Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Slurry based stereolithography: a solid freeform fabrication method of ceramics and composites
(USC Thesis Other)
Slurry based stereolithography: a solid freeform fabrication method of ceramics and composites
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
SLURRY BASED STEREOLITHOGRAPHY: A SOLID
FREEFORM FABRICATION METHOD OF CERAMICS AND
COMPOSITES
By
Xuan Song
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(INDUSTRIAL AND SYSTEMS ENGINEERING)
August 2016
Copyright 2016 Xuan Song
i
Acknowledgements
Since my first day as a PhD student on August 7
th
, 2011, I have continuously received help, guidance
and support for my research from remarkable individuals at USC. As an interdisciplinary research project,
this thesis includes the efforts of dozens of persons, whom I would like to thank sincerely.
First of all, I would like to express my deep appreciation and gratitude to Dr. Yong Chen for his advising
on this research. Dr. Chen has continued to encourage and inspire me, all the way from when I was applying
for his group, to the process of hunting for an academic job. Under Dr. Chen’s supervision, I have obtained
solid training in teaching, research as well as communication. I am truly fortunate to have had the opportunity
to work with him.
I would also like to thank Dr. Berok Khoshnevis for his invaluable assistance and suggestion on my
research. It is his initiative and passion in research that believes me in how a researcher can make a
meaningful difference to human life. Many thanks go to Dr. Satyandra Kumar Gupta for serving on my
thesis committee, to Dr. Qiang Huang, Dr. Qifa Zhou and Dr. Né stor Pé rez-Arancibia for serving on my
qualifying committee.
Next I’d like to recognize Zeyu Chen, Dr. Yang Yang, Dr. Liwen Lei, Dr. Shanghua Wu, Dr. Kirk
Shung, Dr. Steven Nutt, Dr. Will Pannell and Dr. Jay R. Lieberman for the contributions that each of them
made to the completion of this work. Special thanks go to Zeyu Chen who has worked with me on fabricating
ultrasound transducers using slurry-based stereolithography method (Chapter 5); to Dr. Yang Yang who
provided me with the knowledge on advanced materials and helped with the fabrication of high dielectric
capacitors (Chapter 4); to Dr. Liwen Lei and Dr. Shanghua Wu for offering generous help and patient
guidance in post processing of ceramics; to Dr. Kirk Shung and Dr. Steven Nutt for providing us with
ii
instruments for material preparation and characterization; to Dr. Will Pannell and Dr. Jay R. Lieberman for
helping us extend the work to medical research (not included in this thesis).
I am grateful for an opportunity of working at Futurewei as an intern in Summer 2015 under the
mentoring of Dr. Jiafeng Zhu and Dr. Masood Mortazavi. They provided me with all the resources that I
needed for a robot project and taught me about teamwork and collaboration, which I think will benefit me
for the rest of my life.
Others at USC to whom I am grateful are to my lab mates at the additive manufacturing lab: Dr. Chi
Zhou, Dr. Yayue Pan, Dr. Yongqiang Li, Dr. Kai Xu, Dr. Jing Zhang, Dongping Deng, Xiao Yuan, Pu Huang,
Dr. Matthew Petros, Payman Torabi, Behnam Zahiri, Aref Vali, Amir Mann, Dr. Tsz Ho Kwok, Xiang Gao,
Huachao Mao, Xiangjia Li and Jie Jin for sharing many brilliant research ideas and experience with me
which have contributed to my intellectual growth.
My gratitude extends to the undergraduate and graduate students who worked with me during the past
five years for their enthusiasm and willingness to develop new additive manufacturing processes: Xuan
Zheng, Zhuofeng Zhang, Pin-I Wu, Reese Dorrepaal, Kennedy Stine, Tae Woo Lee, Zaid Badwan, Kalline
Tong, Andrew Davidson, Gonzalo Gambino, etc. They are exteremely helpful in the creation of the prototype
systems and basic experiments in this thesis.
This dissertation would not have been possible without funding from the Epstein department of
industrial and systems engineering at USC and the National Science Foundation (NSF-CMMI 1335476).
Their generous financial support is greatly appreciated.
Finally, I especially thank my wife Mrs. Yiqing Zhu, my parents, my parents-in-law, my sister, my
beloved pets and all the other family members in China for their constant support and unconditional love.
They have been always sticking by my side whenever I encountered difficulty in the research. It would have
been impossible for me to complete what I started without their love and faith in me.
i
Table of Contents
Acknowledgements .................................................................................................................................. i
List of Tables .......................................................................................................................................... vi
List of Figures ....................................................................................................................................... vii
Abstract .................................................................................................................................................... i
Chapter 1 Introduction ...................................................................................................................... 1
1.1 Additive manufacturing and its material selection .................................................................... 1
1.2 Ceramic and composite materials .............................................................................................. 4
1.3 Stereolithography ...................................................................................................................... 8
1.4 Problem Formulation ............................................................................................................... 12
1.4.1 High viscosity ............................................................................................................... 13
1.4.2 Slurry homogeneity ...................................................................................................... 15
1.4.3 Low photosensitivity .................................................................................................... 16
1.4.4 Post processing ............................................................................................................. 17
1.5 Research Questions, Hypotheses and Contributions ............................................................... 18
1.6 Outline of this work ................................................................................................................. 23
Chapter 2 Literature Review: Additive Manufacturing from a material selection perspective...... 25
2.1 Composite Fabrication ............................................................................................................ 25
2.1.1 Indirect methods ........................................................................................................... 26
ii
2.1.2 Fused Deposition of Composite (FDC) ........................................................................ 27
2.1.3 Direct Ink Write ........................................................................................................... 28
2.1.4 Stereolithography ......................................................................................................... 31
2.1.5 Binder/Ink Jetting ......................................................................................................... 33
2.1.6 Selective Laser Sintering .............................................................................................. 34
2.1.7 Laminated Object Manufacturing ................................................................................ 34
2.2 Ceramic Fabrication ................................................................................................................ 35
2.2.1 Extrusion Freeform Fabrication ................................................................................... 35
2.2.2 Stereolithography ......................................................................................................... 37
2.2.3 Binder/Ink Jetting ......................................................................................................... 39
2.2.4 SLS/SLM ...................................................................................................................... 40
2.2.5 Slurry-layer casting ...................................................................................................... 42
2.2.6 LOM ............................................................................................................................. 43
2.2.7 Electrophotographic Printing ....................................................................................... 43
2.3 Summary ................................................................................................................................. 45
Chapter 3 Slurry-based SLA: Fabrication of Polymer-based Composites ...................................... 46
3.1 Ceramic composite slurry formulation .................................................................................... 46
3.2 Curing characteristics .............................................................................................................. 48
3.3 Rheological characteristics ...................................................................................................... 51
3.4 Bottom-up projection for Stereolithography ........................................................................... 56
3.5 Tape-casting integrated with Bottom-up Projection based Stereolithography ........................ 57
iii
3.6 Prototype Design and Process Planning .................................................................................. 61
3.7 Summary ................................................................................................................................. 64
Chapter 4 Process Modeling: Parameter Optimization for Slurry-based SLA process ................... 66
4.1 Overview ................................................................................................................................. 66
4.2 Layer recoating via tape casting .............................................................................................. 66
4.2.1 Blade height δ Blade ......................................................................................................... 66
4.2.2 Recoating speed v r ........................................................................................................ 69
4.2.3 Layer Pressing Speed v p ............................................................................................... 70
4.3 Layer separation via sliding mechanism ................................................................................. 73
4.4 Base layers for initial gap compensation ................................................................................. 74
4.5 Material Feeding ..................................................................................................................... 75
4.5.1 Moving distance of dispenser and its plunger .............................................................. 75
4.5.2 Nozzle height ................................................................................................................ 76
4.6 Case Study I: fabrication of glass reinforced composite components ..................................... 77
4.7 Case Study II: fabrication of high dielectric capacitor ............................................................ 79
4.7.1 Starting Materials ......................................................................................................... 79
4.7.2 Results and Discussion ................................................................................................. 81
4.8 Summary ................................................................................................................................. 86
Chapter 5 Post Processing for Ceramic Component Fabrication .................................................... 88
5.1 Overview ................................................................................................................................. 88
5.2 Case Study I: Traditional Ceramics for Structural Purpose .................................................... 90
iv
5.2.1 Green Part Fabrication.................................................................................................. 90
5.2.2 Post Processing ............................................................................................................. 91
5.2.3 Testing Results of Sintered Samples ............................................................................ 93
5.3 Case Study II: Piezoelectric Ceramics for Ultrasound Imaging .............................................. 95
5.3.1 Introduction to Ultrasound Transducer ........................................................................ 95
5.3.2 Related Work ................................................................................................................ 96
5.3.3 Green Part Fabrication.................................................................................................. 99
5.3.4 Post Processing of BTO Green Parts .......................................................................... 102
5.3.5 Testing Results of Sintered Samples .......................................................................... 105
5.4 Summary ............................................................................................................................... 115
Chapter 6 Post Processing for Porous Structure Fabrication............................................................... 116
6.1 Introduction ........................................................................................................................... 116
6.2 Process description ................................................................................................................ 118
6.3 Main process parameters ....................................................................................................... 120
6.4 Sugar removal ....................................................................................................................... 123
6.5 Test cases............................................................................................................................... 127
6.5 Summary ............................................................................................................................... 130
Chapter 7 Conclusions and recommendations .............................................................................. 132
7.1 Answering the Research Questions and Testing Hypothesis ................................................ 132
7.2 Contributions and Intellectual Merit ..................................................................................... 134
7.3 Recommendations for Future Work ...................................................................................... 135
v
Reference ............................................................................................................................................. 137
vi
List of Tables
Table 1.1 AM process classification based on the form of the starting material ............................................ 3
Table 1.2 AM process classification based on the type of the starting material ............................................. 4
Table 1.3 Examples of Applications of Piezoelectric Ceramics ..................................................................... 5
Table 1.4 Information of EnvisionTec SI500 on ingredients .......................................................................... 9
Table 3.1 Physical properties and cure depths of ceramic powders with 65wt% ........................................ 49
Table 4.1 Process parameters used in the fabrication process for glass reinforced composite ..................... 77
Table 5.1 Process parameters used in the fabrication process for alumina composite ................................. 90
Table 5.2 Parameters used in the fabrication process for piezoelectric composites ................................... 101
Table 5.3 Measured properties of BTO samples fabricated by our process ............................................... 107
Table 6.1 Slurry coating with different blade height and solid loadings .................................................... 121
Table 6.2 Sample weight changes after boiling .......................................................................................... 126
vii
List of Figures
Figure 1.1 Basic principle of additive manufacturing processes ..................................................................... 2
Figure 1.2 Applications of Additive Manufacturing ....................................................................................... 3
Figure 1.3 Traditional Processing Method of Ceramics.................................................................................. 6
Figure 1.4 Photopolymerization of a generalized methacrylate monomer ..................................................... 9
Figure 1.5 (a) Laser based SLA (b) Mask image projection based SLA ..................................................... 10
Figure 1.6 Resin Over-cure in SLA processes ............................................................................................. 10
Figure 1.7 Two recoating processes for the top-down projection method. ................................................... 14
Figure 1.8 Deformation of green parts fabricated from diluted slurry after drying ...................................... 14
Figure 1.9 Effects of slurry homogeneity ...................................................................................................... 15
Figure 1.10 Light scattering by solid particles in liquid resin ....................................................................... 17
Figure 1.11 Related chapters to research questions and hypotheses ............................................................. 22
Figure 1.12 Overview of the dissertation ...................................................................................................... 24
Figure 2.1 Schematic of the 3D printer head to produce continuous carbon fiber reinforced composites ... 28
Figure 2.2 Schematic illustration of extruding fiber reinforced suspension via a nozzle system ................ 30
Figure 2.3 Lithoz’s digital light processing system for photosensitive ceramic slurries .............................. 38
Figure 2.4 Schematic of an electrophotographic printing system ................................................................. 44
Figure 3.1 Preparation of ceramic composite slurry ..................................................................................... 48
viii
Figure 3.2 Viscosity of alumina slurries with different concentrations of dilute solvent ............................. 52
Figure 3.3 Viscosity of alumina slurries with different solid loadings ......................................................... 52
Figure 3.4 As-cast slurry tapes with different solid loadings, doctor blade height is 200μm ....................... 53
Figure 3.5 Projection-based SLA process using the bottom-up projection method ...................................... 56
Figure 3.6 (a) Schematic of typical tape caster (b) An automatic tape casting coater ................................ 58
Figure 3.7 Green part fabrication using tape casting integrated SLA ........................................................... 60
Figure 3.8 Process planning for the tape casting integrated SLA process ................................................... 60
Figure 3.9 Fabrication machine of projection-based SLA integrated with tape casting ............................... 61
Figure 3.10 A prototype system using the tape casting integrated SLA process .......................................... 62
Figure 3.11 Flow chart of the tape-casting-integrated SLA process ............................................................. 63
Figure 4.1 Tape casting techniques (a) traditional process (b) used in our process ...................................... 67
Figure 4.2 The velocity profile in the blade recoating .................................................................................. 68
Figure 4.3 PDMS deformation leads to bigger recoated thickness ............................................................... 70
Figure 4.4 Measurement of recoated thickness ............................................................................................. 72
Figure 4.5 Actual layer thickness versus pressing speed .............................................................................. 72
Figure 4.6 Drag force during the layer detachment....................................................................................... 73
Figure 4.7 Parameters in the dispensing system............................................................................................ 76
Figure 4.8 A “Hand” model fabricated with glass reinforced composite ...................................................... 78
Figure 4.9 A “Beethoven” model fabricated with glass reinforced composite ............................................. 78
Figure 4.10 Scheme of the preparation process of the PZT@Ag nanoassemblies ........................................ 80
Figure 4.11 Fabricated capacitors ................................................................................................................. 82
ix
Figure 4.12 Fabricated capacitors (a) Star structure (b) Bowl structure ...................................................... 83
Figure 4.13 (a) Dielectric permittivity (b) Loss tangent of composite film .................................................. 84
Figure 4.14 Electrochemical measurements for the as-printed capacitors .................................................... 85
Figure 5.1 Different methods of bonding ceramic powders in additive manufacturing ............................... 88
Figure 5.2 Schematic of Post Processing for Ceramic Component Fabrication ........................................... 89
Figure 5.3 Alumina Green Parts fabricated by the tape-casting-integrated SLA process ............................. 91
Figure 5.4 Post processing of Alumina green parts. (a) De-binding schedule; (b) sintering schedule ......... 92
Figure 5.5 A test case of a gear model .......................................................................................................... 93
Figure 5.6 The SEM images of different samples. (a) Alumina green part; (b) sintered Alumina part ....... 94
Figure 5.7 An illustration of ultrasound transducers in an ultrasound system .............................................. 96
Figure 5.8 A comparison of piezoelectric component fabrication based on machining and AM processes. 97
Figure 5.9 Curing characteristic of BTO slurry with varying BTO weight ratios ...................................... 100
Figure 5.10 Fabrication result of a 0-3 PZT composite ultrasound transducer array .................................. 101
Figure 5.11 Concave BTO element fabricated by our process .................................................................... 101
Figure 5.12 A 64 element BTO segment annular array fabricated by our process ..................................... 102
Figure 5.13 Temperature schedules for debinding and sintering of BTO green parts ................................ 103
Figure 5.14 The concave transducer element: (a) green part; (b)(c) after sintering .................................... 104
Figure 5.15 The segment annular transducer array: (a) greenpart ; and (b) after sintering ......................... 105
Figure 5.16 Scanning electron microscope images: (a) after debinding; (b) after sintering ...................... 106
Figure 5.17 The X-Ray diffractometer patterns of BTO powders and sintered samples ............................ 107
Figure 5.18 Property measurement of samples sintered at different temperature ....................................... 109
x
Figure 5.19 Polarization–electric field hysteresis loop of sintered BTO samples ...................................... 110
Figure 5.20 Application of a BTO sample in an ultrasound transducer ...................................................... 111
Figure 5.21 Initial pulse and echo generated by the printed focused transducer ........................................ 111
Figure 5.22 (a) Echo signal; (b) Voltage and lateral resolution of the fabricated transducer .................... 113
Figure 5.23 Ultrasonic imaging of porcine eyeball ..................................................................................... 114
Figure 6.1 Applications of various foam structures .................................................................................... 118
Figure 6.2 Schematics of the dynamic sugar templating method based on slurry-based SLA ................... 119
Figure 6.3 Cure depth with different cure time for different solid loadings ............................................... 122
Figure 6.4 Cylinder shell structures with different thickness ...................................................................... 124
Figure 6.5 Different support patterns for thin features ................................................................................ 125
Figure 6.6 Microscope image after sugar removal ...................................................................................... 126
Figure 6.7 Increase porosity by the sugar foaming method ........................................................................ 129
Figure 6.8 Glass composite foam structure fabrication by the slurry-based SLA ....................................... 130
i
Abstract
During the past thirty years, manufacturing community has benefitted from additive manufacturing (AM)
technologies, thanks to many of its advantages including the capability of fabricating components directly
from three dimensional (3D) computer-aided design (CAD) models. However, although it can build any
complex geometry with a relatively rapid speed, its widespread adoption in manufacturing industry is
substantially restricted by its limited material selection. Efforts have been made to fabricate advanced
materials, such as composites and ceramics, with various AM processes, among which projection-based
stereolithography (SLA) process has advantages of faster speed and higher accuracy over the other ones. In
an SLA process for composite materials, a slurry mixture of solid filler and photosensitive resin is photo-
cured layer by layer to form a green part.
Main challenges in the composite slurry based SLA process include high viscosity, low photosensitivity,
homogeneity, etc. Compared with liquid resin that are commonly used, the slurry made by mixing solid
particles and liquid resin has an increased viscosity. High viscosity poses a challenge for layer recoating, in
which a uniform thin layer needs to be created within a reasonable time. The maximum material viscosity
that can be handled by conventional SLA processes is less than 3000mPa•S, whereas composite slurry
usually has a viscosity far beyond this limit (e.g. 5~250Pa•S). Another main challenge in the composite
slurry based SLA process is the reduced cure depth due to the light scattering of solid particles in the liquid
resin. Smaller cure depth not only requires smaller layer thickness to be recoated, which is difficult for the
viscous slurry, but also makes newly built layer more easily detach from previous layers due to a smaller
bonding force. Furthermore, due to the high surface energy of such small particles, it is extremely easy for
the particles in the slurry to aggregate. The inhomogeneous distribution of solid particles in green parts will
ii
lead to non-uniform stress inside the part during the post processing, which will consequently contribute to
failure of the post processing, such as cracking and delamination.
To overcome these challenges, we investigated the existing SLA processes, and presented a modified
SLA process by integrating tape-casting method. The developed slurry-based SLA process has the capability
of recoating uniform thin layer for highly viscous composite suspension. To achieve desired material
properties, various approaches for increasing the solid loading of green composite parts are studied, including
proper preparation of composite suspension, bottom-up image projection, tape-casting based recoating, a
two-channel sliding design for layer separation, etc. Two types of polymer composite materials are tested to
demonstrate the functionality of the new process, including glass reinforced composite and polymer-ceramic
composite. The glass reinforced composite is fabricated from a slurry mixture of glass microspheres and
resin. With glass microsphere as the fillers, the final composite can have improved hardness. In the
fabrication of the polymer-ceramic composites, we used silver decorated lead zirconate titanate (PZT) as the
fillers to enhance the dielectric properties of the composite materials. According to our measurement, the
dielectric permittivity of resin/PZT@Ag composite reaches as high as 120 at 100Hz with 18vol% filler,
which is about 30 times higher than that of pure resin. With the resin/PZT@Ag composite slurry, we built
capacitors in different complex shapes, and measured their specific capacitance as 63 F/g at the current
density of 0.5 A/g.
A promising application of the slurry-based SLA process is to indirectly fabricate ceramic materials: A
polymer-ceramic composite part fabricated by the slurry-based SLA process is heated in a furnace to burn
out the polymer. Since the polymer in the composite part has a much lower melting point than the ceramics,
the ceramic part of interest is left behind and sintered as the final product. The debinding and sintering
processes for both aluminum oxide and barium titante (BTO) were studied. BTO-based ultrasound transducer
arrays have been successfully fabricated using the presented methods.
Another post processing method for the slurry-based SLA was also discussed to fabricate porous
structures. Porous structure has wide application in industry, thanks to some of its special properties such as
iii
low density, low thermal conductivity, high surface area and efficient stress transmission. Both templating
and foaming agent methods are used to fabricate porous structures. However, these methods can only
produce simple geometries. In recent years, many research studies have been done to use SLA in the
fabrication of porous structure, but the porosity that can be achieved is relatively small due to their limited
accuracy in building micro-scale features on a large area. We applied the slurry-based SLA process in the
fabrication of porous polymer and composite structures using sugar as the foaming agent. With a solid
loading of 50wt% of the sugar in the resin, the method can achieve a porosity over 50%. This method can be
used to increase the porosity achieved by current SLA methods by over 100%.
1
Chapter 1 Introduction
1.1 Additive manufacturing and its material selection
Additive manufacturing (AM) has developed for about thirty years since stereolithography (SLA) was
invented in mid 1980s. Over the lifetime of the technical field, numerous terminologies (Bourell, 2009)
have been used to describe it: additive fabrication, additive processes, layer manufacturing, freeform
fabrication, direct digital manufacturing, rapid manufacturing, additive manufacturing, etc. In 2009, the
American Society of the International Association for Testing and Materials (ASTM) founded its committee
F42 in order to develop standards for additive manufacturing technologies. In their standard terminology
F2792-12a (Standard, 2009), “additive manufacturing” is formally defined as a process of joining materials
to make objects from 3D model data, usually layer upon layer, as opposed to subtractive manufacturing
methodologies. To be consistent with the ASTM terminology standard, the Roadmap for Additive
Manufacturing (Bourell, 2009) released by established leaders in the field used the same term for all
applications of the technologies. Since then, “additive manufacturing” has been commonly taken as the
standard term for a wide range of technologies, including selective laser sintering (SLS), fused deposition
modeling (FDM), stereolithography apparatus, inkjet printing, etc. In this dissertation, we use the same
term to describe all the related processes and technologies.
The basic principles of AM technologies are similar: A three dimensional (3D) Computer Aided
Design (CAD) model is first sliced into a set of two-dimensional (2D) layers; A 3D solid model can then
be obtained by fabricating the 2D layers one by one, as depicted in figure 1.1. Comparing with conventional
manufacturing processes, additive manufacturing converts the fabrication of a model from 3D to 2D,
therefore it can theoretically build a component of any geometric complexity in a cost-efficient way. The
manufacturing cost can thus be dramatically reduced in many traditional industries, including space,
2
defense, automotive, architecture, etc. Meanwhile, AM allows for wider application of personally
customized product in industries, such as medicine, fashion and art. Moreover, with the capability of
building shapes of free complexity, AM techniques open the doors for novel structures and provide new
design ideas for industries, such as robots, electronics, materials, energy, food, etc. Other unique benefits
of AM techniques include no tool change, energy efficient, etc.
Figure 1.1 Basic principle of additive manufacturing processes
According to the physical principle that is used to fabricate each 2D layer, there are a wide variety of
AM processes in the field. In terms of the form of the starting materials, these processes can mainly be
classified into four categories, i.e. powder-based, liquid-based, solid-based, slurry-based processes, etc.
Table 1.1 lists some typical AM processes for each category. Powder-based AM processes usually use a
high energy power source, e.g. laser or electron beam, to sinter or melt powders which will consequently
merge into a solid structure. Liquid-based AM processes are based on photo-polymerization of liquid resin,
which will be solidified when exposed to ultraviolet (UV) or visible light. Different from powder-based
and liquid-based processes, solid-based AM processes fabricate a component from material feedstock in
the form of solid, such as laminate and filament. The solid materials are bonded together either by gluing
(Feygin, 1999) or heat melting. In addition to these three forms, the feedstock can also be in the form of
slurry for some AM processes, in which slurry materials are extruded out through a nozzle and solidified
in the air (Lewis, 2006
a
).
3
Figure 1.2 Applications of Additive Manufacturing
Table 1.1 AM process classification based on the form of the starting material
Classification
Examples of Existing AM Processes
Powder-based
Selective Laser Sintering/Melting (SLS/SLM), Direct Metal Laser Sintering
(DMLS), Inkjet 3D Printing (3DP), Electron-beam Melting (EBM), Selective
Inhibition Sintering (SIS, Torabi, 2014), Selective Separation Sintering (SSS)
Liquid-based
Stereolithography, Multijet Printing
Solid-based
Laminated Object Manufacturing (LOM), Fused Deposition Modeling
Slurry-based
Direct Ink Writing (DIW)
4
Existing AM processes can build objects from a wide range of materials, including polymer, metal,
ceramics and composite. Table 1.2 summarizes main AM processes for the fabrication of different types of
materials. It can be seen that most of current AM techniques are mainly focused on polymer or metal.
Compared to polymer and metal, ceramics and composites are still severely limited for the current AM
processes. These two types of materials generally have excellent physical or chemical properties, but are
also difficult to produce with traditional manufacturing processes. The goal of this research is to develop a
low-cost AM processing method to fabricate ceramic and composite materials.
Table 1.2 AM process classification based on the type of the starting material
Materials
Existing AM Processes
Polymer Fabrication
SLS/SLM, 3DP, SIS, SLA, Multi-jet Printing, LOM, FDM
Metal Fabrication
SLS/SLM, DMLS, EBM, SIS, LOM
Ceramic Fabrication
SLS/SLM, 3DP, SLA
Composite Fabrication
3DP, DIW, SLA, FDM
1.2 Ceramic and composite materials
We put a lot of interest in the fabrication of ceramics and composites because of diverse properties
that we can get from these materials. In this section, we will briefly discuss the history, applications and
traditional manufacturing methods of these materials.
As key engineering materials, ceramics can be found everywhere in our daily life. As early as
24,000BC (Richerson, 2005), people began to mold clay mixed with water into a shape and produce man-
made ceramic “earthenware” by firing in kilns at low to moderate temperature. However, traditional
5
ceramic industries produce ceramic product only for storage, construction or heat resistance. Modern
ceramic industries emerge in the second half of the 19
th
century as people began to understand more about
electricity and use ceramic materials for electric insulation. In the first half of the 20
th
century, numerous
efforts were made to learn processing technologies for alumina. Following that, may new ceramics were
developed, such as high-temperature ceramics, high-strength ceramics, piezoelectric ceramics, bio-
ceramics and so on.
Table 1.3 Examples of Applications of Piezoelectric Ceramics (Richerson, 2005)
Quartz watches Charcoal lighters Smart skis
Ultrasonic cleaners Hydrophone Sonar
Medical ultrasound imaging Buzzers, alarms Fish finders
Underwater homing beacon Ocean floor mapping Emulsifiers
Dialysis air bubble detection Wheel balancers Motors
Nondestructive inspection Accelerometers Vibration sensors
Ultrasonic physical therapy Homogenizers Contact sensors
Autofocus camera Underwater decoy Impact sensors
Precise deflection measurement Boat speedometers Loud speakers
Zero-vibration tables
Musical greeting
cards
Microphones
Transformers Actuators Printers
Telescope mirror distortion correction
actuators
Transformers Igniters
6
Figure 1.3 Traditional Processing Method of Ceramics
Modern ceramic product has played a critical role in a wide variety of specialty applications: ranging
from thermal, electrical, optical, medical to energy and so on. The most common application of ceramics
is serving as high temperature insulation components thanks to their high melting point and low thermal
conductivity, e.g. in aerospace industry. Furthermore, high hardness and resistance to wear and corrosion
make ceramics very suitable materials for cutting tools, valves, bearings, etc. Besides, electrical
applications account for the largest economic sector for ceramics (Richerson, 2005), including electrical
insulators, piezoelectric ceramics, semiconductors, capacitors, etc. Table 1.3 shows examples of
applications of piezoelectric ceramics.
Ceramics are brittle, which pose significant challenges for their fabrication. The fabrication of ceramic
components mostly starts with finely ground powder. Fine powders are premixed with other additives, e.g.
binders, and are produced into ceramic particulate compacts by shape-forming processes, including
pressing, casting and plastic forming. An additive removal step follows in order to obtain pure ceramic
compacts. The pure ceramic compact is then densified into strong components through a densification
process—sintering. In order to achieve dimensional tolerances and surface finish, machining processes are
7
required after the densification. Due to the high hardness and brittleness of the materials, machining, usually
with diamond tooling, can be very expensive.
In addition to ceramics, composite materials have also attracted a lot of interest in the manufacturing
research community in the recent years. Composites are materials containing two or more individual
compositions. The individual compositions remain separate within the composites, but yield significantly
different physical properties of the combined material. The individual constituents mainly serve as two
functions: matrix and reinforcement. The matrix maintains the reinforcement materials in position, while
the reinforcement enhances the physical properties of the composite materials. For different applications,
various materials can be used as the reinforcement (Gay, 2014), such as glass, carbon, high strength
polymers, ceramics, etc. Polymer, carbon, metals can all be selected as matrix, among which polymer are
the most popular choice.
The most important purpose for polymeric composite is to improve the properties of polymeric
materials, such as thermal stability, stiffness and strength, by incorporating short fibers or ceramic particles
(Friedrich, 2005), without sacrificing the weight of the materials. Due to their lightweight and enhanced
mechanical properties, polymeric composites have raised many interest in a variety of applications. For
example, structural components in automotive industry or wind turbines have been dominated by polymeric
composites for decades (Friedrich, 2005).
Another significant use of polymeric composites is to form composites of functional ceramics, e.g.
polymer-piezocomposite. In sensors and actuators, a combination of properties such as large piezoelectric
coefficient, low density and high mechanical flexibility (Bhimasankaram, 1998) is desired. Pure
piezoelectric ceramics have high piezoelectric properties, though, they are brittle and lacking flexibility.
Polymeric materials are light and flexible, but hardly generate any piezoelectricity. Even though some
polymers have piezoelectric effects, e.g. polyvinylidene difluoride (PVDF), their piezoelectric properties
are still very weak. A composite of piezoelectric ceramics and polymeric materials will provide an excellent
electromechanical behavior, especially for high frequency applications.
8
The principal processes for the manufacturing of polymeric composite components include molding,
stamping, sheeting forming, etc. During the processes, reinforcement fibers or particles are impregnated or
mixed with resin in a lay-up stage. A solidification process is then conducted to harden the polymer matrix.
These manufacturing processes achieve a high-throughput composite production but restrict, to some extent,
the complexity of product design.
1.3 Stereolithography
SLA was the first additive manufacturing process invented by Charles W.Hull in the mid-1980s (Hull,
1986). It builds a component based on photopolymerization of liquid resin. Liquid resin used in SLA is
typically comprised of monomer, photoinitiator, and other additives which induce desired properties. Table
1.4 gives a list of ingredients in photosensitive resin SI500 produced by EnvisionTec (Gladbeck, Germany).
