 0.284695467
28 62.382 31.072 31.072 0.811541096
29 164.582 69.616 69.616 0.426534184
30 55.746 35.006 35.006 0.571371236
31 96.891 37.869 37.869 0.848604643
32 317.218 92.541 92.541 0.465242184
86
33 353.055 82.766 82.766 0.647333485
34 78.309 41.523 41.523 0.570458257
35 337.128 140.976 140.976 0.21305616
36 142.018 52.648 52.648 0.643530709
37 76.982 49.669 49.669 0.391929539
38 99.546 40.848 40.848 0.749327822
39 53.091 25.789 25.789 1.002631952
40 1473.274 273.575 273.575 0.247240918
41 91.582 58.211 58.211 0.33946081
42 114.146 46.131 46.131 0.673696977
43 74.327 53.602 53.602 0.324918362
44 65.036 41.802 41.802 0.467465258
45 751.237 248.345 248.345 0.152987437
46 50.436 32.422 32.422 0.602629896
47 1376.383 262.401 262.401 0.251071818
48 70.346 31.467 31.467 0.892315356
49 88.927 45.061 45.061 0.550074649
50 69.018 32.701 32.701 0.81064357
51 94.236 45.061 45.061 0.582914464
52 67.691 35.68 35.68 0.667837617
53 59.727 32.701 32.701 0.701517119
54 50.436 28.768 28.768 0.765439615
55 98.218 52.253 52.253 0.451812525
56 1242.328 265.265 265.265 0.221751216
57 108.836 45.177 45.177 0.66977269
58 59.727 27.139 27.139 1.018527264
59 110.164 50.623 50.623 0.539925176
60 106.182 55.232 55.232 0.437178707
61 88.927 48.715 48.715 0.470649788
62 84.946 37.589 37.589 0.755111403
63 107.509 60.12 60.12 0.373591097
64 90.255 52.648 52.648 0.40897537
65 58.4 30.118 30.118 0.808630706
66 201.746 88.887 88.887 0.320714115
67 53.091 30.397 30.397 0.721687359
68 290.673 70.291 70.291 0.73891569
69 184.491 75.853 75.853 0.402734932
70 1789.165 351.732 351.732 0.181641998
71 53.091 26.184 26.184 0.972609619
72 99.546 54.277 54.277 0.424406261
73 84.946 43.152 43.152 0.572968458
87
74 90.255 53.602 53.602 0.39454716
75 204.4 53.043 53.043 0.91246113
76 122.109 83.884 83.884 0.21796113
77 57.073 26.184 26.184 1.045558546
78 253.509 94.45 94.45 0.35692675
79 122.109 89.842 89.842 0.19001089
80 553.473 115.747 115.747 0.518879595
81 66.364 30.118 30.118 0.918903564
82 73 36.914 36.914 0.672868634
83 230.946 103.667 103.667 0.269910029
84 78.309 34.331 34.331 0.834503975
85 193.782 90.912 90.912 0.294483285
86 59.727 37.985 37.985 0.519919468
87 99.546 35.285 35.285 1.004230043
88 116.8 61.469 61.469 0.388257797
89 1874.111 443.435 443.435 0.119708506
90 82.291 38.264 38.264 0.705929358
91 142.018 71.245 71.245 0.3514182
92 188.473 71.245 71.245 0.466369351
93 50.436 31.072 31.072 0.656132967
94 122.109 64.053 64.053 0.373816411
95 99.546 38.939 38.939 0.824601015
96 1190.565 262.17 262.17 0.217558854
97 63.709 35.285 35.285 0.642702788
98 303.946 114.002 114.002 0.293738675
99 94.236 60.794 60.794 0.320246981
100 412.782 79.507 79.507 0.820162025
101 74.327 36.635 36.635 0.69557482
102 124.764 54.952 54.952 0.518933757
103 744.601 162.437 162.437 0.354440495
104 66.364 38.544 38.544 0.561059123
105 90.255 47.085 47.085 0.511323509
106 103.527 48.04 48.04 0.563426502
107 176.527 98.104 98.104 0.230370757
108 211.037 75.853 75.853 0.460683566
109 53.091 26.464 26.464 0.952137279
110 58.4 32.981 32.981 0.674333674
111 54.418 30.397 30.397 0.739725804
112 132.727 45.456 45.456 0.806801106
113 106.182 48.04 48.04 0.577875847
114 975.546 190.645 190.645 0.337121579