University of Southern California Dissertations and Theses

Enhancing the surface quality and dimensional accuracy of SIS-metal parts with application to high temperature alloys

(USC Thesis Other)

Enhancing the surface quality and dimensional accuracy of SIS-metal parts with application to high temperature alloys

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1

Enhancing the Surface Quality and Dimensional
Accuracy of SIS-Metal Parts with Application to
High Temperature Alloys

by

Payman Torabi

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 Payman Torabi
i

To my Parents, the heroes of my life

ii

Acknowledgements
This dissertation would not have been possible without thoughtful guidance, advice and support
of several individuals who provided their sincere and valuable assistance to me.
First and foremost, I would like to express my infinite gratitude to my parents for their
unwavering support, encouragement and prayers. My parents, my brother and my sisters have
always been by my side and will remain close to my heart forever.
I cannot begin to express my gratitude to my talented colleague and amazing friend, Matthew
Petros, for his passionate support, help, and collaboration in all these years. This is a friendship
I’ll treasure for the rest of my life.
I am lucky to have such awesome and supportive friends who have become my extended family
and made this country like a second home to me. I would like to express my deepest gratitude to
them. I am also extremely grateful to my colleague, partner, and friend, Hamed Gholivand, with
whom I started my research in Additive Manufacturing. His friendship is a blessing to me.
Next, I would like to thank all of my friends at the CRAFT Laboratories at USC who made working
there a pleasant experience for me. I owe Maria Yepremian a debt of gratitude for her
hardworking attitude, wisdom, and sense of responsibility.
Last but not least, I would like to offer my utmost gratitude to my advisor, Professor Behrokh
Khoshnevis, who provided me with the great opportunity of working with him. His ingenuity and
creativity as well as his valuable advice and supervision taught me many lessons that will never
be forgotten.
iii

Summary
Selective Inhibition Sintering (SIS) is a disruptive platform Additive Manufacturing (AM) process,
capable of printing parts from polymer, metal and ceramic base materials. The principle idea
behind the SIS process is the prevention of selected segments in each powder layer from
sintering. This research is focused on the enhancement of surface quality and dimensional
accuracy of the SIS-metal parts as well as incorporating new materials into the process. These are
achieved in two phases of the research.
In the first phase, an SIS-metal machine, called SIS-metal beta machine, is developed which
incorporates a high resolution commercial piezo-electric printhead. A systematic procedure is
proposed and implemented to experimentally configure and optimize print parameters. The SIS-
metal beta machine is successfully utilized in fabrication of high resolution bronze parts with
surface qualities and dimensional accuracies comparable to those of commercial metal AM
machines. The presented results suggest a massive improvement in the properties of the
fabricated parts.
The second phase of this research is dedicated to the adaptation of high temperature metals in
the SIS-metal process. For this purpose, a systematic step-by-step approach is developed for
preparing new inhibitor solutions and fine-tuning their fluid properties for optimum jettability
from the printhead and optimized droplet penetrations into the porous powder beds. This
approach is validated by fabricating high resolution stainless steel parts.

iv

Table of Contents
1 Chapter One: Introduction ...................................................................................................... 1
1.1 Additive manufacturing .................................................................................................... 1
1.2 Selective Inhibition Sintering (SIS) ................................................................................... 3
1.3 The Purpose of this Research ........................................................................................... 6
1.4 Organization of the chapters ........................................................................................... 7
2 Chapter Two: Background ....................................................................................................... 9
2.1 SIS-Polymer ...................................................................................................................... 9
2.2 SIS-Metal ........................................................................................................................ 12
2.3 Critique of past approaches ........................................................................................... 15
2.4 Problem Statement ........................................................................................................ 16
2.4.1 Primary problem ..................................................................................................... 17
2.4.2 Secondary Problem ................................................................................................. 17
3 Chapter Three: Literature Review on Droplet Penetration into Porous Powder Beds ......... 19
3.1 Introduction.................................................................................................................... 19
3.2 Droplet Penetration Phenomenon ................................................................................. 20
3.3 Penetration time ............................................................................................................ 26
3.4 The Effects of Fluid Properties on Penetration Time ..................................................... 28
3.5 The Effect of Droplet Size on Penetration Time............................................................. 31
v

3.6 Application in Additive Manufacturing .......................................................................... 33
3.6.1 Vertical and Lateral Penetration ............................................................................. 38
3.6.2 Saturation................................................................................................................ 44
3.7 Application in Selective Inhibition Sintering .................................................................. 46
3.8 Summary ........................................................................................................................ 48
4 Chapter Four: Research Methodology .................................................................................. 51
4.1 Introduction.................................................................................................................... 51
4.2 Inkjet Technology ........................................................................................................... 54
4.2.1 Inkjet Printing in Additive Manufacturing .............................................................. 56
4.2.2 Integration of Inkjet Technology in the SIS-metal process ..................................... 58
4.2.3 Thermal Printheads ................................................................................................. 59
4.2.4 Piezoelectric Printheads.......................................................................................... 61
4.2.5 Conclusion on Inkjet Technologies ......................................................................... 63
4.3 Calibration of Process Parameters ................................................................................. 66
4.4 Fluid-related Parameters ............................................................................................... 68
4.4.1 Inhibitor Solution .................................................................................................... 68
4.4.2 Surface Tension and Viscosity ................................................................................. 71
4.4.3 Surfactants .............................................................................................................. 74
4.4.4 Printer setting adjustments .................................................................................... 76
vi

4.5 Powder-related Parameters ........................................................................................... 78
4.5.1 AM Powder Production........................................................................................... 78
4.5.2 Particle Shape ......................................................................................................... 80
4.5.3 Particle Size and Size Distribution ........................................................................... 81
4.5.4 Powder Density ....................................................................................................... 82
4.5.5 Porosity ................................................................................................................... 82
4.5.6 Powder Flowability ................................................................................................. 84
4.6 Part Fabrication .............................................................................................................. 84
4.7 Application to high-temperature alloys ......................................................................... 88
5 Chapter Five: Enhancing the Resolution of SIS for Metallic Part Fabrication ....................... 90
5.1 Introduction.................................................................................................................... 90
5.2 SIS-Metal Beta Machine ................................................................................................. 90
5.2.1 Hardware ................................................................................................................ 90
5.2.2 Software .................................................................................................................. 93
5.3 Calibration of the Machine Parameters ......................................................................... 96
5.3.1 Inhibitor................................................................................................................... 96
5.3.2 Printer Settings ....................................................................................................... 99
5.4 Part Fabrication Process ............................................................................................... 104
5.4.1 Build Strategies ..................................................................................................... 106
vii

5.4.2 Support Structures ................................................................................................ 110
5.5 Fabrication of Complex 3D Parts .................................................................................. 113
5.6 Benchmark Part Fabrication ......................................................................................... 114
5.7 Surface Quality Measurements .................................................................................... 118
5.8 Conclusion .................................................................................................................... 121
6 Chapter Six: Application to high temperature alloys (Stainless Steel) ................................ 122
6.1 Introduction.................................................................................................................. 122
6.2 Applications of Stainless Steel in Industry ................................................................... 123
6.3 Inhibitor Candidate Determination .............................................................................. 124
6.3.1 Concentration levels ............................................................................................. 127
6.3.2 Droplet Experiments ............................................................................................. 128
6.4 Fluid Properties Fine-Tuning and Jettability ................................................................ 130
6.4.1 Viscosity ................................................................................................................ 131
6.4.2 Surface Tension ..................................................................................................... 134
6.4.3 Droplet Penetration Time ..................................................................................... 137
6.4.4 Droplet Shapes and Sizes ...................................................................................... 140
6.4.5 Surface quality ...................................................................................................... 144
6.4.6 Diluted Surfactant ................................................................................................. 146
6.5 Powder-related Parameters ......................................................................................... 149
viii

6.5.1 Stainless Steel 316L ............................................................................................... 149
6.5.2 Layer Thickness ..................................................................................................... 150
6.5.3 Powder Bed Density .............................................................................................. 150
6.5.4 Print Saturation Level............................................................................................ 151
6.5.5 Sintered Density .................................................................................................... 152
6.6 Part Fabrication ............................................................................................................ 152
6.7 Systematic approach for incorporating new materials in the SIS-metal process ........ 156
7 Chapter Seven: Conclusion and Future Work ..................................................................... 159
7.1 Conclusion .................................................................................................................... 159
7.2 Contributions ................................................................................................................ 160
7.3 Proposed Future Research Directions .......................................................................... 161
7.3.1 In-depth study of the factors affecting SIS-metal process ................................... 162
7.3.2 Adaptation of new materials into the process ..................................................... 162
8 References ........................................................................................................................... 164

ix

List of Figures
Figure 1-1: Generic process of CAD to part in AM processes ........................................................ 1
Figure 1-2: 2D visualization of sintering process ............................................................................ 4
Figure 1-3: The SIS-Metal process .................................................................................................. 5
Figure 2-1: The SIS-Polymer process .............................................................................................. 9
Figure 2-2: The SIS-Polymer Alpha machine ................................................................................. 11
Figure 2-3: Sample parts made by SIS-Polymer Alpha machine .................................................. 11
Figure 2-4: The SIS-Metal process with compression ................................................................... 13
Figure 2-5: The SIS-Metal alpha machine and respective sample parts ...................................... 14
Figure 2-6: SEM pictures of inhibited and uninhibited powder sections ..................................... 15
Figure 3-1: Penetration of a droplet into the pores .................................................................... 21
Figure 3-2: Powder bed modeled as capillaries ........................................................................... 21
Figure 3-3: Contact angle and interfacial energies of a droplet on a powder surface ................. 22
Figure 3-4: Experimental setup for capillary rise measurement ................................................. 24
Figure 3-5: Penetration of PEG200 into AI glass powder ............................................................. 27
Figure 3-6: Experimental measurement of penetration time of different liquids into powder .. 28
Figure 3-7: structural changes observed on the surface of powder caused ............................... 30
Figure 3-8: Spread of a droplet on a larger particle ..................................................................... 31
Figure 3-9: Schematic of the approximation representation of a spreading drop....................... 32
Figure 3-10: The effect of the fluid properties on the formation of a droplet by inkjet printing 36
Figure 3-11: The effect of surface tension on the formation of droplets .................................... 37
Figure 3-12: Penetration and spreading process of a sample droplet into a porous powder ..... 39
x

Figure 3-13: Single binder droplet effect printed using a DOD system on a powder bed ........... 40
Figure 3-14: influence of physical properties of fluids on printed line widths ............................. 40
Figure 3-15: The effect of layer thickness on fluid penetration .................................................. 41
Figure 3-16: Optical micrographs of 3D printed SMA square wire with different saturations. ... 43
Figure 3-17: The effect of particle size on fluid spreading time ................................................... 44
Figure 3-18: The effects of saturation level and layer thickness on printed line width .............. 45
Figure 3-19: Procedure for inhibitor selection ............................................................................. 47
Figure 4-1: Thin layer of inhibitor surrounding the part .............................................................. 51
Figure 4-2: Process routes for the deposition of ceramics using inkjet printing .......................... 55
Figure 4-3: Polyjet
TM
3D printing technology by Strarasys ........................................................... 57
Figure 4-4: Classification of inkjet technologies ........................................................................... 59
Figure 4-5: Droplet generation process in thermal printheads (source: www.aldertech.com) ... 60
Figure 4-6: Droplet generation process in piezoelectric printheads (source: Epson.com) .......... 61
Figure 4-7: Epson WorkForce 30 (source: Epson.com) ................................................................. 65
Figure 4-8: Three rows of nozzles in Epson WF30 printhead ....................................................... 65
Figure 4-9: Fishbone diagram showing important parameters in SIS-metal process .................. 68
Figure 4-10: SEM micrograph of "printed" vs. "unprinted" sections .......................................... 70
Figure 4-11: Schematic illustration of a surfactant....................................................................... 75
Figure 4-12: Epson WorkForce 30 control software ..................................................................... 77
Figure 4-13: Gas atomization process for metal powder production (source: LPW Technology) 79
Figure 4-14: Plasma atomization process for metal powder production ..................................... 80
Figure 4-15: Powder particle shapes (source: thelibraryofmanufacturing.com) ......................... 81
xi