When exposed to visible or UV light, the photoinitiator absorbs the light energy of a certain wavelength
and generates free radicals in the resin, refer to figure 1.4. The highly reactive free radicals, each of which
has an unpaired electron, then attach carbon-carbon double bond contained in a metharylate monomer (Cai,
2004) by drawing one electron from the bond and leave the other electron in the double bond with a carbon.
This reaction produces a carbon radical which will continue to take an electron from monomers until all
monomers run out. Through this free radical photopolymerization process, polymer chains can be formed
consequently.
In a conventional SLA machine, as shown in figure 1.5a, a tank is filled with liquid photosensitive
resin. The fabrication begins when the platform surface is positioned at one layer-thickness under the liquid
resin surface. During the fabrication, A UV laser dot is focused on the liquid resin surface and scanning
throughout the cross section region of a given 3D model. The position of the UV laser dot is accurately
controlled by a two-axis reflecting mirror. As discussed above, when exposed to the UV light energy, the
photoinitiator in the liquid resin will decompose into free radicals and cross link the monomers in the liquid
to form hardened polymeric material. We call this hardening procedure as photocuring. After one layer is
9
cured, the Z platform moves down for another layer-thickness and a new layer is successively fabricated in
the same manner. Since very thin layer thickness (e.g. 25μm) can be used in laser-based SLA, the surface
finish can be very smooth.
Table 1.4 Information of EnvisionTec SI500 on ingredients (EnvisionTec, 2012)
Components Approximate % by weight
1. Methacrylated monomer 60-90%
2. 1,6-Hexanediol acrylate 5-10%
3. Acrylated monomer 5-15%
4. Titanium Dioxide 1-2%
5. Photoinitiator 0.1 - 1%
Figure 1.4 Photopolymerization of a generalized methacrylate monomer
10
Figure 1.5 (a) Laser based SLA (b) Mask image projection based SLA
Figure 1.6 Resin Over-cure in SLA processes
The SLA process can use other lighting approaches instead of a laser, such as image projection. The
projection-based SLA is an AM process that also uses photopolymerization to fabricate 3D shapes (refer to
(a)
(b)
11
figure 1.5b). However, in the projection-based SLA process, a Digital Micro-Mirror Device (DMD) is used
to dynamically project a mask image of sliced layers onto the surface of photosensitive resin. The DMD
chip consists of millions of micro mirrors, which can be turned on and off at a frequency of 5kHz. Hence,
the shape of a whole thin layer can be solidified simultaneously. When a mask image is projected onto the
liquid surface, each pixel of the image can be independently controlled. This allows a very high resolution
of the fabricated components, and makes the process applicable in micro-scale fabrication. Comparing with
the other processes, projection-based SLA has the following advantages:
Fast speed: since a whole image can be projected at a high frequency, this process can potentially
achieve a faster production speed than the other processes. An example can be found in (Tumbleston, 2015).
Surface finish: Projection-based SLA solidifies an entire area at the same time, and can use a small
layer thickness (e.g. ~1 micron) in the fabrication. These could achieve better surface finish comparing with
the other processes (Pan, 2012
a
).
Manufacturing cost: The mask image in the projection-based SLA process is implemented by a DMD
chip, which is relatively inexpensive comparing with the laser beam in SLS/SLM processes. As a matter of
fact, projection-based SLA is becoming one of the most popular processes on the desktop 3d printer market.
Applicable scale: Projection-based SLA is capable of fabricating micro-scale components by reducing
mask images to a small size (e.g. ~1mm). It has been widely used in the fabrication of microstructures
(Choi, 2006 & 2009; Cheng, 2009; Sun, 2005; Park, 2011).
In projection-based SLA process (Song, 2012), when a mask image is projected on to the surface of
liquid resin, the material that is exposed to the energy exceeding a critical energy exposure threshold (E c)
will be cured. The projected light beam is not a uniformly focused ideal square. Instead it follows a
Gaussian distribution whose center is at the center of the pixel. Both the light beam radius (XY- plane) and
the depth of penetration of resin (Z-plane) are bigger than the cured geometry (Zhou, 2009). The depth that
the light can penetrate into the resin is defined as cure depth, as shown in figure.1.6 The cure depth
12
characterizes the maximum thickness of the materials in SLA which can be solidified under the same light
intensity and is the most significant factor to determine the layer thickness. We will further discuss it in the
later chapters.
As one of the main AM fabrication methods for polymer materials, SLA is a potential process that can
be used in the fabrication of ceramic and polymeric composite materials. A promising way of processing
non-polymeric materials by SLA process is to premix reinforcement fillers with photosensitive resin. The
resin in the slurry can be selectively solidified by controlled light-induced photopolymerization, and finally
serves as matrix material in the final component to bind filler particles together. The final component will
thereby have enhanced properties introduced by the filler particles or fibers. Post-processing of the
fabricated green parts is needed to reinforce the bonding interface between the particles and polymer, or
remove the organics in the green part in order to indirectly yield materials which are difficult to manufacture.
1.4 Problem Formulation
The materials used in conventional SLA processes are restricted to photocurable resin, from which
only polymer components can be made. Their physical properties such as mechanical strength and hardness
are limited for many industry applications. Ceramic and composite fabrication using SLA process is
potentially a method to achieve enhanced material properties along with complex geometry. The materials
used in the fabrication process are generally prepared by mixing photocurable resin with solid particles. By
adding reinforcing fillers into the photocurable resin, the properties of the final components fabricated by
the method can be greatly improved. With different types of solid particles, we can obtain significantly
different properties for the green parts. However, adding solid fillers into the liquid resin will obviously
change the fabrication characteristics of the materials, such as rheological behavior and photosensitivity,
and this introduces a lot of challenges to the additive manufacturing processes. That is, the different
rheological and photosensitive properties of the mixed slurry make it more difficult use existing liquid-
resin-based SLA method to perform the fabrication. In this section, we will discuss the problems involved
13
in the SLA process for ceramic and composite fabrication and present our method to solve the main
challenges in the process.
1.4.1 High viscosity
In projection-based SLA processes, light is projected onto a free liquid surface and new layers are
solidified on the top of previously cured layers. After a layer is built, the building platform is moved down
for one layer thickness δ (e.g. ~100μm) to allow the liquid resin to refill the gap. This procedure is called
layer recoating. Low viscosity of the resin can facilitate the layer recoating with a fast speed and guarantee
the gap is fully refilled without voids generated inside. It is recommended by most SLA processes that the
viscosity of the used materials should be less than 3000mPa•S (Jacob, 1992). When viscosity is higher than
this limit, uniform layer recoating becomes increasingly difficult.
Compared with liquid resins that are commonly used, the slurry made by mixing solid powders and
liquid resin has a dramatically increased viscosity. Higher solid loading (e.g. above 60wt% alumina slurry)
in the composite slurry will result in a larger viscosity that can substantially exceed the maximum viscosity
limit. With such a high viscosity, the slurry can’t flow into the gap within a reasonable time. In a top-down
projection based SLA process (Jacobs, 1992) shown in figure 1.7b, a blade is employed to aid in forming a
new layer by moving across the free surface of the slurry, but ultrathin layers such as 10μm layer thickness
still cannot be achieved for viscous slurry. In addition, since the fabricated part is immersed in a tank filled
with slurry, tiny features can be easily broken by drag force of the viscous slurry when the platform moves
up and down.
Efforts can be made to reduce the viscosity of the slurry. For example, the use of diluents and a heating
system may reduce the slurry’s viscosity to certain extent (Jang, 2000). However, the addition of diluents
may lead to issues such as the large shrinkage and fragility of the structures (Hinczewski 1998
a
; Melchels,
2009), as shown in figure 1.8. Increasing the temperatures of composite slurry will require an additional
heating system; in addition, the temperature increase has a limited range since resin may be cured by
14
increased temperature as well. Hence the resulting viscosity may still be too high to ensure a good layer
recoating (Hinczewski, 1998
b
).
Figure 1.7 Two recoating processes for the top-down projection method.
Figure 1.8 Deformation of green parts fabricated from diluted slurry after drying
(
b)
(
a)
(
c)
(a)
(b)
(c)
15
1.4.2 Slurry homogeneity
Homogeneity is another characteristic of the slurry that should be considered in the fabrication process.
As we discussed above, final properties of a component depend on the solid content of the slurry. The
particle sizes of the solid filler used in the SLA processes are usually less than 5 μm. Due to the high surface
energy of such small particles, it is extremely easy for the particles in the slurry to aggregate. The particle
aggregation will reduce the homogeneity of the slurry, and hence influence the property isotropy of final
green parts. Furthermore, the inhomogeneous distribution of solid particles in green parts will lead to non-
uniform stress inside the part during the post processing, which will consequently contribute to failure of
the post processing, such as cracking and delamination (refer to figure 1.9).
Figure 1.9 Effects of slurry homogeneity
In the top-down projection-based method, the volume of the slurry in the tank must be large enough
to ensure the building volume is completely immersed. The particles in the slurry tend to become aggregated
after the well-mixed composite suspension has been stored in the tank for a while. In order to ensure
homogeneous composite slurry for the fabrication, storage of a large amount of slurry in a tank should be
avoided.
16
Instead of using a slurry-filled tank, another recoating approach is to directly deposit slurry on top of
the previously built layers as shown in figure 1.7c. The recoating approach eliminates the needs of a large
amount of slurry, and avoids particle aggregation in the suspension. However, due to the surface tension of
the slurry, the recoated layers by a blade tend to have defects, which will adversely affect the surface finish
of green parts. A test example based on the recoating approach is shown in figure 1.7c.
1.4.3 Low photosensitivity
Another main challenge in the ceramic composite slurry based SLA process is the reduced cure depth.
That is, when light travels through composite slurry, the solid particles will absorb and scatter the incident
light. The light energy that can access the photosensitive resin is decreased by an order of magnitude. Hence,
the cure depth will be significantly reduced. For example, pure photocurable resin typically has cure depth
up to 1000 μm, but ceramic composite only has cure depth less than 100 μm under the same light intensity.
Small cure depth brings two technical issues to the fabrication process. First, as discussed in section
1.4.1, a smaller layer thickness is needed to accommodate the reduced cure depth. For example, for ceramic
composite mentioned above, a layer thickness as small as 10μm may be required to ensure sufficient
overcure between layers. However, this will impose difficulty in recoating such ultra thin layers, especially
for highly viscous slurry. The second issue is layer detachment becomes more difficult. In the SLA process,
a transparent cover is commonly used to make the recoated layer more uniform. After the layer is built, it
needs to be detached from the cover to allow a new layer to be recoated. To achieve this, the attaching force
between the layer and the transparent cover has to be smaller than the bonding force of the layer with the
previous layers. A smaller attaching force can be achieved by coating the cover with a film (e.g. Teflon),
but it is still too large with respect to the bonding force in the case of composite slurry. The bonding force
between layers is mainly determined by light overcure (see figure 1.10), which is the light penetration depth
beyond the layer thickness. Since composite slurry has a relatively small cure depth, its overcure range is
accordingly small, such that a fresh layer tends to stick to the transparent cover during the layer detachment.
17
Figure 1.10 Light scattering by solid particles in liquid resin
1.4.4 Post processing
The properties of the fabricated green part can be further improved by proper post processing. Heat
treatment is a way that can be chosen to enhance the interface between solid particles and photocured
polymer. It has been studied (Deng, 2012) that the temperature has an effect on the final properties of
specific compositions.
Another post processing method involves both polymer burn-out and high temperature sintering of the
fabricated composite green part, in order to obtain dense organic-free component. This processing method
has been extensively studied for ceramic fabrication, in which a polymer-ceramic composite object is heated
in a furnace to burn out the resin. Since the resin has a much lower melting temperature than the ceramic,
the ceramic part of interest is left behind as the final product. Sintering then follows as a densification
process to reduce the porosity of the final part. For both debinding and sintering, temperature, rate and
dwell time have to be carefully designed especially for low-photosensitive slurry materials, otherwise
polymer will not be able to be completely removed, and cracks and delamination may easily occur during
the sintering.
18
In contrast with debinding and sintering, some post processing methods remove the filler particles in
the composite green parts instead of the matrix. For example, in foam agent method, a component
containing foam agent particles and ceramics or composites is heated such that foam agent particles can be
decomposed and a foam structure can be obtained.
1.5 Research Questions, Hypotheses and Contributions
The principal research goal in this dissertation is to extend material selection options for additive
manufacturing, especially ceramics and composites. The AM process that we are focused on is projection-
based SLA due to many of its advantages over the other processes. The primary research questions for the
dissertation are stated as follows:
RQ1. Can stereolithography method be used to fabricate materials other than plastics?
RQ2. How can post processing be conducted to obtain materials or structures with desired physical
properties?
We made the following two hypotheses in response to the primary research questions above. These
two hypotheses describe the basic methods that we investigate in the fabrication of ceramics and composites.
Hypothesis 1 Projection-based Stereolithography can fabricate polymer-based composite components
from a slurry mixture of photocurable resin and filler particles.
Hypothesis 2 Post processing of polymer-based composites, such as debinding and sintering, sugar
foaming, etc., can be used to obtain high performance materials (e.g. ceramics) or novel structures (e.g.
porous structures).
The challenges in composite fabrication using SLA, as discussed in the previous sections, include high
viscosity, low photo-sensitivity of composite slurry as well as slurry homogeneity. The following three
questions illustrate the main problems that need to be solved in the fabrication of polymer-based composite
19
components using SLA. Three corresponding hypotheses are investigated to answer each of these research
questions.
RQ1.1. How should the slurry materials be prepared for the projection-based SLA process to achieve the
best properties?
RQ1.2. How can viscous slurry materials be recoated during the SLA process?
RQ1.3. How can a thin layer be detached from constrained surface?
Hypothesis 1.1 A slurry mixture without any dilute solvent added will help avoid large deformation and
big shrinkage in the post processes.
Hypothesis1.2 Recoated layer thickness is restricted by slurry viscosity, which can be reduced by
increasing shear rate on the slurry.
Hypothesis1.3 Separation force becomes less by using sliding mechanism than by direct pull-up.
When the slurry is viscous, it becomes increasingly difficult to recoat a thin layer. Therefore, it is
essential to reduce the slurry viscosity during the process in order to achieve thin layer recoating. In the
case of polymer-ceramic composite slurry with a non-Newtonian rheological behavior, the viscosity can be
changed by controlling the parameters of recoating process without adding diluents into the suspension.
The reason that we want to avoid the use of diluent solvent in the slurry is we expect that geometries can
be better controlled in the post processes by using less solvent.
When detaching a newly fabricated layer from substrate, we can either choose to directly pull the
platform up, or slide the substrate aside. By direct pull-up, the new layer has to overcome the adhesive
force from the substrate and can easily detach from the previous layers. Sliding mechanism works well for
liquid resin (Zhou, 2013), and can potentially be used to fabricate viscous materials.
20
According to hypothesis H2, ceramic components can be fabricated through post-processing polymer-
ceramic composite parts (also called green parts), which can be built by slurry-based SLA. If the polymer
in the green parts can be completely burned out, a sintering process can follow to densify the ceramic
components. Moreover, porous structures can be fabricated through post-processing polymer-foam-agent
composite parts. Water-soluble particles, such as sugar, can be chosen as the foam agent in order to enable
an efficient foam agent removal without destroying the other constituents in the composite parts.
Research questions and corresponding hypotheses are stated below to support research question RQ2
and hypothesis H2:
RQ.2.1. How can polymers in a polymer-ceramic composite part be removed in order to obtain pure and
dense ceramic materials?
RQ.2.2. What properties of a ceramic component can be achieved by post processing a polymer-ceramic
composite part fabricated by slurry-based SLA?
RQ.2.3. How can solid particles in a polymer-based composite part be removed in order to obtain porous
structures?
Hypothesis 2.1 The debinding process should be conducted slowly in order to avoid cracks and
delamination, while the sintering process should be in an air atmosphere.
Hypothesis 2.2 Properties for both structural (e.g. strength and hardness) and functional (e.g.
piezoelectricity) purposes can be achieved by debinding and sintering a green part fabricated by slurry-
based SLA.
Hypothesis 2.3 Sugar foaming method can help increase the porosity of a foam structure fabricated by
SLA.
21
To solve the aforementioned problems and verify individual hypothesis, slurry-based SLA methods
have been studied in the dissertation. The related chapters to each question and hypothesis are shown in
figure1.11. First of all, a bottom-up projection based SLA process has been developed. In our method, a
tape-casting system is integrated to recoat thick slurry layers on a transparent glass substrate; uniform thin
layers are then formed by moving the platform down and leaving a desired gap between the previously
cured layers and the glass substrate. After the layer recoating, mask images are projected upwards onto the
bottom of the substrate. The newly cured layer is separated from the substrate using a two-channel sliding
mechanism. Compared with other processes, our method can achieve a thin recoating layer of viscous
slurry (as small as 10μm). By using the bottom-up projection approach, it also removes the need for a large
amount of slurry. In addition, a higher shear rate in our tape-casting system is able to reduce the viscosity
of suspension during the recoating due to the shear thinning behavior of composite suspension (Chartier,
1999). Hence our slurry-based SLA process can fabricate green parts using slurries that have significantly
higher solid loadings. Furthermore, two types of post processing methods were discussed, based on the
composite green parts fabricated by the slurry-based SLA process: debinding/sintering for ceramic
fabrication and sugar foaming for porous structure fabrication. These two methods work in an opposite way:
binding/sintering removes the matrix material and leaves the filler particles behind in the final part, while
sugar foaming removes the filler particles but leaves the matrix material behind.
22
Figure 1.11 Related chapters to research questions and hypotheses
23
1.6 Outline of this work
The aim of this research is to study the fabrication methods of ceramics and composites using a slurry-
based SLA process. More specifically, the major goals of this research include:
(1) Develop a projection-based SLA process to fabricate polymer-based composite material with high
viscosity and low light penetration depth;
(2) Study the associated process parameters that affect the layer recoating and separation mechanism
during the fabrication;
(3) Study post processing methods for ceramic fabrication and explore the application of the fabricated
ceramic and composite components;
(4) Investigate a sugar foaming process to increase the porosity that can be achieved by conventional SLA
processes.
Figure 1.12 gives an overview of all chapters in the dissertation. Chapter 1 introduces research
background and primary research goals of the dissertation. In this chapter, research problems are discussed;
several hypotheses are derived based on the research questions. Chapter 2 reviews the state-of-the-art of
the AM technologies in the fabrication of ceramics and composites.
Chapter 3 and chapter 4 presents the development, modeling and optimization of a slurry-based SLA
process. The developed process is introduced in chapter 3, in which a prototype system is demonstrated.
Chapter 4 analyzes each step involved in the process, establishes the analytical models for the fabrication
process and presents applications of the process in composite fabrication.
Chapter 5 and chapter 6 discusses post processing methods for the polymer-based composite
components fabricated by the slurry-based SLA. Chapter 5 introduces the post processing methods for
ceramic fabrication. Their application in fabricating piezoelectric components, i.e. ultrasound transducer,
is presented in this chapter. Chapter 6 introduces the post processing method for porous structure
24
fabrication. Its application in increasing the porosity of scaffold structures in tissue engineering is
investigated. Chapter 7 summarizes the dissertation and gives some suggestions for future research.
Figure 1.12 Overview of the dissertation
25
Chapter 2 Literature Review: Additive Manufacturing from a
material selection perspective
Various additive manufacturing processes have been invented and applied into production since 1980s.
Most of these processes are focused on plastics and metals and many research activities are devoted to
improving their fabrication capabilities in terms of manufacturing quality, speed, cost, etc. Adding
reinforcement phases into current AM processes is one of the routes to enhance the mechanical properties
of traditional plastic or metal components. A lot of research papers have been reported in the recent decade
to fabricate fiber or particle reinforced composite components. In this chapter, we will review main methods
that have been developed for the fabrication of composites, in particular, polymer-based composites.
Advances in AM of composites have led to the development in ceramic fabrication using AM.
Conventional ceramic fabrication processes use ceramic-organic-based composite as the feedstock
materials and create dense ceramics through thermally processing the composites. Inspired by the
traditional processing methods, efforts have been made to produce ceramics by combining polymer-ceramic
composite AM processes with common post processing techniques. In this chapter, we will also review the
development in AM technologies of ceramics.
2.1 Composite Fabrication
In this section, we mainly review the state-of-the-art of polymer-based composite fabrication using
AM, including direct ink write, stereolithography, binder/ink jetting, selective laser sintering, laminated
object manufacturing, etc. Main purposes of these composites are to enhance the mechanical properties of
polymers, which are the most common materials available for additive manufacturing. Some composite
26
fabrication techniques will also be discussed in the next section, but they are used to produce ceramic-based
composites, and the reinforcement phases are incorporated in order to reduce the porosity of the ceramic
components printed by AM processes.
2.1.1 Indirect methods
Rapid prototyping is originally adopted to fabricate composite materials by injection molding. Taboas
et al. (Taboas, 2003) used this method to fabricate porous polymer-ceramic composite scaffolds. They
created a ceramic mold by casting Hydroxyapatite (HA)-based slurry in a wax or polysulphonamide (PSA)
inverse mold built by a SolidScape 3d printer and sintered in a furnace. These molds were then used to cast
Poly(l)lactide (PLA) materials such that a HA/PLA composite can be obtained.
Love et al. (Love, 2014) used another simple way to build fiber reinforced 3D model. They didn’t
really build a composite component using additive manufacturing, but assemble a 3d printed plastic piece
with a long beam wound with continuous carbon fiber filaments.
Another simple method to indirectly create a composite 3D printed part is infiltration. A part is first
made by additive manufacturing, which is then put in thermosetting materials such as acrylates or epoxy to
fill the pores inside via a capillary effect. This method has been utilized in fabricating polymer-ceramic
composites, such as polymer-piezo-composites (Safari, 1997; Bandyopadhyay, 1998), and provides a
relatively inexpensive way to produce a fully dense and stiff polymer matrix composite part (Evans, 2005).
In his work on piezoelectric composites, Bandyopadhyay and coworkers immersed a ceramic part
fabricated by fused deposition in epoxy for one hour for thorough infiltration. After the epoxy was cured,
the sample was removed from the epoxy tank and polished such that a piezoelectric ceramic/polymer
composite part can be created.
27
2.1.2 Fused Deposition of Composite (FDC)
The most cost-efficient method of 3d printing composites is through fused deposition modeling (FDM)
by replacing conventional thermoplastic filament with composite filament. Carbon has been extensively
studied as a filler to reinforce thermoplastic filament for FDM process in the form of particle or fiber.
In 3d printing of polymer-based composites, carbon particle usually serves as conductive fillers to
improve the conductivity of plastic components. Leigh et al. (Leigh, 2012) presented a low cost conductive
composite material for fabrication of electronic sensors using FDM process. In their method, a regular low-
cost desktop FDM machine can be used without any modification. The conductive composite material is
prepared from Carbon Black as the filler and polycaprolactone (PCL) as the thermoplastic matrix. Filament
for the FDM process is produced by a rolling method. A filler concentration of 15wt% was identified as the
optimal setting for a usable electrical conductivity and smooth nozzle extrusion. The extruded material
exhibited piezoresistive and capacitive behaviors, which enable its application as smart sensors.
Carbon fiber reinforced polymer composites attracted many attentions recent years due to their unique
thermal, electrical and mechanical properties (Shofner, 2003). The polymer-composite filaments for FDM
are prepared as follows: Vapor-grown carbon fibers were homogeneously dispersed in poly(acrylonitrile–
butadiene–styrene) (ABS) by Banbury mixing. The mixture was shaped into bulk sheets by compression
molding process. The bulk sheets were further granulated and extruded into FDM filaments via a single-
screw extruder. Thanks to the shearing force in the extrusion process, the carbon fibers can be highly aligned
along the axial direction of the extrusion. The carbon fiber content and fiber length play an important role
in improving the mechanical properties of the specimen. Ning et al. (Ning, 2015) reported that a fiber
content around 5wt% achieved the optimal properties, including flexural modulus, tensile strength, etc.
Although strength and modulus significantly increased in the printed samples, the fiber breakage and
pore formation during the material preparation and part fabrication still limit the actual application of this
method (Tekinalp, 2014). One potential solution is to add fibers continuously during the extrusion
28
(Matsuzaki, 2016). Matsuzaki and coworkers developed a new fused deposition modeling process to
produce continuous fiber reinforced polymeric composites. Carbon fibers or twisted yarns of natural jute
fibers were used as the reinforcement fillers, with the upper limit of fiber volume fraction around 40~50%.
In their printing process, fibers are thermoplastic filament are fed into a printing head separately. Before
the fiber was impregnated into the nozzle, it was heated by a nichrome wire to enhance its interface with
plastic filament. The fiber and the thermoplastic filament are bonded into a single composite filament in
the nozzle. Property tests show that the Young’s moduli and tensile strengths of the printed composites are
improved.
Figure 2.1 Schematic of the 3D printer head to produce continuous carbon fiber reinforced
composites (Matsuzaki, 2016)
2.1.3 Direct Ink Write
Direct ink write (DIW) is an extrusion process similar to FDC, but deposits colloidal inks on a platform
instead of melted thermoplastic-based composites. After the ink is extruded out of a nozzle, ultraviolet
(UV) curing is most commonly used to solidify the materials. For example, de Hazan et al. (de Hazan, 2012)
29
fabricates ceramic/polymer nanocomposite from high loaded, solvent free UV curable ceramic colloidal
inks. They used a 3d robotic deposition method, in which an extrusion system is used to deliver the ink.
UV curing was manually done every 2 layers with a UV lamp. Lebel et al. (Lebel, 2010) developed an
ultraviolet-assisted direct-write fabrication method to build continuous 3D micro coils. The materials
consist of a blend of single walled carbon nanotubes and polymer matrix. The carbon nanotube filler in the
composite allows for desired spring rigidity in the coils. Bakarich et al. (Bakarich, 2014) used a commercial
extrusion system to deposit different gels on a platform, including alginate/acrylamide gel, an epoxy based
UV-curable adhesive and alginate support gel. The gels are cured by a UV light source after each layer is
deposited. In the final composite component, the alginate/acrylamide gel serves as reinforcement fiber and
the epoxy serves as matrix.
Some research groups tailor the alignment direction of fillers before photocuring the ink, such that
anisotropical properties within the final parts can be obtained. Gladman and Lewis (Gladman, 2016)
demonstrated a 3d printed hydrogel composite structure with locally aligned cellulose fibrils which can
mimic morphology changes exhibited by plants in response to environmental stimuli. In their method,
hydrogel composite ink was printed through a nozzle-based system and photopolymerized by a UV light
source. The cellulous fibrils were aligned along the extrusion direction under the shear forces that arose as
the ink flowed through the nozzle. The resulted filament will consequently swell anisotropically and cause
the printed flat structure morph into a pre-defined 3D shape. A similar method was also used in their work
on the fabrication of lightweight cellular composite parts (Compton, 2014). Kokkinis et al. (Kokkinis, 2015)
used magnet to align the fillers. They fabricated composite materials by combining a direct writing 3d
printer with a photopolymerization-based process. Four syringes on the 3d printer are loaded with inks with
different formulations. The inks were prepared by mixing photocurable resin with magnetically responsive
alumina platelets in different ratios. During the fabrication, different inks are dispensed on a substrate. A
magnet was used to align alumina platelets along a desired direction (refer to Le Ferrand, 2015 for similar
methods). Following that, a light source was turned on to solidify the entire layer. Since a variety of
30
materials can be selected in the process, it provides a novel way for the design and fabrication of functional
heterogeneous composite materials.
Figure 2.2 Schematic illustration of extruding fiber reinforced suspension via a nozzle system
(Compton, 2014)
Other than UV curing, ink can also be solidified through the evaporation of volatile solution in the
deposited layer. Postiglione et al. (Postiglione, 2015) built conductive 3D microstructures by depositing
conductive nanocomposite dispersion through a syringe-based dispensing system. The conductive
nanocomposite ink consists of multiwall carbon nanotubes (MWCNTs), PLA and a highly volatile solution.
Fast evaporation of the volatile solution after liquid deposition leads to the solidification of a wet layer.
Similar method has also been used to create polylactic acid (PLA) and calcium phosphate based composite
scaffold (Serra, 2013).
In all the direct ink write methods described above, composite feedstock materials are prepared before
the fabrication by mixing matrix materials with fillers. Different from these methods, multi-nozzle direct
write system can deposit each composition independently to form a composite. Lee and coworkers (Lee,
2014) demonstrates a method to build composite tissues using such a multi-nozzle direct write system.
Human ear is comprised of both the auricular cartilage and fat tissue. This complex composition along with
31
its complex shape make it difficult to regenerate an ear using traditional methods. The authors first built an
ear scaffold from poly-caprolactone (PCL), with poly-ethylene-glycol (PEG) as the support materials. After
the scaffold was fabricated, PEG was removed by soaking the structure in aqueous solutions. To obtain a
complete tissue for the in vitro test, chrondrocyte laden alginate hydrogel (CLH) and adipocyte laden
alginate hydrogel (ALH) were dispensed on the final PCL scaffold using the same multi-nozzle printing
system. Since CLH and ALH were dispensed at different locations on the scaffold, post processing was
performed afterwards to crosslink these two compositions.
Although direct ink write process can achieve a high filler concentration, e.g. over 55vol% in (de
Hazan, 2012), and consequently relatively good properties in the composites, the accuracy is still limited
by the diameter of the extrusion nozzle. Therefore, it probably has more application in the fabrication of
lattice structure.
2.1.4 Stereolithography
Stereolithography of composite parts is simply performed by incorporating reinforcement phases into
photocurable resin to improve the mechanical properties of the polymer parts. According to the actual
purposes, the reinforcement can be fibers or particles. For example, Zak et al. (Zak, 1996) used short
discontinuous glass fibers as reinforcements to improve the mechanical properties of polymeric objects.
Liu and coworks (Liu, 2010) modified the photocurable resin with silica nanoparticles in order to enhance
mechanical and thermal properties of the materials. Kumar et al. (Kumar, 2012) reported the addition of
small amounts of cellulose nanocrystals into curable resin to reinforce the mechanical properties of the
printed objects by SLA. Several advantages of cellulose nanocrystals, including biosustainability,
biorenewability, low production cost, simple functionalization and dispersibility, make them a very ideal
choice for composite reinforcement.
As discussed in chapter 1, the addition of reinforcement phases in SLA will dramatically influence the
rheological behaviors of the materials. Zak et al. (Zak, 1996) studied the effects of different factors on the
32
rheological behavior of fiber-based composites, including the volume fraction of the fibers, their aspect
ratio, surface coating, etc. and concluded that the volume fraction of fiber has to be controlled at a low level,
i.e. less than 15vol% in order to maintain a low viscosity. Liu et al. (Liu, 2010) tested different solid loadings
and found that the optimal mechanical properties, including flexural modulus, tensile strength etc., were
obtained with a solid loading of 1-3wt%, since a higher ratio will give rise to agglomerates and rapid
increase in the viscosity. Kumar et al. (Kumar, 2012) only used a low filler concentration below 5% of
cellulose nanocrystals in the mixture.