Figure 4-16: Simple shape designs used in SIS-metal process ..................................................... 87
Figure 4-17: Benchmark part used in the evaluation of SIS-metal process ................................. 88
Figure 5-1: 3D desing of the SIS-metal beta machine .................................................................. 92
Figure 5-2: The SIS-Metal Beta machine ....................................................................................... 93
Figure 5-3: User interface software developed for SIS-metal beta machine ............................... 95
Figure 5-4: STL model of a part and one layer generated by SIS-metal slicing software ............. 96
Figure 5-5: VISCOlab 3000: Laboratory Viscometer ..................................................................... 98
Figure 5-6: The fabricated experimental part............................................................................. 103
Figure 5-7: Penetration depth vs. number of passes under different print qualities ................ 104
Figure 5-8: 3D model of the simple shapes used for studying build strategies ......................... 106
Figure 5-9: Part fabrication using HSA ....................................................................................... 107
Figure 5-10: Part fabrication using HSB ..................................................................................... 107
Figure 5-11: Part fabrication using HSC ..................................................................................... 108
Figure 5-12: Part fabrication using HSD ..................................................................................... 108
Figure 5-13: Gear fabrication using HSD ..................................................................................... 109
Figure 5-14: Part fabrication using HSE ..................................................................................... 109
Figure 5-15: 3D models of support structure study test parts ................................................... 110
Figure 5-16: Failed parts due to lack of support structures ....................................................... 111
Figure 5-17: Test parts with support structures ......................................................................... 111
Figure 5-18: Test part covered in fine ceramic powder for support .......................................... 112
Figure 5-19: Successful fabrication of the test parts using ceramic powder support ................ 112
Figure 5-20: Digital model of Einstein's head beside the fabricated part in SIS-metal process . 113
xii

Figure 5-21: Fabricated jet engine model (3D design provided by GE) ...................................... 114
Figure 5-22: Fabricated part for testing the resolution of the machine..................................... 115
Figure 5-23: Modified benchmarking part design ...................................................................... 116
Figure 5-24: The benchmarking part (a) digital model (b) fabricated part ................................ 118
Figure 5-25: (a) Bottom surface quality (b) top surface quality of the SIS-Metal test part ....... 119
Figure 5-26: Surface roughness experiment test part and the locations of measurements ..... 120
Figure 5-5: Top and bottom surface pore sizes .......................................................................... 120
Figure 6-1: Droplet experiments with 10 different concentration levels .................................. 130
Figure 6-2: Inhibitor fluid preparation process ........................................................................... 131
Figure 6-3: Brookfield AMETEK digital viscometer ..................................................................... 132
Figure 6-4: Viscosity measurements for solutions with different concentration levels ............ 133
Figure 6-5: Surface tensiometer manufactured by Central Scientific Company, Inc ................. 134
Figure 6-6: Surface of film at breaking point. ............................................................................. 135
Figure 6-7: Surface tension measurements for solutions with different concentration levels . 136
Figure 6-8: SciLogex single channel variable micropipette......................................................... 138
Figure 6-9: Experimental setup for droplet penetration time measurements .......................... 138
Figure 6-10: Droplet penetration time vs. concentration level .................................................. 139
Figure 6-11: Inkjet-printed droplets for inhibitor solutions ....................................................... 141
Figure 6-12: 500x500 px square printed for droplet size measurements .................................. 142
Figure 6-13: Printed surface quality of the solutions with different concentration levels ........ 145
Figure 6-14: Surface tension measurements versus the amount of diluted surfactant ............ 148
Figure 6-15: 316L stainless steel powder used in SIS-metal research ........................................ 149
xiii

Figure 6-16: 50x50x3 mm rectangular prism printed to determine bed density ....................... 151
Figure 6-17: A 1x1x0.12 mm shape element of a powder layer ................................................. 151
Figure 6-18: Cross section design of the first SIS-metal stainless steel part .............................. 153
Figure 6-19: (Left) Stainless steel gear- first trial - (Right) Microscopic image of the surface ... 154
Figure 6-20: (Left) Stainless steel gear- second trial-(Right) Microscopic image of the surface 154
Figure 6-21: 3D complex part model to be fabricated by SIS-metal (source: GE) ...................... 155
Figure 6-22: Sintered compressor part with inhibited regions around it ................................... 156
Figure 6-23: Final fabricated 3D stainless steel part................................................................... 156
Figure 6-24: Systematic approach for incorporating new materials in the SIS-metal process .. 158

xiv

List of Tables
Table 4-1: General characteristics of major AM processes using inkjet systems ......................... 58
Table 4-2: Thermal printheads vs. piezoelectric printheads ........................................................ 63
Table 4-3: Viscosity values of some common liquids ................................................................... 73
Table 4-4: Typical values of surface and interfacial tensions (mN/m) ......................................... 75
Table 5-1: Epson WF30 color printer characteristics .................................................................... 92
Table 5-2: Viscosity test results for magenta ink .......................................................................... 98
Table 5-3: Chemical Composition of the used bronze powder .................................................. 100
Table 5-4: Important factors and their levels used in the DOE .................................................. 101
Table 5-5: Factor combinations for the squares ......................................................................... 102
Table 5-6: Comparison of different print qualities ..................................................................... 104
Table 5-7: Feature size measurements ....................................................................................... 115
Table 5-8: Benchmarking part features ...................................................................................... 117
Table 6-1: Common metal oxides and their respective melting points ..................................... 125
Table 6-2: Potential chemical precursors to desired metal oxides ........................................... 126
Table 6-3: Magnesium Nitrate Concentration levels .................................................................. 128
Table 6-4: Droplet positions and their corresponding concentration level .............................. 129
Table 6-5: Droplet size measurements for different concentration levels ................................ 143
Table 6-6: DeMULS DLN-2314 characterisitcs (source: DeForest website) ................................ 146
Table 6-7: Chemical composition of 316L powder ..................................................................... 149

1

1 Chapter One: Introduction
1.1 Additive manufacturing
Additive manufacturing (AM) processes, also known as 3D printing, fabricate physical objects
directly from three-dimensional models and other digital data sources through layer by layer
manufacturing methods. All different AM processes share four general common steps (Figure
1-1):
1) A 3D model of the part is prepared by either modeling it using a CAD software in computer
or scanning a physical part.
2) The 3D model is digitally sliced into different layers with specific layer thicknesses.
3) The physical part is built layer by layer in the AM process.
4) The part is post-processed to achieve the desired aesthetics and/or mechanical
properties.

Figure 1-1: Generic process of CAD to part in AM processes [1]
2

AM technologies have applications in different industries such as automotive, aerospace, mold
making, jewelry, medicine, and art. These applications include prototyping, reverse engineering,
design optimization, functional part manufacturing, tool making, etc. Every AM technology and
process has strengths and weaknesses over other processes. Among various types of AM
technologies, none can claim to cover all different applications of 3D printing on its own. Each
application or industrial part needs to be studied separately so that the right technology for its
manufacturing can be determined.
However, AM technologies as a group can be compared to traditional manufacturing processes
leading to the following advantages and disadvantages of 3D printing in industry.
Advantages of AM Technologies include:
 Customized part manufacturing
 Design flexibility
 Fast and inexpensive prototyping
 No need for customized tooling
 Reduced manufacturing lead times for low-volume productions
 Reduced labor cost
 Reduced raw material waste
Disadvantages of AM technologies include:
 Expensive raw material
3

 Reduced mechanical performance
 Lower surface quality
 Limited choice of material
 Limited size
Different AM processes have been developed and commercialized to fabricate parts from
plastics, metals, and ceramics. AM metal processes are of interest to this research. The most
common commercial metal AM machines use either laser beam or electron beam to sinter or
melt metal powder particles and bond them together. The third type of metal machines use
inkjet technology to print binder liquid on each layer of metal powder and glue the particles
together. These industrial machines cost generally more than $250K and can fabricate parts
with delicate features and meet high tolerances. However, they are limited by several other
factors such as the purity of the material, speed of the process, in-part residual thermal
stresses, cost of raw material and maintenance, and size limitation. This research aims to tackle
some of these issues and provide a more viable solution to industrial metal additive
manufacturing through Selective Inhibition Sintering (SIS) process.
1.2 Selective Inhibition Sintering (SIS)
SIS is a novel AM process developed by the CRAFT Laboratories team at the University of
Southern California [2-5]. Sintering is the thermal process by which adjacent powder particles
are consolidated and bonded to each other by being heated to a temperature lower than the
material’s melting point. Figure 1-2 demonstrates a 2D visualization of sintering process. As
4

shown in the figure, pores exist in the process. For fixed powder particle sizes and shapes, the
pore sizes depend on the sintering temperature and the amount of heat received by the
particles. Lower values of porosity can be achieved by increasing temperature and sintering
hold time.

Figure 1-2: 2D visualization of sintering process
Inhibition, on the other hand, is the retardation of sintering process. Several factors and
methods can be used to inhibit sintering. In this research, the inhibition is achieved by the
introduction of an external high temperature material in between base powder particles which
will be explained in more details in the following chapters.
The core principle behind SIS is inhibiting selected regions of the powder to prevent sintering.
For fabricating a metallic part using SIS, a layer of powder is first spread onto the build tank. An
inhibitor solution is then deposited onto selected regions of the layer. The inhibitor is only
deposited on the boundary of the part without affecting internal regions. The next layer is then
spread over the previous one and the inhibitor is deposited again. These steps are repeated
until all the layers for the part are completed. The part is then bulk sintered in a furnace under
appropriate atmosphere. In this stage, all areas of the part will be sintered except the ones
5

treated by the inhibitor. Finally, the inhibited regions are manually removed, revealing the part.
An analogy would be to think of the part as if it were resting inside a sacrificial mold. The SIS-
Metal steps are summarized in Figure 1-3.

Figure 1-3: The SIS-Metal process
SIS-Metal has several advantages over other commercially available metal processes, such as:
 The machine will be much more economical due to the use of a commercial printhead
instead of expensive laser or electron beam generators [3].
 The process is faster since only the boundary of the part needs to be treated. This is in
contrast with other methods in which the whole area inside the part should be scanned
or treated [6].
6

 SIS-Metal is far more scalable regarding the size of the printed parts when compared to
beam based technologies. Larger parts can be built without reducing the resolution of
the build.
 Since any needed supports are attached to the “sacrificial mold”, and not the part itself,
the part surface quality is not affected by support structure.
1.3 The Purpose of this Research
Industrial AM processes are in need of a method to fabricate high quality parts without
sacrificing production speed. With this end in mind, inkjet technologies present a tremendous
opportunity to AM processes due to their high resolution and high speed print abilities.
Therefore, integrating inkjet printing into a potential process can vastly increase its capabilities.
The previous SIS research served as the proof-of-concept for SIS-Metal process and paved the
road for the current research. The goal of this research is to further develop the process,
increase the resolution, surface quality and dimensional accuracy of the parts, and provide a
systematic approach to incorporate high temperature alloys into the process to extend its
industrial applications.
This goal will be achieved in two steps. The first step is the development of a high resolution
SIS-metal machine using a piezo-electric printhead and calibration of its parameters so that the
final parts meet acceptable quality and dimensional accuracy requirements. The second step is
the detailed study of the governing mechanisms of liquid penetration into powder beds in both
7

vertical and lateral directions in order to come up with a systematic algorithm for adaptation of
a new material in the SIS-metal process. The latter will be validated through fabrication of
stainless steel parts.
1.4 Organization of the chapters
The organization of the chapters is as follows:
Chapter 1 provides a brief introduction to Additive Manufacturing and Selective Inhibition
Sintering (SIS) process, and states the purpose of the study.
Chapter 2 presents the background of the SIS research with an introduction to the applications
of SIS in fabricating polymeric and bronze alloy parts. The shortcomings of the previous
research and the problem statement will also be discussed in this chapter.
Chapter 3 is dedicated to literature review on the study of droplet penetration into porous
powder beds. Understanding this phenomenon is crucial in characterizing the inhibitor solution
for the SIS process.
Chapter 4 explains the research methodology in detail and defines important parameters that
are studied in this research.
The design and manufacturing of a high resolution SIS-metal machine using inkjet technology is
discussed in chapter 5. The calibration of the machine parameters and the capability of the
machine in fabricating high quality bronze parts will also be presented in this chapter.
8

Chapter 6 demonstrates the development of a systematic approach on how the process can be
extended to a different material. Stainless steel alloy will be used as a case study to validate the
method.
Finally, conclusion and future work is discussed in chapter 7. SIS is a relatively new AM process
and there are still a lot of unknowns associated with it. Future directed research can turn SIS
into a potentially competitive process to other industrial AM technologies

9

2 Chapter Two: Background
2.1 SIS-Polymer
In 2003, Khoshnevis et al. [2] introduced a new AM process called Selective Inhibition Sintering
(SIS) for plastic parts based on powder sintering. SIS-Polymer follows the following four steps in
its part fabrication process (Figure 2-1).

Figure 2-1: The SIS-Polymer process [2]
 Spreading a thin layer of powder
 Deposition of sintering inhibitor
 Minimizing radiation frame
 Sintering the layer by thermal radiation
An alternative to sintering each layer would be bulk sintering the entire powder volume after
inhibitor is applied to all layers [2].
10

The inhibition process plays the major role in SIS. There are four theories behind the sintering
inhibition phenomenon:
 Macroscopic mechanical inhibition: The powder particles are physically displaced and
separated from each other.
 Microscopic mechanical inhibition: Inhibitor droplets penetrate into the powder
without disturbing the powder surface. The inhibition occurs in microscopic level.
 Thermal inhibition: Inhibitor droplets penetrate into the powder and cool down the
particles during sintering step.
 Chemical inhibition: Inhibitor droplets penetrate into the powder and chemically react
with the powder particles and increase their resistance to heat.
An SIS-Polymer machine was developed based on microscopic mechanical inhibition (Figure
2-2) [2]. In this machine a continuous inkjet (CIJ) nozzle capable of delivering droplets as small
as 5 nl was utilized for the deposition of inhibitor on the powder layer. This machine proved the
capability of SIS-Polymer in building three-dimensional plastic parts.
11

Figure 2-2: The SIS-Polymer Alpha machine
Later in 2003, Asiabanpour et al. [7] investigated surface quality and dimensional accuracy of
the parts using statistical tools such as design of experiments and response surface
methodology. In their research, many different process operating factors and their influence on
part properties were studied and modeled. The model was finally validated through fabrication
of sample parts and different part property values were compared to the one achieved from
the model. Figure 2-3 demonstrates parts fabricated by the SIS-Polymer process.

Figure 2-3: Sample parts made by SIS-Polymer Alpha machine [7]
12

2.2 SIS-Metal
In 2005, the fabrication of metallic parts using SIS was investigated by Mojdeh [8]. The high
temperatures required for metal sintering process necessitates modification to the SIS-Polymer
sintering process. Mojdeh utilized bulk sintering instead of layer-by-layer sintering in the
fabrication process. Besides, he also added a compression step to the process in order to
increase the density of the parts. The modified SIS-Metal process is illustrated in Figure 2-4
which consisted of the following steps:
 Deposition of inhibitor: Sintering inhibitor in form of ceramic slurry is delivered by an
extrusion nozzle.
 Delivering powder to the tank: A layer of powder is spread by a roller over the build
tank.
 Creation of a boundary to contain part: An adhesive is deposited on the powder at the
periphery of the part profile.
 Compression: The created powder layer is compacted by a press mechanism to near its
maximum density.
 Bulk sintering: After all layers have been completed, the block is extracted from the
build tank and place in a sintering furnace. After sintering, the part is revealed.
13

Figure 2-4: The SIS-Metal process with compression
The SIS-Metal process was later modified by Khoshnevis et al. in 2012 [5]. The SIS-Metal alpha
machine was developed based on their research. In this machine the compression step was
replaced by a heating step. First, a thin layer of metal powder is spread on the build tank. Then
the inhibitor solution is printed on the powder bed using a single nozzle printhead. After the
deposition of the inhibitor is complete, a heater is used to evaporate water and other liquid
additives in the printed sections. The whole build tank is then placed into a furnace and the part
is bulk sintered. Finally, the part is revealed by removing the inhibited sections. The SIS-Metal
alpha machine and respective sample parts fabricated with it are shown in Figure 2-5.
14

Figure 2-5: The SIS-Metal alpha machine and respective sample parts [5]
Yoozbashizadeh [9] investigated the inhibition mechanisms by comparing SEM pictures of
inhibited and non-inhibited powder beds before and after sintering. The result is illustrated in
Figure 2-6. In this figure, it is shown that the salt which is contained in inhibitor decomposes
into ceramic particles which inhibit sintering of metal particles.
15

Figure 2-6: SEM pictures of inhibited and uninhibited powder sections before and after sintering
2.3 Critique of past approaches
The previous SIS research was mainly focused on the inhibition mechanisms, sintering of bronze
and characterization of the mechanical properties of the parts. In the former SIS-Metal
machine, a single drop-on-demand (DOD) solenoid nozzle with an orifice size of 0.005”
(0.127mm) was used for depositing the inhibitor onto powder layers [9]. Due to the lack of
control over droplet size and jetting force, the penetration of the droplets into the powder and
consequently the layer thickness were relatively large (800 µm). The droplet size along with the
16

low resolution print mechanism resulted in parts with poor surface quality. The shortcomings of
the past approaches can be summarized in the following statements:
 The lack of control over the droplet sizes and eventually the penetration onto the
powder layer due to the use of a single DOD nozzle.
 Low resolution print and positioning control of the nozzle as well as 800-micron layer
thickness would result in the parts with poor surface qualities.
 Parts with finer details and small features couldn’t be fabricated with the previous
setup.
 No research has been conducted on improving the surface quality of the parts
fabricated by SIS.
 A systematic approach needs to be developed to address the inhibitor-powder
interaction issues such as: vertical and lateral penetration extents, saturation of
inhibitor on the powder layer, and determining the optimized layer thickness.
2.4 Problem Statement
Surface quality and dimensional accuracy of the parts are important factors in evaluating an AM
process. In the AM processes that deal with the deposition of a fluid on powder, the primary
parameters in determining the quality and resolution of the parts are layer thickness and
precision of the fluid deposition. Two main problems are addressed in this study.
17

2.4.1 Primary problem
The inhibitor deposition mechanism as well as the inkjet system used in the SIS-Metal alpha
machine were the main causes of the poor surface quality and dimensional accuracy of the
parts. Therefore, the primary problem was determined to be low-resolution prints and large
droplet sizes. This led to excessive depths of penetration of the droplets resulting in large layer
thickness.
This problem will be addressed by the design and manufacture of an SIS-Metal beta machine
which uses a high resolution piezoelectric commercial printhead. This printhead has a
maximum print resolution of 5760*1440 dpi and can deposit droplets as small as 14 microns in
diameter [10]. The development process, calibration of machine parameters, and fabrication of
high resolution bronze parts with the SIS-metal beta machine are discussed in Chapter 5.
2.4.2 Secondary Problem
The SIS-Metal beta machine signifies an order of magnitude improvement in the resolution and
quality of SIS-metal parts, comparable to those fabricated by commercial AM processes. This is
validated by fabricating 3D complex parts using bronze powder. Bronze is widely available and
affordable in powder form and its processing is not very challenging. However, it has limited
industrial applications. A high temperature alloy such as stainless steel can be of great interest
to different industries. Fabricating stainless steel parts and developing a systematic approach
for adaptation of new materials into the SIS process is the focus of the second part of this
18

research. This requires thorough study of different fluid properties and their effects on
jettability and on the behavior of droplets on porous powder beds.
The development of the systematic approach and fabricating stainless steel parts will be
discussed in chapter six.
19

3 Chapter Three: Literature Review on Droplet Penetration into Porous
Powder Beds
3.1 Introduction
In SIS-metal, the inhibitor fluid and its interaction with the powder bed is one of the most
important parameters affecting the strength, quality and dimensional accuracy of fabricated
green parts. Thus, in order to study and optimize SIS-metal part characteristics, two fluid
behaviors must be studied and improved upon. First, the jettability and droplet formation as
the inhibitor fluid is being ejected from a print head must be characterized and fine-tuned.
Next, the behavior of the inhibitor fluid when in contact with the powder bed must be studied
and improved upon to ensure penetration.
As mentioned earlier, SIS-metal process uses drop-on-demand (DOD) inkjet technology to print
inhibitor fluid onto the powder layers in the fabrication process. The printed fluid consists of
thousands of microscopic droplets that come into contact with the powder bed and penetrate
into it. Therefore, a single droplet-powder interaction study can serve as an entry point into
studying the important parameters affecting SIS-metal; such as: inhibitor fluid properties,
powder particle sizes, pores sizes, and layer thickness.
The study of droplet penetration provides researchers with precious knowledge on fluid
properties that need to be considered in experiments. It also greatly aids in reaching one of the
20

goals of this study which is to develop a systematic approach and a methodology in studying
different inhibitor candidates to be used as a guideline in the SIS process.
In literature, much research has been conducted on the interaction of a droplet with a loosely
packed powder bed. Both theoretical and empirical approaches have been utilized to this end.
The theoretical studies are focused on the principal physics behind the penetration of the
droplets. These studies use capillary rise concept in simplifying the porous powder bed effects
and try to come up with governing formulas for this complicated physical phenomenon.
In empirical studies, on the other hand, the shape and behavior of the droplet is observed from
the exact moment it touches the powder surface through the point where it has completely
penetrated into the powder bed. In these studies, many different types of fluids and powders
are used in experiments and different theories are created. Both approaches will be addressed
in this literature review and the most applicable method for the SIS-Metal process will be
chosen and utilized to study the powder-inhibitor interaction.
3.2 Droplet Penetration Phenomenon
Penetration depth and time of a droplet into the powder bed is a complicated physical
phenomenon. A droplet starts penetrating into the pores of a powder surface due to generated
capillary forces, assuming that the penetration conditions, to be discussed later, are met.
(Figure 3-1).
21

Figure 3-1: Penetration of a droplet into the pores [11]
This observation leads to a simplification of the phenomenon by considering the porous
powder surface to be consisting of capillary tubes of effective radius 𝑅 (Figure 3-2). The porous
material can thus be modeled as a solid material with an array of parallel, right cylindrical
pores. While the assumption of constant radius might be inaccurate, it can be effectively
compensated for by incorporating the effective pore parameters [12].

Figure 3-2: Powder bed modeled as capillaries [13]
The depth of penetration of the droplet into the powder can be calculated by determining the
average pore size and consequently the driving capillary forces. Derivation of the governing
22

fluid penetration formulas allows for better analysis of the important characteristics of the fluid
that play an important role in the penetration interaction.
One important parameter which plays a crucial role in the study of a droplet penetration is the
contact angle created between the droplet and the powder bed. A droplet sitting on a solid
surface is associated with three surface (interfacial) energies that are caused by the interaction
of different phases in the contact point (Figure 3-3) [14]:
 Solid-liquid interfacial energy (𝛾 𝑠𝑙
)
 Liquid-vapor interfacial energy (𝛾 𝑙𝑣
)
 Solid-Vapor interfacial energy (𝛾 𝑠𝑣
)

Figure 3-3: Contact angle and interfacial energies of a droplet on a powder surface
Liquid-vapor interfacial energy is also called “surface tension.” In the case of a static droplet,
the contact angle (𝜃 ) is such that the horizontal components of surface energies balance
(Young’s equation) [14]:
23

cos
sv sl lv
    
Different values of contact angles dictate changes in behavior of the droplet on a solid surface.
In general, a fluid wets a solid if the contact angle is less than 90
𝑜 . Alternatively, a fluid spreads
on a solid when the contact angle equals 0
o
. Similarly, the contact angle and also the surface
tension affect the kinetics of liquid incorporation into the pores of powder surfaces.
Hence, the measurement of the contact angle for a specific fluid-powder interaction study is
essential. Although contact angle can be readily measured for liquids on ideal smooth surfaces,
there are many questions associated with the measurement of contact angle on powders due
to the inherent surface roughness of particles and the pore structure of powder compacts [15].
The Washburn test is one approach to measure the contact angle on a porous surface. In this
method, an experimental setup similar to the one shown in Figure 3-4 is utilized. A tube is filled
with a known mass of powder that is brought into contact with a reservoir filled with the
desired fluid. The penetration of fluid into the column of loosely packed powder is then
observed.
24

Figure 3-4: Experimental setup for capillary rise measurement [15]
The capillary pressure difference (𝑃 𝑐 ) provides the driving force for the transport of liquids into
powder. From analysis of the wetting of an ideal cylindrical pore of radius R, the capillary
pressure is”
2 cos
lv
P
R


where
lv
 is liquid-vapor interfacial energy [𝑁 /𝑚 ], 𝜃 is the contact angle [𝑟𝑎𝑑 ].
When the fluid starts to rise up the capillary, it will be resisted by viscous losses. From the
Hagen-Poiseuille equation, the viscous pressure drop at some height h is:
25