Although the mechanical properties can be improved by introducing the reinforcing agents, the
improvement is still limited by the low solid loading that is available for current SLA process. The final
composites are insufficient in major properties that are required by composite industries, hence they are far
from an actual application. Attempts have to made to further improve the properties of SLA-fabricated
composite parts either by modifying the interface between filler and matrix or by increasing the solid
loading of reinforcement fillers in the mixture without increasing the viscosity.
Modifying the filler/matrix interface may work for functional composite fabrication. Kim and
coworkers (Kim, 2014) presented a method to fabricate piezoelectric composite part with a low solid
loading by projection based stereolithography method. In their approach, only 10wt% of BTO nanoparticles
were added into photocurable polymer solutions. The low mass loading of BTO nanoparticles leads to
insufficient piezoelectric properties of the composites. In order to enhance the mechanical-to-electrical
conversion of nanocomposites, the BTO nanoparticles were chemically modified for generating tight bonds
with the polymer matrix. The experiment showed a piezoelectric constant of ~40Pc/N in the fabricated
piezoelectric composite.
Adding diluent solvent is a general way to achieve relatively high solid loading with a reasonable
viscosity. In his research on biocompatible orthopaedic implant produced from ceramic/polymer composite
materials, Lee (Lee, 2002) added a solid loading of greater than 40vol.% of alumina into biocompatible
monomer solution, to match the elastic modulus of bone, i.e. ~16GPa. They used a reactive monomer
33
(Methyl methacrylate, MMA) as the diluent to both lower the viscosity and improve the reaction rate of the
biocompatible monomer solution. Although the viscosity can be reduced to a reasonable level for SLA
fabrication, the formation of a rigid ceramic “skin” due to surface evaporation of the solvent can still result
in shearing or incomplete coverage of the cured layer with fresh suspension during the process.
In addition to adding fibers or particles, composite component can be produced by placing fibers on
the building part during the process. Karalekas (Karalekas, 2004) used nonwoven glass fiber mats as the
reinforcement to improve the mechanical properties of SLA products. The fiber mat was manually inserted
between two neighboring layers during the SLA process. The interlayer adhesion of photopolymers to
fibrous mats brought a lot of challenges to the process. The authors didn’t really solve this problem but only
embedded a single layer of reinforcing fabric into the mid-plane of a sample. However, experiment tests
still reveal higher tensile properties in the reinforced sample compared to non-reinforced one.
2.1.5 Binder/Ink Jetting
Although Lewis (Lewis, 2006b) classified Binder or ink jetting as one of the direct ink write processes,
we discuss them separately in this section in that they are quite different from nozzle-based methods in
terms of accuracy and material requirement. Binder jetting deposits adhesive binder onto a powder bed to
bond powder particle together. Composite powder is simply used in the fabrication of composites by binder
jetting. Christ et al. (Christ, 2015) studied the fabrication of short-fiber-reinforced composite materials by
binder jetting process. Limited by the capability of the process in spreading a uniform powder layer, only
1% short fibers with a maximum length of 1-2mm were investigated. Cellulose-modified gypsum powder
was used as the matrix. Powders were mixed with reinforcement fibers for 10min and printed by a
commercial binder jetting printer. By comparing with non-reinforced printed samples, the bending strength
of the fiber-reinforced samples were improved by 180% and work of fracture were enhanced up to 10 times.
Ink jetting directly deposits composite materials onto a region of cross section, which will be solidified
either by UV curing or drying. Elliott et al. (Elliott, 2013) used quantum dots (QD) as nanoparticle fillers
34
in a polymeric matrix and studied their fabrication using poly-jet printing process. QD is a type of
nanoparticle which has unique optical properties, such as size-dependent photoluminescence property. The
addition of QD nanoparticles into the jetting resin will affect the viscosity of the materials and consequently
the jetting ability of the printing head. It will also influence the rate and depth of photocuring. In their
research, only a small percentage of QDs (up to 0.5wt%) is used in order to guarantee the materials within
the printable region of the process.
2.1.6 Selective Laser Sintering
Selective laser sintering (SLS) can build composites by mixing powders with reinforcement fillers.
Kenzari and coworkers (Kenzari, 2012) used quasicrystals (QC) as filler particles to produce polyamide-
based composite parts by SLS process. The QC particles are mixed with nylon particles with an appropriate
concentration (i.e. 30vol% QC). The fabricated quasicrystal-polymer composite parts can have reduced
friction, improved wear resistance and are leak-tight.
2.1.7 Laminated Object Manufacturing
Laminated object manufacturing (LOM) is an approach of building 3d models by laminating 2d sheets.
When composite sheets are used instead, composite 3D model can be obtained by LOM. Klosterman et al.
(Klosterman, 1999) used this method to fabricate glass fiber/epoxy composites. The method lay-ups green
composite laminates and cuts the layers into shapes using a CO2 laser. The stacked layers are then
consolidated in one step via a vacuum bag/oven cure method. The dimension accuracy of the process is
within 1% of the design specification except height, which is around 8%. The bigger error along the height
direction may be caused by the lamination-related errors. The fiber volume fraction that can be achieved
by this method is around 41-45% and a shear strength of 3600psi is reported in the research, which is
significantly lower than that of general high performance composites (i.e. 12,000-18,000psi). A problem
with LOM in the fabrication of composites is the poor surface quality. The edge burning effect, due to
charring from the cutting action of the CO2 laser, can be clearly seen on the fabricated parts.
35
2.2 Ceramic Fabrication
As one of the three basic material categories, ceramics have attracted a lot of attentions both in the
ceramic and additive manufacturing research community. A complete AM process of ceramics involves
material formulation, green part fabrication and post processing. Several reviews have been given on the
additive manufacturing of ceramics in the recent years. The interested readers are referred to (Lewis, 2006b)
(Travitzky, 2014) and (Zocca, 2015). According to these reviews, main green part fabrication methods for
ceramics include extrusion freeform fabrication (EFF), SLA, binder/ink jetting, SLS/SLM, fused deposition
of ceramics, LOM, etc. The formulation of feedstock materials replies on the actual fabrication process,
while post processing for most of current AM processes of ceramics contains debinding and sintering. In
this section, we will provide a brief review of the latest AM technologies of ceramics and give an analysis
from material and process perspectives.
2.2.1 Extrusion Freeform Fabrication
Some AM techniques of ceramics form 3D geometries by controlling a nozzle. Different terms have
been defined to characterize these techniques, such as fused deposition of ceramics (FDC), filament-based
writing (FBW), etc. Their common principle is to deposit a mixture of ceramic powder and solidification
agent (e.g. binder) via a nozzle extrusion system and solidify the materials immediately after the extrusion.
In the earliest research on extrusion fabrication of ceramics, the ceramic mixture is prepared as filament.
For example, McNulty et al. (McNulty, 1998) developed a new ceramic filament formulated entirely from
commercially available constituents. Jafari et al. (Jafari, 2000) prepared continuous ceramic filaments from
a mixture of ceramic powders, precoated with stearic acid, and binders.
As ceramic suspension is well studied, some extrusion processes directly extruded the mixture without
formulating a filament at beginning. In the research of Scheithauer and coworkers (Scheithauer, 2014),
stainless steel powder and zirconia powder were mixed with a binder system at a temperature of 100° C. An
extrusion system is used to dispense the mixture onto a substrate at a temperature of about 80° C. The
36
mixture solidifies immediately after printing on the substrate due to the temperature difference. At the end,
debinding and sintering processes are used to decompose the organic matter in the samples and densify the
structures.
Different routes have been investigated to solidify the ceramic suspension after its extrusion.
Scheithauer et al (Scheithauer, 2015) prepared a thermoplastic ceramic suspension with very high solid
loading (over 65vol% alumina) and extrude the material using a heatable dispensing unit. The viscosity of
the suspension can be reduced to a lower range, i.e. 5-10Pa•s, by heating up it to a temperature of 80°C.
The high viscosity could make the extruded material retain its geometry. After debinding and sintering, a
final part with density up to 99% can be obtained. Huang and Leu (Huang, 2009a) developed a ceramic
fabrication technique called the freeze form extrusion fabrication (FEF). Similar to FDC method, FEF
extruded aqueous ceramic paste through a nozzle and deposited the materials on a platform layer by layer.
The difference of this process is it has fabrication process performed below the past freezing temperature.
Under such as low temperature, the extruded material becomes solid immediately after the extrusion. In
their process, a higher solid loading (e.g. 50vol% alumina) and a larger part size can be made to achieve a
density up to 98% of theoretical density after binder removal and sintering. Morissette et al. (Morissette,
2000) developed aqueous alumina-poly(vinyl alcohol) (Al2O3-PVA) gelcasting suspension in order to
improve the mechanical strength of the deposited layers. Gelcasting of each layer is based on
polymerization of monomers or cross-linking of polymers in the solution and only requires a low organic
content (≤5vol%). The AM of the suspension was conducted using a two nozzle extrusion system, one of
which delivered the alumina suspension, and the other one dispensed cross-linking agent solution.
However, a big disadvantage of extrusion freeform fabrication techniques is their relative low accuracy.
Due to issues such as inaccurate geometric representation and system control, process voids may be
generated in each layer, which lead to poor density of the final parts (Jafari, 2000).
37
2.2.2 Stereolithography
By adding ceramic powders into photocurable binder, stereolithography can be used to produce dense
ceramic green parts. Unlike extrusion freeform fabrication techniques, SLA can yield a higher resolution
in the final parts. This offers a great potential to create ceramic parts with very delicate structures, such as
scaffolds (Bian, 2012). In traditional SLA process for ceramic fabrication, diluent has to be added and
process temperature has to be controlled to tailor the viscosity to be suitable for the SLA process
(Hinczewski, 1998). Most ceramic suspensions for SLA used deionized water as the diluent solution, since
deionized water can have lower viscosity. A disadvantage of these aqueous ceramic suspensions is the low
strength of the as-gelled green body. Resin-based ceramic suspension has higher cured strength than the
aqueous ceramic suspension, but its higher viscosity greatly limits the volume fraction of ceramic powders
in the suspension. In order to solve the contradiction between the cured strength and viscosity, Zhou and
coworkers (Zhou, 2010) replaced deionized water in aqueous ceramic suspension with silicasol, which has
low viscosity and can improve the strength of ceramic green parts. The achieved suspension (50vol% silica)
in their process has a viscosity less than 1800mPa•s and has a cure depth larger than 200μm. However,
support structures can still be easily damaged by the viscous suspension during the fabrication, and special
support structures have to be designed to avoid the defects.
Another way to overcome high viscosity of highly loaded ceramic suspension is adding a doctor blade
to spread the viscous suspension. Lithoz (Schwentenwein, 2015; Hatzenbichler, 2012; Tesavibul, 2012) is
one of the most successfully commercialized ceramic AM technologies that adopt a doctor blade to spread
slurry layers. They developed a sterelithography-based ceramic manufacturing system based on a Digital
Micro-Mirror Device (DMD) chip. Alumina suspension was prepared with a viscosity between 12 and 14
Pa•s. Such a high viscosity necessitates a wiper blade for the recoating of each layer. A rotary vat along
with the blade provides continuous slurry flow into the production area but also makes the machine much
bigger than the actual building size.
38
Figure 2.3 Lithoz’s digital light processing system for photosensitive ceramic slurries (Tesavibul,
2012)
Some recent research activities (Echel, 2016; Zanchetta, 2016) have focused on the development of
UV-active preceramic monomers for ceramic SLA. These materials convert to preceramic polymers under
UV light and can further pyrolyze into SiC, SiOC, or other compositions upon heat treatment. Although
this method can yield ceramic parts with very excellent mechanical properties, it can only produce Si-based
ceramics and is limited in the fabrication of other ceramics, such as alumina, barium titanate, etc.
39
2.2.3 Binder/Ink Jetting
Ink jet printing of ceramics works in the same way as regular 2D ink printer, except that preceramic
polymer or ceramic suspension is used as the ink. Preceramic polymer and its ceramic suspension have
been printed by direct inkjet printing process in (Mott, 2001). Silicon carbide was obtained by pyrolyzing
the deposited preceramic polymer. Ink jetting of ceramic suspension is more flexible with the ceramics that
can be selected, but the contradiction between solid loading and viscosity remains a challenge for the
process. Ebert et al. (Ebert, 2009) discussed a direct ink jet printing method for dental prostheses made of
zirconia. An empty standard HP printer was modified by injecting the ceramic suspension into its cartridge.
Limited by the maximum viscosity of ink jetting, only 27vol% of zirconia powder is mixed within the
suspension. A density of 96.9% of the theoretical density was obtained after heat treatment.
Most research activities on jetting related techniques have focused on binder jetting, but the density
that has been achieved is limited by the density of powder bed. One solution to improve the final parts is
to add reinforcement fillers. Suwanprateeb et al. (Suwanprateeb, 2009) studied the improvement of
mechanical properties of hydroxyapatite-based bone implant printed by binder jetting. Hydroxyapatite is a
type of excellent biocompatible material used as bone implant. The authors employed apatite–wollastonite
(A–W) glass as reinforcing phase and fabricated A-W glass ceramic/hydroxyapatite composite with
standard binder-jetting process. After debinding and sintering, a bioactive composite implant can be
produced with improved flexural modulus and strength. Inzana et al. (Inzana, 2014) explored the fabrication
of bone scaffolds based on binder jetting printing of composite calcium phosphate and collagen. It has been
shown that adding collagen into bones could be beneficial for tissue strength and toughness. In this research,
a calcium phosphate structure is printed by a binder-jetting process, where a phosphoric acid solution was
used as the binder. This binder solution eliminated the need of binder removal, as what was usually done
in calcium phosphate printing, and thereby enabled the incorporation of collagen (enhance strength and
toughness) into the scaffold structures. By dissolving collagen into the phosphoric acid solution and
dispensing it together with the binder into the calcium phosphate powders, a composite calcium phosphate
40
and collagen scaffold can be produced eventually. A similar AM method was also used in (Bergmann,
2010), but heat treatment still followed to increase the mechanical strength of the scaffolds. Different from
Inzana’s work, the feedstock materials used in this research are granules consisting of 60wt% bioactive
glass and 40wt% calcium phosphate. The composite granule was prepared by a blending and melting
process, and a grain size of d50=41μm was achieved in the final granule. Since calcium phosphate and
bioactive glass have different biodegradation properties, the presented method provides a possibility for
tailored biodegradation capabilities of the implant in an in vivo application.
Another route for improving the density of final parts in binder jetting processes is through infiltration.
Melcher et al. (Melcher, 2006) used binder-jetting process to build porous alumina preforms for
pressureless infiltration with copper alloy. Yin and coworkers (Yin, 2007) used a similar method to
indirectly fabricate a dense composite of Ti3AlC2-TiAl3-Al2O3-Al. A TiO2/TiC powder mixture is 3d
printed into a “green” preform, which was then pyrolyzed at 800° C and consolidated at temperatures up to
1400 ° C in Argon. The porous ceramic preform is infiltrated by reactive Al, resulting in the form of different
compositions in the final composite. According to the measurement, the toughness of the materials is
improved greatly comparing with that of ceramics. Lipke et al. (Lipke, 2010) explored the fabrication of
ZrC/W composite parts by combining a binder-jetting process and Displacive Compensation of Porosity
(DCP) process. ZrC/W composites possesses higher stiffness, reduced weight and higher resistance to
fracture relative to monolithic tungsten or zirconium carbide. A WC preform was first built by a binder-
jetting process followed by conventional de-binding and sintering processes. The porous WC preform was
then infiltrated by molten Zr2Cu at 1150-1300° C. In the infiltration, the Zr reacts with WC to yield ZrC
and W products for the ZrC/W composites.
2.2.4 SLS/SLM
The simplest way in which SLS/SLM is used to produce ceramic components is through injection
molding. Rapid prototyping assisted molding has been an effective method to produce highly dense ceramic
41
parts. Guo et al. (Guo, 2004) used SLS to build a polymer mold and cast aqueous PZT suspension with it.
The polymer mold was then removed during the sintering of the PZT part. However, this method can’t
really be considered as ceramic SLS/SLM, so we won’t discuss it in details in this section.
Selective laser sintering or melting has been mainly used to deal with metallic materials, but has limited
capability in fabricating ceramics. Some research papers have been focused on the application of SLS/SLM
in ceramic manufacturing. During the fabrication processes, a powder layer is selectively heated and
sintered or melted by a focused laser beam, which will induce stresses inside the ceramic part due to the
thermal gradients. These stresses will consequently cause crack formation in the material. In order to avoid
crack formation, Wilkes et al. (Wilkes, 2013) preheated the ceramic powder bed to 1600° C with a CO2-
laser (output power 1,000W) and then used an ND:YAG-laser (output power 150W) to selectively melt the
ceramic powder. By preheating the powder bed to a higher temperature, thermal gradients during the
fabrication are reduced and therefore mechanical stresses can be significantly relieved. However, the
surface finish achieved by this method is rough due to the difficulty in the control of melting pool. Crack
formation may still occur during SLM, caused by other reasons such as deposition of cool powder layers
on the preheated layer.
Compared with direct SLS/SLM of ceramics, indirect SLS is probably a more feasible way to be
applied in actual production of ceramics. In indirect SLS processes, a mixture of ceramics and polymers is
used, in which polymer phase serves as a sacrificial binder. Shahzad et al. (Shahzad, 2013; Shahzad, 2014)
produced polymer-ceramic composite particles via a thermally induced phase separation method. The
synthesized composite particles were then fabricated by conventional SLS process. The green parts were
debinded and sintered in the post processes and only a density of 38.5% of true alumina was obtained. The
low sintered density can be explained by the low green density and inhomogeneous green microstructure.
The concentration of ceramics in the composite powder can’t be high, otherwise the green parts will be
fragile and delaminated.
42
It has been reported (Subramanian, 2015) that SLS-fabricated green parts have a porosity which is
quite similar to that of the powder bed from which they are formed. That means the density of ceramic parts
fabricated by indirect SLS is limited by the low density of its green parts. Infiltration and isostatic pressure
have been employed to increase the density of the final parts (Shahzad, 2013; Deckers, 2012), but the
resulted density was still less than 90%.
2.2.5 Slurry-layer casting
Many recent work on ceramic AM particularly explored the usage of slurry casting in ceramic green
part fabrication. These research papers (Yen, 2015; Tang, 2015; Muhler, 2015) argued that appreciable
density can be reached by the use of ceramic slurries. Both Tian (Tian, 2012) and Tang (Tang, 2005) studied
slurry-based AM processes for ceramic fabrication. Tian claimed that layer-wise deposited slurry could
achieve a relatively high green density of ceramic film after drying compared with powder-based or organic
binder-based ceramic processing methods. They used water-based ceramic slurry as the feedstock material
and deposited slurry layer through a deposition system similar to (Yen, 2015). After drying for about 2min,
the ceramic tape was sintered by a 100W laser system. The final part was post processed in a furnace for
improved density. The process developed by Tang et al. works in a similar way. A layer of ceramic slurry,
consisting of water, silica sol (40-50nm) and silica powder (10μm), was deposited onto a platform. The
slurry was scanned by a laser beam and became solid due to the “gelling effect” of the silica sol.
The fabrication of ceramic slurry can also be accomplished in an opposite way, as described in (Tang,
2015). When a dried layer of ceramic slurry was scanned by a laser beam, the binder (polyvinyl alcohol) in
the scanned region (usually the boundary of a 2D cross section) evaporated, leaving a weaken area with
only ceramic powder along the scanning track. A green part could thus be separated from the dried slurry
block along the weaken area. A benefit of this method is that no support structure was required in the
process since the weaken area could serve as the support.
43
A challenge for slurry-based AM processes is to stably spread thin and uniform slurry layers. Yen et
al. (Yen, 2015) developed a slurry-layer casting system for additive manufacturing of ceramics. The
presented slurry-layer casting system is comprised of a slurry-feeding mechanism, a casting device and a
cleanup mechanism for the casting device. The tested slurry only has a viscosity less than 650cP, which is
far below that of the ceramic slurry used in our method. Waetjen et al. (Waetjen, 2009) used an airbrush
spraying method to deposit aqueous ceramic suspension instead of conventional doctor blade. The spraying
can handle alumina suspension with a solid loading of 60-76wt%. After the spraying, the slurry layer was
dried with an infrared heater, followed by SLS method. An average density of 98.5% of alumina parts was
achieved by their method.
2.2.6 LOM
Tape casting and LOM process has been investigated in the last decade to produce 3D ceramic parts.
Aqueous ceramic suspension is cast with a doctor blade system into ceramic tapes, which are used as
starting materials in the LOM process (Gomes, 2009). LOM process cuts one layer of ceramic tape with a
laser and laminates it on top of the previous layers using a binder solution as adhesive agent. Post processing
for LOM is the same as other ceramic AM methods, including binder pyrolysis and sintering. The biggest
issues with LOM process are the relatively large amount of material waste during the fabrication and poor
surface finish limited by the thickness of ceramic tapes.
2.2.7 Electrophotographic Printing
Electrophotography process has been used in AM of ceramics (Kumar, 2004). In this process, a
photoconductor drum is used to electrophotographically create a layer of part powder or binder powder in
a given image pattern and transfer it onto a building platform. For ceramic fabrication, depositing binder
powder on part powder appears to be a better choice. That is, binder powder is first electrophotographically
printed on uniformly deposited part powder and then thermally fused to bind the part powder together. A
44
resolution up to 600dpi can be achieved in each layer but the instability of powder transfer prevents its
further application in ceramic fabrication.
Pascall and coworkers (Pascall, 2014) also utilized electrophotographic principle to process ceramic
powders but their method works in a different way. Ceramic suspension was filled into a space between a
photoconductive electrode and a counter electrode. The charged region on the photoconductive electrode
was dynamically changed by illuminating different light patterns such that the ceramic powder in the
suspension can be selectively deposited onto a desired region on the electrode. This is a promising method
that can allow for the fabrication of multiple types of ceramics simultaneously, but the low density of
deposited ceramic layer and the small building volume are two big issues that need to be solved for the
process.
Figure 2.4 Schematic of an electrophotographic printing system (Kumar, 2004)
45
2.3 Summary
In this chapter, we outlined major AM technologies that have been studied in the fabrication of
composites and ceramics, including DIW, SLA, LOM, SLS/SLM, etc. In the review of composite AM
processes, particular emphasis was given to polymer-based composites. According to the review on ceramic
fabrication, most of current AM processes of ceramics are based on the fabrication of ceramic-polymer
composites (i.e. green parts) and heat treatment. In other words, to fabricate a ceramic part, a ceramic-
polymer composite green part is first built and then heat-treated for consolidation. Therefore, many
processes can be utilized in the production of both composites and ceramics, such as fused deposition, SLA
and SLS.
AM processes for polymer-based composites are almost the same as those for polymers, except that
polymeric feedstock materials are replaced by composite ones. The composite feedstock materials are either
prepared before or during the process. When composite materials are formed in situ during the process,
continuous fillers can be achieved in the final parts. Otherwise, only discontinuous fillers can be imparted
into the parts.
To the best of our knowledge, slurry-based AM processes can achieve better properties in the final
ceramic parts, such as extrusion freeform fabrication, SLA and slurry-layer-casting-based SLS. According
to limited reports on powder-based AM processes for ceramics, they usually yield inferior properties to
slurry-based ones, which can be explained by the low density of the powder bed in these processes.
46
Chapter 3 Slurry-based SLA: Fabrication of Polymer-based
Composites
Recognizing the importance of ceramic slurry in ceramic processing, we presented a slurry-based
stereolithography (SLA) process by replacing liquid photosensitive resin with ceramic-resin composite
slurry. Photosensitive resin is the material used in conventional SLA processes, from which only polymeric
components can be fabricated. The presented slurry-based SLA process fabricates polymer-ceramic
composite components by photocuring ceramic composite slurry. Ceramic composite slurry is prepared by
mixing ceramic filler powders with photocurable resin. In the fabrication process, the resin serves as binder
to bond ceramic powders into given 3D shapes. The physical properties of the polymer-ceramic composite
parts largely depend on the amount of ceramic particles that are mixed in the slurry. However, a higher
percentage of ceramic fillers in the slurry mixture will dramatically decrease the curing sensitivity of the
slurry and increase its viscosity. Hence both layer recoating and curing would be more difficult than those
in liquid resin-based SLA. To overcome these challenges, we developed a projection-based SLA process
by integrating slurry tape-casting and a sliding motion design for layer separation. Compared with other
processes, our method can achieve a thin recoating layer (as small as 10μm) of ceramic composite slurry
and can fabricate green parts using slurries with higher solid loadings.
3.1 Ceramic composite slurry formulation
In our study, different fillers were tested using the presented slurry-based AM process, including solid
glass microsphere (Z-CEL, 8054, Potters, PA), Aluminum Oxide (type CR3, Baikalox, Charlotte, NC),
Zirconia (TZ-3YS-E, Tosoh, Tokyo, Japan), Barium Titanate (BTO, Sigma-Aldrich, St.Louis, MO) and
47
Lead Zirconate Titanate (PZT, Piezoelectric Technology Inc., Indianapolis, IN). In order to obtain
homogenous suspension, the powders were first deagglomerated in an azeotropic mixture of
methylethylketone (66 vol%, MEK, Sigma-Aldrich, Saint Louis, MO) and ethanol (34vol%, Sigma-Aldrich,
Saint Louis, MO) with dispersant by ball milling (Fritsch GmbH, Idar-Oberstein, Germany) at the room
temperature for 12 hours. The same ratio of azeotropic mixture was also used in (Chartier, 1999; Mikeska,
1988). The solid loading of the dispersion is 25vol%. Phospholan PS-131 (AkzoNobel, Chicago, IL) and
Triton x-100 (Dow Chemical Co., Midland, MI) were selected as dispersant due to their good dispersion
properties (refer to Mikeska, 1988; Paik, 1998; Jang, 2000). Dispersant with 0.5-0.8wt% concentration was
added into the mixture, on a dry weight basis of ceramic powders.
The dispersion is then dried at 50° C for 12 hours. After the evaporation of the solvent in the dispersion,
dry ceramic powders with dispersant adsorbed onto their surface can be obtained. The deagglomerated
ceramic powders (containing 0.5-0.8wt% dispersant) were dissolved in commercial photocurable resin
(SI500, EnvisionTec Inc., Ferndale, MI) in a mortar based on a designed ratio. The photocurable resin has
a viscosity of 0.18Pa•s at 30°C and a density of 1.10g/cm
3
at 25° C. Its glass transition temperature is 61° C.
The resin is selected in our study due to its excellent photosensitivity. It contains 60-90wt% Methacrylated
monomer, 5-10wt% 1,6-Hexanediol acrylate, 5-15wt% Acrylated monomer, 1-2wt% Titanium dioxide and
0.1-1wt% photoinitiator. Additional 0.2wt% dispersant was added into the suspension to ensure full
dispersions of particles. The suspension was then milled at 200rpm for 1-2 hours with stainless steel beads
to break down the agglomerates formed during solvent evaporation.
Similar to other SLA processes (Dufaud, 2003; Ventura, 2000; Cheverton, 2012; Chabok, 2012) for
ceramic composite fabrication, solvent such as IPA can also be added to reduce the slurry viscosity. In our
process, less diluent solvent is required, due to the integrated tape casting system and two-channel sliding
design. Finally, the mixture is degassed in a vacuum chamber (Bel-Art Products, Wayne, NJ) for 0.5-5
hours to remove the air bubbles that may be generated during the ball milling process.
48
Figure 3.1 Preparation of ceramic composite slurry
3.2 Curing characteristics
Different from pure photocurable resin, curing characteristics of ceramic composite slurry are
significantly influenced by the light scattering effect of the ceramic particles in it. In other words, when
light travels through highly concentrated ceramic composite slurry, its propagation direction will be
changed by the ceramic particles. Therefore, penetration depth of the incident light, or cure depth of the
slurry, is reduced dramatically. Table 3.1 shows the cure depths of four different types of composite slurry
that are measured in a prototype system of the slurry-based SLA process developed in our lab. A single-
layer-curing method is used to measure the cure depth of ceramic composite slurry. In the measurement, a
layer of ceramic composite slurry is first coated on a film collector by a doctor blade, whose height is set
to be 3~5 times bigger than the expected cure depth. A digital image of a circle with a diameter of 10mm
is then projected onto the coated layer of ceramic composite slurry. After the mask image is projected for
10 seconds, a thin layer of ceramic composite will be solidified. The thickness of the solidified layer is
49
measured by a micrometer caliper, which is considered as the cure depth of the material. The cure depth is
a significant parameter that determines the thickness of each layer during the fabrication of a 3D model.
Empirically, layer thickness has to be set less than one half of the cure depth in order to ensure tight bonding
force between neighboring layers.
Table 3.1 Physical properties and cure depths of ceramic powders with 65wt%
Type
Cure depth
( μm )
d50
( μm )
np@
632.8nm
Density
(g/cm3)
Resin
Cure time
(s)
PZT 39 3 2.397 7.7
SI500
10
BTO 97 3 2.4043 6.08 SI500 10
Alumina 340 1.1 1.766 3.95 SI500 10
Glass 590 5 1.5151
†
1.8 SI500 10
Pure resin 1087 NA
1.382-
1.441
††
1.10 SI500 10
Note: †refer to (data for glasses on http://refractiveindex.info/) ††refer to (Griffith, 1996).
The relation between cure depth and energy input is defined by the Jacob's equation (Jacobs, 1992):
ln( )
dp
c
E
CD
E
(3.1)
where C d is the cure depth; E is the energy density of incident light; E c is the critical energy density,
i.e. the minimum energy for the photocurable resin to be solidified; and D p is the resin sensitivity. The light
intensity of our projection system is ~31.6mW/cm
2
measured by an illumination level meter (Simpson
50
electric, WI). The light energy absorbed by ceramic composite slurry is smaller than the light source energy
due to the energy loss when light passing through the Polydimethylsiloxane (PDMS)-coated film collector
(refer to section 3.5). In our setup the thicknesses of the glass and the PDMS is ~2.38mm and ~1mm,
respectively. The cure depth may be slightly different if their thicknesses are modified.