22
8 8 ( / ) P v dh dt
h R R
 

where µ is the viscosity of the fluid [𝑃𝑎 .𝑠 ] and 𝜐 ̅ is droplet penetration velocity [𝑚 /𝑠 ].
Combining the last two equations and integrating gives the height of liquid in the capillary as a
function of time:
cos
2
lv
R
ht







The contact angle can then be calculated from the slope of a plot of ℎ
2
vs. 𝑡 , provided 𝑅 , µ, and
lv
 are known.
The average pore radius (𝑅 ) can be estimated in many ways. One way is to assume the particles
are approximately spherical. The surface mean particle diameter 𝑑 32
is defined as the diameter
of a sphere with the same surface to volume ratio as the irregular particle. Therefore, the 𝑑 32

particle size together with a shape factor 𝜑 (which accounts for non-spherical particle shape)
can be used. This results in the following expression relating particle size to pore size:
32
31
d
R
 




where ε is the loose packed powder bed voidage.
Alternatively, the effective pore size could be calculated from thermodynamic principles. White
defines the effective capillary radius 𝑅 𝑒𝑓𝑓 as [16]:
26

2(1 )
p
eff
p s p
R
A




where
p
 is the volume fraction occupied by the solid in the powder bed,
s
 is the mass density
of the solid material [𝑘𝑔 /𝑚 3
] and
p
A is the specific surface area [𝑚 2
/𝑘𝑔 ]. 𝑅 𝑒𝑓𝑓 above is
equivalent to 𝑅 in calculating ℎ.
3.3 Penetration time
In addition to the depth of penetration, the penetration time can also be considered as an
important measure to predict how well a droplet can penetrate into a powder bed. The
penetration time also directly affects the extent of spreading of the printed droplet [17].
Penetration time is dependent on the powder particle size, viscosity, surface tension, and
contact angle.
When a droplet of volume 𝑉 𝑑 hits a powder bed, the flow rate of it into the powder is:
2 d
d
dV
Q r v
dt
 
Where 𝑟 𝑑 is the radius of the droplet and  is the powder bed voidage (porosity) and v is the
average velocity in the pore which is given by the differential form of the Washburn equation.

cos
8
lv
R dh
v
dt t



27

Combining the two equations and integrating gives the time for the total volume of the droplet
to penetrate the bed (penetration time) [11, 18]:
2/3
2
1.35
cos
d
p
lv
V
t
R

  

It can be seen that the penetration time depends on both the wetting thermodynamics
(𝛾 𝑙𝑣
cos𝜃 ) and the wetting kinetics (strongly affected by 𝜇 and 𝑅 𝑝𝑜𝑟𝑒 ).
This result can be validated through experimentation. In 2002, Hapgood et al. [11] studied the
kinetics of droplet penetration by filming single droplets of several different fluids as they
penetrated into loosely packed powder beds and comparing the results with the values gained
from theoretical calculations. Figure 3-5 shows the penetration of a sample material (PEG200)
into AI glass powder.

Figure 3-5: Penetration of PEG200 into AI glass powder [11]
In their experiments, they have shown that for different powder beds, the penetration time
varies linearly with expression [
cos
lv


] as predicted by the formula above for 𝑡 𝑃 (Figure 3-6).
28

Figure 3-6: Experimental measurement of penetration time of different liquids into various powder beds [11]
As shown earlier, the penetration time is directly proportional to the viscosity and inversely
proportional to the surface tension. If the contact angle remains less than 90
o
(to ensure
penetration) the dominant factor in penetration time is the binder viscosity since it can vary by
1-2 orders of magnitude.
3.4 The Effects of Fluid Properties on Penetration Time
Hapgood et al. [11] also investigated the effect of surface tension by comparing the penetration
time of two different fluids with similar contact angles and viscosities but with different surface
tensions. Penetration time is expected to increase as
lv
 decreases, since the capillary pressure
( P  ) reduces to force flow into the capillaries. However, the result showed just a slight
increase in the penetration time while there was a huge difference in surface tension values.
This was observed to be due to the decrease in the droplet volume produced by the needle as
the surface tension decreases. Therefore, the decrease in droplet volume approximately
29

compensates for the slower fluid penetration and no significant effect of surface tension was
found.
Popovich et al. [15] studied the effect of physical characteristics on the wetting and spreading
of various fluids on powders. In the research, four interesting structural changes on the surface
of powders caused by droplet penetration were observed:
 The surface severely cracked (Figure 3-7(a))
 The surface cracked into large chips (Figure 3-7(b))
 Single chips were produced that could be lifted (Figure 3-7(c))
 The wetted area shrank and sunk into the compact (Figure 3-7(d))
30

Figure 3-7: structural changes observed on the surface of powder caused by droplet penetration [19]
A simple physical explanation for this phenomenon is offered by Popovich et al. [15]. The
capillary forces produced by the pores of the powder affects both the droplets and the particles
around them. Fluids with lower viscosity easily penetrate into the pores; however, higher
viscosity fluids resist the movement and cause the powder particles to move instead. Thus, as a
high viscosity fluid penetrates a pore, the particles constituting the sides of that pore may be
pulled inward in some regions. This could result in the formation of cracks in other portions of
the compact.
31

For the case shown in Figure 3-7(d), the fluid infiltration moves the particles to such an extent
that the entire wetted area collapses and no cracks are seen.
3.5 The Effect of Droplet Size on Penetration Time
The aforementioned studies consider the case where a droplet penetrates into powder bed
with much smaller particles. In the current study, the parameters are different in that inkjet
printing of the fluid creates printed droplets that are smaller in size than the powder particles.
Typical nozzles deposit small droplets (diameter of less than 70 microns) with low kinetic
energy in order to maximize feature fidelity and resolution. Incomplete penetration will reduce
layer-to-layer connectivity, and lead to lamination defects. Larger or faster droplets can
penetrate thicker layers, at the expense of print resolution. Furthermore, larger incremental
layer thickness will result in correspondingly large vertical z-steps along external curved
surfaces.
In the proposed research, the behavior of the droplet is different. As illustrated in Figure 3-8,
the droplet will spread over the particle rather than penetrate into the pore.

Figure 3-8: Spread of a droplet on a larger particle [19]
32

Chua et al. [19] studied this phenomenon through theoretical calculations and also empirical
experimentations and validated the results using computer simulations. They proposed that the
spreading process was driven by the effective surface tension, cos
lv
 , and resisted by the
viscosity of the fluid,  . In their analysis, they simplified the problem by ignoring the curvature
of the granule and considering the droplet as a cylinder with radius a (Figure 3-9).

Figure 3-9: Schematic of the approximation representation of a spreading drop
This resulted in the following equation for calculating the spreading time:
4
0
3
(K 1)
64 cos
ww
d




where 𝐾 𝑤 is a constant given as a function of contact angle and 𝑑 0
is the initial droplet
diameter. Comparing the following two formulas,
2/3
2
1.35
cos
d
p
lv
V
t
R

  
 it can be concluded
that the physical properties of the fluid have almost the same effect on the penetration time
and spreading time of a droplet.
33

{

4
0
3
(K 1)
64 cos
ww
d





This can provide a fine approximation on the behavior of a small droplet in contact with a larger
particle and can be used in the prediction of an inkjet-printed droplet.
Another argument that can be made here regarding the size of a printed droplet is that in inkjet
printing, a single printed dot will consist of several smaller droplets that are placed closely
beside each other. Since the printing process is fast, it is safe to assume that the droplets will
join together before penetrating into the pores. In this way, they form larger droplets that that
will indeed be greater in size than the particles; these printed droplets follow the physical rules
associated with the stated equations.
3.6 Application in Additive Manufacturing
There are a few AM processes that benefit from the fluid-powder interaction in their
fabrication process. Therefore, the droplet penetration phenomenon can directly apply to them
and provide valuable information regarding the behavior of the fluid in contact with the powder
bed. The three-Dimensional Printing (3DP) process is one example that inkjet prints binder on
powder to form one layer of the part [6]. In literature, extensive study has been carried out on
parameter optimization in 3DP in order to increase the quality of the printed parts as well as
their precision and dimensional accuracy. In many cases, the results of 3DP research can be
2/3
2
1.35
cos
d
p
lv
V
t
R

  

34

modified and applied to the SIS process due to the similarities of the processes. Therefore, a
careful study of 3DP literature is essential.
Various factors such as powder particle sizes, droplet sizes, fluid viscosity, surface tension, fluid
saturation level, contact angle, layer thickness and printing conditions play important roles in
the strength, surface quality and dimensional accuracy of the 3DP process [20]. These factors
and variables can be classified into three categories: 3D printer-related variables, powder-
related variables, and fluid-related variables [21]. 3D printer variables include powder
spreading speed, layer thickness, droplet sizes and print quality. Powder-related variables refer
to variables such as: powder particle sizes, wettability of the powder, particle shapes and
porosity. Finally, fluid-related factors are vertical penetration depth, lateral penetration (lateral
spreading) and saturation level.
Vertical and lateral penetrations directly affect the quality of top and side surfaces as well as
the dimensional accuracy of the printed features. By understanding the penetration
phenomenon, the layer thickness and the amount of deposited fluid for each layer can be
matched and optimized in order to get the desired saturation level and also control the
spreading and lateral penetration of the printed fluid.
Most of the research conducted on 3DP takes the experimental approach in addressing the
fundamental issues of surface finish and dimensional accuracy of the parts.
35

The addition of a printhead to the process introduces several new limitations to fluid
properties. The fluid not only needs to penetrate into the pores with acceptable velocity, it also
needs to be jettable and compatible with the materials used in the printhead. Moon et al. [22]
studied the inkjet printing of binder for ceramic components. In the study, it was stated that a
water-based binder system is preferred over a solvent-based binder for its compatibility with
materials in the printhead as well as improved jet stability. The water-based binder can also be
tailored to have a wide range of surface tension and viscosity, whereas the solvent-based
binder tends to have relatively low surface tension.
Based on their research, the most important properties controlling the size and shape of the
fluid droplet and jet reliability are surface tension and rheology of the fluid which is similar to
the result of theoretical studies on droplet penetration. It also confirms that the physical
properties of the fluid play an important role in accurate and stable ink-jet printing behavior.
Figure 3-10 shows the effect of the fluid properties on the formation of a droplet produced by
inkjet printing [23]. In the figure, the upper row shows the printing sequence of a Newtonian
fluid. The droplet is formed with a perfect spherical shape and additional droplets are printed.
The bottom row of the picture illustrates the result of addition of a small amount of high
molecular weight polymer to the fluid resulting in increase in its viscosity. It is shown that in the
printing of the more viscous fluid, a filament is formed attached to the droplet which may affect
the precision of the print.
36

Figure 3-10: The effect of the fluid properties on the formation of a droplet produced by inkjet printing. Upper
row shows the behavior of a Newtonian fluid, lower row shows the effect of increasing the viscosity of the
fluid[23]
The other important factor affecting the droplet size and shape is surface tension. Figure 3-11
shows the formation of droplets by a piezoelectric DOD printhead for a solution with different
surface tensions. It is shown that if the surface tension is inadequately low, small satellites will
be formed (Figure 3-11 (a)). On the other hand, the droplets formed by inkjet printing of higher
surface tension fluids are spherical and have equal size, shape, and spacing (Figure 3-11 (b)).
37

Figure 3-11: The effect of surface tension on the formation of droplets (a) inadequately low surface tension (b)
fluid with appropriate surface tension [26]
As discussed, it can be seen that fluid properties directly affect print parameters such as droplet
sizes, penetration depth, layer thickness, and saturation level (fluid to pore volume ratio for a
given printing volume). Layer thickness and fluid saturation level are identified as the primary
variables needing systematic evaluation in order to achieve high structural integrity and
minimum dimensional variation [21]. Layer thickness defines the precision and the resolution of
the machine or process and has an influence on properties of the green part such as surface
quality, dimensional accuracy and strength. The important factors in determining layer
thickness in 3DP are powder particle sizes and penetration depth of fluid into the layer. Since
the particle size used in this research is fixed, as determined by the manufacturer, the vertical
penetration depth should basically be used to fine-tune the layer thickness.
38

Lateral spreading of the fluid also is a dominant factor in dimensional accuracy of the fabricated
parts. Vertical and Lateral penetration will be discussed in the following section.
3.6.1 Vertical and Lateral Penetration
The impact and spreading of liquid droplets on a porous surface is of great interest for inkjet
printing applications [17]. First, the degree of liquid spreading on the powder after depositing
directly affects the resolution of the print. Second, the presence of porosity in the surface
affects the final printed dot size.
Briefly, when a liquid droplet impacts a loose powder surface, it decelerates and deforms from
a sphere to a larger diameter disk [24]. After being deposited, the droplet remains temporarily
on the surface of the powder due to its high-packing density. It will then start spreading out and
simultaneously penetrating into the powder bed due to capillary forces generated by the
interaction of the droplet and the pores in the powder. Depending on inertial forces, the
droplet may splash during initial impact and/or create a crater in the loose powder bed. At
typical inkjet droplet dimensions (< 10
-4
𝑚 ), velocity (~ 1 𝑚 /𝑠 ), density (~ 1 𝑔 /𝑚𝑙 ) and surface
tension (~10
-2
𝑁 /𝑚 )splashing is unlikely to happen [24].
The droplet-powder interaction is influenced by microstructure and surface finish of the
powder bed as well as physical properties of the fluid and printing conditions [22], all of which
determines the geometry of the wetted area. Figure 3-12 shows the penetration and spreading
process of a sample droplet into a porous powder surface.
39

Figure 3-12: Penetration and spreading process of a sample droplet into a porous powder surface [17]
The extent of fluid spreading is governed by the printing conditions (such as droplet impact
velocity), and liquid properties (such as viscosity, and surface tension of the fluid). Under the
constant printing conditions of jet velocity, droplet size, and surface tension, the fluid impact
size is inversely proportional to approximately the square root of viscosity due to viscous
energy dissipation [22].
Viscosity and surface tension both affect the dimensional accuracy of the green part. An
increase in the viscosity of the fluid reduces the initial impact diameter of the droplet due to
viscous dissipation. This leads to a decrease in the printed line thickness and increase in the
dimensional accuracy. The effect of the surface tension needs to be considered here as well.
Surface tension determines the contact angle of the fluid on the powder bed. The penetration
of the high surface tension droplet forms a spherical cap-shaped wetted volume. The fluid with
low surface tension, on the other hand, forms a cylindrical disk-shaped wetted volume by
spontaneous spreading. The resulting impact size of the fluid is much larger due to low surface
energy of the fluid (Figure 3-13) [22].
40

Figure 3-13: Single binder droplet effect printed using a DOD system on a powder bed: (a) side view
(b) bottom view
Moon et al. [22] examined the influence of physical properties of a number of fluids on printed
line widths. In their experiment, four different fluids with different viscosity and surface
tensions were printed on the same powder with the same settings and saturation levels. Figure
3-14 shows the printed line width vs. fluid physical properties. It can be concluded that line
width increases with decreasing viscosity and surface tension.

Figure 3-14: influence of physical properties of fluids on printed line widths: (a) surface tension (b) viscosity [22]
41

The effect of layer thickness on fluid penetration is also illustrated in Figure 3-15. The figure
illustrates three typical cases when fluid is printed onto a powder bed:

 The printing layer is thin and excessive fluid spreading happens which leads to poor
dimensional accuracy and poor side surface quality.
 Optimal layer thickness and minimal fluid spreading occurs which translates to the least
dimensional variations and optimum surface qualities.
 The printing layer is thick and incomplete vertical fluid spreading is seen. This leads to
insufficient binding among layers which may cause the part to break and will result in
poor surface qualities.

Figure 3-15: The effect of layer thickness on fluid penetration [21]
With a decrease of layer thickness, the fluid penetrates quickly to the bottom of the layer.
However, the previous layer printed prevents the fluid from further spreading. In the lateral
direction, the fluid spreads without such limitation [25].
42

The extent of lateral spreading of a fluid droplet is always larger than that of vertical
penetration. This is because the fluid lateral spreading occurs in multiple directions, and the
vertical spreading is in one direction only. When the fluid lateral spreading rate is high, the
spreading time is shorter. This results in the droplet not having enough time to penetrate
vertically into the powder. Therefore, more lateral spreading occurs [26]. The incomplete
vertical penetration case happens especially when the printing layer thickness is large
Figure 3-16 illustrates the results of experiments conducted by Lu et al. [21] for the AM of a thin
Shape Memory Alloy (SMA) wire using the 3DP process. Various wire meshes are printed under
different layer thicknesses and saturation levels. It can be seen that under the same fluid
saturation level, the wire meshes printed with 20 μm printing layer thickness show more
dimensional variation in comparison to those printed with 35 and 50 μm printing layer
thicknesses. This is due to the lateral penetration and spreading of the fluid in thinner layers.

43

Figure 3-16: Optical micrographs of 3D printed SMA square wire with fluid saturation level at (a) 55%, (b) 110%,
and (c) 170% for which the layer thickness is 20 microns and (d) 55%, (b) 110%, and (c) 170% for which the layer
thickness is 35 microns [21].
Particle size also affects the rate of lateral spreading. For larger powder particles, there is faster
fluid lateral spreading and larger printing layer thickness. Since the fluid is the dominant factor
to impart green strength in the 3DP process, some powder particles may not be bonded in
certain layers for the larger powder bed. This subsequently results in low green strength for the
corresponding structures [26].
44

The effect of particle size on fluid spreading time is demonstrated in Figure 3-17.

Figure 3-17: The effect of particle size on fluid spreading time [26]
It is shown that as the particle size increases, the spreading time decreases, meaning faster
lateral spreading happens. This causes incomplete vertical penetration and weak inter-layer
bindings.
3.6.2 Saturation
Saturation is defined as the ratio of the deposited fluid volume to pore volume available within
the powder target. In the study conducted by Popovich et al. [15], it is shown that the higher
the viscosity of the wetting fluid, the higher the level of saturation that will be achieved.
Although higher viscosity fluids take longer to infiltrate the powder, they result in a greater
opportunity to fill a larger fraction of the pore space. On the other hand, lower viscosity fluids
penetrate relatively quickly, entrapping more air in the process.
Fluid saturation level is closely related to powder packing density, fluid-particle interaction, and
thus fluid distribution in the 3D printed meshes. Under the same printing layer thickness, higher
45

saturation level results in higher volume of sprayed fluid and stronger bonds between particles,
but too high a saturation causes lateral spreading and poor dimensional accuracy [21].
Figure 3-18 shows the result of experiments carried out by Lu et al. [21] in which a line with 200
micron width was printed under various saturation levels and layer thicknesses.

Figure 3-18: The effects of saturation level and layer thickness on printed line width [21]
It can be concluded that decreasing layer thickness and increasing saturation level increase
lateral spreading and affect the dimensional accuracy.
Therefore, it can be said that build resolution can be improved by decreasing saturation, such
as increasing spreading powder layer thickness, but is limited at the upper end by lamination
defects. Fluid migration can be decreased by increasing fluid viscosity, but this approach is
limited by the highest viscosity that the printhead will print reliably without clogging [24].
46

3.7 Application in Selective Inhibition Sintering
Drop penetration study can contribute to the SIS process for the following two purposes:
 Optimization of the current process and inhibitor parameters in order to increase the
surface quality and dimensional accuracy of the parts
 Introduction of a methodology and procedure for selection of new inhibitors suitable for
various metal base materials
Candidate materials need to go through the procedure shown in Figure 3-19 in order to be
tested in the machine for part fabrication. The experiments carried out in the process can be
very time consuming and costly. Theoretical studies can provide precious information and
predictions on the properties of the fluid that tremendously reduce the cost and time of the
experiments.

47

Figure 3-19: Procedure for inhibitor selection

Candidate
Inhibitors
• Candidate Inhibtors are chosen through chemical studies
Chemical
Stability
• Chemical stability and compatibilty with the materials used in the
printhead needs to be considered
Inhibition
Drop Test
• A droplet of inhibitor is deposited on the powder bed and sintered to
ensure inhibition happens
Jettability
• The viscosity and surface tension of the fluid needs to be calibrated
to be in the range of printhead characteristics and to ensure good
penetration into the powder
Inkjet
Printing
• The inhibitor is inkjet printed on powder and the penetration and
print quality is observed
Fine Tuning
• The viscosity and surface tension is fine-tuned to increase the surface
quality and vertical penetration and decrease lateral penetration
Parameter
Optimization
• Print parameters such as: layer thickness, saturation level, and
printed line thickness are optimized based on the conducted
experiments
48

The results of this study directly applies to several steps of the procedure, namely the inhibition
droplet test, jettability, fine-tuning, and parameter optimization steps.
In inhibition droplet test, a primary study on the rheology of the solution should be done so
that the drop penetrates into the powder with acceptable speed.
After the occurrence of inhibition is confirmed, the fluid properties need to be carefully
calibrated in order to make it jettable and printable and prevent clogging of the nozzle.
The inhibitor is then inkjet printed on the powder and its behavior is observed. Different factors
such as: penetration time, vertical and lateral penetration extents, and surface quality of the
final sintered part need to be analyzed and studied.
Based on the inkjet printing observations, the properties of the fluid will be fine-tuned to
increase the quality of the prints.
Finally, different machine parameters such as: layer thickness, saturation level, and printed line
thickness need to be optimized.
3.8 Summary
In this chapter, the effects of fluid properties on the behavior of droplets on porous powder
beds and the governing theories were investigated. Based on the reported research in the
literature, fluid viscosity, surface tension, and contact angle play important roles in the
49

penetration of liquid droplet into dry powder. The penetration time was also studied to be used
as a measure to compare the influence of different fluid properties on the droplet -powder
interaction.
A thorough study of the MIT’s three-dimensional printing process was conducted due to similar
features shared with the SIS-Metal process. Both processes utilize inkjet technology to deposit
fluid onto powder to fabricate one layer of the part. The interaction of the fluid with powder is
the most important factor affecting the surface quality and dimensional accuracy of the parts.
The following significant conclusions were drawn from the review of literature:
 Saturation of the inhibitor solution in the pore of the powder layer should be high
enough to ensure printed layers bind together and prevent layering effect. However,
too high of a saturation level will cause a decrease in dimensional accuracy due to
lateral spreading of the fluid. This typically results in incorrect feature size.
 Build resolution can be improved upon by decreasing saturation through increasing
powder layer thickness. However, this is limited at the upper end by lamination
defects. Fluid migration can be decreased by increasing fluid viscosity, but this
approach is limited by the highest viscosity that the printhead can print reliably
without clogging [24].
 The viscosity and surface tension can be calibrated to form a cylindrical-shape
penetration instead of a spherical-shape one. This will result in less lateral spreading
and improved feature size printing.
50

 There are trade-offs that need to be considered in choosing proper fluid properties.
For instance, a higher viscosity fluid has less impact size which means thinner lines
can be printed, but it will have difficulty penetrating into the powder. On the other
hand, a low surface tension will result in cylindrical-shape droplet, but increases
lateral spreading. Therefore, high viscosity and low surface tension is preferred for
more precise penetration (smaller impact and cylindrical penetration [22]).
However, high viscosity may cause cracks due to viscous forces dragging particles
[15].
 Top surface quality is directly affected by the penetration behavior of the fluid onto
the powder surface.

51

4 Chapter Four: Research Methodology
4.1 Introduction
The SIS-metal process works based on depositing droplets of inhibitor fluid on the periphery of
the part with a layer-by-layer approach. The powder particles on the boundary of the part of
each layer will be in contact with the inhibitor fluid droplets. When the print process for all
layers is completed, the thin printed areas around the part act as temporary mold containing
the desired part geometry. Figure 4-1 illustrates the visualization of a cube and a thin layer of
printed inhibitor around it.