For ceramic composite slurry, the parameter D p is closely related to the absorption by the photoinitiator,
the light absorber (e.g. dye) in the resin, as well as the light scattering by mixed particles (Tomeckova,
2010
a
). According to Griffith and Halloran's scattering equation for ceramic composite suspension
(Tomeckova, 2010
a
; Griffith, 1997; Abouliatim, 2009; Tomeckova, 2010
b
), D p can be determined by the
following equation:
𝐷 𝑃 =
2 𝑑 50
3 𝑄 ̃
𝑛 0
2
∆ 𝑛 2
(3.2)
where d 50 is the average particle size; Δn is the refractive index difference between the solid particle
(n p) and the liquid resin (n 0, i.e. Δn
2
= (n p – n 0)
2
). 𝑄 ̃
is the scattering efficiency term, which embodies a
complex physics for scattering behavior in a dense system. Empirically it depends on volume fraction ϕ,
d 50, and the light wavelength λ.
Equations have been studied in (Gentry, 2013) to describe the effect of scattering on cure width 𝜔 𝑐𝑢 𝑟 𝑒 :
𝜔 𝑐𝑢 𝑟 𝑒 = 𝜔 𝑏𝑒 𝑎𝑚 + 2 𝜔 𝑒𝑥
= 𝜔 𝑏𝑒 𝑎𝑚 + 2 𝑆 𝑤 ln (
𝐸 𝐸 𝑤 ) (3.3)
where cure width 𝜔 𝑐𝑢 𝑟 𝑒 is obtained by adding the illumination width ( 𝜔 𝑏𝑒 𝑎𝑚 ) and excess width ( 𝜔 𝑒𝑥
)
due to light scattering in the horizontal direction; and excess width 𝜔 𝑒𝑥
is related to the width sensitivity
𝑆 𝑤 , incident light energy E and the width critical energy dose 𝐸 𝑤 .
Compared with regular photocurable resins, the cure characteristic of ceramic composite slurry is
affected by the slurry sensitivity defined in equation (3.2) and (3.3). Specifically, the cure depth depends
on the refractive index difference between the filler particle and the resin, the particle size, and the solid
51
loading of solid powders in the suspension. Table 3.1 shows the measured cure depth of PZT, BaTiO3,
Alumina and glass microsphere with the same solid loadings - 65% weight concentration.
Our experiments show that the cure depth of ceramic composite slurry is dominated by the refractive
index difference Δn. Among the three types of ceramics that are tested, PZT has the largest refractive index,
while Alumina has the smallest one. Consequently, the cure depth of PZT is the smallest, while the cure
depth of Alumina is the largest. Glass microspheres that are tested have good optical transparency, and its
refractive index is smaller than alumina, and slightly higher than the SI500 resin. The d50 as shown in the
table is the size from the vendors’ datasheets. The actual particle size will be smaller due to the grinding
process.
3.3 Rheological characteristics
We studied the rheological behaviors of alumina slurries using a Brookfield dial reading viscometer
equipped with a small sample adapter (SC4-14/6R). Alumina slurries containing 65wt% alumina powder
and 0.5wt% dispersant on a dry weight basis of the powders were first analyzed. Deionized (DI) water was
added as dilute into the resin with a weight concentration of 40~60wt% based on the total weight of the
resin and the solvent. Figure 3.2 shows the viscosity of three different dilute concentrations at the shear rate
1~8s
-1
. When the dilute concentration is below 45wt% or above 55wt%, the viscosity exceeded the force
range of the instrument and can’t be recorded. The measurement results indicate that the viscosity at the
shear rate of ~8s
-1
was reduced to 8,000cP when 55wt% dilute solvent was added, which can be used in
conventional SLA process.
As we discussed in chapter 1, adding dilute solvent will reduce the strength of the green parts and
consequently contribute to their large deformation during post processing. For this reason, we studied the
viscosity of slurries at different solid loadings without any dilute solvent and plotted the measurement
results in figure 3.3. It is shown in figure 3.3 that the slurries exhibited shear thinning behavior with a
Newtonian plateau at high shear rate. The viscosity of 50wt% slurry at high shear rate exceeds the range of
52
the instrument and were not recorded. The slurry with a solid loading of 55wt% or higher is too viscous to
be measured. The viscosity of 50wt% slurry at the shear rate ~4s
-1
is over 80,000cP.
Figure 3.2 Viscosity of alumina slurries with different concentrations of dilute solvent
Figure 3.3 Viscosity of alumina slurries with different solid loadings
53
Figure 3.4 As-cast slurry tapes with different solid loadings, doctor blade height is 200μm
In a tape casting process, the viscosity of the slurries can’t be large in order to achieve a smooth and
homogeneous surface of a tape. For this reason, conventional tape casting process usually limits the
viscosity of the suspensions in the range of 0.8-1.2Pa•s at the shear rate imposed by the blade during casting
(Chartier, 1999). When tape casting is applied into our projection-based SLA method, the surface quality
of a cast layer is not as important as that in the manufacturing of ceramic tapes, since a new layer for the
fabrication is formed by pressing the cast tape into a thinner layer. Even if the original cast layer has a poor
surface quality ，the quality of the final layer can be improved after pressing the cast tape. Therefore, a
bigger viscosity is allowed in the tape casting of our AM process as long as a solid slurry layer can be
54
obtained without void islands in it. Figure 3.4 illustrates the layer recoating results of different highly loaded
slurries via the tape casting system in our process. The solid loadings that can yield a relatively uniform
layer without pores can be used in the final fabrication. According to figure 3.3 and 3.4, the 65wt% loaded
slurry that can be successfully coated by the doctor blade into thin layers (e.g.~200μm) has a low shear
viscosity >500Pa•s.
The rheological behaviors of ceramic slurries are determined by the type and concentration of powder,
organic additives and solvent (Olhero, 2005). Many mathematical models have been developed to describe
the dependence of rheological behaviors of ceramic slurry on solid loading, shear rate, etc., such as power-
law model (Lewis, 1996), Herschel-Bulkley model (Hinczewski, 1998
b
; Chartier, 1997), Casson model
(Gurauskis, 2007; Bitterlich, 2002; Zupancic, 1998; Song, 2004).
The Herschel-Bulkley model is expressed in a form of
𝜏 = 𝜏 0
+ 𝐾 ∙ 𝛾 ̇ 𝑛 (3.4)
where 𝜏 is the shear stress (N/m
2
), 𝛾 ̇ is the shear rate (s
-1
), 𝜏 0
is the yield stress (N/m
2
), n is the shear
rate exponent and K is a constant. For slurries that exhibit shear thinning behavior, n<1.
The Casson model is given as:
√ 𝜏 = √ 𝜏 0
+ 𝐾 √ 𝛾 ̇ (3.5)
The influence of solid loading on the viscosity have been extensively studied by researchers. A simple
model of relative viscosity 𝜂 𝑟 (Lewis, 1996) described that 𝜂 𝑟 is power-law dependent on the ceramic
volume fraction:
𝜂 𝑟 =
𝜂 𝑎 𝑝𝑝 , 0
𝜂 𝑠 𝑜𝑙 𝑛 = 𝐾 Φ
𝑛 (3.6)
where 𝜂 𝑎𝑝𝑝 , 0
is the low shear Newtonian viscosity, 𝜂 𝑠 𝑜 𝑙 𝑛 is the viscosity of the solution, i.e. resin and
dilute.
55
A well-known Krieger-Dougherty model (Krieger, 1959) gives the relative viscosity dependence of
hard sphere system on volume fraction as:
𝜂 𝑟 = ( 1 −
Φ
Φ
𝑚 )
− [ 𝜂 ] Φ
𝑚 (3.7)
In the equation, Φ
𝑚 is the maximum volume fraction and [ 𝜂 ] is the intrinsic viscosity ( [ 𝜂 ]=2.5 for
sphere particle). As for soft sphere systems (Bergströ m, 1996; Lewis, 2000), such as the slurry used in our
process, scaled volume fraction Φ
𝑒 𝑓𝑓 for spherical system can be applied in the Krieger-Dougherty model
by following:
Φ
𝑒 𝑓𝑓 = Φ ( 1 +
Δ
𝑎 )
3
(3.8)
in which Δ is the adlayer thickness around each particle (a layer thickness of 10nm is used in many
reports), a is the particle radius.
No matter which model is chosen, the parameters in the above equations are calculated by a
mathematical fit to experimental data such that any viscosity of a specific slurry at a given shear rate and
solid loading can be predicted.
Due to the limited shear rate range of our viscometer, we can use the limiting form of the Cross model
(Bergströ m, 1996) to predict the viscosity at the high shear rates. In the high shear rate limit, the Cross
model takes the form:
𝜂 𝑟 = 𝜂 ∞
+
𝜂 0
− 𝜂 ∞
𝑏 𝛾 ̇ 𝑝 (3.9)
which has three parameters, namely 𝜂 ∞
,
𝜂 0
− 𝜂 ∞
𝑏 and p. To fit the high shear rate viscosity, set p=0.6.
To obtain the viscosity of slurry with higher solid loadings, we can apply a modified Krieger-
Dougherty model presented in (Bergströ m, 1996).
𝜂 𝑟 = ( 1 −
Φ
Φ
𝑚 )
− 𝑛 (3.10)
56
which is a two-parameter empirical equation. With the slurry viscosity results of different
concentrations at the same shear rate, the viscosity of a higher solid loading of alumina can be estimated at
the corresponding shear rate.
3.4 Bottom-up projection for Stereolithography
High solid loadings in ceramic composite slurry are desired such that good physical properties can be
obtained after the fabricated green parts are sintered. However, as discussed in chapter 1, recoating ceramic
composite slurry with high viscosity is difficult in the top-down projection based SLA process. To address
the challenges, we developed a tape-casting-integrated SLA process based on the bottom-up projection
approach, which can build polymer-ceramic composite green parts using slurries with high solid loadings.
DMD
Projector
Building
platform
tank
PDMS
resin
part
Reflecting
mirror
Figure 3.5 Projection-based SLA process using the bottom-up projection method
Different from the top-down projection based approach as shown in figure 1.7, the bottom-up
projection based SLA process is shown in figure 3.5. Mask images are projected through a transparent resin
57
tank to cure layers that are hung upside-down on the building platform. In addition, the bottom of the resin
tank is coated with materials such as PDMS or Teflon to facilitate the separation of newly cured layers
(Zhou, 2013). The benefits of the bottom-up projection based approach include:
(1) Only a small amount of building material is needed. It is not required to have a tank filled with
slurry during the fabrication process. Hence the problem of slurry sitting in the tank for a long time can be
overcome, and the damage of tiny features caused by the viscous slurry can also be avoided.
(2) The recoating of ultrathin layers is enabled. In the bottom-up projection based system, a new layer
is sandwiched between previously cured layers and the bottom of the tank. Hence its thickness can be
accurately controlled by adjusting the height of the building platform.
3.5 Tape-casting integrated with Bottom-up Projection based
Stereolithography
Tape casting is a widely used ceramic manufacturing process that can produce thin ceramic tapes from
various types of slurries and has been applied in manufacturing a variety of ceramic products, including
solid oxide fuel cells (SOFCs), ceramic substrates, etc. Initially developed by Glenn Howatt during World
War II (Mistler, 2000), tape casting is also known as doctor blading and knife coating. A typical tape casting
process starts with a slurry consisting of ceramic powders, binder, solvent, surfactant, etc. A wet tape is
formed by casting the slurry onto a flat surface, or film collector, by a doctor blade. Heated air is supplied
to remove the solvent in the wet tape and obtain dried tape. For ceramic tapes, a firing process is required
after the drying to remove the binder and densify the tapes. Figure 3.6b shows an automatic ceramic film
casting coater. Its capability of coating uniform thin layers provides an approach for recoating viscous slurry
in the ceramic composite SLA process.
Figure 3.7 shows the main components of the developed tape-casting-integrated projection-based
SLA process. The figure shows the moment when an initial layer is cured and has not been detached yet.
58
The developed system consists of a Z stage, a Digital Micro-Mirror Device (DMD) based projection system,
and a tape casting module. The tape casting module is made up of a film collector and a linear guide to
move it in the X axis. The film collector is mainly an embedded glass sheet coated with a PDMS or Teflon
film.
Figure 3.6 (a) Schematic of typical tape caster (Diamond & Related Materials Laboratory n.d.) (b)
An automatic tape casting coater (Haikutech n.d.)
59
In our previous research on multi-material fabrication (Zhou, 2013), it is shown the separation force is
considerably large when the Z stage is directly pulled up even with the coated PDMS film on the film
collector. In addition, the separation force will significantly increase when the resin viscosity is increased.
Hence, for viscous ceramic composite slurry that is used in the ceramic composite SLA process, the
separation force will be large. If such a separation force is larger than the bonding force between the newly
cured layer and the previously cured layers, the newly cured layer will be detached from the part and the
building process will fail. To facilitate the layer separation, a sliding mechanism based on a two-channel
design is presented in (Zhou, 2013). Compared to the separation force during a direct pulling-up, the
shearing force during a sliding motion along the X axis is much smaller. The two-channel design as
discussed in (Zhou, 2013) is also used in our tape-casting-integrated SLA process.
As shown in figure 3.7, the film collector can be moved along the X-axis, and the Z stage can be moved
along the Z-axis. The height of a doctor blade is set to be δ Blade. Note that δ Blade is usually much larger than
the layer thickness δ (i.e. δ Blade > δ). The film collector is divided into two channels, one has the PDMS
coating and another one has not. The mask images controlled by a DMD are only projected onto the channel
that has the PDMS coating.
After a layer is cured, the Z stage first moves up for a small distance (d release) to release the stress in the
cured layer generated in the polymerization process. Right after the lift-up, the film collector slides from
left to right with a speed of V s for a distance d feed. As discussed in (Zhou, 2013), the force in this sliding
movement is relatively small since a thin oxygen-aided inhibition layer (~2.5 μm) is formed near the PDMS
film, which can provide a non-polymerized lubricating layer for easy sliding. As soon as the cured layer
slides to the channel that has no PDMS, the vacuum between the cured layers and the film has been broken.
The built layers thus can be detached from the film collector. The peak separation force based on the sliding
mechanism is only 4-5% of that based on the direct pulling-up approach (Zhou, 2013).
60
Figure 3.7 Green part fabrication using tape casting integrated SLA
Figure 3.8 Process planning for the tape casting integrated SLA process
61
The film collector moves all the way to the right under the dispenser to add new slurry. Before the film
collector moves back, the building platform is moved up for a distance d lift (0.5~1mm) to allow the slurry
that is spread by the blade to be conveyed underneath the built layers. After the film collector returns with
a speed of v r, the Z stage moves the building platform down for a distance δ - d release - d lift. Thus a gap of
one-layer thickness is left above the PDMS film. Hence, based on a thick layer that is formed by the tape
casting module (i.e. δ Blade), the designed motion of the platform can further form a thinner layer of slurry
(i.e. δ). Even for rather viscous slurry, our system can achieve a layer thickness δ as small as 10 m. Such
ultrathin layers are difficult to be formed if only the doctor-blade-based tape casting process is used.
3.6 Prototype Design and Process Planning
A testbed including both hardware and software systems have been built to demonstrate the presented
polymer-ceramic composite fabrication process.
Figure 3.9 Fabrication machine of projection-based SLA integrated with tape casting
62
The hardware system design is shown in figure 3.9. A motorized micro positioning linear stage
(Danaher Corporation, Washington, D.C., US) is used to drive the building platform in the Z axis. A clear
glass sheet (3/32” or 2.38mm thickness) coated with PDMS (Sylgard 184, Dow Corning) is embedded in a
film collector. The film collector moves along a linear guide through four linear ball bearing slide bushings.
A belt-pulley mechanism was used to drive the film collector in the X axis. A micrometer adjustable film
applicator (MTI Corporation, Richmond, CA) is used to conduct material spreading. Its blade height can
be adjusted by a micrometer head with a resolution of 10 m. Hence the ceramic composite suspension
layer that is recoated can be well controlled to be larger than the set layer thickness. A second blade is
mounted at the right of the film doctor blade but in the opposite direction in order to recollect the slurry on
the substrate.
Figure 3.10 A prototype system using the tape casting integrated SLA process
63
A 60cc syringe is fixed behind the blades to dispense a certain amount of ceramic composite slurry
after each motion cycle. The syringe plunger is driven by a linear actuator (Eastern Air Devices, Dover,NH).
The dispensing is performed by moving the syringe along the y axis through a belt-pulley mechanism, while
pushing the plunger down for a controlled distance at the same time. A prototype machine is shown in
figure 3.10.
Figure 3.11 Flow chart of the tape-casting-integrated SLA process
64
A process control software system has been developed using Microsoft Visual C++. Figure 3.11 shows
the graphical user interface (GUI) of the developed software system. The software system can slice a given
CAD model and dynamically output mask images to the designed projection system. The projection image
size of the prototype system is designed to be 43.18mm × 24.384mm. The resolution of the DMD chip
(Texas Instrument, Dallas, TX) is 1280 × 720. The output light intensity of the projector is ~31.6mW/cm
2
.
The software system can also send motion control commands to a 4-axis motion controller (Dynomotion
Inc, Calabasas, CA).
The planned process is also given in the figure 3.11. Before a 3D model starts to be fabricated, a base
is first built to compensate the contact clearance between the Z platform and the surface of the film collector.
The light exposure time for base layers is usually larger than that for regular layers, unless different
materials are used in base and regular layers. Layer separation and layer recoating are performed
sequentially right after process parameters are initialized. Main parameters include light exposure time for
base and regular layers, base layer thickness, doctor blade height, recoating speed, pressing speed, minimum
sliding height, sliding speed, dispensing amount, etc. In the next chapter, all these process parameters are
analyzed and parameter design methods are discussed.
3.7 Summary
In this chapter, we presented a modified bottom-up projection based SLA process by integrating with
tape casting. The method can be used to process polymer-based composite materials including ceramic-
polymer composites and provides an indirect route for the processing of ceramic components.
Serving as the feedstock material for the presented process, ceramic composite slurry is prepared by
mixing ceramic powder, photosensitive resin and other organic additives. Four steps are involved in the
slurry preparation, including powder deagglomeration, premixing, ball milling and degassing.
65
Different ceramic powders were introduced and their slurries were prepared. The analysis of their
curing characteristics indicates the great impact of factors, including incident light, powder type, solid
loading, liquid resin, etc., on the cure depth and width of the final slurry materials.
Rheological properties of the ceramic slurry were also experimentally and analytically studied, using
alumina slurry as an example. The viscosity of 65wt% solid loaded slurries with different dilute
concentrations was first measured. Alumina slurries with different solid loadings were then studied by both
viscosity measurement and layer recoating. It suggests that a solid loading of 65wt% can achieve the best
layer recoating without adding any dilute solvent. Different mathematical models were presented to study
the dependence of the viscosity on shear rate and solid loading. The limiting form of the Cross model and
a modified Krieger-Dougherty model can be used to predict the viscosity of different solid loadings at the
high shear rate.
The tape-casting-integrated SLA process was presented for fabricating viscous ceramic slurries.
Uniform ultrathin layers can be recoated by integrating the bottom-up projection approach and the ceramic
tape casting process. A doctor blade first spreads viscous slurry into thick layers. Platform pressing motion
is then used to achieve a desired layer thickness that can be much smaller than the blade-recoated layer
thickness. The design can also facilitate the layer separation after each layer is fabricated. Benefits of using
the bottom-up projection and tape-casting methods were analyzed for the ceramic slurry based SLA process.
An experimental prototype system has been developed to test the capability of the presented ceramic
composite fabrication process.
66
Chapter 4 Process Modeling: Parameter Optimization for Slurry-
based SLA process
4.1 Overview
In chapter 3, we presented an additive manufacturing (AM) process for fabricating viscous composite
slurry by integrating a projection-based stereolithography (SLA) process with tape casting technique.
Composite slurries are prepared by mixing filler particles with photocurable resin. When different filler
powders are added, the slurry will exhibit distinct curing and rheological behaviors. Since process
parameters can be significantly influenced by these behaviors of the materials, the dependence of process
parameters on the material properties have to be studied in order to successfully fabricate a composite
component. In this chapter, experimental and analytical methods (Song, 2015) were presented for the design
of each process parameter, including doctor blade height, recoating speed, layer pressing speed, separation
speed, etc. Two test cases were fabricated to demonstrate the capability of our process in the fabrication of
composites, including glass reinforced composites and dielectric polymer-ceramic composites.
4.2 Layer recoating via tape casting
4.2.1 Blade height δBlade
A fluid flow process in a traditional tape casting system (figure 4.1a) is a combination of pressure flow
and Couette (drag) flow in the parallel channel of a casting head (Chou, 1987; Tok, 2000). For the pressure
flow, its flow rate results from an applied pressure, i.e. pressure p in figure 4.1a, while the Couette flow is
caused by the drag force acting on the fluid by the doctor blade.
67
Figure 4.1 Tape casting techniques (a) traditional process (b) used in our process
Unlike the traditional tape casting system, the one in our process only dispenses a small amount of
slurry behind the doctor blade for each layer (see figure 4.1b). That means the pressure p in the dispensed
slurry is very small. Since the viscosity of the materials is relatively high and the pressure in the dispensed
slurry is negligible, the blade recoating procedure can be modeled a plane Couette flow pattern, as shown
in figure 4.2, with the bottom plate moving at a constant speed of v r relative to the blade.
For a steady and fully developed fluid flow with negligible gravity and z-direction flow, the Navier-
Stokes equation can be simplified as follows (Joshi, 2002):
0 = −
𝜕𝑝
𝜕𝑥
+ 𝜂 𝜕 2
𝑢 𝑥 𝜕 𝑦 2
（4.1 ）
The boundary conditions follow:
𝑢 𝑥 = 𝑣 𝑟 , 𝑎𝑡 𝑦 = 0;
𝑢 𝑥 = 0 , 𝑎𝑡 𝑦 = 𝛿 𝐵 𝑙 𝑎𝑑 𝑒 (4.2)
68
𝑝 = 0 , 𝑎𝑡 𝑥 = 0 𝑎 𝑛𝑑 𝑥 = 𝐿
𝜂 is the viscosity. L is the length of the film collector. 𝛿 𝐵 𝑙 𝑎𝑑 𝑒 is the doctor blade height. By solving the
Navier-Stokes equations under the given boundary conditions, one can get the velocity profile:
𝑢 𝑥 = 𝑣 𝑟 ( 1 −
𝑦 𝛿 𝐵 𝑙 𝑎 𝑑 𝑒 ) (4.3)
Figure 4.2 The velocity profile in the blade recoating
The volumetric flow rate of the fluid can be calculated as:
𝑄 𝑐 = ∫ 𝑢 𝑥 𝑊 𝑑𝑦
𝛿 𝐵 𝑙 𝑎 𝑑 𝑒 0
=
1
2
𝛿 𝐵 𝑙 𝑎𝑑 𝑒 𝑊 𝑣 𝑟 (4.4)
where W is the doctor blade width.
It is known that the mass of the flow passing the doctor blade in a unit time is identical to the mass of
the coated layer. So the following equation can be established:
𝜌 𝑄 𝑐 = 𝜌 ′ 𝛿 ′ 𝑊 𝑣 𝑟 (4.5)
y
Slurry
u
x
δ
Blade
0
u
x
=v
r
u
x
=0
69
where is the slurry density and ’ is the density of the newly formed layer. A simplified equation of
the thickness ’ of the wet recoated layer can thus be derived as:
𝛿 ′
=
1
2
𝜌 𝜌 ′
𝛿 𝐵 𝑙 𝑎𝑑 𝑒 (4.6)
If we consider the side flow occurring during the tape casting, which extend the width W to W’, we
have:
𝑊𝛿 ′ = 𝑊 ′ 𝛿 " (4.7)
and the thickness 𝛿 " of the final recoated layer becomes:
𝛿 " =
𝑊 𝑊 ′
𝛿 ′ =
𝛼 2
𝜌 𝜌 ′
𝛿 𝐵 𝑙 𝑎𝑑 𝑒 (4.8)
where 𝛼 =
𝑊 𝑊 ′
. To ensure tight interlayer bonding, " must satisfy 𝛿 "
> 𝛽𝛿 , where β is a safety factor
to ensure two neighboring layers are tightly bonded (β> 1) and δ is the layer thickness chosen for the
fabrication.
Hence the inequation for blade gap δ Blade must hold:
𝛿 𝐵 𝑙 𝑎𝑑 𝑒 >
2 𝛽𝜌 ′ 𝛿 𝛼𝜌
(4.9)
4.2.2 Recoating speed vr
Shear rate 𝛾 ̇ is calculated by recoating speed v r and blade gap δ Blade at 𝛾 ̇ = 𝑣 𝑟 𝛿 𝐵 𝑙 𝑎𝑑 𝑒 − 1
. Due to the shear
thinning behavior of composite slurry , a higher shear rate will result in a lower viscosity of the slurry when
the slurry passes the blade. Thus the sedimentation of filler particles in the recoated layer can be avoided,
and a good homogeneity of the fabricated green parts can be achieved. With the same blade gap, a higher
recoating speed v r will generate a bigger shear rate 𝛾 ̇, and accordingly a smaller slurry viscosity.
70
4.2.3 Layer Pressing Speed vp
As mentioned in chapter 3, conventional tape casting process requires the viscosity of the slurries in
the range of 0.8-1.2Pa•s at the shear rate imposed by the blade during casting (Chartier 1999) in order to
achieve a smooth and homogeneous surface of a tape. Composite slurry is too viscous to be spread by a
doctor blade into a uniform layer which is thin enough (10~100 μm) to tightly bond with previous ones. In
our process, a thick layer (100 ~ 500μm) of slurry is first coated on the substrate by the tape casting system.
Since the cast layer is much thicker than the required thickness, its surface doesn’t have to be smooth as
long as the building area is fully covered after casting. Therefore, a higher viscosity can thus be used in our
process.
Figure 4.3 PDMS deformation leads to bigger recoated thickness
Based on this thick layer δ”, a thinner thickness of the layer is then achieved by a pressing movement.
Assume the required layer thickness is set to be δ, and cure depth of the tested material is C d. As shown in
figure 4.3, after a layer with a thickness of δ” is cast, the platform is moved down towards the
Polydimethylsiloxane (PDMS) until its distance to PDMS surface is equal to δ. Ideally extra slurry in the
gap will flow outwards under the pressure of the platform, and a thin layer with the desired thickness δ can
71
be obtained between the platform and PDMS. However, due to the high viscosity of the slurry, the materials
can't flow out freely, but push the substrate down and lead to the elastic deformation of the PDMS film.
Therefore, the final slurry layer thickness, which is equal to the gap δ
*
between the two surfaces, will be
larger than the desired value δ and can result in the failure of the layer attachment to the previous layers.
The actual layer thickness depends on a lot of factors, including PDMS thickness, slurry viscosity,
blade height, and pressing speed. Thanks to the shear thinning behavior of polymer-ceramic composite
slurry, its viscosity will decrease as the shear rate increases. This allows for a way of controlling the actual
layer thickness by selecting a proper pressing speed. To achieve the desired layer thickness δ, we
investigated the relationship between the actual layer thickness δ
*
and pressing speed v p. In the experiments,
all the other factors are fixed: the PDMS film has a thickness of 1mm, blade height is 250μm, and tested
material is 70wt% Barium Titanate (BTO) with d50 = 3μm.
The procedures of measuring recoating thickness are described in figure 4.4. Three layers of bases are
built with pure resin to compensate for the clearance between the platform and PDMS surface. Resin has
a cure depth as high as 1000μm which can ensure its attachment to the platform. After base layers are built,
one layer of composite material is fabricated with the given pressing speed v p. If this layer fails to attach to
the base layers, the thickness δ
*
of the recoated layer under the speed v p must be bigger than cure depth C d
of the composite material. We then use C d as the actual layer thickness with respect to the pressing speed
v p for simplicity. If the layer attaches to the bases successfully, the actual thickness δ
*
should be smaller
than cure depth C d, and a wrapping method will be adopted to measure the thickness δ
*
. That is, after the
first layer of composite is built, all the fabricated layers, including the bases and the first layer, are covered
by a plastic wrap (thickness ~10μm). The fabrication of the second composite layer is then continued with
the same pressing speed v p. Due to the existence of the plastic wrap, the second layer will not attach to the
previous layers and can be measured separately by a length gauge (Heidenhain, Schaumburg, IL, US). The
actual layer thickness δ
*
is equal to the sum of the thickness of the plastic wrap and that of the second layer.
72
Figure 4.4 Measurement of recoated thickness
Figure 4.5 Actual layer thickness versus pressing speed. Note: (1) The cured layer at point A failed
to attach to the platform, so the actual layer thickness is larger than the measured thickness. (2) The cured
layer
0
20
40
60
80
100
120
0.098 0.195 0.391 0.586 0.781 0.977
Layer thickness( μm)
Pressing Speedv
p
(mm/s)
Actual layer thickness
A
B
73
The measurement results are shown in figure 4.5. It can be seen that the actual layer thickness decreases
as the pressing speed increases until a minimum thickness (equal to designed layer thickness) is obtained.
The curve is measured with fixed blade height, PDMS thickness and material viscosity. If different blade
height, PDMS coating or materials are used, the curve should be re-measured and pressing speed can be
accordingly identified based on the curve to minimize the actual layer thickness.
4.3 Layer separation via sliding mechanism
Sliding speed v s. After a layer is cured, the platform goes up for a small distance from the PDMS. Then
the film collector moves in the X axis with a speed v s to detach the newly cured layer from the PDMS film.
Suppose a cylindrical shape is being fabricated, whose diameter is d and the current building height is L.
As shown in figure 4.6, when the part slides from left to right, the portion of the cylinder immersed in the
slurry layer will experience a drag force F in the slurry moving direction as the average force q ( 𝑞 =
𝐹 𝛿 "
).
Figure 4.6 Drag force during the layer detachment
Based on the drag force equation, we know:
74
𝐹 =
1
2
𝜌 𝑣 𝑠 2
𝐶 𝐷 𝐴 (4.10)
where ρ is the mass density of the slurry; C D is the drag coefficient that is related to Reynolds number
R e of the slurry and can be identified by experiments; A is the reference area. In our case, 𝐴 =
𝜋 𝑑𝛿 "
2
.
Substitute equation (4.10) into the flexure formula of a cylinder 𝜎 𝑚 𝑎𝑥 =
𝑀𝑐
𝐼 ( 𝐼 =
𝜋 𝑑 4
64
is the area
moment of inertia for solid circular cross section), the maximal normal stress σ can be derived as:
𝜎 =
8 𝜌 𝑣 𝑠 2
𝐶 𝐷 𝛿 " 𝐿 𝑑 2
< [ 𝜎 ] (4.11)
Hence,
0 < 𝑣 𝑠 < √
[ 𝜎 ] 𝑑 2
8 𝜌 𝐶 𝐷 𝛿 " 𝐿 (4.12)
The material used in our tests has a minimal bending stress [σ] = 65MPa.