Figure 4-1: Thin layer of inhibitor surrounding the part [27]
Next, the printed part is transferred to a vacuum furnace. In the furnace, the inhibitor fluid
decomposes and creates high temperature ceramic particles which inhibit the sintering process
52

of the metal particles in contact with them. This creates a thin inhibited layer around the part.
This inhibited layer will later be removed by sandblasting which reveals the final part.
The precision of the part fabrication with SIS-metal depends on the accuracy and resolution of
the print mechanism by which the droplets are positioned on the powder layer and also on the
size of each droplet. As the resolution of the fluid deposition system increases, the precision in
defining the part boundary also increases which provides better dimensional accuracy.
Although these are the main factors affecting the resolution and dimensional accuracy of the
parts, there are other factors that play important roles as well, such as inhibitor fluid
properties, metal powder properties, and process parameters.
Inhibitor fluid properties such as viscosity and surface tension determine how well and how
much the inhibitor will penetrate into the powder in both vertical and lateral directions. The
vertical penetration affects the resolution of the print in z direction and determines how well
the layers bond with each other. The lateral penetration, on the other hand, affects the
dimensional accuracy of the parts. To achieve the desired surface quality and dimensional
accuracy, these properties need to be carefully studied and fine-tuned in regards with the
metal powder and fluid deposition system used.
Metal powder properties include powder particle size distribution, wettability, flowability,
sintering temperature, and mechanical properties. Each property plays an important role in
the final part characteristics and process parameters. For instance, particle size and flowability
53

define layer thickness and porosity of the part, wettability affects the vertical and lateral
penetrations, and sintering temperature affects the mechanical properties.
Process and machine parameters will be calibrated after determining and fixing inhibitor and
powder properties. Different machine parameters such as print resolution, droplet sizes, and
print speed govern the part fabrication process.
In this research, these parameters will be studied and improved in two steps. The first step is to
design and manufacture a machine with highly precise motion control systems, a high
resolution inkjet printhead, and precise droplet positioning mechanisms. A commercial inkjet
printhead with drop-on-demand piezoelectric technology will be utilized for precise deposition
of inhibitor fluid droplets which satisfies the desired resolution and accuracy for this process.
This modification introduces a major improvement in the surface quality and dimensional
accuracy of the parts.
The second phase of the research includes systematic fine-tuning of other process parameters
to achieve higher levels of dimensional accuracy and resolution which leads to the development
of systematic approach in adapting new materials in the SIS-metal process.
These two steps will first be tested on bronze alloy powder which has already been proven to
be compatible with the SIS-metal process. Next, the method will be generalized and applied to
higher temperature alloys. Stainless steel powder will be used as a successful case study. For
54

the final results, high resolution complex parts from bronze and stainless steel alloys will be
presented.
4.2 Inkjet Technology
Dimensional accuracy and surface quality of the parts are directly influenced by the inhibitor
droplet generation and deposition systems. Therefore, an extensive research has been done on
the selection of the fluid deposition system for SIS-metal and the most suitable one for the
purpose of this paper has been identified.
The most common technologies commercially available for the precise deposition of liquid
droplets on substrates are inkjet technologies. Inks in these technologies refer to the liquids
being deposited which can have different chemical compositions and can contain a wide range
of materials such as metals, ceramics, and polymers. The important restrictions here are that
the material needs to be in liquid form and it has to satisfy the necessary rheological properties.
In this research, the microscopic mechanical inhibition approach is studied and utilized. In this
method, ceramic particles (in nano-scale size) with high melting temperature are created
between metal particles during sintering which results in sintering inhibition.
In order to introduce ceramic particles to the sintering process using inkjet technologies, two
different approaches can be utilized. As shown in Figure 4-2, these routes include either direct
deposition (inkjet printing) of ceramic particles in a suspension form or inkjet printing ceramic
precursors and create ceramics afterwards.
55

Figure 4-2: Process routes for the deposition of ceramics using inkjet printing (source: modified from [28])
In the first route, a suspension solution is created using fine ceramic powder particles. The
particles need to be small enough (nanoparticles) to avoid clogging and blockage of the nozzles.
Typically a size smaller than 1/50th of the nozzle diameter is chosen for the particles [28]. These
suspensions having the correct viscosity and surface tension can be directly jetted. The
deposited particles then need to be sintered and fused together to form the final desired
product. This method can create ceramic parts with densities more than 40%. Fully dense parts
can also be achieved depending on the sintering conditions. The creation of the right
56

suspension with desired viscosity and surface tension requires thorough research on the
chemistry of the materials and different fabrication processes.
The second route, however, would be more straight-forward and favorable for the purpose of
this research. In this method, chemical precursor solutions are prepared and the properties are
fine-tuned for inkjet printing. After being printed, a thermal process decomposes the
precursors and turns them into ceramic particles which act as inhibitors for the metal particles
in the furnace. The density of the ceramic particles here is lower than the previous route.
However, the creation of the ink and the selection of the suitable print technology will be less
challenging for precursor solutions than suspensions.
4.2.1 Inkjet Printing in Additive Manufacturing
Since inkjet technologies can handle many different materials, a vast variety of applications can
also be named for them. These include both consumer and industrial applications. From low
volume photo and document printing at homes to high volume production lines and high-tech
research labs, the footprints of inkjet printing can be traced.
The idea of utilizing inkjet printing in additive manufacturing for the creation of physical objects
has been existing for decades. The main AM processes in this regard can be categorized into
liquid-based processes and powder-based processes. In liquid-based techniques, a polymeric
ink is deposited by the printhead. The ink can be either a molten thermoplastic that solidifies
after being jetted, a wax, or a photopolymer that cures and solidifies by an external source of
57

energy such as a UV lamp. The latter is used in Polyjet
TM
3D printing technology by Strarasys as
shown in Figure 4-3.

Figure 4-3: Polyjet
TM
3D printing technology by Strarasys (Photo by Daniel Dikovsky from Stratasys)
The technologies in the powder-based group of inkjet-based AM processes utilize inkjet printing
to deposit liquid material on powder beds and fabricate the part with the resulting interactions.
One such process is called three-dimensional printing or the binder jetting process. This process
jets liquid binder on powder to aid selective consolidation of powder particles in a layer by layer
approach to create the part.
Table 4-1 summarizes the process parameters such as build envelop, print resolution, and
materials for four major commercial AM processes that utilize inkjets in their systems.
58

Table 4-1: General characteristics of major AM processes using inkjet systems (modified from [28])
System Build Envelope
(mm)
Print Resolution Material
PolyJet
Stratasys
(XYZ) 490 x 390
x 200
(XY): 600 x 600 dpi
(Z): 16 microns
Acrylate
photopolymers:
colored, opaque,
flexible, rigid
ProJet
3D Systems Inc.
(XYZ) 298 x 185
x 203
(XY): 656 x 656 dpi
(Z): 16 microns
Acrylate
photopolymers:
colored, opaque,
waxes
Solidscape Inc. (XYZ) 152 x 152
x 100
(XY): 5000 x 5000
dpi
(Z): 13 microns
Polyester
thermoplastic
Casting wax
Zprinter
3D Systems Inc.
(XYZ) 254 x 381
x 203
(XY): 600 x 600 dpi
(Z): 89 - 102
microns
Composites
Foundry sand mixes
Elastomerics

4.2.2 Integration of Inkjet Technology in the SIS-metal process
Inkjet technologies can be classified into two main categories of continuous inkjet (CIJ) and
drop-on-demand (DOD) technologies (Figure 4-4). Both groups of technologies print liquid
through small orifices called nozzles. In a CIJ setup, there is a continuous flow of ink out of the
nozzle. The flow can be stopped and started again, but droplets cannot be precisely generated.
Conversely, DOD systems are impulsive meaning small droplets are formed which can be
positioned separately as needed [28].
59

Figure 4-4: Classification of inkjet technologies
Since the precise formation and positioning of the droplets are essential for improving the
surface quality and accuracy of SIS-metal parts, the selection of the inkjet technology in this
research is limited to the DOD group. The DOD printheads are further classified into two major
categories of thermal printheads and piezoelectric printheads which have been thoroughly
studied to find the best option for this research.
4.2.3 Thermal Printheads
In thermal inkjet printing a small heater, in form of a microscopic resistor behind the print
nozzle, heats a portion of the ink in the nozzle chamber. The heat creates vapor bubbles that
60

eject drops of ink from the orifice (Figure 4-5). It takes a few microseconds for a droplet to form
during which the ink gets to more than 300 degrees centigrade and returns to room
temperature.

Figure 4-5: Droplet generation process in thermal printheads (source: www.aldertech.com)
The major thermal printhead manufacturers include HP, Canon, and Memjet. HP and Canon
were the pioneers in thermal inkjet printing while Memjet is relatively new to the market. They
use MEMS technology in their nozzles and offer page-wide printheads that can cover the whole
width of a paper and print as fast as 1 page per second.
Among these manufacturers, HP offers solutions for technology developers that allow them to
integrate commercial printheads into their systems and gain the ability to inkjet print different
types of inks. However, the required fluid properties are very limited. Therefore, the inhibitor
solution formulations of this research would have to fit within a narrow range of surface
tensions and viscosities which was not suitable for the purpose of this paper.
The other issue with using a thermal printhead was the nature of the SIS-metal inhibitor
solutions. They consist of high concentration levels of dissolved precursors into the water.
61

Thermal inkjet printing of these solutions would result in the deposition of salt particles on the
nozzles and clogging. The other option would be to use piezoelectric technology.
4.2.4 Piezoelectric Printheads
In piezoelectric printing technology, a microscopic bending cantilever or membrane applies
pressure on the ink inside the chamber and forces a droplet out of the nozzle. The pressurizing
element is made of piezoelectric material that changes shape by receiving electric pulses. In this
technology, different precise droplet sizes can be achieved for each nozzle by changing the
frequency of the pulses which consequently affects the vibration frequency of the piezo. Figure
4-6 illustrates the principle behind droplet generation using piezoelectric technology.

Figure 4-6: Droplet generation process in piezoelectric printheads (source: Epson.com)
On average, piezoelectric printheads have higher droplet generation speeds and are dominating
the market for high-speed industrial printing [29].
62

For optimum performance, the surface tension and viscosity of the inks needs to fit within a
narrow window. However, the mechanical vibration of the piezo and the pulse shape
adjustments allow for the use of solutions with a wider range of fluid properties. This useful
characteristic makes piezoelectric technology a better choice for the current lab setup.
Companies manufacturing piezoelectric printheads include Epson, Fujifilm Dimatix, Xaar, Seiko
Epson, Konica Minolta, Ricoh, Kyocera, and a dozen more companies whose products are not as
widely used. Some of these companies such as Dimatix and Ricoh offer commercial printheads
that can be purchased separately and be utilized in lab setups. These printheads usually cost
thousands of dollars excluding drive electronics, software, additional components, and
maintenance. In a lab setup, a variety of chemicals with different properties might be used in
experiments which may cause damage to the printhead. Therefore, the expensive setup
needed for commercial piezoelectric printheads makes them unfavorable for this research.
The more sensible option for a laboratory setup is to start with a consumer level printer and
then switch to a commercial one. Therefore, in this research, a consumer piezoelectric photo
printer was utilized. Control software and related electronics were developed to control the
printheads motion to print on a powder bed instead of a paper. Along with the cost benefits of
this approach, the consumer printer also offers advantages such as reduced complexity of the
setup, high speed and high resolution printing, and precise droplet positioning.
63

4.2.5 Conclusion on Inkjet Technologies
Both types of printing technologies have some advantages and some disadvantages over the
other one. For instance, piezoelectric printheads allow the droplet size to be adjusted during
printing by controlling the pulse. They can also print with a higher speed and have higher
lifetime. Besides, the ink used in the piezoelectric inkjet technology is not heated and can have
a wider range of rheological properties. This property makes piezo printheads a better choice
for experimental purposes.
On the other hand, thermal inkjet printheads are usually much more affordable, and the heads
are disposable and don’t need extensive cleaning cycles and maintenance. The pulse generation
and electronics to drive thermal heads are also more straight-forward and inexpensive.
Table 4-2 summarizes the comparison between thermal and piezoelectric technologies.
Table 4-2: Thermal printheads vs. piezoelectric printheads
Thermal Printheads Piezoelectric Printheads
Pros Less Expensive
Less maintenance
Less complexity
Precise and variable droplet sizes
Wide range of ink solutions
High speed
High lifetime
Cons Limited number of droplet sizes
Limited ink options
Not very durable
Expensive technology
Requires Intense cleaning and maintenance
Complex electronics
64

Based on the study on different types of inkjet printheads, the Epson WorkForce 30 (WF30), a
consumer photo printer, has been chosen to be used in the SIS-metal machine (Figure 4-7). This
high resolution printer uses micro-piezo technology and can eject droplets ranging from 6pL to
26pL [30].
However, the actual droplet sizes may vary for custom-made solutions and inks. WF30 has an
optimized resolution of 5760 x 1440 dpi and can print up to four different colors (black, cyan.
Magenta, and yellow) through five different channels (two for black and one channel for other
colors each). This translates into the ability to print and control five different solutions
individually. This will be useful for studies that may need to deposit more than one chemical in
their process. In this study, the same inhibitor solution is used in two different colors to
increase the speed of the process by depositing more inhibitor on one pass of the printhead.
65

Figure 4-7: Epson WorkForce 30 (source: Epson.com)

The WF30 printhead has three rows of nozzles each containing 180 orifices, two rows for black
inks and one row divided into three equal sections for other colors (Figure 4-8).

Figure 4-8: Three rows of nozzles in Epson WF30 printhead
It will be shown that by utilizing the high resolution inkjet printhead, the quality and
dimensional accuracy of the parts has significantly increased.
66

4.3 Calibration of Process Parameters
The SIS-metal process is in its early stages of development. Prior studies had been focused on
the basics of the process and understanding the principles behind the inhibition mechanisms.
This study is built upon the previous work and tends to further develop the process and take
the quality of SIS-metal parts to a level comparable with other commercial metal AM processes.
Thus, determining important parameters affecting the process is a crucial step in the
development and optimization of SIS-metal.
Extensive research and numerous experimentations have been carried out to strengthen the
understanding of the process and to extract features defining the desired final product which is
a high resolution 3D part.
The fishbone diagram illustrated in Figure 4-9 has been created to show the most important
parameters affecting the SIS-metal process.
67

68

Figure 4-9: Fishbone diagram showing important parameters in SIS-metal process
Among all these parameters, the focus of this research will be on the ones related to the
surface quality, resolution and dimensional accuracy of the parts. These are classified into two
main groups: fluid-related parameters and powder-related parameters.
4.4 Fluid-related Parameters
This group of parameters are mostly influenced by chemistry of the inhibitor solution and its
rheological properties. The inhibitor solution directly interacts with both the printhead and the
powder bed. Therefore, the solution needs to be studied with regards to both. On one hand,
the inhibitor fluid needs to be compatible with the materials used in the printhead and also its
fluid properties should be fine-tuned to be in the desired range of the head. On the other hand,
the droplets of the solution need to have a controlled amount of penetration into the powder
bed and inhibit the sintering of the powder to a desired degree. These aspects counteract each
other in most cases and a trade-off is inevitable.
4.4.1 Inhibitor Solution
The inhibition mechanism in the SIS-metal process has been studied by Yoozbashizadeh [9] and
Petros [31]. In SIS-metal, the goal is to slow down or stop the sintering process of the selected
areas of the powder bed. As stated in the previous chapters, there are four major theories
defining the inhibition mechanism [2, 4]:
 Macroscopic mechanical inhibition: The powder particles are physically displaced and
separated from each other.
69

 Microscopic mechanical inhibition: Inhibitor droplets penetrate into the powder
without disturbing the powder surface. The inhibition occurs in microscopic level.
 Thermal inhibition: Inhibitor droplets penetrate into the powder and cool down the
particles during the sintering step.
 Chemical inhibition: Inhibitor droplets penetrate into the powder and chemically react
with the powder particles and increase their resistance to heat.
In SIS-metal utilized in this research, microscopic mechanical inhibition is dominant. In this
method, the inhibitor solution consists of chemical precursors that surround powder particles
on the boundary of the parts. During the sintering process, these precursors decompose and
turn into high temperature materials such as carbon or different ceramics and cover the metal
particles in touch with the inhibitor solution. The high temperature particles retard bonding of
the coated metal particles and inhibit their sintering. Figure 4-10 illustrates this phenomenon.
70

Figure 4-10: SEM micrograph of "printed" vs. "unprinted" sections [9]
The effectiveness of the inhibition process depends on several factors such as the thermal
decomposition process of the inhibitor material, the generated high temperature particles and
the density of them in the printed areas, sintering temperature and heating profile of the
metal, and the base metal material used in the process.
Among these factors, the density of the inhibitor particles is of interest to the current research
which is mainly defined by the amount of precursors present in the inhibitor solution. The
inhibitor solutions used in this study consist of precursors dissolved in water for inhibition
purposes and surfactants for calibrating fluid properties. Different amounts of precursors in the
solution define how well inhibition occurs and how easy it would be to remove inhibited areas
after sintering. In this paper, different amounts of precursors in the solution and the
effectiveness of inhibition are respectively referred to as concentration levels and degree of
inhibition.
71

Concentration level of the inhibitor solution directly affects the rheology of the fluid and its
jettability as a result. As the concentration level increases, a better degree of inhibition is
achieved. But, the fluid properties go further apart from the desired range of inkjet values. This
can partially be compensated by the addition of surfactants to the solution in order to calibrate
the surface tension and viscosity values. However, the initial concentration level needs to be
chosen properly in order to get the desired results. This will experimentally be achieved by a
design of experiments (DOE) approach where the viscosity and surface tension of solutions with
different concentration levels are measured before and after addition of surfactants. The
solutions with the desired values will then be chosen for test in the printhead.
4.4.2 Surface Tension and Viscosity
Surface tension and viscosity dominate droplet interactions with powder particles as well as the
amount and speed of vertical and lateral penetration of inhibitor onto the powder layer.
Besides, jettability of the fluid and droplet sizes are also determined by rheology of the
solution. Therefore, understanding the physics behind these properties is helpful in optimizing
the inhibitor characteristics.
Viscosity is a property of a fluid that describes its resistance to flow or deform under shear
stress. Viscosity (η) is the ratio of shearing stress to the velocity gradient in a fluid shown with
the following equation:
η=
𝐹̅
𝐴 ⁄
∆𝑣 𝑥 ∆𝑧 ⁄

72

or
η=
𝐹 𝐴 ⁄
𝑑 𝑣 𝑥 𝑑 𝑧 ⁄

There is a similarity to the Newton’s second law of motion:
𝑭 𝑨 =𝛈 𝒅 𝒗 𝒙
𝒅 𝒛 ↔𝑭 =𝒎 𝒅 𝒗 𝒅𝒕

The SI unit of viscosity is pascal second [Pas]. However, the more common unit is [dyne.s/cm
2
]
or poise [P]. The relationships between these units is as follows:
{
𝟏 𝑷𝒂 𝒔 =𝟏𝟎 𝑷
𝟏 𝒎𝑷𝒂 𝒔 =𝟏 𝒄𝑷

There are several factors that affect viscosity. The nature of the material and temperature are
among the most important ones. Viscosity is a property and a function of the material. For
instance, most liquids have viscosities on the order of 1 to 1000 mPas, while gases have
viscosities on the order of 1 to 10 µPas. For a specific material, temperature affects viscosity the
most. As temperature increases, the average speed of the molecules in a liquid increases and
the amount of time they spend in contact with their nearest neighbors decreases. Thus, as
temperature increases, the average intermolecular forces decrease and viscosity decreases.
Table 4-3 demonstrates the viscosities of some common liquids under stated temperatures.
73

Table 4-3: Viscosity values of some common liquids
Fluid T (
o
C) η ( cP)
alcohol, ethyl 20 1.1
blood 37 3–4
glycerin 20 1420
milk 25 3
water 0 1.79
water 20 1
water 100 0.28

The surface tension of a fluid, on the other hand, considers the fact that atoms or molecules at
a free surface have a higher energy than those in the bulk. Surface tension is the elastic
tendency of a fluid surface which makes it acquire the least surface area possible. It causes
liquids to always form into shapes that minimize their total energy and thus their free surface
area. For instance, in the absence of other forces, a free droplet liquid will take the shape of a
sphere which has the lowest surface area, and therefore requires the lowest surface energy.
Surface tension (γ) is defined as the ratio of the surface force (F) to the length (d) along which
the force acts:
𝜸 =𝑭 /𝒅
Surface tension is measured in SI unit of N/m, although the more common unit is dyn/cm.
𝟏 𝒅𝒚𝒏 /𝒄𝒎 = 𝟏 𝒎𝑵 /𝒎
74

The importance of viscosity and surface tension of inhibitor fluid in the SIS-metal process and
also in defining the behavior of inhibitor droplets on powder beds will be discussed in the
coming chapters.
4.4.3 Surfactants
Optimization of SIS-metal requires fine-tuning of inhibitor fluid properties. The adjustment of
surface tension will be obtained by the addition of surface active agent or surfactants to
inhibitor solutions.
Surfactants are substances that are active at surface areas and are attracted to phase
boundaries between immiscible materials or interfaces. The term surface is used when one of
the two immiscible phases of an interface is a gas. Surfactants act at surfaces to lower the free
energy of that phase boundary. When the boundary is covered by surfactant molecules, the
surface tension is reduced. As the number of surfactant molecules at the interface increase, the
reduction in surface tension gets larger. This continues until the liquid is saturated with
surfactants. Table 4-4 presents the effect of surfactant on lowering surface tensions of sample
solutions [32].
75

Table 4-4: Typical values of surface and interfacial tensions (mN/m)
Air-water 72-73
Air-10% aqueous NaOH 78
Air-aqueous surfactant solution 40-50
Aliphatic hydrocarbon-water 28-30
Aromatic hydrocarbon-water 20-30
Hydrocarbon-aqueous surfactant solution 1-10

The molecular structure of surfactants is amphiphilic, meaning it consists of two parts. When
the fluid is water the two parts are: hydrophilic head which is soluble in a water, and
hydrophobic tail which is insoluble (Figure 4-11) [33].

Figure 4-11: Schematic illustration of a surfactant
Detergents are common surfactants that are specialized to dispersing oils or oily particles in
water. They can effectively reduce the surface tension of water and water solution. Since they
are quite inexpensive and widely accessible to consumers, they have been candidates for the
preparation of inhibitor in the SIS-metal process.
76

For the fabrication of bronze parts using SIS-metal, a common detergent has been chosen
which proved effective in the process. However, for stainless steel parts, it has been replaced
by an industrial surfactant due to its incompatibility with the ceramic precursor used in the
solution. This will be discussed in more details in chapters five and six.
4.4.4 Printer setting adjustments
The other set of fluid-related parameters to be adjusted is printer setting parameters. Since a
consumer photo printhead is used in the process along with its software and control board, the
available settings need to be investigated to achieve best results. Epson WorkForce 30 printer
provides several options for photo printing settings. These settings include paper type, paper
size, print quality, color controls, and print speed. Changing each parameter will result in
different nozzle activation outcomes. These settings will be experimentally studied and the best
combination of them will be chosen for part fabrication experiments. Figure 4-12 illustrates
some the options available for controlling the print quality and other mentioned settings
through Epson WorkForce 30 software.
77

Figure 4-12: Epson WorkForce 30 control software
In order to make the results of these experiments applicable to other printheads, these settings
further translate into two major process parameters of droplet sizes and fabrication speed.
Droplet size is determined by print quality (resolution of the print) and paper type. Fabrication
speed is affected by print quality as well and also color controls. Higher resolution and higher
DPI means smaller droplets and better print quality, however, it will reduce print speed. This
fact as well as other factors such as droplet penetration into the powder need to be considered
in the setting adjustments.
In addition to print speed, vertical penetration of the droplets and surface quality also play
important roles in the determination of optimal print settings. Assuming fluid properties
adjusted, when droplets are placed closer to each other (higher resolution), a more uniform
78

surface will be obtained. However, too much fluid on the powder will flood the surface and
saturate the printed layer. Vertical penetration also is affected by how much fluid is printed on
the powder.
4.5 Powder-related Parameters
The success of any powder-based metal AM process depends significantly on the
characterization of its powder material. Powder in an AM process can be quantified by several
different measures, such as:
 Material which can be a pure element or an alloy
 Mean particle size and size distribution
 Particle shape
 Flowability
Powder-related parameters are of great importance in SIS-metal process. Studying and
understanding these parameters are crucial in optimizing the surface quality and dimensional
accuracy of the fabricated parts.
4.5.1 AM Powder Production
AM metal powders are typically manufactured by either gas atomization method or plasma
atomization. Gas atomized powders are the most commonly used raw materials for the additive
manufacturing of metal parts [34]. In gas atomization method, molten metal stream is
79

disrupted by gas jets and the generated small metal particles are cooled and gathered in a
collection chamber (Figure 4-13).