4.4 Base layers for initial gap compensation
In order to enable the initial layer to attach to the building platform, the platform plane has to be strictly
parallel to the substrate. In practice, exact alignment between the two surfaces can't be guaranteed due to
manufacturing and assembly errors. For composite slurry with relatively high cure depth (e.g. 200μm), this
error can be ignored, since the clearance can be filled up after the first layer is built. However, for composite
slurry with smaller cure depth (e.g. 50 μm), even a small clearance between the two planes (e.g. 20 μm)
will lead to unsuccessful attachment of the initial layers to the building platform. Furthermore, directly
removing a fabricated part from the building platform could cause defects at the bottom of the part. As a
result of the bottom defects, crack propagation will occur during the post processing.
For these reasons, we first build 5~10 layers of bases using pure photosensitive resin. This material,
serving as the base layers, has a lower melting point compared to the composite layers, such that the base
can be melted in the post processing without contaminating the final parts. By this base coating method, a
75
part can thus tightly stick to the building platform, even with a small cure depth (e.g. 50 μm), and can also
be easily removed from the platform after the fabrication is finished without damaging the bottom surface
of the green part.
4.5 Material Feeding
4.5.1 Moving distance of dispenser and its plunger
At the beginning of fabrication, the film substrate is moved to the dispenser, which drops a slurry line
onto the substrate surface. The doctor blade will then spread the slurry line out and form a thin slurry film
with a thickness of δ”, as derived in equation (4.8). Assume α=1 and the density before and after the casting
is the same, hence 𝛿 " =
𝛿 𝑏 𝑙 𝑎 𝑑 𝑒 2
. After one layer is fabricated, only a small portion of the recoated slurry layer
is cured, and all the other slurries are left on the substrate. In the next cycle, the film collector moves to a
wiper instead of the dispenser, to sweep the remaining materials to the leftmost end of the substrate. The
same wiping process will be done in the following cycles until the recoated slurry can't cover the projection
region. Assume the inner diameter of the dispenser chamber is D, the inner diameter of its nozzle is d (refer
to figure 4.7). The projection region is designed to be L by W, based on the bounding box of the 3D model.
Then we have the equation for the moving distance of the dispenser:
𝑊 𝑒 𝑥 𝑡 = 𝑊 (4.13)
The moving distance ℎ
𝑒 𝑥 𝑡 of the dispenser plunger for each material feeding can be calculated by
𝑆 ∙
𝛿 𝑏 𝑙 𝑎 𝑑 𝑒 2
= (
𝐷 2
)
2
𝜋 ℎ
𝑒 𝑥 𝑡 (4.14)
where S is the length of the recoating range. The left hand side of the equation means the volume of
the recoated slurry, and the right hand size is the volume of the extruded material. So extrusion distance is
decided by:
76
ℎ
𝑒 𝑥 𝑡 =
2 𝑆 𝛿 𝑏 𝑙 𝑎 𝑑 𝑒 𝐷 2
𝜋 (4.15)
Assume the projection region is positioned at the center of the recoating range. For the maximum
number of layers 𝑛 𝑚 𝑎𝑥 that material can fill up without new material dispensed, the equation should hold:
𝐷 2
4
𝜋 ∙ ℎ
𝑒 𝑥 𝑡 − 𝑛 𝑚 𝑎𝑥 𝑊𝐿
𝛿 𝑏 𝑙 𝑎 𝑑 𝑒 2
=
𝛿 𝑏 𝑙 𝑎 𝑑 𝑒 2
𝑊 𝑒 𝑥 𝑡 𝑆 + 𝐿 2
(4.16)
Then the number of layers recoated by wiper should satisfy:
1 ≤ 𝑛 ≤ 𝑛 𝑚 𝑎𝑥 =
𝐷 2
𝜋 ℎ
𝑒𝑥 𝑡 −
𝛿 𝑏 𝑙 𝑎 𝑑 𝑒 𝑊 𝑒𝑥 𝑡 ( 𝑆 + 𝐿 )
2 𝑊𝐿 𝛿 𝑏 𝑙 𝑎 𝑑 𝑒 (4.17)
Figure 4.7 Parameters in the dispensing system
4.5.2 Nozzle height
The distance between the nozzle tip of the dispenser and the substrate should also be controlled within
a suitable range, such that the cross section of the extruded slurry lines can maintain an arch shape with a
contact angle larger than 45° . As a consequence of the shear thinning behavior, the slurry will not spread
77
out after it leaves the nozzle. To avoid a squeezing effect (Wang, 2005), i.e. the slurry is forced to flow
beyond the volume formed by the nozzle height, the nozzle diameter and the distance the nozzle travelled
in unit time, the nozzle height h must be bigger than the critical nozzle height ℎ
𝑐 , and
ℎ
𝑐 =
𝑣 𝑝𝑙 𝑢 𝑛 𝑔 𝑒𝑟 𝜋 𝐷 2
4
𝑣 𝑑 𝑖 𝑠 𝑝 𝑑 =
𝜋 𝐷 2
ℎ
𝑒𝑥 𝑡 4 𝑑 𝑊 𝑒𝑥 𝑡 (4.18)
Then the nozzle height should follow:
ℎ
𝑐 < ℎ < 𝑑 (4.19)
4.6 Case Study I: fabrication of glass reinforced composite components
The capability of the slurry-based SLA process in fabricating viscous materials extends the breadth of
material selection. More materials, especially polymer-based composites, can be chosen in the SLA process
other than liquid resin. Low cost solid powders, such as glass fibers, carbon fibers, etc., can be employed
as a reinforcing phase in the resin matrix. The added reinforcing fillers can not only improve the mechanical
properties of the final parts, but also reduce the required amount of liquid resin in the process, and
consequently reduce the cost on the materials which are usually rather expensive.
Table 4.1 Process parameters used in the fabrication process for glass reinforced composite
Curing time
for base/normal
layers (s)
Layer
thickness
(μm)
Blade
gap δ blade
(μm)
Recoating
speed v r
(m/s)
Pressing
speed v p
(mm/s)
Sliding
speed v s
(m/s)
Glass 2/1 100 400 0.0254 0.078 0.0254
78
Figure 4.8 A “Hand” model fabricated with glass reinforced composite (a) CAD model (b-c) Two
views of the built object
Figure 4.9 A “Beethoven” model fabricated with glass reinforced composite (a) CAD model (b-c)
Two views of the built object
In this research, we investigate polymer based composite slurry by mixing sodium aluminosilicate
glasses (Z-CEL 8054, Potters, Valley Forge, PA) with liquid resin. Sodium aluminosolicate glasses are
white microspheres, originally developed as filler for paints, coatings and films. It has been used to improve
hardness and abrasion performance of final product.
5mm 5mm
5mm 5mm
(b)
(c)
(a) (b)
(c)
(a)
79
Composite slurry with 65% weight ratio of the glasses was prepared and used as the feedstock materials.
The properties of the composite slurry are shown in table 3.1. All the associated process parameters
discussed above are designed as in the table 4.1. Good optical transparency of the composite slurry allows
a layer thickness as large as 100μm. Two sample models are fabricated with the given parameters, and the
built objects are shown in the figure 4.8 and 4.9.
4.7 Case Study II: fabrication of high dielectric capacitor
Dielectric capacitors have attracted great interest as electrical energy storage devices, due to their
capability of charging within milliseconds and high fatigue and retention properties (Yang, 2016).
Polymer/ferroelectric ceramic composites and polymer/conductive filler percolative dielectric composites
are two types of materials that have been widely studied for dielectric capacitors (Ramajo, 2007;
Rahimabady, 2012; Mikolajek, 2015). Traditional fabrication methods for these devices embody drop
casting, spin coating, hot pressing, roll coating, etc., which are significantly restricted by their incapability
of producing complex structures. In this work, we fabricated 3D structured capacitors with the presented
slurry-based SLA process. Ag decorated lead zirconate titanate (PZT) was used as fillers to enhance the
dielectric properties of the composite material. This method offers the potential to reliably produce
capacitive components for printed circuit board, high-k gate dielectrics and embedded passive components.
4.7.1 Starting Materials
PZT powders (DeL. Piezo Specialties, LLC, West Palm Beach, FL) with an average particle size of
3μm were deagglomerated in a ball mill for 24 hours. Photocurable resin from Makerjuice Lab (Flex) was
selected as the matrix material in this study due to its excellent photosensitivity. Other used materials
include aqueous solution of H 2O 2 (30 wt%), toluene, ethylene glycol, silver nitrate (AgNO3), 3-
aminopropyltriethoxysilane, etc. All the chemicals were used as received unless particularly specified.
80
Although PZT powders can have a dielectric constant ranging from 300 to 3850 (wiki PZT 2016), the
dielectric properties of the PZT/polymer composites are still poor, due to the low dielectric constant of their
resin matrix. Surface decoration of the PZT particles with Ag (denoted as PZT@Ag) is one of the methods
to improve the compatibility and dielectric properties of the composites (Lu, 2008; Dang, 2013). By
introducing a conductive interlayer around a dielectric particle, polarization and charge aggregates, so
called Maxwell-wagner-sillars effect, can be enhanced (Panda 2008; Huang 2009b), such that the dielectric
permittivity can be consequently improved. PZT@Ag hybrid particles were synthesized via a seed-
mediated growing process by a redox reaction between silver nitrate and ethylene glycol, as shown in figure
4.10. Three steps were involved in the seed-mediated growing process: hydroxylation of as-received PZT
particles, functionalization of PZT-OH and grafting of nano-Ag particles (Yang 2016).
Figure 4.10 Scheme of the preparation process of the PZT@Ag nanoassemblies
Hydroxylation of as-received PZT particles: 10 g PZT particles were added into 50 ml aqueous solution
of H2O2 (30 wt%) in a round-bottomed flask. The mixture was sonicated for 30 min and then heated at
105
o
C for 4 h. The nanoparticles were recovered by centrifugation. The obtained PZT particles were washed
with deionized water for three times and then dried under vacuum at 80
o
C for 12 h. The hydroxylated PZT
particles were named as PZT-OH.
Functionalization of PZT-OH by chemical groups: 10 g PZT-OH nanoparticles and 2 g (3-aminopropyl)
triethoxysilane (APTES) were first added into 80 ml toluene in a round-bottomed flask and sonicated for
30 min, and then the mixture was heated at 100
o
C for 2h. The nanoparticles were by centrifugation. The
obtained nanoparticles were washed with toluene two times and were dried under vacuum at 80
o
C for 12
h and named as PZT-SH.
81
Grafting of nano-Ag particles onto the surface of PZT-SH: 10 g PZT-SH was added into 50 ml ethylene
glycol in a flask and sonicated for 30 min. Then ethylene glycol containing 0.5 g AgNO3 was slowly added.
The mixture was heated to 160
o
C and stirred under nitrogen for 2 h. The particles were recovered by
centrifugation and were washed with water and acetone several times. The obtained particles were denoted
as PZT@Ag.
PZT@Ag-polymer composites are prepared by mixing the synthesized PZT@Ag particles with the
Flex resin. The selected solid loading of the particles in our case is 18vol% or less, with the minimum filler
aggregates. The mixing is conducted in an ultrasonic bath for 1 hour and then stirred for 5 hours before use.
The introduction of conductive layer around PZT particle could improve the electric field distortion, but at
the same time it increases the refractive index of the fillers and hence decrease the cure depth to be less
than 40um. By making use of the advantages of our process in fabricating viscous materials with low
photosensitivity, we are able to fabricate capacitors with complex geometry from the developed materials.
4.7.2 Results and Discussion
For the PZT@Ag-polymer composite, a layer thickness of 25 µ m and cure time of 10s were selected
to fabricate capacitors, considering the 40µ m cure depth under the same cure time. Figure 4.11a.1
demonstrates several hexagonal patterns with different size. The maximum edge width is 1mm (figure
4.11a.2) and the minimum edge width is 200µ m (Figure 4.11a.4). Figure 4.11b.1 and figure 4.11b.2 show
two 3D capacitors with solid hexagon pillars or holes. The hexagon unit in both models has an average
diameter of 400 µ m and height of 600µ m (about 24 layers). In figure 4.12, we built a star and a bowl using
Flex and 18 vol% PZT@Ag. It can be seen that there are no macroscopically observed cracks in the parts
and the surface of each part is smooth, which are important for the electrical properties of the composites.
82
Figure 4.11 Fabricated capacitors (a) A cured layer of hexagonal pattern with different sizes (b) 3D
hexagonal patterns of Flex/PZT@Ag (18 vol%) composites
(a)
(
b)
(b)
83
Figure 4.12 Fabricated capacitors (a) Star structure (b) Bowl structure
(
a)
(
b)
(a)
(b)
84
Figure 4.13 (a) Dielectric permittivity of composite film w.r.t frequency (b) Loss tangent of
composite film w.r.t. frequency
(
a)
(
b)
(a)
(b)
85
Figure 4.14 Electrochemical measurements for the as-printed capacitors (a) Cyclic voltammetry
(CV) curves for the solid hexagonal pillar array (figure 4.11b.1) at different scan rates (5,10,50,100,200
mVS-1), (b) charge-discharge curves of the solid hexagonal pillar, (c) specific capacitances of different
types of 3D printed capacitors, (d) Capacitance retention of the solid hexagonal pillar array capacitor after
1000 cycles under current density of 5 Ag
−1
.
Dielectric permittivity (ε) and loss tangent (tan δ) of PZT@Ag-polymer composite films with different
volume fractions were measured by a precision impedance analyzer (Agilent 4294a, Keysight Technologies,
Santa Rosa, USA), as depicted in figure 4.13. The measurement was conducted in the frequency range of
100Hz-100MHz at room temperature. It can be seen in the figure that both ε and tanδ increase gradually
with filler content. The composite film with a 18vol% solid loading has the largest dielectric permittivity
of 120 (100Hz) and a low dielectric loss (0.028 at 100Hz). The dielectric permittivity of this composite film
(c)
(d)
(a)
(b)
86
is about 30 times higher than that of the pure Flex resin (ε=4). The loss tangent is slightly higher than the
other solid loadings though, it is still low enough, especially at a high frequency, to use in a capacitor for
reduced energy loss.
To measure the capacitance of each capacitor shown in figure 4.11b and figure 4.12, both sides of the
printed parts were sputter-coated with a gold electrode. Figure 4.14a shows cyclic voltammetry (CV) curves
of the 3D printed solid hexagonal pillar array capacitor, which exhibits nearly rectangular shapes at different
scan rates from 5 mVs-1 to 200 mVs-1, indicating this kind of capacitor possesses low resistance and ideal
capacitive properties. Figure 4.14b displays the gavanostatic (GA) charge/discharge curves of the capacitor
at different current densities. The specific capacitances of different types of 3D printed capacitor were
calculated from their charge/discharge curves according to the following equation:
sp
It
C
V
(4.20)
where C sp is the specific capacitance, I is the constant discharging current density, Δt is the discharging
time, and ΔV is the voltage window (Lu, 2012; Wang, 2012). The calculated specific capacitance of our 3D
printed capacitor is about 63 F g
-1
at the current density of 0.5 A g
-1
. Capacitance retention with respect to
cycle number was measured with the solid hexagonal pillar array capacitor. Figure 4.14d shows the result:
little degradation was observed for the tested capacitor after 1000 cycles, which indicates an excellent
electrochemical stability of the 3D printed capacitors.
4.8 Summary
In this chapter, we discussed the main parameters involved in the slurry-based SLA process, including
doctor blade height, recoating speed, separation speed, etc. These parameters are mainly determined by the
curing and rheological properties of the materials, and have to be changed if a different type of slurry is
used.
87
The tape casting system in our process is modeled as a Couette flow pattern, in which the flow rates
result from the drag force between the doctor blade and the film collector. Based on the boundary conditions
of this flow pattern, the dependence of the recoated layer thickness on the doctor blade height was derived.
Furthermore, a higher speed was suggested for the slurry recoating which can yield a lower viscosity of the
slurry.
An experimental method was presented to determine the layer pressing speed. The speed can reduce
the error of actual layer thickness arising from the deformation of PDMS film under the pressure of Z
platform.
A selection range for sliding speed was derived to enable the separation of each layer by the sliding
mechanism without breaking tiny features.
Equations are derived for the material feeding system, including the moving distance of the dispenser
and its plunger and the nozzle height, in order to guarantee a desired amount of slurry is delivered for each
layer.
Finally, two test cases were presented to demonstrate the application of our process in fabricating
composite components. The first case is to produce glass reinforced polymer composites. The fabricated
samples using our process suggest a good accuracy of our method. The second case is to build high
dielectric capacitors from resin/PZT@Ag composites. According to our measurement, the dielectric
permittivity of resin/PZT@Ag composite reaches as high as 120 at 100Hz with 18vol% filler, which is
about 30 times higher than that of pure resin. Although the loss tangent increases with the filler loading, it
still remains at a low level around 0.028. Additionally, we built capacitors in different shapes, and measured
their specific capacitance as 63 F/g at the current density of 0.5 A/g.
88
Chapter 5 Post Processing for Ceramic Component Fabrication
5.1 Overview
An important application of the presented slurry-based stereolithography (SLA) process is to fabricate
high-performance structural or functional ceramic materials, such as aluminum oxide and piezo-ceramic
component. Researchers have studied AM processes fabricating high-performance ceramic powders. A
developed method based on Selective Laser Sintering (SLS) is to use high energy laser beam to locally
sinter ceramic particles such that they will bond with each other. However, due to the high sintering
temperature of ceramics, low melting point media such as polymer powder is usually added into ceramic
powders. As shown in figure 5.1a, the low-melting point particles serve to bind ceramic particles together
in the SLS process. Finally, some post-processing procedures are performed to burn out the added binder
and to fuse ceramic particles.
Figure 5.1 Different methods of bonding ceramic powders in additive manufacturing
Ceramic Particles
Low Melting
Point Binder
Photocurable Binder
(a)
(b)
89
Similarly, slurry-based SLA process adds photocurable resin into ceramic powders, which serves to
bind ceramic particles together after being solidified by a light source (refer to figure 5.1b). Post processes
were also performed, during which cured resin is burned out, ceramic particles of interested are left behind
and grow together such that the sintered ceramic part becomes much denser.
Figure 5.2 Schematic of Post Processing for Ceramic Component Fabrication
As shown in figure 5.2, there are three steps involved in the ceramic fabrication by firing polymer-
ceramic composites. First ceramic powder is mixed with photocurable binder. The obtained viscous mixture
is fabricated into a polymer-ceramic composite component via the slurry-based SLA process. We call this
component as green part. In the second step, the green part is heated in a furnace, until the acrylate polymer
in the green part is completely decomposed into char, flammable liquids and gas. The flammable liquids
and gas are removed after the debinding, while the residual char in the debinded part serves as binders to
retain the shape of the part. Finally, the part is sintered in a furnace. During the sintering, all the residual
char reacts with the oxygen and turns into carbon dioxide. As a result, a pure and dense ceramic component
can then be obtained. In the following sections of this chapter, we introduced the fabrication of two popular
90
types of ceramics using the method described above: Aluminum Oxide and Barium Titanate (BTO), which
can be categorized as structure and functional ceramics respectively.
5.2 Case Study I: Traditional Ceramics for Structural Purpose
Aluminum oxide, commonly called alumina, is a crystalline raw material for a broad range of ceramic
product. It is widely used for its low electric conductivity, chemical inertness, high strength and hardness.
However, such excellent physical and chemical properties make it hardly be produced with conventional
machining processes. By using our slurry-based SLA process along with post processing, we successfully
generate alumina components with a relatively high density.
5.2.1 Green Part Fabrication
Alumina slurry with a solid loading of 65wt% is made by mixing the alumina powders with SI500
resin. The viscosity of the mixed slurry is around 5~250Pa•s with no diluent added in the suspension. The
main process parameters that are used in the fabrication process are listed in Table 5.1. A longer curing
time is used when building the initial 5-10 base layers; for other layers, a shorter curing time is used.
Table 5.1 Process parameters used in the fabrication process for alumina composite
Curing time
for base/normal
layers (s)
Layer
thickness
(μm)
Blade gap
δ blade (μm)
Recoating
speed v r
(m/s)
Pressing
speed v p
(mm/s)
Sliding
speed v s
(m/s)
Alumina 2/1 50 300 0.0254 0.293 0.0254
Figure 5.3 shows some of the fabrication results based on the selected process parameters. Figure 5.3a
is a gear with a diameter of 15mm and a thickness of 3.78mm. Figure 5.3b shows a model of 25 hexagon
91
array. Both the length and width of the part are 15mm. The height of the base is 1mm, and the height of
the hexagon is 1.26mm. Figure 5.3c shows a cutting tool in a diamond shape with a Φ3mm hole in the
center. In figure 5.3d is a tooth model.
Figure 5.3 Alumina Green Parts fabricated by the tape-casting-integrated SLA process
5.2.2 Post Processing
The green parts fabricated by the slurry-based SLA process is an alumina-resin composite part
consisting of organic compound (cured resin) and ceramic powders. In this section, we briefly discuss the
post processing of the fabricated green parts in order to achieve fully dense ceramic components. The main
post processing steps include organic binder burning-out and high-temperature sintering. A test case of a
gear model is used to demonstrate the post-processing procedures. The solid loading of alumina powders
is 65wt%. Based on the thermal analysis (TG-DSC) of the composite materials, the de-binding and sintering
procedures for alumina are designed as follows.
In the de-binding step, the organic composition in the green part is fully burned out. To avoid
delamination and pores, the green sample is put in a vacuum muffle furnace and debinded under a
temperature of 600° C for 3 hours. After the polymers in the green part are burned out, a high-temperature
92
sintering process is carried out to improve its density. The sintering of the debinded sample is conducted in
a regular muffle furnace at 1650° C for 1 hour. The heating schedules for both de-binding and sintering of
the Alumina gear part are shown in figure 5.4.
Figure 5.4 Post processing of Alumina green parts. (a) De-binding schedule; (b) sintering schedule
(a)
(b)
93
5.2.3 Testing Results of Sintered Samples
The Alumina green part and the corresponding sintered part are shown in figure 5.5a and 5.5b,
respectively. The diameter of the green part is 15mm. After the de-binding and sintering procedures, the
diameter of the gear part is ~11.6mm. The shrinkage of the presented ceramics fabrication process for the
gear model is ~22.7%. Such shrinkage can be compensated by enlarging the original CAD model by an
accordingly estimated ratio.
Figure 5.5 A test case of a gear model. (a) The green part of an alumina gear model; (b) the
fabricated alumina gear after the de-binding and sintering procedures
Samples in the shape of a simple block are fabricated using the same materials and process settings as
the ones used in the fabrication of the gear model. The cross-sections of the block samples before and after
the post processing are prepared and put under a Scanning Electron Microscope (SEM). Figure 5.6 shows
the SEM images of green and sintered parts. Figure 5.6a shows that the polymer and alumina particles are
mixed uniformly throughout the entire volume, although some agglomerates in the green part may still exist.
The average size of the alumina particles in the green part is less than 1μm due to the ball milling process.
Figure 5.6b shows large alumina grains with obvious boundaries are generated after the de-binding and
sintering processes. Improvement in the post-processing can be made to remove pores in the final part by
(a) (b)
94
adding isostatic pressure during the sintering process. A density of 93.2% of true alumina density
(3.95~3.98g/cm
3
) is obtained in the sintered block samples
Figure 5.6 The SEM images of different samples. (a) Alumina green part; (b) sintered Alumina part
(
a)
(b)
(a)
95
5.3 Case Study II: Piezoelectric Ceramics for Ultrasound Imaging
5.3.1 Introduction to Ultrasound Transducer
Ultrasonic imaging is an important medical imaging technique. Since ultrasound poses no known risks
to patients, this technology has become one of the most widely used diagnostic tools in modern medicine
(Shung, 2005). An ultrasonic imaging system requires an ultrasound probe, as shown in figure 5.7a. One
of the core components of the ultrasound probe is a transducer array, which can produce mechanical energy
in response to electrical signals and produce electrical signals in response to mechanical stimulus
conversely. Ultrasound transducer arrays have been used to detect and visualize muscles, tendons, and many
internal organs due to their advantages, such as high bandwidth, fast response, and high sensitivity.
Piezoelectric components such as ultrasound transducer arrays are generally made of piezoelectric
ceramic materials, e.g. Barium Titanate and Lead Zirconate Titanate (PZT). Since these piezoelectric
materials have relatively poor machinability, transducer arrays are typically fabricated into simple shapes,
such as square or rectangle. In recent decades, new transducer designs based on aperiodic and non-
rectangular array shapes have been investigated in the ultrasound transducer industry to achieve more
efficient energy conversion (Akhnak, 2002; Thomenius, 2005). One of such design examples is shown in
figure 5.7e, in which a hexagonal pattern is used (Thomenius, 2005). The new design is found to have less
lateral mode coupling, contributing to a better acoustic efficiency. However, complex geometric designs
with small dimensions poses significant challenges on the fabrication of an ultrasound transducer array.
Current manufacturing methods have great difficulty in fabricating such complex piezoelectric components.
For example, the dice-and-fill process prepares piezo-components by cutting a piezo-ceramic plate into an
orthogonal array and filling the kerfs with polymer material, so it can only produce square arrays. To
address such a challenge, we investigated the fabrication of piezoelectric ceramic components using
additive manufacturing processes.
96
Figure 5.7 An illustration of ultrasound transducers in an ultrasound system
5.3.2 Related Work
Current approaches to fabricating piezoelectric components are mainly based on machining. Main
steps of machining approaches are shown in figure 5.8a. Functional ceramics are first built in bulk and then
cut into shapes by machining tools such as a dicing saw used in dice-and-fill techniques (Smith, 1991; Liu,
2001). However, cutting bulk piezoelectric materials becomes increasingly difficult as current trends on
piezoelectric component design require more complex geometries to enhance their performance (Smith,
1986). Moreover, machining processes usually have relatively big feature resolutions which are limited by
97
their machining tools. Some other machining processes such as laser dicing techniques (Lukacs, 1999;
Farlow, 2001) have been developed to fabricate smaller features. However, the ablation side effects and
conic shape of ceramic arrays fabricated by these techniques have been observed, which will affect the
piezoelectric performance.
Figure 5.8 A comparison of piezoelectric component fabrication based on machining and AM
processes.
Another approach to fabricating piezoelectric components is based on molding and additive
manufacturing (refer to figure 5.8b). In these approaches, piezocomposite slurry is first made by mixing
piezo-ceramic powders with polymers and solutions in certain mixture ratios. Various techniques have been
developed to define the desired geometry in green-parts. An example is composite micro molding and lost
silicon molding techniques (Hirata, 1997; Cochran, 2004), which consist of following steps: a silicon (or
98
plastic) mold is first made using the LIGA (lithography, galvano-forming and plastic molding) process
(Becker, 1986); then the piezocomposite slurry is cast into the mold; and finally the mold is removed after
applying a high temperature. The lost silicon molding or injection molding processes follow similar steps
(Gentilman, 1994; Chu, 1998). However, these methods are indirect processes with multiple steps and each
step requires significant effort. Other examples include chemical vapor deposition (CVD) (Aota, 2007)
and tape casting methods (Chartier, 1997), which have been developed to directly deposit piezo-ceramic
atoms or very thin tapes on a semiconductor substrate. However, the resulted geometry by these methods
is usually simple and it is difficult to control the processes for more complex shapes.
During the past thirty years, many novel additive manufacturing (AM) processes such as
Stereolithography, Selective Laser Sintering, and Fused Deposition Modeling (FDM) have been
successfully developed and commercialized (Bourell, 2009). These AM processes have been investigated
to fabricate piezo-ceramic parts before. For example, the fused deposition of ceramics (Lous, 2000; Safari,
2006) and robocasting processes (Cesarano, 1998) can directly fabricate piezo-ceramic parts by extruding
piezocomposite slurry from a controlled nozzle. However, since these processes generally have a limited
resolution and building speed, they are not suitable for the fabrication of ultrasound transducer arrays.
Several variations of stereolithography (Brady, 1997; Dufaud, 2002; Sun, 2002; Bertsch, 2004) have also
been developed by using a highly focused laser beam to scan over the ceramic slurry, but these processes
are usually slow and require the viscosity of the materials to be small, hence they can only fabricate
materials with low solid loadings. Digital projection devices such as Digital Micromirror Devices (DMDs)
provide powerful tools that can dynamically control the energy input of a projection image. By using these
digital devices in stereolithographic AM processes, a whole layer can be fabricated simultaneously and the
fabrication speed can thus become substantially faster. Several research and commercial projection systems
based on DMD have been developed (Farsari, 1999; Monneret, 1999; Bertsch, 2000; Sun, 2005; Lu, 2006;
Pan, 2012b; Song, 2015). However, most of the research are focused on the fabrication of photocurable
resin or structural ceramics such as alumina. Some of the previous work that considered piezo ceramics
99
(e.g. Dufaud, 2002) only studied the green-part fabrication; the heat treatment procedure and related
material property measurements were not discussed. In this section we presented our investigation on both
green-part fabrication and post processes. In addition to material property measurements, a functional
device was built to demonstrate the capability of the projection-based stereolithographic AM process in
piezocomposite fabrication.
5.3.3 Green Part Fabrication
BTO slurry used in our study is prepared as described in chapter 3.1. Tests on the curing characteristics
of BTO slurry have been performed. An image of a square with a length of 8.57mm was projected into the
system. The thickness and width of the cured films with different weight ratios were measured respectively
with a length gauge (Heidenhain, Schaumburg, IL) and a caliper. The relations between cure depth/width
and BTO weight ratio under different curing time (i.e. 1s, 2s, 8s and 16s) based on our experimental setup
is shown in figure 5.9. For simplicity, slurry samples with weight ratio 30%-50% were prepared by
ultrasonic mixing for 30 minutes. Slurry samples with weight ratios 60% to 80% were prepared by directly
mixing photocurable resin with BTO powders in a ball mill. The light intensity of our projection system is
~31.6mW/cm
2
measured by an illumination level meter (Simpson electric, WI). It can be seen from figure
5.9 that cure depth decreases as more BTO powders are added. The slight increase in cure depth from 70%
to 80% can be explained by non-homogenous mixing of 80% slurry due to its high viscosity, and this further
indicates the significance of the presented slurry preparation method. The cure depth curves suggest bigger
cure depth can be obtained by increasing curing time, but a longer curing time will also yield bigger
overcure in width. The experimental result shows the optimal weight ratio for BTO fabrication through our
system is 60-80wt% and a curing time of 2-8s, under which a reasonable cure depth can be obtained with
the minimum overcure width.
100
Figure 5.9 Curing characteristic of BTO slurry with varying BTO weight ratios
Piezoelectric ceramic powders have high refractive index with respect to the photosensitive resin in
the slurry. This results in a light penetration depth as low as 40μm. To ensure enough over cure beyond the
layers, we chose the layer thickness for the building process as 20μm. Then according to the doctor blade
height equation, the doctor blade height ’is set to 100μm with a safety factor β = 2.
101
Table 5.2 Parameters used in the fabrication process for piezoelectric composites
Figure 5.10 Fabrication result of a 0-3 PZT composite ultrasound transducer array with hexagon
pillars (a) CAD model (b) Green part fabricated by our process
Figure 5.11 Concave BTO element fabricated by our process
Curing time
for base/normal
layers (s)
Layer
thickness
(μm)
Blade gap
δ blade (μm)
Recoating
speed v r
(m/s)
Pressing
speed v p
(mm/s)
Sliding
speed v s
(m/s)
PZT 2/10 20 100
0.0254 0.195 0.0254
BTO 2/2 20 100
(a)
(b)
102
Figure 5.12 A 64 element BTO segment annular array fabricated by our process
With a light intensity of ~31.6mW/cm
2
, suitable cure time is chosen for each powder at which a
maximum cure depth and a minimum width overcure can be achieved. Based on the pressing speed curve
for 3 micron BTO, a pressing speed v p of 0.195mm/s was chosen in the process. All the selected parameters
are listed in the table 5.2.
The CAD model of a 0-3 PZT composite ultrasound transducer array with 25 hexagon elements is
shown in figure 5.10a. The size of the array is 6.5× 6.5× 1.0mm, and the width of each pillar is 0.85mm. A
concave ultrasound transducer element and an annular transducer array are fabricated with these parameter
settings for BTO as shown in figure 5.11 and figure 5.12. The transducer array is designed to have 64
elements that can be dynamically excited to achieve desired ultrasonic imaging shapes. Each element is in
a fan shape, with two edges intersecting at the center of the array.
5.3.4 Post Processing of BTO Green Parts
BTO green-parts fabricated by the AM process are a mixture of polymer and BTO particles. In the
mixture, BTO particles are separated by polymers, which prevent efficient transmission of
compression stress between the BTO particles when an external force is added. In order for the
fabricated component to obtain piezoelectric properties, the polymer has to be removed and the left-
behind parts need sintering to fuse BTO particles together. The debinding process is conducted first
103
to burn out the polymer in the samples. Following the debinding process, the sintering process is performed
to convert the debinded BTO green-parts into fully dense ceramic components with desired piezoelectric
properties. In this section the heat treatment procedures of the debinding and sintering processes are
discussed. The material properties, especially piezoelectric properties, of sintered BTO components are
measured and presented in section 5.3.5.
The temperature curves of both debinding and sintering processes of BTO green parts are shown in
figure 5.13:
Figure 5.13 Temperature schedules for debinding and sintering of BTO green parts
104
(1) In the debinding process (refer to figure 5.13a), an argon furnace is used to fire the fabricated green
parts. A batch of green parts can be heated at the same time in the furnace. The temperature rises from the
room temperature at a rate of 1° C/min, and is held at 200° C, 300° C, 400° C and 500° C for 30 minutes,
respectively. The polymer composition in the samples is fully pyrolyzed after the temperature were held at
600° C for three hours. During the debinding process, we use carefully designed conditions, including Argon
atmosphere, low heating rate, and 30-minute dwell time at different temperatures, to slow down the
chemical reaction and consequently avoid part damage due to the vapors that are generated in the pyrolysis
of polymer.
(2) After green parts are debinded in the argon furnace, the sintering process (refer to figure 5.13b) is
carried out in a regular furnace under a higher temperature (1330° C) with a dwell time of 6 hours. The ramp
up rate in the sintering process is set to 3° C/min. We tested different sintering temperatures within 1200-
1500° C, and finally choose 1330° C as the sintering temperature for BTO parts. The fabricated green parts
of concave transducer element (refer to figure 5.14) and segment annular transducer array (refer to figure
5.15) are debinded and sintered using the aforementioned heating procedures. The sintered components are
shown in figure 5.14bc and figure 5.15b. For the segment annular transducer array, the shrinkage during
the debinding and sintering processes is about 26.7% along the X and Y axes and about 34.3% along the Z
axis. The density of the sintered component is 5.64g/cm
3
, or ~93.7% of bulk BTO material (density
6.02g/cm
3
), measured by standard density test method ASTM B962-15.
Figure 5.14 The concave transducer element: (a) green part; (b)(c) after sintering
(a) (b) (c)
105
Figure 5.15 The segment annular transducer array: (a) greenpart ; and (b) after sintering
5.3.5 Testing Results of Sintered Samples
5.3.5.1 Material Property Measurements
In order to characterize the piezoelectric properties of sintered materials, a series of samples in a
cylindrical shape (diameter 10mm, thickness 3mm) were fabricated and post-processed. The debinded and
sintered samples were first analyzed using SEM. Figure 5.16a shows the sample surface after the debinding
process, and figure 5.16b characterizes the surface of a sample after 6 hours sintering. It can be seen that
polymers were burned out after the debinding process, which consequently left a lot of pores inside the
samples and made the sample become fragile. After the aforementioned sintering process, however, the
sample becomes much denser, as shown in figure 5.16b.
The structure of the samples was also examined using a Rigaku X-Ray diffractometer (Rigaku
Corporation, Tokyo, Japan). As shown in figure 5.17, the fabricated material has a relatively well-
crystallized perovskite phase and is suitable for multi-ferroic applications such as ultrasound transducers.
106
Figure 5.16 Scanning electron microscope images: (a) after debinding; (b) after sintering
(b)
(a)
107
Figure 5.17 The X-Ray diffractometer patterns of BTO powders and sintered samples
Table 5.3 Measured properties of BTO samples fabricated by our process compared with bulk BTO
material
Piezoelectric and Dielectric properties Printed BTO
Bulk BTO(Pardo, 2011;
Kim, 1998; Bechmann,
1956)
Curie Temperature T (° C) 122 120
Electromechanical coupling factor K t 0.474 0.35
Piezoelectric constant d 33 (pC N
-1
) 160 190
Dielectric constant ε (1kHz) 1350 1700
Dielectric loss tangent tanδ(1kHz) 0.018 0.03
108
To understand the performance of the developed AM process on fabricating BTO components, the
dielectric and ferroelectric properties of the sintered samples were measured. Circular Cr/Au electrodes
with a diameter of 10 mm were first deposited by sputtering onto the sintered BTO samples as top electrodes.
The dielectric properties were then measured using an Agilent 4294A impedance analyzer. The values of
dielectric constant and dielectric loss tan of the samples at 1 kHz are 1350 and 0.018 respectively.
The measured piezoelectric constant (d 33) of the fabricated samples is around 160pC/N. The measured
electromechanical coupling coefficient (K t) is around 0.474. The measured properties were compared with
those of bulk BTO material in Table 5.3. As can be seen in the table, both curie temperature and
electromechanical coupling factor of the printed BTO component are close to the real values of BTO
material. The piezoelectric constant is smaller than the true value but is enough for the component to
display piezoelectricity. Both dielectric constant and dielectric loss tangent are influenced by the existing
pores in the sintered parts and could be further improved by increasing the final density. Furthermore, it
should be noted that the properties of the final materials largely rely on the sintering conditions, such as
temperature and dwell time. In figure 5.18, the measured properties of samples sintered at different
temperatures were compared, including density, piezoelectric constant d33, dielectric constant and loss
tangent. It can be seen that the 1330
o
C sample owns the best overall properties. In addition, our experiments
also indicate that the properties of samples sintered under a shorter dwell time (e.g. 4 hours) are not as good
as the ones sintered for 6 hours. The interested reader is referred to (Song, 2017).
Polarization field (P–E) hysteresis properties were also evaluated using a radiant precision materials
analyzer (Radiant Technologies, Albuquerque, NM). Figure 19 shows the ferroelectric hysteresis (P-E loop)
of the sintered samples under different electric fields. It can be observed that the P-E loop exhibits good
symmetry, which suggests satisfactory ferroelectricity of the fabricated samples. The remnant polarization
(P r) are 2.2μC/cm
2
, 2.4μC/cm
2
and 7.0μC/cm
2
,
respectively, when the electric fields are 10kV/cm, 20kV/cm
and 30kV/cm.
109
Figure 5.18 Property measurement of samples sintered at different temperature (a) Density and
piezoelectric constant (b) dielectric constant and loss tangent
(a)
(b)
110
Figure 5.19 Polarization–electric field hysteresis loop of sintered BTO samples
5.3.5.2 Ultrasound Transducer Fabrication and Testing
To further demonstrate the developed AM process, a simple ultrasound transducer was designed and
fabricated by using the sintered concave BTO transducer element (figure 5.14) as the piezo layer. The
structure of the transducer is shown in figure 20ab.
Poling of the BTO sample was conducted by applying a 2V/μm polarization voltage field on the
transducer at 120° C for 30 min. The acoustic performance of the transducer was measured through a pulse-
echo test in a deionized water bath at room temperature. In the pulse-echo test (refer to figure 5.20c), the
transducer is excited by broadband negative pulses emitted from a pulser/receiver unit (Panametrics
PR5900, Olympus NDT Inc., Waltham, MA) and generates ultrasonic signal in the degassed water. This
ultrasonic signal is transmitted in the water, and finally reflected by a quartz. After the echo signal is
received by the transducer, it will be converted into electrical signal, which will consequently be received
by the same pulser/receiver unit. Figure 5.21 shows the received echo signal reflected by a quartz at a
distance of about 5mm away from the transducer.
111
Figure 5.20 Application of a BTO sample in an ultrasound transducer: (a) Schematic diagram of the
ultrasound transducer; (b) the transducer assembled with the fabricated BTO sample; (c) pulse-echo test
of the fabricated transducer
Figure 5.21 Initial pulse and echo generated by the printed focused transducer
112
The echo electrical signal is then digitized by a 500MHz oscilloscope (LC534, LeCroy Corp., Chestnut
Ridge, NY). Figure 5.22a shows the time-domain echo signal and its frequency spectrum. Echo response
with an amplitude of 0.3V (± 0.15V) can be seen from 19.8us to 21us, which indicates the fabricated
transducer can effectively achieve the conversion between ultrasonic signal and electrical signal. Center
frequency (F c) of the transducer can be calculated from frequency F l and F h at the magnitude of -6 dB in
the frequency spectrum as: 𝐹 𝑐 =
𝐹 𝑙 + 𝐹 ℎ
2
=6.28MHz. The bandwidth (BW) of the transducer can be calculated
as: 𝐵𝑊 =
( 𝐹 ℎ
− 𝐹 𝑙 )
𝐹 𝑐 × 100% = 41 . 28%. According to the center frequency 6.28MHz and bandwidth
41.28%, it can be suggested that the BTO piezo based transducer that was fabricated by the AM process
has a good potential for clinic ultrasonic imaging.
Comparing with a flat piezoelectric element, a concave element can focus ultrasound signal at its focus
point, at which the strongest signal and the best resolution can be obtained. This property has been proved
by our experiment, as shown in figure 5.22b. At the focused location of 15.5mm, the tested sample has the
minimum lateral resolution (i.e. ~770μm) and the maximum echo signal strength (i.e. ~0.301V).
The principle purpose of the printed 6.28MHz ultrasound transducer is for ultrasound imaging. Figure
5.23 shows an ultrasound microscopy image of a pig eyeball using the fabricated transducer. From the
picture one can clearly see the inner structure of the eyeball, including the cornea, its constituent layers,
and the anterior and posterior chambers.
113
Figure 5.22 (a)Time-domain response and frequency spectrum of echo signal of the fabricated
transducer; (b) Voltage and lateral resolution as the function of depth between the transducer and the
target
(a)
114
Figure 5.23 (a) The 6.28MHz ultrasonic scan through porcine eyeball using the printing focused
transducer. (b) Ultrasonic imaging of porcine eyeball
(a)
115
5.4 Summary
In the chapter, the tape-casting-integrated projection-based SLA process was applied in the fabrication
of ceramic components, including alumina and piezoelectric components. The developed process
overcomes the problems of high viscosity and low photosensitivity associated with high solid loading slurry
by using a slurry recoating method and a sliding motion design. The debinding and sintering processes have
been studied for the alumina and BTO green parts. The debinding process is only dependent on the polymers
in the green parts. Therefore, the temperature profiles in both types of materials are the same, except that
the debinding of alumina is conducted in a vacuum furnace and that of BTO in an argon furnace. The reason
that we used a different furnace to process BTO samples is that they have a weaker inter-layer bonding
force than alumina, due to a smaller cure depth of BTO material, and the vacuum degree of the vacuum
furnace used in our process is not high enough to prevent crack and delamination of a green part during its
debinding.
After debinding and sintering, the final alumina part has a density of 93.2% of true alumina density
with a dimensional shrinkage of ~22.7%. The final BTO annular transducer array has a density of 93.7%
with a dimensional shrinkage of ~26.7%. Measured dielectric and piezoelectric properties of the final BTO
samples suggest great potentials of the developed AM process in fabricating piezoelectric components that
can be applied in functional devices such as ultrasound transducers. The new fabrication process would
enable the development of novel piezoelectric sensors and actuators through the use of 3d printed BTO
components in much more complex shapes.
116
Chapter 6 Post Processing for Porous Structure Fabrication
6.1 Introduction
Porous structure, also known as foam structure, is defined as a multi-scale structure which has a lot of
interconnected micro pores randomly distributed inside a relatively big dimension of volume. These
interconnected micro pores introduce many benefits to the structure, such as low density, low thermal
conductivity, high surface area and efficient stress transmission. By making use of these properties, porous
structures have a wide variety of industrial application such as oil absorption (Choi, 2011; Zhang, 2013),
heating/electromagnetic/sound shielding (Chen, 2013a; Yang, 2005), sensing and energy harvesting
(Marselli, 1999; Dai, 2012; Cha, 2011; Kara, 2003; Boumchedda, 2007; Lin, 2011), tissue engineering
(Hutmacher, 2000; Rodriguez, 2012) and sandwich structure (Daniel, 2011; Hung, 2014; Patrick, 2012), as
described in figure 6.1.
Due to their considerably wide range of usage, plenty of research papers have been reported on novel
processing technologies to fabricate foam structures with desired porosity. The most common approaches
of making foam structures include templating method (Chen, 2013b; Shastri, 2000; Lee, 2005) and foaming
agent-based method (Choi, 2011; Chen,2013a; Marselli, 1999; Ramay, 2003; Fukasawa, 2001; Liao, 2002;
Mikos, 1994; Hou, 2003; McCall, 2014; Lin, 2013; Dey, 2011; Bai, 2014). The templating method builds
a foam structure from an existing porous template. The template is then removed by dissolving the entire
structure in a certain type of solution. The foaming agent-based method is the most widely used approach
for the preparation of foam structures with high porosity and is usually employed to process polymer and
ceramic materials. In the foaming agent-based method, a specific type of foaming agent is mixed with the
matrix materials and a green part is fabricated from the mixture. The green part is then post-processed to
remove the foaming agent in the volume, such that interconnected pores can be generated in place of
117
foaming agent. Foaming agent varies in different research, but generally employs heat-decomposable or
soluble particles. For example, Choi et al.(Choi, 2011) and McCall (McCall, 2014) use commercially
available sugar as foaming agent to make highly porous polymer composite. In their method, free sugar is
added directly to uncured polymer composite. After a thin layer is cured from the mixture, the sugar
particles are removed by soaking the cured composite layer in hot water, leaving a three-Dimensional (3D)
isotropic network of air channels in the polymer composite. Both templating and foaming agent methods
can achieve high porosity in the final components. However, these methods can only deal with simple
porous structures, such as a thin layer, and have limited capability of fabricating complex 3D geometries.
Vigorous research studies have been conducted to directly make novel porous structures by Additive
Manufacturing (AM) technologies, but the porosity that can be achieved by AM processes is still limited
by their relatively small resolution in the fabrication of multi-scale features. For example, for
Stereolithography (SLA) process, it is extremely difficult to fabricate a macro-scale component with a large
number of micro-scale pores, due to the large overcure in both depth and width (Song, 2012).
In this research, we presented a SLA-based sugar foaming method to fabricate complex 3D porous
structures. In our method, micro sugar particles, as foaming agent, are mixed with photosensitive resin or
composites and thus form a viscous mixture of sugar and resin/composites. A desired geometry is then
fabricated from the mixture with the presented slurry-based SLA process. Similar to ceramic fabrication,
the fabricated component is called green part. The resin or composite is called body material. To obtain
micro pores in the green part, a boiling process is performed afterwards. In the boiling process, the green
part is soaked in hot water for a certain amount of time. The sugar particles inside the green part will
gradually dissolve in the hot water until a porous polymeric or composite structure is obtained. To facilitate
sugar removal in the hot water, special patterns are designed on the fabricated parts with thin yet strong
features.
The remainder of this chapter is organized as follows. The green part fabrication process is described
in the section 6.2. Process parameters such as solid loading, blade coating height and curing time are
118
analyzed in section 6.3. Section 6.4 discusses the minimum thin feature that can be fabricated by the
presented process and demonstrates some pattern designs to support the thin features. Section 6.5 analyzes
the application of this method in the fabrication of porous scaffolds in tissue engineering.
Figure 6.1 Applications of various foam structures
6.2 Process description
In the fabrication process of porous structures using slurry-based SLA, sugar particles serve as foam
agent. The sugar used in the process is bakery sugar, purchased from King Arthur Flour. The size of each
particle is around ~150µ m, which can dissolve quickly and completely in water. The photosensitive resin
(EnvisionTEC SI500) is used as the matrix material in the experiments. The resin has a viscosity of 180cP
and a density of 1.1g/cm
3
under the room temperature (EnvisionTec, 2012). The resin-sugar mixture is
prepared in the following steps: The sugar is first deagglomerated using a mortar and pestle. A certain
weight ratio of sugar is added into resin and mixed uniformly with a stirring rod for 30 minutes.
119
The addition of sugar into the resin will lead to dramatic increase in the viscosity. As discussed in
chapter 1, current SLA processes require the viscosity of the material to be smaller than 3000mPa•s. A
larger viscosity will make the refill of the material extremely difficult, such that new layers cannot be
fabricated successfully. Therefore, it is necessary to use the slurry-based SLA process to fabricate the
viscous mixture.
Figure 6.2 Schematics of the dynamic sugar templating method based on slurry-based SLA
As depicted in figure 6.2, when a fabrication cycle of one layer begins, a few amount of resin-sugar
mixture is dispensed onto the glass substrate behind the doctor blade. A green part can be consequently
built layer by layer by following the procedure described in chapter 3.5. After a green part is finished, it is
soaked in boiling water for a certain amount of time to remove the sugar, and a 3D porous structure can
thus be obtained.
120
6.3 Main process parameters
In this section, we discuss main process parameters associated with the green part fabrication in the
SLA-based sugar foaming method. The fabrication of a green part is greatly influenced by the process
parameters, such as solid loading of sugar in the mixture, doctor blade height, and curing time used in the
SLA process. The effect of each parameter on the process as well as their design methods are presented as
follows.
A tape casting process is conducted to coat the resin-sugar mixture into a slurry layer. The recoated
slurry layer is required to be thin and uniform, such that enough light can penetrate through the material to
solidify the layer and the layer can ultimately bond tightly with the previous layers. Furthermore, after a
layer of resin-sugar mixture is recoated on the glass substrate, the Z platform moves down to press the
recoated layer into a desired layer thickness. When the thickness of the recoated slurry layer is large, it will
break small features on the part that has been fabricated more easily during the pressing process. To avoid
the damage of small features, a smaller thickness of layer recoating is preferred. Doctor blade height is a
parameter that determines the thickness of the recoated layer. It is suggested in chapter 4.2 that the doctor
blade height should be two times bigger than the recoated layer thickness.
In order to obtain a component with higher porosity, a higher solid loading of sugar is desired for the
resin-sugar mixture. However, more sugar gives rise to an increase in the viscosity of the mixture and makes
the layer recoating more difficult, especially when the height of the doctor blade is small. To identify the
optimal solid loading and blade height for the resin-sugar mixture, we investigated the slurry recoating
under different combinations of sugar ratio and blade height, as shown in table 6.1. Four different solid
loadings are chosen from 40wt% to 70wt%. The blade height is set to 0.2mm, 0.4mm, 0.6mm and 0.8mm
respectively. The same amount of material was dispensed for each combination of solid loading and blade
height and the glass substrate moves underneath the doctor blade with the same speed. As can be seen in
the table, a blade height of 0.2mm is too small to achieve a uniform layer recoating of any solid loading
used in our tests. As the blade height increases, the recoated layer gets thicker but the uniformity gets
121
improved. Although a bigger solid loading (e.g. 70wt%) is desired to yield higher porosity, it can’t be coated
with the doctor blade successfully (refer to solid loading 60wt% and 70wt% in table 6.1). Among all the
combinations of sugar ratio and doctor blade height, 50wt% and 0.4mm are the optimal settings with the
biggest sugar percentage and the smallest recoated layer thickness. In the following experiments, 50wt%
and 0.4mm are chosen as the solid loading and doctor blade height, with which a thin and uniform layer
can be recoated.
Table 6.1 Slurry coating with different blade height and solid loadings
Sugar ratio/
Blade
height(mm)
40% 50% 60% 70%
0.2
0.4
0.6
0.8
In order to decide the layer thickness for slicing the Computer-Aided Design (CAD) model, curing
characteristics of the resin-sugar mixture with different weight ratios are studied, as depicted in figure 6.3.
122
In the experiments, enough slurry with different weight ratios was uniformly spread on the glass substrate.
An image of a circle with a diameter of 10mm was projected onto the bottom of the substrate for a certain
amount of time from 1s to 5s to cure the spread slurry. The thickness of the cured layer was measured with
a micro caliper and is recorded as cure depth of the tested material under the given cure time. The measured
thickness (or cure depth) of different materials with respect to different cure time was plotted in the figure
6.3.
It can be seen in figure 6.3 that the pure resin has the smallest cure depth, comparing with the resin-
sugar mixture. For the same material, cure depth increases with the curing time, since more light energy is
exposed under a longer cure time. As the solid loading of sugar in the mixture increase, the cure depth
becomes bigger. This is because the amount of energy that is lost when light pass through a sugar particle is
smaller than the energy that is absorbed when the light cures the same volume of resin. Hence the light can
go deeper into the slurry with more sugar particles, resulting in a bigger cure depth of the mixture with a
bigger solid loading.
Figure 6.3 Cure depth with different cure time for different solid loadings
123
6.4 Sugar removal
After the part is fabricated by the presented process, it contains both sugar and body material. In next
step, a boiling process should be used to dissolve the sugar in hot water and generate pores in place of the
sugar particles inside the part. In order to remove the sugar particles efficiently, the printed features should
be thin enough; otherwise some of the sugar particles would be trapped in the body and could not be removed
during the boiling process. In this section, we will discuss design of porous structures with the optimal feature
size, which can be built by the process.
We use a cylindrical shell structure (refer to figure 6.4a) and pure resin to study the minimum feature
that can be fabricated by our process. The outer diameter of the shape is 8mm. The height of the model is
3mm. The thickness of the shell changes from 0.2mm to 0.6mm. All the fabrication results are shown in
figure 6.4b. It can be seen in the figure that both shell structures at the thickness of 0.4mm and 0.6mm can
be made, while some portions of the fabricated shells at the thickness of 0.2mm and 0.3mm got broken, as
shown in the last image of figure 6.4b. That is because the parts with the shell thickness of 0.2mm and 0.3mm
were too weak to survive when they moved inside the viscous slurry. The experiment suggests that the
minimum feature that can be fabricated by our process without any damage is around 0.4mm. However,
since the particle size is around 150µ m, a smaller feature thickness (e.g. 0.2mm) is desired to completely
remove the sugar. To avoid the damage of thinner features below 0.4mm during the fabrication, we designed
different patterns to support the weak structures, as shown in figure 6.5.
The thickness of the outer shells of the three parts in figure 6.5 is 0.2mm. The shells have the same
diameter as the one in figure 6.4a, but were inserted three different types of pattern designs inside. The
pattern structure can support the thin features and retain the shape of the shell during the fabrication.
124
8mm
t
3mm
(a)
Figure 6.4 Cylinder shell structures with different thickness
125
To demonstrate the sugar removal process, we used the first pattern design in figure 6.5 and fabricated
three samples with a shell thickness of 0.2mm, 0.3mm and 0.4mm respectively. All the fabricated parts were
cleaned with alcohol in an ultrasonic cleaning machine for 10 minutes. The cleaned components were then
dried in the air for 1 hour. After that, all the three samples were put in boiling water to remove the sugar
particles for 2 hours. The weight of each sample was measured before (w 1) and after (w 2) the boiling. Assume
the weight ratio of the sugar in the mixture is Φ, the density of the sugar and resin is ρ s and ρ r respectively.
Then the porosity φ can be calculated as
𝜑 =
( 𝑤 1
− 𝑤 2
) ∙ 𝜌 𝑠 𝑤 1
∙ 𝛷 ∙ 𝜌 𝑠 + 𝑤 1
∙ ( 1 − 𝛷 ) ∙ 𝜌 𝑟 (6.1)
Figure 6.5 Different support patterns for thin features
The measured weights of each sample before and after the boiling are listed in table 6.2. The porosity
after 2 hours’ boiling is calculated using equation (6.1). The density of the sugar we used in the calculation
is 1.587 g/cm
3
. As shown in table 6.2, the achieved porosity is approximately 58%. Although the weight
change may contain loss of the polymer portion in the sample during the water boiling and will thus introduce
errors in the calculated porosity, the result could still suggest an increase of porosity in the final parts. Similar
ratios of weight loss in the three samples also suggest the weight loss of polymer should be small. The pores
that were generated after the removal of the sugar are shown in the figure 6.7b.
126
Figure 6.6 Microscope image after sugar removal
Table 6.2 Sample weight changes after boiling
Sample Thickness
(mm)
Weight before
boiling w 1 (g)
Weight after
boiling w 2 (g)
Porosity after 2h
0.2 0.057 0.029 58.0%
0.3 0.088 0.051 49.7%
0.4 0.122 0.064 56.15%
127
6.5 Test cases
An important application of our method is to increase the porosity of scaffolds fabricated by
stereolithography process in tissue engineering. Tissue engineering emerged in the early 1990s to conduct
tissue/organ repair by transplanting a biofactor (cells, genes and/or proteins) within a porous degradable
material known as a scaffold (Hollister, 2005). Numerous research studies have indicated that more tissue
ingrowth and new bone formation in vivo occurred in areas with higher porosity (Karageorgious, 2005).
Although SLA process has been used in the fabrication of porous scaffolds (Gauvin, 2012; Cooke, 2003),
the porosity that can been achieved is limited by the resolution of the process.
A scaffold design is given in figure 6.8. The edge length of pores is a and the width of a strut in the
scaffold is b. The thickness of a single layer of the scaffold is set to b, and the distance between two scaffold
layers is a. Assume the number of pores along all directions is n. Theoretically the porosity φ of the scaffold
can be calculated as:
𝜑 =
𝑎 3
𝑛 3
+ 3 𝑎 2
𝑏 𝑛 2
( 𝑛 + 1 )
( 𝑛 ∙ 𝑎 + 𝑛 ∙ 𝑏 + 𝑏 )
3
(6.2)
When fabricated by our tape casting integrated slurry-based SLA process in a bottom-up projection
manner, the actual pores will be smaller than the desired size both in XY plane and along Z direction, due to
the width and depth overcure of the light. For simplicity, we assume the overcure in the XY plane is
proportional to the length of a feature with a ratio of η xy, and the overcure along Z, which occurs at the
overhung features, is proportional to the thickness of the overhung feature with a ratio of η z, as depicted in
the figure 8. Then in the actual structure fabricated by our process, strut width will increase from b to b(1+
η xy), and the edge length of each pore will decrease from a to (a-b η xy). The thickness of the first layer of the
scaffold keeps constant at b, but all the other layers increase from b to b(1+ η z) and the distance between two
neighboring layers decrease from a to (a-b η z). Then the actual porosity of the scaffold fabricated by our
process can be calculated as:
128
𝜑 =
( 𝑎 − 𝑏 η
xy
)
2
( 𝑏 𝑛 2
+ 𝑏 𝑛 3
+ 𝑎 𝑛 3
) + 2 ( 𝑎 − 𝑏 η
z
) ( 𝑎 − 𝑏 η
xy
) 𝑏 ( 1 + η
xy
) 𝑛 2
( 𝑛 + 1 )
( 𝑛 ∙ 𝑎 + 𝑛 ∙ 𝑏 + 𝑏 + 𝑏 η
xy
)
2
( 𝑛 ∙ 𝑎 + 𝑛 ∙ 𝑏 + 𝑏 )
(6.3)
Pore size used in scaffolds is typically designed in the range of 100-600μm in order to optima for bone-
related outcomes (Karageorgious, 2005). Then the following constraints should be satisfied for the designed
scaffold:
𝑎 − 𝑏 η
xy
∈ [ 100 , 60 0 ]𝜇𝑚
𝑎 − 𝑏 η
z
∈ [ 100 , 6 00 ]𝜇𝑚
η
xy
, η
z
∈ [ 0 , 1 ]
η
xy
= 𝑓 ( 𝑏 )
η
z
= 𝑔 ( 𝑏 )
The overcure in the SLA process gives rise to a lower porosity in the scaffold directly fabricated by the
process. For example, in the case of a scaffold with designed pore size a=300μm, designed strut width
b=300μm, n=100, and η xy can be identified by an experiment as 30%, the actual porosity 𝜑 is calculated as
26.5% and a pore size of 200μm is achieved in the final part.
By using the sugar foaming method, we can further increase the porosity of the final scaffold. When
the actual porosity that can be achieved by the SLA process is φ a. As shown in the section 6.4, a porosity of
φ sugar =50% can be obtained using our sugar foaming method. Then after applying the sugar foaming method
to the designed scaffold structure, the combined porosity in the final scaffold is:
φ
a
+ φ
s u g ar
− φ
a
φ
s u g ar
(6.4)
Then the incremental ratio δ of the final porosity with respect to the one directly fabricated through the
SLA process is:
𝛿 =
φ
s u g a r
φ
a
− φ
s u g ar
(6.5)
129
When φ sugar =50% and φ a = 26.5% as calculated in the above example, the ratio δ can be as high as
138.7%.
Figure 6.7 Increase porosity by the sugar foaming method
We used the 65wt% glass composite discussed in chapter 4.6 as an example to further demonstrate the
sugar foaming method in the fabrication of composite material. The reason that we selected this composite
is that it has a relatively large cure depth (~590μm) compared with the particle size of sugar. The glass
composite was mixed with sugar particles at a weight ratio of 40%. The layer thickness during the fabrication
is selected as 200μm. The green part was soaked in hot water (80
o
C) for about 40 minutes. Figure 6.9c and
130
6.9d show the microstructures of the green part and figure 6.9e and 6.9f are the structures after the post
processing. The weight of the composite foam decreases from 1.004g to 0.783g after the boiling. The density
of the glass composite is calculated as 1.472 g/cm
3
. From equation 6.1 we can compute the porosity of the
final scaffold as 23%. A higher porosity may be obtained by increasing the sugar concentration and the
boiling time.
Figure 6.8 Glass composite foam structure fabrication by the slurry-based SLA
6.5 Summary
A sugar foaming method is used to post process the green parts fabricated by the slurry-based SLA and
generate porous structures. The material is prepared by mixing sugar particles with photosensitive resin or
composites. The mixture is then used in the tape-casting integrated slurry-based SLA process to produce 3D
components. Different process parameters were discussed to allow for the fabrication of resin-sugar slurry,
including solid loading, blade height, curing time, layer thickness, etc. In order to remove the sugar particles
efficiently, a part is printed as thin features with specifically designed internal patterns as the support. The
131
sugar particles within the printed components are dissolved in boiling water, such that pores can be generated
in the final part. Our experiment indicates a porosity of 50% can be achieved with the current process
parameters. The application of this method in tissue engineering is discussed to increase the porosity of
scaffolds that are fabricated by current SLA process.
132
Chapter 7 Conclusions and recommendations
Ceramic and composite components are materials of particular interest to engineers due to their unique
physical properties. The development of additive manufacturing (AM) has led to technological advances in
the fabrication processes of these materials. In this dissertation, particular emphasis is directed to slurry-
based stereolithography (SLA) process, which offers the potential to reliably produce accurate ceramic and
composite components through careful control of viscous slurry flow. We examined technique problems
associated with this method in the fabrication of ceramic and composite materials and demonstrated its
capability of building complex geometries with desired properties. In this chapter, we reviewed the research
questions and hypothesis as presented in chapter 1. We also discussed main contributions of our work in the
dissertation and gave recommendations for future research that could move the technology forward.
7.1 Answering the Research Questions and Testing Hypothesis
To verify two main hypotheses, we subdivided the two main research questions into three sub-questions
and come up with a sub-hypothesis corresponding to each sub-question. Each hypothesis is tested in prior
chapters as follows:
Hypothesis 1.1 A slurry mixture without any dilute solvent added will help avoid large deformation and
big shrinkage in the post processes.
Hypothesis1.2 Recoated layer thickness is restricted by slurry viscosity, which can be reduced by
increasing shear rate on the slurry.
Hypothesis1.3 Separation force becomes less by using sliding mechanism than by direct pull-up.
133
To verify hypothesis 1.1, the slurry formulation and its preparation method was presented in chapter
3.1. Several test cases were shown in chapter 4 and chapter 5, including composite capacitors, alumina
components and piezo-electric elements. Both the composite components and the ceramic green parts
indicate enough strength compared with the ones fabricated from diluted slurry.
To verify hypothesis 1.2, both curing and rheological characteristics of the slurries were studied in
chapter 3.2 and 3.3. In particular, the relationship between solid loading (viscosity) and blade recoating were
presented. A bottom-up projection based SLA process was integrated with tape casting process to facilitate
the layer recoating of viscous slurry. Process parameters including doctor blade height, recoating speed, layer
pressing speed, separation speed, etc., were discussed in chapter 4.
To verify hypothesis 1.3, the sliding mechanism was implemented in the tape-casting integrated SLA
process. The sliding mechanism was described in chapter 3.5 and the related parameters (i.e. separation speed)
were analyzed in chapter 4.3.
Hypothesis 2.1. The debinding process should be conducted slowly in order to avoid cracks and
delamination, and the sintering process should be in an air atmosphere.
Hypothesis 2.2. Properties for both structural (e.g. strength and hardness) and functional (e.g.
piezoelectricity) purposes can be achieved by debinding and sintering a green part fabricated by slurry-
based SLA.
Hypothesis2.3. Sugar foaming method can help increase the porosity of a foam structure fabricated by
SLA.
To verify hypothesis 2.1, different debinding and sintering conditions were studied to achieve high
density in the final parts without any crack and delamination. The debinding and sintering processes for
alumina and barium titanate (BTO) were given in chapter 5.2 and 5.3. In particular, the piezoelectric
properties of BTO components fabricated with different debinding and sintering conditions were tested and
compared in chapter 5.3.5.
134
To verify hypothesis 2.2, both structural and functional ceramic materials were tested, including alumina
and BTO. Their processing methods can be extended to many other ceramics, such as calcium phosphate,
Lead zirconate titanate (PZT), zirconia, etc. A density of 93% or more was achieved in the final ceramic
components.
To verify hypothesis 2.3, different process parameters were discussed to allow for the fabrication of
resin-sugar slurry. Thin features with special internal patterns were designed in order to remove the sugar
particles efficiently. A porosity of 50% was achieved with the current process parameters. The application
of this method in tissue engineering is discussed to increase the porosity of scaffolds that can be achieved by
current SLA processes.
7.2 Contributions and Intellectual Merit
Main contributions of our work can be summarized as follows:
(1) Used bottom-up projection-based SLA to process viscous slurry materials. Integrated tape casting
process into the system for layer recoating.
(2) Solved the technical problems associated with the developed process in fabricating viscous slurries,
including layer curing, layer recoating, layer detachment, etc.
(3) Established analytical models for main process parameters and provided parameter design methods for
the process. Applied the process in the fabrication of high dielectric capacitors.
(4) Provided post processing methods for the fabrication of structural and functional ceramics, including
alumina and barium titanate. Specifically, ultrasound transducer was fabricated using the presented
methods.
(5) Provided post processing methods for the fabrication of porous structures.
135
7.3 Recommendations for Future Work
Slurry-based AM process is a promising approach of fabricating ceramic and composite materials with
desired physical properties. In future, more work can be done in the following aspects.
(1) Surface finish and fabrication speed. Solid particles in the composite slurry diffuse the incident light
outwards in all directions, and reduce the resolution at the boundary of a layer. A big cure time will
yield gel around the boundary. The gel along the contour of a cured layer contributes to obvious
“staircase effect” on the surface of green parts. Using a smaller cure time could avoid the gel, but can
also reduce cure depth at the same time. In the next step, work can be done to improve the surface finish
by optimizing the image projection and developing new fabrication strategies. In addition, in the current
process, it takes around one minute to fabricate one layer. That is because both material dispensing and
layer recoating are required for each single layer. Future work can also include improving the process
to make it faster.
(2) Quality control for slurry-based SLA. An issue with ceramic fabrication using slurry-based SLA method
is its big shrinkage. Actually shape deformation happens in most AM processes. In future, geometric
shape error control methods can be studied to optimize the final geometric accuracy.
(3) Consolidation for functional ceramic fabrication. Compared with the piezoelectric components that are
fabricated by the traditional manufacturing processes, the BTO components fabricated the AM process
still have inferior piezoelectric properties. At this stage, the final part density is restricted by the
maximum solid loading of BTO in photocurable resin that will allow light to pass through and be able
to recoat. To further improve the piezoelectric properties, future work can 1) investigate different
methods to increase the maximum solid loading of BTO slurry, such as using more powerful light energy,
enhancing photosensitivity of BTO slurry, etc; 2) improve poling and electrode plating processes; and
3) study the benefits of using complex geometry in piezo component design to enhance transducers’
ultrasonic performance.
136
(4) Process-property-structure relationships. By adding solid fillers with diverse properties into liquid resin,
it becomes possible to obtain green parts with desired functionality. Thanks to the benefits of additive
manufacturing in fabricating complex geometries, special structures can now be designed to achieve
customized properties and functions. Further work can be done to explore the relationships between
special structures and properties. In addition, our process has great potentials in fabricating various
fiber-based polymeric composites with enhanced physical properties, whose dependence on the
fabrication process can be further studied in the future.
137
Reference
[1] Abouliatim, Y., Chartier, T., Abelard, P., Chaput, C. and Delage, C., 2009. Optical characterization of
stereolithography alumina suspensions using the Kubelka–Munk model. Journal of the European
Ceramic Society, 29(5), pp.919-924.
[2] Akhnak, M., Martinez, O., de Espinosa, F.M. and Ullate, L.G., 2002. Development of a segmented
annular array transducer for acoustic imaging. Ndt & E International, 35(7), pp.427-431.
[3] Aota, Y., Tanifuji, S., Oguma, H., Kameda, S., Nakase, H., Takagi, T. and Tsubouchi, K., 2007,
October. P1H-4 FBAR Characteristics with AlN Film Using MOCVD Method and Ru/Ta Electrode.
In Ultrasonics Symposium, 2007. IEEE (pp. 1425-1428). IEEE.
[4] Bai, C.Y., Deng,X.Y., Li,J.B., Jin,Y.N., Jiang,W.K., Liu,Z.M. and Li,Y., 2014. Fabrication and
properties of cordierite–mullite bonded porous SiC ceramics, Ceramics International, 40(4), pp.6225-
6231.
[5] Bakarich, S.E., Gorkin III, R., in het Panhuis, M. and Spinks, G.M., 2014. Three-dimensional printing
fiber reinforced hydrogel composites. ACS applied materials & interfaces, 6(18), pp.15998-16006.
[6] Bandyopadhyay, A., Panda, R.K., McNulty, T.F., Mohammadi, F., Danforth, S.C. and Safari, A., 1998.
Piezoelectric ceramics and composites via rapid prototyping techniques. Rapid Prototyping
Journal, 4(1), pp.37-49.
[7] Bechmann, R., 1956. Elastic, piezoelectric, and dielectric constants of polarized barium titanate
ceramics and some applications of the piezoelectric equations. The Journal of the Acoustical Society
of America,28(3), pp.347-350.
[8] Becker, E., Ehrfeld, W., Hagmann, P., Maner, A. and Mü nchmeyer, D., 1986. Fabrication of
microstructures with high aspect ratios and great structural heights by synchrotron radiation
lithography, galvanoforming, and plastic moulding (LIGA process). Microelectronic
engineering, 4(1), pp.35-56.
138
[9] Bergmann, C., Lindner, M., Zhang, W., Koczur, K., Kirsten, A., Telle, R. and Fischer, H., 2010. 3D
printing of bone substitute implants using calcium phosphate and bioactive glasses. Journal of the
European Ceramic Society,30(12), pp.2563-2567.
[10] Bergströ m, L., 1996. Rheological properties of concentrated, nonaqueous silicon nitride
suspensions. Journal of the American Ceramic Society, 79(12), pp: 3033-3040.
[11] Bertsch, A., Bernhard, P., Vogt, C. and Renaud, P., 2000. Rapid prototyping of small size
objects. Rapid Prototyping Journal, 6(4), pp.259-266.
[12] Bertsch, A., Jiguet, S., Hofmann, H. and Renaud, P., 2004. Ceramic microcomponents by
microstereolithography. In Micro Electro Mechanical Systems, 2004. 17th IEEE International
Conference on.(MEMS) (pp. 725-728). IEEE.
[13] Bhimasankaram, T., Suryanarayana, S.V. and Prasad, G., 1998. Piezoelectric polymer composite
materials. CURRENT SCIENCE-BANGALORE-, 74, pp.967-976.
[14] Bian, W., Li, D., Lian, Q., Li, X., Zhang, W., Wang, K. and Jin, Z., 2012. Fabrication of a bio-inspired
beta-Tricalcium phosphate/collagen scaffold based on ceramic stereolithography and gel casting for
osteochondral tissue engineering. Rapid Prototyping Journal, 18(1), pp.68-80.
[15] Bitterlich, B., Lutz, C. and Roosen, A., 2002. Rheological characterization of water-based slurries for
the tape casting process. Ceramics international,28(6), pp.675-683.
[16] Boumchedda, K., Hamadi,M. and Fantozzi, G., 2007. Properties of a hydrophone produced with
porous PZT ceramic, Journal of the European Ceramic Society, 27(13), pp.4169–4171.
[17] Bourell, D.L., Leu, M.C. and Rosen, D.W., 2009. Roadmap for additive manufacturing: identifying
the future of freeform processing. The University of Texas at Austin, Austin, TX.
[18] Brady, G.A. and Halloran, J.W., 1997. Stereolithography of ceramic suspensions. Rapid Prototyping
Journal, 3(2), pp.61-65.
[19] Cai, Y. and Jessop, J.L., 2004. Photopolymerization, Free Radical. Encyclopedia Of Polymer Science
and Technology, 10, pp.807-837.
[20] Cesarano III, J., King, B.H. and Denham, H.B., 1998. Recent developments in robocasting of ceramics
139
and multimaterial deposition (No. SAND--98-2195C; CONF-980826--). Sandia National Labs.,
Albuquerque, NM (United States).
[21] Cha, S.N., Kim,S.M., Kim,H.J., Ku, J.Y., Sohn,J.I., Park,Y.J., Song,B.G., Jung,M.H., Lee,E.K.,
Choi,B.L., Park,J.J., Wang,Z.L., Kim,J.M. and Kim,K., 2011. Porous PVDF As Effective Sonic Wave
Driven Nano-generators, Nano Letter, 11(12), pp. 5142–5147.
[22] Chabok, H., Zhou, C., Chen, Y., Eskandarinazhad, A., Zhou, Q., Shung, K., 2012. Ultrasound
transducer array fabrication based on additive manufacturing of piezocomposites. In: Proceedings of
the ASME 2012 international symposium on flexible automation, pp.433-444.
[23] Chartier, T., Hinczewski, C. and Corbel, S., 1999. UV curable systems for tape casting. Journal of the
European Ceramic Society, 19(1), pp.67-74.
[24] Chartier, T., Penarroya, R., Pagnoux, C. and Baumard, J.F., 1997. Tape casting using UV curable
binders. Journal of the European ceramic society,17(6), pp.765-771.
[25] Chen, Z.P., Xu,C., Ma,C.Q., Ren,W.C. and Chen,H.M., 2013a. Lightweight and Flexible Graphene
Foam Composites for high-performance electromagnetic interference shielding, Advanced Material,
25(9), pp.1296–1300.
[26] Chen, S.L., He,G.H., Hu,H., Jin,S.Q., Zhou,Y., He,Y.Y., He,S.J., Zhao,F. and Hou,H.Q., 2013b.
Elastic carbon foam via direct carbonization of polymer for flexible electrodes and organic chemical
absorption, Energy & Environmental Science, 6(8), pp.2435-2439.
[27] Chen, Y., Zhou, C. and Lao, J., 2011. A layerless additive manufacturing process based on CNC
accumulation. Rapid Prototyping Journal, 17(3), pp.218-227.
[28] Cheng, Y.L. and Lee, M.L., 2009. Development of dynamic masking rapid prototyping system for
application in tissue engineering. Rapid Prototyping Journal, 15(1), pp.29-41.
[29] Cheverton, M, Singh, P., Smith, L.S. and Chan, K.P., 2012. Ceramic polymer additive manufacturing
system for ultrasound transducer. In: Proceedings of Solid Freeform Fabrication Symposium, pp. 863-
875.
[30] Choi, J.W., Ha, Y.M., Lee, S.H. and Choi, K.H., 2006. Design of microstereolithography system based
140
on dynamic image projection for fabrication of three-dimensional microstructures. Journal of
mechanical science and technology, 20(12), pp.2094-2104.
[31] Choi, J.W., Wicker, R., Lee, S.H., Choi, K.H., Ha, C.S. and Chung, I., 2009. Fabrication of 3D
biocompatible/biodegradable micro-scaffolds using dynamic mask projection
microstereolithography. Journal of Materials Processing Technology, 209(15), pp.5494-5503.
[32] Choi, S.J., Kwon,T.H. Im,H., Moon,D.I., Baek,D.J., Seol,M.L., Duarte,J.P. and Choi,Y.K., 2011. A
Polydimethylsiloxane (PDMS) Sponge for the Selective Absorption of Oil from Water, ACS Appl.
Mater. Interfaces, 3(12), pp. 4552–4556.
[33] Chou, Y.T., Ko, Y.T. and Yan, M.F., 1987. Fluid flow model for ceramic tape casting. Journal of the
American Ceramic Society, 70(10).
[34] Christ, S., Schnabel, M., Vorndran, E., Groll, J. and Gbureck, U., 2015. Fiber reinforcement during
3D printing. Materials Letters, 139, pp.165-168.
[35] Chu, G.T., Brady, G.A., Miao, W. and Halloran, J.W., 1998. Ceramic SFF by direct and indirect
stereolithography. In MRS Proceedings (Vol. 542, p. 119). Cambridge University Press.
[36] Cochran, S., Abrar, A., Kirk, K.J., Zhang, D., Button, T.W., Su, B. and Meggs, C., 2004, August. Net-
shape ceramic processing as a route to ultrafine scale 1-3 connectivity piezoelectric ceramic-polymer
composite transducers. In Ultrasonics Symposium, 2004 IEEE (Vol. 3, pp. 1682-1685). IEEE.
[37] Compton, B.G. and Lewis, J.A., 2014. 3D ‐printing of lightweight cellular composites. Advanced
Materials, 26(34), pp.5930-5935.
[38] Cooke, M.N., Fisher,J.P., Dean,D., Rimnac,C. and Mikos,A.G., 2003. Use of Stereolithography to
Manufacture Critical-Sized 3D Biodegradable Scaffolds for Bone Ingrowth. Journal of Biomedical
Materials Research. 64(2), pp. 65-69.
[39] Dai, K., Ji,X., Xiang,Z.D., Zhang,W.Q., Tang,J.H. and Li, Z.M., 2012. Electrical Properties of an
Ultralight Conductive carbon nanotubepolymer composite foam. Polymer-Plastics Technology and
Engineering, 51(3), pp.304–306.
141
[40] Dang, Z.M., Yuan, J.K., Yao, S.H. and Liao, R.J., 2013. Flexible nanodielectric materials with high
permittivity for power energy storage. Advanced Materials, 25(44), pp.6334-6365.
[41] Daniel, I.M. and Cho,J.M., 2011. Characterization of Anisotropic Polymeric Foam under Static and
Dynamic Loading. Experimental Mechanics, 51(8), pp.1395-1403.
[42] de Hazan, Y., Thä nert, M., Trunec, M. and Misak, J., 2012. Robotic deposition of 3d nanocomposite
and ceramic fiber architectures via UV curable colloidal inks. Journal of the European Ceramic
Society, 32(6), pp.1187-1198.
[43] Deckers, J., Shahzad, K., Vleugels, J. and Kruth, J.P., 2012. Isostatic pressing assisted indirect
selective laser sintering of alumina components. Rapid Prototyping Journal, 18(5), pp.409-419.
[44] Deng, D., Chen, Y. and Zhou, C., 2012. Investigation on PEEK Fabrication Using Mask-image-
projection-based Stereolithography. Solid Freeform Fabrication Symposium. Austin, Texas, pp. 606-
616.
[45] Dey,A., Kayal, N. and Chakrabarti,O., 2011. Preparation of porous SiC ceramics by an infiltration
technique. Ceramics International, 37(1), pp.223-230.
[46] Diamond & Related Materials Laboratory, n.d. Available at:
http://drm.kist.re.kr/cerapedia/process/images/jgl00303.gif [Accessed May.17.2016]
[47] Dufaud, O. and Corbel, S., 2002. Stereolithography of PZT ceramic suspensions. Rapid Prototyping
Journal, 8(2), pp.83-90.
[48] Dufaud, O., Gall, HL and Corbel, S., 2003. Stereolithography of lead zirconate titanate ceramics for
MEMS applications. In: Microtechnologies for the New Millennium 2003 (pp. 28-37). International
Society for Optics and Photonics.
[49] Ebert, J., Özkol, E., Zeichner, A., Uibel, K., Weiss, Ö., Koops, U., Telle, R. and Fischer, H., 2009.
Direct inkjet printing of dental prostheses made of zirconia. Journal of dental research, 88(7), pp.673-
676.
[50] Eckel, Z.C., Zhou, C., Martin, J.H., Jacobsen, A.J., Carter, W.B. and Schaedler, T.A., 2016. Additive
manufacturing of polymer-derived ceramics. Science, 351(6268), pp.58-62.
142
[51] Elliott, A.M., Ivanova, O.S., Williams, C.B. and Campbell, T.A., 2013. Inkjet printing of quantum
dots in photopolymer for use in additive manufacturing of nanocomposites. Advanced Engineering
Materials, 15(10), pp.903-907.
[52] Envisiontec SI500, 2012. Available from: http://envisiontec.com/envisiontec/wp-
content/uploads/MSDS-SI500.pdf
[53] Evans, R.S., Bourell, D.L., Beaman, J.J. and Campbell, M.I., 2005. Rapid manufacturing of silicon
carbide composites. Rapid Prototyping Journal,11(1), pp.37-40.
[54] Farlow, R., Galbraith, W., Knowles, M. and Hayward, G., 2001. Micromachining of a piezocomposite
transducer using a copper vapor laser. Ultrasonics, Ferroelectrics, and Frequency Control, IEEE
Transactions on,48(3), pp.639-640.
[55] Farsari, M., Huang, S., Birch, P., Claret-Tournier, F., Young, R., Budgett, D., Bradfield, C. and
Chatwin, C., 1999. Microfabrication by use of a spatial light modulator in the ultraviolet: experimental
results. Optics Letters, 24(8), pp.549-550.
[56] Feygin, M. and Pak, S.S., Helisys, Inc., 1999. Laminated object manufacturing apparatus and method.
U.S. Patent 5,876,550.
[57] Friedrich, K., Fakirov, S. and Zhang, Z. eds., 2005. Polymer composites: from nano-to macro-scale.
Springer Science & Business Media.
[58] Fukasawa, T., Ando,M., Ohji,T. and Kanzaki, S., 2001. Synthesis of Porous Ceramics with Complex
Pore Structure by freeze-dry processing, Journal of the American. Ceramic Society., 84(1), pp.230–
232.
[59] Gauvin, R., Chen,Y.C., Lee,J.W., Soman,P., Zorlutuna,P., Nichol,J.W., Bae,H., Chen,S. and
Khademhosseini, A., 2012. Microfabrication of complex porous tissue engineering scaffolds using 3D
projection stereolithography. Biomaterials, 33(15), pp.3824-3834.
[60] Gay, D., 2014. Composite materials: design and applications. CRC press.
[61] Gentilman, R.L., Fiore, D.F., Pham, H.T., French, K.W. and Bowen, L.J., 1994. Fabrication and
properties of 1-3 PZT-polymer composites (No. CONF-931142--). American Ceramic Society,
143
Westerville, OH (United States).
[62] Gentry, S.P. and Halloran, J.W., 2013. Depth and width of cured lines in photopolymerizable ceramic
suspensions. Journal of the European Ceramic Society, 33(10), pp.1981-1988.
[63] Gladman, A.S., Matsumoto, E.A., Nuzzo, R.G., Mahadevan, L. and Lewis, J.A., 2016. Biomimetic
4D printing. Nature materials.
[64] Gomes, C.M., Rambo, C.R., De Oliveira, A.P.N., Hotza, D., Gouveˆa, D., Travitzky, N. and Greil, P.,
2009. Colloidal processing of glass–ceramics for laminated object manufacturing. Journal of the
American Ceramic Society,92(6), pp.1186-1191.
[65] Griffith, M.L. and Halloran, J.W., 1996. Freeform fabrication of ceramics via
stereolithography. Journal of the American Ceramic Society, 79(10), pp.2601-2608.
[66] Griffith, M.L. and Halloran, J.W., 1997. Scattering of ultraviolet radiation in turbid
suspensions. Journal of applied physics, 81(6), pp.2538-2546.
[67] Guo, D., Li, L.T., Cai, K., Gui, Z.L. and Nan, C.W., 2004. Rapid prototyping of piezoelectric ceramics
via selective laser sintering and gelcasting. Journal of the American Ceramic Society, 87(1), pp.17-22.
[68] Gurauskis, J., Baudí n, C. and Sá nchez-Herencia, A.J., 2007. Tape casting of Y-TZP with low binder
content. Ceramics international, 33(6), pp.1099-1103.
[69] Haikutech, n.d. Available at: http://www.haikutech.com/App_Themes/Default/7A_steel-belt-tape-
caster_S.jpg [Accessed May.17.2016]
[70] Hatzenbichler, M., Geppert, M., Gruber, S., Ipp, E., Almedal, R. and Stampfl, J., 2012, February.
DLP-based light engines for additive manufacturing of ceramic parts. In SPIE MOEMS-MEMS (pp.
82540E-82540E). International Society for Optics and Photonics.
[71] Hinczewski, C., Corbel, S. and Chartier, T., 1998a. Ceramic suspensions suitable for
stereolithography. Journal of the European Ceramic Society,18(6), pp.583-590.
[72] Hinczewski, C., Corbel, S. and Chartier, T., 1998b. Stereolithography for the fabrication of ceramic
three-dimensional parts. Rapid Prototyping Journal,4(3), pp.104-111.
[73] Hirata, Y., Nakaishi, H., Numazawa, T. and Takada, H., 1997, October. Piezocomposite of fine PZT
144
rods realized with synchrotron radiation lithography. In Ultrasonics Symposium, 1997. Proceedings.,
1997 IEEE (Vol. 2, pp. 877-881). IEEE.
[74] Hollister, S.J., 2005. Porous scaffold design for tissue engineering. Nature materials, 4(7), pp.518-
524.
[75] Hou, Q.P., Grijpma,D.W. and Feijen, J., 2003. Porous polymeric structures for tissue engineering
prepared by a coagulation, compression moulding and salt leaching technique. Biomaterials, 24(11),
pp.1937-1947.
[76] Huang, T., Mason, M.S., Zhao, X., Hilmas, G.E. and Leu, M.C., 2009a. Aqueous-based freeze-form
extrusion fabrication of alumina components. Rapid Prototyping Journal, 15(2), pp.88-95.
[77] Huang, X., Jiang, P. and Xie, L., 2009b. Ferroelectric polymer/silver nanocomposites with high
dielectric constant and high thermal conductivity. Applied Physics Letters, 95(24), p.242901.
[78] Hull, C.W., Uvp, Inc., 1986. Apparatus for production of three-dimensional objects by
stereolithography. U.S. Patent 4,575,330.
[79] Hung, T.C., Huang,J.S., Wang,Y.W. and Lin,K.Y., 2014. Inorganic polymeric foam as a sound
absorbing and insulating material. Construction and Building Materials, 50, pp.328-334.
[80] Hutmacher, D.W., 2000. Scaffolds in tissue engineering bone and cartilage. Biomaterials, 21(24), pp.
2529-2543.
[81] Inzana, J.A., Olvera, D., Fuller, S.M., Kelly, J.P., Graeve, O.A., Schwarz, E.M., Kates, S.L. and Awad,
H.A., 2014. 3D printing of composite calcium phosphate and collagen scaffolds for bone
regeneration. Biomaterials,35(13), pp.4026-4034.
[82] Jacobs F.P., 1992. Rapid prototyping and manufacturing: fundamental of stereolithography. Dearborn:
Society of Manufacturing Engineer.
[83] Jafari, M., Han, W., Mohammadi, F., Safari, A., Danforth, S.C. and Langrana, N., 2000. A novel
system for fused deposition of advanced multiple ceramics. Rapid Prototyping Journal, 6(3), pp.161-
175.
[84] Jang, J.H., Wang, S., Pilgrim, S.M. and Schulze, W.A., 2000. Preparation and characterization of
145
barium titanate suspensions for stereolithography. Journal of the American Ceramic Society, 83(7),
pp.1804-1806.
[85] Joshi, S.C., Lam, Y.C., Boey, F.Y.C. and Tok, A.I.Y., 2002. Power law fluids and Bingham plastics
flow models for ceramic tape casting. Journal of Materials Processing Technology, 120(1), pp.215-
225.
[86] Kara, H., Ramesh,R., Stevens,R. and Bowen,C.R., 2003. Porous PZT Ceramics for receiving
transducers. IEEE Transactions On Ultrasonics, Ferroelectrics, and Frequency Control, 50(3),
pp.289-296.
[87] Karageorgious,V. and Kaplan ,D., 2005. Porosity of 3D Biomaterial Scaffolds and Osteogenesis.
Biomaterials, 26(27), pp.5474-5491.
[88] Karalekas, D. and Antoniou, K., 2004. Composite rapid prototyping: overcoming the drawback of
poor mechanical properties. Journal of materials processing technology, 153, pp.526-530.
[89] Kenzari, S., Bonina, D., Dubois, J.M. and Fourné e, V., 2012. Quasicrystal–polymer composites for
selective laser sintering technology. Materials & Design, 35, pp.691-695.
[90] Kim, B. J., Park,T. G. and Kim, M. H. 1998. Temperature and frequency dependence of dielectric
properties of (Ba, Sr, Mg) TiO3 ceramic capacitors. Journal of the Korean Physical Society, 32:
pp.289-291.
[91] Klosterman, D., Chartoff, R., Agarwala, M., Fiscus, I., Murphy, J., Cullen, S. and Yeazell, M., 1999.
Direct Fabrication of Polymer Composite Structures with Curved LOM. Solid Freeform Fabrication
Symposium.
[92] Kokkinis, D., Schaffner, M. and Studart, A.R., 2015. Multimaterial magnetically assisted 3D printing
of composite materials. Nature communications, 6.
[93] Krieger, I.M. and Dougherty, T.J., 1959. A mechanism for non ‐Newtonian flow in suspensions of
rigid spheres. Transactions of The Society of Rheology (1957-1977), 3(1), pp.137-152.
[94] Kumar, A.V., Dutta, A. and Fay, J.E., 2004. Electrophotographic printing of part and binder
146
powders. Rapid Prototyping Journal, 10(1), pp.7-13.
[95] Kumar, S., Hofmann, M., Steinmann, B., Foster, E.J. and Weder, C., 2012. Reinforcement of
stereolithographic resins for rapid prototyping with cellulose nanocrystals. ACS applied materials &
interfaces, 4(10), pp.5399-5407.
[96] Le Ferrand, H., Bouville, F., Niebel, T.P. and Studart, A.R., 2015. Magnetically assisted slip casting
of bioinspired heterogeneous composites. Nature materials, 14(11), pp.1172-1179.
[97] Lebel, L.L., Aissa, B., Khakani, M.A.E. and Therriault, D., 2010. Ultraviolet ‐Assisted Direct-Write
Fabrication of Carbon Nanotube/Polymer Nanocomposite Microcoils. Advanced Materials, 22(5),
pp.592-596.
[98] Lee, D., Lee,J., Kim,J.B., Kim,J., Na,H.B., Kim,B., Shin,C.H., Kwak,J.H., Dohnalkova,A., Grate,J.W.,
Hyeon,T. and Kim,H.S., 2005. Simple Fabrication of a Highly Sensitive and Fast Glucose Biosensor.
Advanced Materials, 17(23), pp.2828-2833.
[99] Lee, J.H. 2002, Design and synthesis of biocompatible ceramic/polymer composites via
stereolithography, ProQuest Dissertations Publishing.
[100] Lee, J.S., Hong, J.M., Jung, J.W., Shim, J.H., Oh, J.H. and Cho, D.W., 2014. 3D printing of composite
tissue with complex shape applied to ear regeneration. Biofabrication, 6(2), p.024103.
[101] Leigh, S.J., Bradley, R.J., Purssell, C.P., Billson, D.R. and Hutchins, D.A., 2012. A simple, low-cost
conductive composite material for 3D printing of electronic sensors. PloS one, 7(11), p.e49365.
[102] Lewis, J.A., 2000. Colloidal processing of ceramics. Journal of the American Ceramic Society, 83(10),
pp.2341-2359.
[103] Lewis, J.A., 2006a. Direct ink writing of 3D functional materials. Advanced Functional
Materials, 16(17), pp.2193-2204.
[104] Lewis, J.A., Blackman, K.A., Ogden, A.L., Payne, J.A. and Francis, L.F., 1996. Rheological property
and stress development during drying of tape ‐cast ceramic layers. Journal of the American Ceramic
Society, 79(12), pp.3225-3234.
147
[105] Lewis, J.A., Smay, J.E., Stuecker, J. and Cesarano, J., 2006b. Direct ink writing of three ‐dimensional
ceramic structures. Journal of the American Ceramic Society, 89(12), pp.3599-3609.
[106] Liao, C.J., Chen,C.F., Chen, J.H., Chiang,S.F., Lin,Y.J. and Chang,K.Y., 2002. Fabrication of porous
biodegradable polymer scaffolds using a solvent merging/particulate leaching method, Journal of
biomedical materials research, 59(4), pp.676-681.
[107] Lin, C.T., Liao,L.D., Liu,Y.H., Wang,I.J., Lin,B.S. and Chang,J.Y., 2011. Novel Dry Polymer Foam
Electrodes for Long-Term EEG measurement, IEEE Transaction On Biomedical Engineering, 58(5),
pp. 1200-1207.
[108] Lin, Q.L., Luo,B., Qu,L., Fang, C.Q. and Chen,Z.M., 2013. Direct preparation of carbon foam by
pyrolysis of cyanate ester resin at ambient pressure. Journal of analytical and applied pyrolysis, 104,
pp.714-717.
[109] Lipke, D.W., Zhang, Y., Liu, Y., Church, B.C. and Sandhage, K.H., 2010. Near net-shape/net-
dimension ZrC/W-based composites with complex geometries via rapid prototyping and displacive
compensation of porosity. Journal of the European Ceramic Society, 30(11), pp.2265-2277.
[110] Liu, H. and Mo, J., 2010. Study on Nanosilica Reinforced Stereolithography Resin. Journal of
Reinforced Plastics and Composites, 29(6), pp.909-920.
[111] Liu, R., Harasiewicz, K.A. and Foster, F.S., 2001. Interdigital pair bonding for high frequency (20-50
MHz) ultrasonic composite transducers. Ultrasonics, Ferroelectrics, and Frequency Control, IEEE
Transactions on, 48(1), pp.299-306.
[112] Lous, G.M., Cornejo, I.A., McNulty, T.F., Safari, A. and Danforth, S.C., 2000. Fabrication of
piezoelectric ceramic/polymer composite transducers using fused deposition of ceramics. Journal of
the American Ceramic Society,83(1), pp.124-28.
[113] Love, L.J., Kunc, V., Rios, O., Duty, C.E., Elliott, A.M., Post, B.K., Smith, R.J. and Blue, C.A., 2014.
The importance of carbon fiber to polymer additive manufacturing. Journal of Materials
Research, 29(17), pp.1893-1898.
148
[114] Lu, J., Moon, K.S., and Wong, C.P., 2008. Silver/polymer nanocomposite as a high-kpolymer matrix
for dielectric composites with improved dielectric performance. Journal of Materials Chemistry, 18,
pp. 4821–4826
[115] Lu, X., Wang, G., Zhai, T., Yu, M., Xie, S., Ling, Y., Liang, C., Tong, Y. and Li, Y., 2012. Stabilized
TiN nanowire arrays for high-performance and flexible supercapacitors. Nano letters, 12(10),
pp.5376-5381.
[116] Lu, Y., Mapili, G., Suhali, G., Chen, S. and Roy, K., 2006. A digital micro ‐mirror device ‐based
system for the microfabrication of complex, spatially patterned tissue engineering scaffolds. Journal
of Biomedical Materials Research Part A, 77(2), pp.396-405.
[117] Lukacs, M., Sayer, M., Lockwood, G. and Foster, S., 1999. Laser micromachined high frequency
ultrasonic arrays. In Ultrasonics Symposium, 1999. Proceedings. 1999 IEEE (Vol. 2, pp. 1209-1212).
IEEE.
[118] Marselli, S., Pavia,V., Galassi,C., Rocari, E., Craciun,F., and Guidarelli,G., 1999. Porous piezoelectric
ceramic hydrophone, J. Acoust. Soc. Am. 106 (2), pp.733-738.
[119] Matsuzaki, R., Ueda, M., Namiki, M., Jeong, T.K., Asahara, H., Horiguchi, K., Nakamura, T.,
Todoroki, A. and Hirano, Y., 2016. Three-dimensional printing of continuous-fiber composites by in-
nozzle impregnation. Scientific reports, 6.
[120] McCall, W.R., Kim,K., Heath,C., Pierre,G.L. and Sirbuly,D.J., 2014. Piezoelectric
Nanoparticle−Polymer Composite Foams. ACS Appl. Mater. Interfaces, 6(22), pp.19504-19509.
[121] McNulty, T.F., Mohammadi, F., Bandyopadhyay, A., Shanefield, D.J., Danforth, S.C. and Safari, A.,
1998. Development of a binder formulation for fused deposition of ceramics. Rapid Prototyping
Journal, 4(4), pp.144-150.
[122] Melchels, F.P., Feijen, J. and Grijpma, D.W., 2009. A poly (D, L-lactide) resin for the preparation of
tissue engineering scaffolds by stereolithography. Biomaterials, 30(23), pp.3801-3809.
[123] Melcher, R., Martins, S., Travitzky, N. and Greil, P., 2006. Fabrication of Al 2 O 3-based composites
149
by indirect 3D-printing. Materials Letters, 60(4), pp.572-575.
[124] Mikeska, K.R. and Cannon, W.R., 1988. Non-aqueous dispersion properties of pure barium titanate
for tape casting. Colloids and Surfaces, 29(3), pp.305-321.
[125] Mikolajek, M., Friederich, A., Kohler, C., Rosen, M., Rathjen, A., Krü ger, K. and Binder, J.R., 2015.
Direct inkjet printing of dielectric ceramic/polymer composite thick films. Advanced Engineering
Materials, 17(9), pp.1294-1301.
[126] Mikos,A.G., Thorson,A.J., Czernwonka,L.A., Bao,Y., Langer,R., Winslow, D.N. and Vincati,
J.P.,1994. Preparation and characterization of poly(l-lactic acid) foams. Polymer, 35(5), pp.1068-1077.
[127] Mistler, R.E. and Twiname, E.R., 2000. Tape Casting: Theory and Practice, John Wiley & Sons, New
York.
[128] Monneret, S., Loubere, V. and Corbel, S., 1999, March. Microstereolithography using a dynamic mask
generator and a noncoherent visible light source. In Design, Test, and Microfabrication of
MEMS/MOEMS(pp. 553-561). International Society for Optics and Photonics.
[129] Morissette, S.L., Lewis, J.A., Cesarano, J., Dimos, D.B. and Baer, T., 2000. Solid freeform fabrication
of aqueous alumina–poly (vinyl alcohol) gelcasting suspensions. Journal of the American Ceramic
Society, 83(10), pp.2409-2416.
[130] Mott, M. and Evans, J.R., 2001. Solid freeforming of silicon carbide by inkjet printing using a
polymeric precursor. Journal of the American Ceramic Society, 84(2), pp.307-13.
[131] Mü hler, T., Gomes, C.M., Heinrich, J. and Gü nster, J., 2015. Slurry ‐Based Additive Manufacturing
of Ceramics. International Journal of Applied Ceramic Technology, 12(1), pp.18-25.
[132] Ning, F., Cong, W., Qiu, J., Wei, J. and Wang, S., 2015. Additive manufacturing of carbon fiber
reinforced thermoplastic composites using fused deposition modeling. Composites Part B:
Engineering, 80, pp.369-378.
[133] Olhero, S.M. and Ferreira, J.M.F., 2005. Rheological characterisation of water-based AlN slurries for
the tape casting process. Journal of materials processing technology, 169(2), pp.206-213.
150
[134] Paik, U., Hackley, V.A., Choi, S.C. and Jung, Y.G., 1998. The effect of electrostatic repulsive forces
on the stability of BaTiO 3 particles suspended in non-aqueous media. Colloids and Surfaces A:
Physicochemical and Engineering Aspects, 135(1), pp.77-88.
[135] Pan, Y., Zhao, X., Zhou, C. and Chen, Y., 2012a. Smooth surface fabrication in mask projection based
stereolithography. Journal of Manufacturing Processes, 14(4), pp.460-470.
[136] Pan, Y., Zhou, C. and Chen, Y., 2012b. A fast mask projection stereolithography process for
fabricating digital models in minutes. Journal of Manufacturing Science and Engineering, 134(5),
p.051011.
[137] Panda, M., Srinivas, V. and Thakur, A.K., 2008. Surface and interfacial effect of filler particle on
electrical properties of polyvinyledene fluoride/nickel composites. Applied Physics Letters, 93(24),
p.242908.
[138] Pardo, L. and Ricote, J. eds., 2011. Multifunctional polycrystalline ferroelectric materials: processing
and properties (Vol. 140). Springer Science & Business Media.
[139] Park, I.B., Ha, Y.M. and Lee, S.H., 2011. Dithering method for improving the surface quality of a
microstructure in projection microstereolithography. The International Journal of Advanced
Manufacturing Technology, 52(5-8), pp.545-553.
[140] Pascall, A.J., Qian, F., Wang, G., Worsley, M.A., Li, Y. and Kuntz, J.D., 2014. Light ‐Directed
Electrophoretic Deposition: A New Additive Manufacturing Technique for Arbitrarily Patterned 3D
Composites. Advanced Materials, 26(14), pp.2252-2256.
[141] Patrick,J.F., Sottos, N.R. and White, S.R., 2012. Microvascular based self-healing polymeric foam.
Polymer, 53(19), pp.4231-4240.
[142] Postiglione, G., Natale, G., Griffini, G., Levi, M. and Turri, S., 2015. Conductive 3D microstructures
by direct 3D printing of polymer/carbon nanotube nanocomposites via liquid deposition
modeling. Composites Part A: Applied Science and Manufacturing, 76, pp.110-114.
[143] Rahimabady, M., Yao, K., Arabnejad, S., Lu, L., Shim, V.P. and Chet, D.C.W., 2012. Intermolecular
151
interactions and high dielectric energy storage density in poly (vinylidene fluoride-
hexafluoropropylene)/poly (vinylidene fluoride) blend thin films. Applied Physics Letters, 100(25),
p.252907.
[144] Ramajo, L.A., Reboredo, M.M. and Castro, M.S., 2007. Characterisation of epoxy/BaTiO3
composites processed by dipping for integral capacitor films (ICF). Journal of Materials
Science, 42(10), pp.3685-3691.
[145] Ramay, H.R. and Zhang, M., 2003. Preparation of porous hydroxyapatite scaffolds by combination of
the gel-casting and polymer sponge methods. Biomaterials, 24(19), pp.3293-3302.
[146] Richerson, D., Richerson, D.W. and Lee, W.E., 2005. Modern ceramic engineering: properties,
processing, and use in design. CRC press.
[147] Rodriguez, J.N., Yu,Y.J., Miller,M.W., Wilson,T.S., Hartman,J., Clubb,F.J., Gentry,B. and
Maitland,D.J., 2012. Opacification of shape memory polymer foam designed for treatment of
intracranial aneurysms. Annals of biomedical engineering, 40(4), pp.883-897.
[148] Safari, A. and Janas, V.F., 1997. Processing of fine-scale piezoelectric ceramic/polymer composites
for transducer applications. Ferroelectrics,196(1), pp.187-190.
[149] Safari, A., Allahverdi, M. and Akdogan, E.K., 2006. Solid freeform fabrication of piezoelectric sensors
and actuators. Journal of materials science, 41(1), pp.177-198.
[150] Scheithauer, U., Bergner, A., Schwarzer, E., Richter, H.J. and Moritz, T., 2014. Studies on
thermoplastic 3D printing of steel–zirconia composites. Journal of Materials Research, 29(17),
pp.1931-1940.
[151] Scheithauer, U., Schwarzer, E., Richter, H.J. and Moritz, T., 2015. Thermoplastic 3D Printing—An
Additive Manufacturing Method for Producing Dense Ceramics. International Journal of Applied
Ceramic Technology, 12(1), pp.26-31.
[152] Schwentenwein, M. and Homa, J., 2015. Additive Manufacturing of Dense Alumina
Ceramics. International Journal of Applied Ceramic Technology,12(1), pp.1-7.
[153] Serra, T., Planell, J.A. and Navarro, M., 2013. High-resolution PLA-based composite scaffolds via 3-
152
D printing technology. Acta biomaterialia, 9(3), pp.5521-5530.
[154] Shahzad, K., Deckers, J., Kruth, J.P. and Vleugels, J., 2013. Additive manufacturing of alumina parts
by indirect selective laser sintering and post processing. Journal of Materials Processing
Technology, 213(9), pp.1484-1494.
[155] Shahzad, K., Deckers, J., Zhang, Z., Kruth, J.P. and Vleugels, J., 2014. Additive manufacturing of
zirconia parts by indirect selective laser sintering. Journal of the European Ceramic Society, 34(1),
pp.81-89.
[156] Shastri,V.P., Martin,I. and Langer,R., 2000. Macroporous polymer foams by hydrocarbon templating.
PNAS, 29(5), pp.1970-1975.
[157] Shofner, M.L., Lozano, K., Rodrí guez ‐Mací as, F.J. and Barrera, E.V., 2003. Nanofiber ‐reinforced
polymers prepared by fused deposition modeling. Journal of applied polymer science, 89(11),
pp.3081-3090.
[158] Shung, K.K., 2015. Diagnostic ultrasound: Imaging and blood flow measurements. CRC press.
[159] Smith, W.A. and Auld, B.A., 1991. Modeling 1-3 composite piezoelectrics: thickness-mode
oscillations. Ultrasonics, Ferroelectrics, and Frequency Control, IEEE Transactions on, 38(1), pp.40-
47.
[160] Smith, W.A., 1986, June. Composite piezoelectric materials for medical ultrasonic imaging
transducers-a review. In Applications of Ferroelectrics. 1986 Sixth IEEE International Symposium
on (pp. 249-256). IEEE.
[161] Song, X. and Chen, Y., 2012. Joint Design for 3-D Printing Non-Assembly Mechanisms. ASME 2012
International Design Engineering Technical Conferences and Computers and Information in
Engineering Conference. pp. 619-631.
[162] Song, X., Chen, Y., Lee, T.W., Wu, S. and Cheng, L., 2015. Ceramic fabrication using Mask-Image-
Projection-based Stereolithography integrated with tape-casting. Journal of Manufacturing
Processes, 20, pp.456-464.
153
[163] Song, X., Chen, Z., Lei, L., Shung, K., Zhou, Q., Chen, Y., 2017. Piezoelectric Component Fabrication
Using Projection-based Stereolithography of Barium Titanate Ceramic Suspensions. Rapid
Prototyping Journal. 23(1). In press.
[164] Song, Y., Liu, X. and Chen, J. 2004. The maximum solid loading and viscosity estimation of ultra-
fine BaTiO3 aqueous suspensions. Colloids and Surfaces A: Physicochemical and Engineering
Aspects, 247(1-3), pp.27-34.
[165] Standard F2792-12a, 2009. Available from:
http://www.astm.org/FULL_TEXT/F2792/HTML/F2792.htm
[166] Subramanian, K., Vail, N., Barlow, J. and Marcus, H., 1995. Selective laser sintering of alumina with
polymer binders. Rapid Prototyping Journal, 1(2), pp.24-35.
[167] Sun, C. and Zhang, X., 2002. The influences of the material properties on ceramic micro-
stereolithography. Sensors and Actuators A: Physical, 101(3), pp.364-370.
[168] Sun, C., Fang, N., Wu, D.M. and Zhang, X., 2005. Projection micro-stereolithography using digital
micro-mirror dynamic mask. Sensors and Actuators A: Physical, 121(1), pp.113-120.
[169] Suwanprateeb, J., Sanngam, R., Suvannapruk, W. and Panyathanmaporn, T., 2009. Mechanical and in
vitro performance of apatite–wollastonite glass ceramic reinforced hydroxyapatite composite
fabricated by 3D-printing. Journal of Materials Science: Materials in Medicine, 20(6), pp.1281-1289.
[170] Taboas, J.M., Maddox, R.D., Krebsbach, P.H. and Hollister, S.J., 2003. Indirect solid free form
fabrication of local and global porous, biomimetic and composite 3D polymer-ceramic
scaffolds. Biomaterials, 24(1), pp.181-194.
[171] Tang, H.H. and Liu, F.H., 2005. Ceramic laser gelling. Journal of the European Ceramic
Society, 25(5), pp.627-632.
[172] Tang, H.H. and Yen, H.C., 2015. Slurry-based additive manufacturing of ceramic parts by selective
laser burn-out. Journal of the European Ceramic Society, 35(3), pp.981-987.
[173] Tekinalp, H.L., Kunc, V., Velez-Garcia, G.M., Duty, C.E., Love, L.J., Naskar, A.K., Blue, C.A. and
Ozcan, S., 2014. Highly oriented carbon fiber–polymer composites via additive
154
manufacturing. Composites Science and Technology, 105, pp.144-150.
[174] Tesavibul, P., Felzmann, R., Gruber, S., Liska, R., Thompson, I., Boccaccini, A.R. and Stampfl, J.,
2012. Processing of 45S5 Bioglass® by lithography-based additive manufacturing. Materials
Letters, 74, pp.81-84.
[175] Thomenius, K., R. A. Fisher, D. M. Mills, R. G. Wodnicki, C. R. Hazard, L. S. Smith, 2005. Mosaic
Arrays Using Micromachined Ultrasound Transducers, US Patent No.: 6,865,140.
[176] Tian, X., Li, D. and Heinrich, J.G., 2012. Rapid prototyping of porcelain products by layer-wise slurry
deposition (LSD) and direct laser sintering. Rapid Prototyping Journal, 18(5), pp.362-373.
[177] Tok, A.I., Boey, F.Y. and Lam, Y.C., 2000. Non-Newtonian fluid flow model for ceramic tape
casting. Materials Science and Engineering: A, 280(2), pp.282-288.
[178] Tomeckova, V. and Halloran, J.W., 2010a. Predictive models for the photopolymerization of ceramic
suspensions. Journal of the European Ceramic Society, 30(14), pp.2833-2840.
[179] Tomeckova, V. and Halloran, J.W., 2010b. Cure depth for photopolymerization of ceramic
suspensions. Journal of the European Ceramic Society, 30(15), pp.3023-3033.
[180] Torabi, P., Petros, M. and Khoshnevis, B., 2014. Selective inhibition sintering: the process for
consumer metal additive manufacturing. 3D Printing and Additive Manufacturing, 1(3), pp.152-155.
[181] Travitzky, N., Bonet, A., Dermeik, B., Fey, T., Filbert ‐Demut, I., Schlier, L., Schlordt, T. and Greil,
P., 2014. Additive Manufacturing of Ceramic ‐ Based Materials. Advanced Engineering
Materials, 16(6), pp.729-754.
[182] Tumbleston, J.R., Shirvanyants, D., Ermoshkin, N., Janusziewicz, R., Johnson, A.R., Kelly, D., Chen,
K., Pinschmidt, R., Rolland, J.P., Ermoshkin, A. and Samulski, E.T., 2015. Continuous liquid interface
production of 3D objects. Science, 347(6228), pp.1349-1352.
[183] Ventura, S., Narang, S., Guerit, P., Liu, S., Twait, D., Khandelwal, P., Cohen, E. and Fish, R., 2000.
Freeform Fabrication of Functional Silicon Nitride Components by Direct Photo Shaping. In: MRS
Proceedings, 625(81).
155
[184] Waetjen, A.M., Polsakiewicz, D.A., Kuhl, I., Telle, R. and Fischer, H., 2009. Slurry deposition by
airbrush for selective laser sintering of ceramic components. Journal of the European Ceramic
Society, 29(1), pp.1-6.
[185] Wang, G., Lu, X., Ling, Y., Zhai, T., Wang, H., Tong, Y. and Li, Y., 2012. LiCl/PVA gel electrolyte
stabilizes vanadium oxide nanowire electrodes for pseudocapacitors. ACS nano, 6(11), pp.10296-
10302.
[186] Wang, J. and Shaw, L.L., 2005. Rheological and extrusion behavior of dental porcelain slurries for
rapid prototyping applications. Materials Science and Engineering: A, 397(1), pp.314-321.
[187] Wiki PZT. Available at: https://en.wikipedia.org/wiki/Lead_zirconate_titanate#cite_note-1 [Accessed
May.23.2016]
[188] Wilkes, J., Hagedorn, Y.C., Meiners, W. and Wissenbach, K., 2013. Additive manufacturing of ZrO2-
Al2O3 ceramic components by selective laser melting. Rapid Prototyping Journal, 19(1), pp.51-57.
[189] Yang, Y., Chen, Z., Song, X., Zhu, B., Hsiai, T., Wu, P.I., Xiong, R., Shi, J., Chen, Y., Zhou, Q. and
Shung, K.K., 2016. Three dimensional printing of high dielectric capacitor using projection based
stereolithography method. Nano Energy, 22, pp.414-421.
[190] Yang, Y.L., Goopta,M.C., Dudley, K.L. and Lawrence,R.W., 2005. Conductive carbon nanofiber-
polymer foam structures. Advanced Material, 17(16), pp.1999-2003.
[191] Yen, H.C., 2015. Experimental studying on development of slurry-layer casting system for additive
manufacturing of ceramics. The International Journal of Advanced Manufacturing Technology, 77(5-
8), pp.915-925.
[192] Yin, X., Travitzky, N. and Greil, P., 2007. Three ‐Dimensional Printing of Nanolaminated Ti3AlC2
Toughened TiAl3–Al2O3 Composites. Journal of the American Ceramic Society, 90(7), pp.2128-
2134.
[193] Zak, G., Chan, A.Y.F., Park, C.B. and Benhabib, B., 1996. Viscosity analysis of photopolymer and
glass-fibre composites for rapid layered manufacturing. Rapid Prototyping Journal, 2(3), pp.16-23.
156
[194] Zanchetta, E., Cattaldo, M., Franchin, G., Schwentenwein, M., Homa, J., Brusatin, G. and Colombo,
P., 2016. Stereolithography of SiOC Ceramic Microcomponents. Advanced Materials, 28(2), pp.370-
376.
[195] Zhang, A.J., Chen,M.J., Du,C., Guo, H.Z., Bai,H. and Li,L., 2013. Poly(dimethylsiloxane) Oil
Absorbent with a Three-Dimensionally Interconnected Porous Structure and Swellable Skeleton. ACS
Appl. Mater. Interfaces, 5(20), pp. 10201−10206.
[196] Zhou, C., Chen, Y. and Waltz, R.A., 2009. Optimized mask image projection for solid freeform
fabrication. Journal of Manufacturing Science and Engineering, 131(6), p.061004.
[197] Zhou, C., Chen, Y., Yang, Z. and Khoshnevis, B., 2013. Digital material fabrication using mask-
image-projection-based stereolithography. Rapid Prototyping Journal, 19(3), pp.153-165.
[198] Zhou, W., Li, D. and Wang, H., 2010. A novel aqueous ceramic suspension for ceramic
stereolithography. Rapid Prototyping Journal, 16(1), pp.29-35.
[199] Zocca, A., Colombo, P., Gomes, C.M. and Gü nster, J., 2015. Additive Manufacturing of Ceramics:
Issues, Potentialities, and Opportunities. Journal of the American Ceramic Society, 98(7), pp.1983-
2001.
[200] Zupancic, A., Lapasin, R. and Kristoffersson, A., 1998. Influence of particle concentration on
rheological properties of aqueous α-Al 2 O 3 suspensions. Journal of the European Ceramic
Society, 18(5), pp.467-477.
Abstract (if available)
Abstract
During the past thirty years, manufacturing community has benefitted from additive manufacturing (AM) technologies, thanks to many of its advantages including the capability of fabricating components directly from three dimensional (3D) computer-aided design (CAD) models. However, although it can build any complex geometry with a relatively rapid speed, its widespread adoption in manufacturing industry is substantially restricted by its limited material selection. Efforts have been made to fabricate advanced materials, such as composites and ceramics, with various AM processes, among which projection-based stereolithography (SLA) process has advantages of faster speed and higher accuracy over the other ones. In an SLA process for composite materials, a slurry mixture of solid filler and photosensitive resin is photocured layer by layer to form a green part. ❧ Main challenges in the composite slurry based SLA process include high viscosity, low photosensitivity, homogeneity, etc. Compared with liquid resin that are commonly used, the slurry made by mixing solid particles and liquid resin has an increased viscosity. High viscosity poses a challenge for layer recoating, in which a uniform thin layer needs to be created within a reasonable time. The maximum material viscosity that can be handled by conventional SLA processes is less than 3000mPa•S, whereas composite slurry usually has a viscosity far beyond this limit (e.g. 5~250Pa•S). Another main challenge in the composite slurry based SLA process is the reduced cure depth due to the light scattering of solid particles in the liquid resin. Smaller cure depth not only requires smaller layer thickness to be recoated, which is difficult for the viscous slurry, but also makes newly built layer more easily detach from previous layers due to a smaller bonding force. Furthermore, due to the high surface energy of such small particles, it is extremely easy for the particles in the slurry to aggregate. The inhomogeneous distribution of solid particles in green parts will lead to non-uniform stress inside the part during the post processing, which will consequently contribute to failure of the post processing, such as cracking and delamination. ❧ To overcome these challenges, we investigated the existing SLA processes, and presented a modified SLA process by integrating tape-casting method. The developed slurry-based SLA process has the capability of recoating uniform thin layer for highly viscous composite suspension. To achieve desired material properties, various approaches for increasing the solid loading of green composite parts are studied, including proper preparation of composite suspension, bottom-up image projection, tape-casting based recoating, a two-channel sliding design for layer separation, etc. Two types of polymer composite materials are tested to demonstrate the functionality of the new process, including glass reinforced composite and polymer-ceramic composite. The glass reinforced composite is fabricated from a slurry mixture of glass microspheres and resin. With glass microsphere as the fillers, the final composite can have improved hardness. In the fabrication of the polymer-ceramic composites, we used silver decorated lead zirconate titanate (PZT) as the fillers to enhance the dielectric properties of the composite materials. According to our measurement, the dielectric permittivity of resin/PZT@Ag composite reaches as high as 120 at 100Hz with 18vol% filler, which is about 30 times higher than that of pure resin. With the resin/PZT@Ag composite slurry, we built capacitors in different complex shapes, and measured their specific capacitance as 63 F/g at the current density of 0.5 A/g. ❧ A promising application of the slurry-based SLA process is to indirectly fabricate ceramic materials: A polymer-ceramic composite part fabricated by the slurry-based SLA process is heated in a furnace to burn out the polymer. Since the polymer in the composite part has a much lower melting point than the ceramics, the ceramic part of interest is left behind and sintered as the final product. The debinding and sintering processes for both aluminum oxide and barium titante (BTO) were studied. BTO-based ultrasound transducer arrays have been successfully fabricated using the presented methods. ❧ Another post processing method for the slurry-based SLA was also discussed to fabricate porous structures. Porous structure has wide application in industry, thanks to some of its special properties such as low density, low thermal conductivity, high surface area and efficient stress transmission. Both templating and foaming agent methods are used to fabricate porous structures. However, these methods can only produce simple geometries. In recent years, many research studies have been done to use SLA in the fabrication of porous structure, but the porosity that can be achieved is relatively small due to their limited accuracy in building micro-scale features on a large area. We applied the slurry-based SLA process in the fabrication of porous polymer and composite structures using sugar as the foaming agent. With a solid loading of 50wt% of the sugar in the resin, the method can achieve a porosity over 50%. This method can be used to increase the porosity achieved by current SLA methods by over 100%.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Deformation control for mask image projection based stereolithography process
PDF
Hybrid vat photopolymerization: methods and systems
PDF
Scalable polymerization additive manufacturing: principle and optimization
PDF
Selective separation shaping: an additive manufacturing method for metals and ceramics
PDF
Energy control and material deposition methods for fast fabrication with high surface quality in additive manufacturing using photo-polymerization
PDF
Hybrid vat photopolymerization processes for viscous photocurable and non-photocurable materials
PDF
Selective Separation Shaping (SSS): large scale cementitious fabrication potentials
PDF
Multi-scale biomimetic structure fabrication based on immersed surface accumulation
PDF
Additive manufacturing of piezoelectric and composite for biomedical application
PDF
Enhancing the surface quality and dimensional accuracy of SIS-metal parts with application to high temperature alloys
PDF
Metallic part fabrication wiht selective inhibition sintering (SIS) based on microscopic mechanical inhibition
PDF
Contour crafting construction with sulfur concrete
PDF
The extension of selective inhibition sintering (SIS) to high temperature alloys
PDF
Analyses of strength of layered structures fabricated by Contour Crafting
PDF
Fabrication of functional porous membranes via polymerization of solid monomer by a vapor-phase initiator
PDF
Motion-assisted vat photopolymerization: an approach to high-resolution additive manufacturing
PDF
Material and process development and optimization for efficient manufacturing of polymer composites
PDF
Fabrication of ultrasound transducer and 3D-prinitng ultrasonic device
PDF
Experimental and numerical techniques to characterize structural properties of fresh concrete
PDF
Developing efficient methods for the manufacture and analysis of composites structures
Asset Metadata
Creator
Song, Xuan
(author)
Core Title
Slurry based stereolithography: a solid freeform fabrication method of ceramics and composites
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Industrial and Systems Engineering
Publication Date
07/26/2016
Defense Date
06/17/2016
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
additive manufacturing,Ceramics,composites,digital light processing,digital micro-mirror device,OAI-PMH Harvest,slurry,stereolithography
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Chen, Yong (
committee chair
), Gupta, Satyandra Kumar (
committee member
), Khoshnevis, Berok (
committee member
)
Creator Email
soongx@gmail.com,xuansong@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-279140
Unique identifier
UC11281067
Identifier
etd-SongXuan-4620.pdf (filename),usctheses-c40-279140 (legacy record id)
Legacy Identifier
etd-SongXuan-4620-0.pdf
Dmrecord
279140
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Song, Xuan
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
additive manufacturing
composites
digital light processing
digital micro-mirror device
slurry
stereolithography