Figure 4-13: Gas atomization process for metal powder production (source: LPW Technology)
Plasma atomization is a relatively new technology capable of producing high quality and
extremely spherical powders. In this method, a rotating wire feedstock is arced with gas
plasma, and the molten metal is atomized as it is flung off the wire. The molten metal droplets
then cool into spherical powder particles. This method is limited to materials that can be
formed into wires.
80

Figure 4-14: Plasma atomization process for metal powder production (source: LPW Technology)
Depending on the production methods, different particle shapes and size distributions can be
achieved.
4.5.2 Particle Shape
The shape of the particles is a major factor in powder-based processes. It affects several
process parameters. Particle shape plays an important role in bed density, flowability, and
sintering characteristics.
Figure 4-15illustrates several types of ideal powder particle shapes. In reality the shapes may
vary slightly from the ideal cases.
81

Figure 4-15: Powder particle shapes (source: thelibraryofmanufacturing.com)
Spherical powders are more common in powder AM processes. They possess good flowability
and acceptable sintering characteristics that make them ideal for SIS-metal.
4.5.3 Particle Size and Size Distribution
Powder particle size is a factor that basically defines the minimum possible layer thickness that
can be achieved in the process. In reality, in a batch of powder produced with the same
method, particles can be of the same shape, but they will not have the exact same size [35]. In
other words, there always exists a size distribution with regards to powders. Size distribution
curves are typically provided by the manufacturers and can be accounted for in characterization
of the process parameters.
In this research, powders with the maximum particle size of 40 microns are used. Layer
thickness will also be chosen to be three times the maximum particle size which provides a
smooth powder spread. Besides layer thickness, particle size and size distribution also affect the
density and porosity of the printed part.
82

4.5.4 Powder Density
There are different density measurements that are of interest to AM processes:
 Apparent density
 Tap density
 Bed density
Apparent density (AD) of powder is the mass of a unit volume of powder under standard
conditions without any further processing applied to it. AD is a function of particle shape and
strongly influences the strength of the parts.
Tap density is the apparent density of the powder in a container that has been tapped under
specified conditions [36].
Bed density is an important factor in determining the density of the “green”, or unsintered part
in an AM process. It is the density of a volume of powder consisting of different layers of
powders laid down on each other.
In the SIS-metal process, these densities need to be measured and characterized in order to
understand the porosity of the green part and decide on the amount of inhibitor to be
deposited on each layer.
4.5.5 Porosity
Porosity is defined as the amount of pores (voids) expressed as a percentage of the total
volume of the part [36]. It is an important measure in metal powder sintering that shows how
83

well the particles have bonded with each other and how strong the resulted part might be.
Pores can be either interconnected or isolated. Interconnected pores are connected to the
surface and are permeable to the atmosphere and fluids such as liquid lubricants and water.
Isolated pores are either dead end or closed. Closed pores are not open to the outside
atmosphere and do not contribute to the permeability of the part.
In a sintered part with a density less than 75% of theoretical density, it is safe to assume that
nearly all of the porosity is interconnected. As the overall density increases from 75% to 92%,
the percentage of isolated pores increases rapidly and the functional interconnected porosity is
depleted. [37].
Lower porosities are preferred in AM processes for structural parts that require strong
mechanical properties. However, there are cases where porosity may be favorable, including
but not limited to: fabrication of foam structures, filters, bone implants (to allow bone growth
inside the pores), and bearings for better lubrications.
Archimedes test is the common method in measuring the porosity of sintered parts. In this
method, the pores in the part are sealed using oil impregnation and the volume is measure by
dipping the part into water and using Archimedes principle. Calculating the porous part density
and dividing it to the theoretical density results in the density percentage and consequently the
porosity.
84

In the SIS-metal process, porosity plays an important role in determining sintering profile in
regards with the degree of inhibition. Higher sintering temperature means less porosity which
results in decreasing the degree of inhibition.
4.5.6 Powder Flowability
In SIS-metal, good powder flowability is required for better powder surface qualities after
spreading and uniform layer thicknesses. This helps with uniform penetration of inhibitor
solution into the powder bed which results in better top and side surface qualities.
Flowability is correlated to the Hausner ratio which is calculated using tap density (𝜌 𝑇 ) and
apparent density (𝜌 𝐴 ) as follows [38]:
𝐻 =
𝜌 𝑇 𝜌 𝐴
There are many factors involved in the process and some of them are in direct opposition to
each other. Therefore, there always needs to be a compromise on the choice of powder
characteristics.
4.6 Part Fabrication
After preparing the inhibitor solution with the desired viscosity and surface tension, and also
selecting metal powder with preferred characteristics, they will be utilized for experimentations
using the SIS-metal machine.
85

The first material to be used in the study is bronze for several reasons. First, bronze has been
previously studied and characterized in SIS research and a good amount of knowledge has been
generated in sintering it. Second, using bronze in the SIS-metal beta machine provides a base
for better comparison of the results with previous researches. The third reason is that
experiments with bronze are much less expensive than high temperature materials. Bronze
doesn’t require severe vacuum conditions and high temperatures. A simple and inexpensive
furnace would be good enough for the experimentations.
The SIS-metal beta machine experiments start with the calibration of layer thickness. Layer
thickness is chosen such that layers are spread with smooth surfaces. The smoothness of the
spread is important to achieve a uniform deposition of inhibitor and also to avoid disrupting the
previous printed layers.
The next step is characterizing the amount of penetration of inhibitor droplets onto the
powder. As explained earlier, the Epson WF30 printhead produces very fine droplets. These
droplets have diameters less than 20 microns which are smaller than the maximum powder
particle sizes used in the process. Therefore, more than one pass of printing needs to be
performed in order to achieve the desired depth of penetration onto the layer.
The number of passes alongside other printer settings will be calibrated using a design of
experiments (DOE) approach. In these experiments, different factors are considered for each
variable. Each combination of factors will result in a different depth of penetration and surface
86

quality. Finally, the best result will be chosen based on the desired goals and the corresponding
factors will be selected as process parameters for part printing.
The amount of inhibitor fluid deposited on the layer should be enough to fill all the voids in the
printed areas and also to reach and connect with the previous printed layers. This is known as
saturation level. In inkjet based AM processes, a saturation of more than 100% is necessary to
provide binding between the layers.
Printed line thickness is another parameter that needs careful study and calibration. Thinner
lines printer on the boundary of the parts produce better surface qualities. However, if the line
is chosen too thin, the strength of the sacrificial mold printed around the part will not be
sufficient for part handling after printing and before putting into the furnace. Thick lines on the
other hand, cause more side penetration of the inhibitor into the part and reduce the surface
quality and dimensional accuracy of the parts. The actual line thickness is also chosen through
experimentation.
By finalizing all important parameters, they will be put together in part fabrication experiments.
The first of such experiments is fabricating simple shapes. The goal is to reduce the
complication and the number of variables (de-aggregation) so that the source of the problem
can be identified much more easily. The simple parts consist of a few layers of simple shapes
extruded in a 2.5D form (Figure 4-16). Different build strategies and also parameters such as
line thickness, base layer formation, top layer quality, and green part strength are tested
through simple shape part fabrications.
87

Figure 4-16: Simple shape designs used in SIS-metal process [27]
After process evaluation and validation through initial experiments, more complex 3D parts are
chosen to be fabricated. Different 3D features bring new challenges to the process. These
features include curved surfaces, overhanging sections, small features, stair stepping effects,
internal channels and numerous others that bring up issues which need to be dealt with as they
appear.
If complex part fabrications are successful, the machine and process limitations in fabricating
different feature is characterized. A benchmark part design used to study several other AM
processes will be modified and adapted to study SIS-metal as well. The 3D design of the
benchmark part is shown in Figure 4-17.
88

Figure 4-17: Benchmark part used in the evaluation of SIS-metal process [39]
4.7 Application to high-temperature alloys
Fabricating high resolution 3D parts with bronze material serves as the proof of concept for SIS-
metal beta machine. Since bronze has limited industrial applications, moving to a high
temperature alloy such as stainless steel with numerous industrial applications would be the
next reasonable step.
Adaptation of a new material to the SIS-metal process starts with chemical studies on inhibition
and selecting a few inhibitor ceramic precursors as candidates. The effectiveness of these
candidates is measured by droplet tests. A droplet test includes depositing a droplet of the
candidate inhibitor solution on metal powder and sintering it. The inhibition is then tested by
scratching the deposited areas. The easier the materials in the inhibited regions are removed,
89

the more effective the inhibitor is. The most effective inhibitor is then chosen for the inhibitor
solution preparation and fluid properties fine-tuning.
The viscosity and surface tension of the solution need to be in the desired range of printhead
specifications an also provide good penetration characteristics of the inhibitor onto the
powder. The prepared solution will then be tested on the printhead and final fine-tuning will be
performed on it based the results of the inkjet printing.
After characterizing the inhibitor solution, other process parameters such as layer thickness,
saturation level, and printed line thickness will be calibrated. Finally, part fabrication
experiments follow with 2.5D extruded shapes and then more complex, three-dimensional
parts.

90

5 Chapter Five: Enhancing the Resolution of Selective Inhibition
Sintering (SIS) for Metallic Part Fabrication
5.1 Introduction
Selective Inhibition Sintering (SIS) is a disruptive platform Additive Manufacturing (AM) process
capable of printing parts from polymer, metal and ceramic base materials. In this chapter, the
implementation of a commercial piezo-electric printhead in the fabrication of metal parts using
SIS-metal process is studied. The current system replaces the single-nozzle solenoid valve
previously used in the process and allows for the fabrication of high resolution metallic parts.
This is a result of smaller droplet sizes as well as high resolution printing mechanisms. A Design
of Experiments (DOE) approach is utilized to study the effects of important factors in inhibitor
deposition. These factors include: composition of the inhibitor, quality of the print and the
amount of fluid deposited for each layer. Based on the results of these experiments,
parameters have been identified for the creation of three-dimensional parts. Finally, high
resolution parts are presented and results are discussed.
5.2 SIS-Metal Beta Machine
5.2.1 Hardware
Precise control of droplet sizes is necessary in the printing of high resolution metallic parts. This
is due to droplet sizes and amount of fluid directly affecting powder bed penetration depth. For
this purpose, the use of a DOD high resolution inkjet printhead is necessitated. As mentioned in
91

the previous chapter, the most economical options in the market are commercial home-use
photo printers that utilize either thermal or piezoelectric printheads. In a thermal printhead, a
heated resistor evaporates a small portion of the ink forming an air bubble which forces a
droplet onto the medium. On the other hand, with a piezoelectric (PZT) head a signal is sent to
micro-piezo actuators causing them to vibrate and form a droplet by mechanically pushing fluid
out of the orifice [40]. The advantage in selecting a PZT printhead for this research is a higher
tolerance for variances in the properties of the fluid being printed such as viscosity, surface
tension and density which play important roles in the printability of the fluid. This makes the
calibration of the composition of the fluid easier due to a wider acceptable range of property
values, which in turn enables the use of different candidate materials for inhibition.
The Epson Workforce 30 (WF30), a high resolution and high performance piezoelectric printer
[10], was chosen and incorporated into the SIS-Metal beta machine. This printhead provides
different droplet sizes, resolutions and print qualities that need to be calibrated to meet the SIS
process requirements. The calibration of the parameters will be discussed in following sections.
The WF30 printhead contains three rows of nozzles with each row containing 180 nozzles. Two
rows are dedicated to black color cartridges, and the last is divided into three 60-nozzle
sections: Cyan, Magenta and Yellow [41]. Table 5-1 summarizes WF30 color printer
characteristics as provided by Epson.
92

Table 5-1: Epson WF30 color printer characteristics
Printing Technology 4-color (CMYK) drop-on-demand MicroPiezo® inkjet technology

Ink Palette Cyan, Magenta, Yellow and Black
Ink Cartridge Configuration 5 individual ink cartridges
Original Ink Type Pigment ink
Print Longevity Up to 105 years
Minimum Droplet Size 3 droplet sizes, as small as 3 picoliters
Maximum Print Resolution 5760 x 1440 optimized dpi

The SIS-Metal beta prototype machine consists of a three-axis linear motion mechanisms with
four degrees of freedom (X and Y motion plus two Z motions for the tanks), Motion control
electronics, a piezoelectric printhead, printhead controller board, motion control software and
user-interface software. The 3D design of the machine and different sections of it can be seen
in Figure 5-1.

Figure 5-1: 3D desing of the SIS-metal beta machine
93

Figure 5-2 presents the final manufactured machine.

Figure 5-2: The SIS-Metal Beta machine
5.2.2 Software
SIS-metal beta machine software consists of four major sub sections as follows:
 Printhead control software
 Motion control software
 User interface
 Slicing software
94

The printhead control electronics and software is developed by Epson. In this research, the
same controller and software is modified and utilized for controlling the print process. The
software provides different options for different print parameters that affect resolution,
quality, speed, and output color of the print. These options are studied here to fine-tune
process parameters.
Arduino-based electronics and software are developed for controlling the motion of the
machine. The control board communicates with user through the USB port of the computer and
performs different movements based on the command received by the user.
The user interface is developed in Visual Studio C# platform and is responsible for connecting
the hardware of the machine to user commands and also send different slices to printer for
part fabrication. It provides user with different options for adjusting process parameters such
as layer thickness, movement speed, amount of inhibitor deposition, and positioning of the
axes. Figure 5-3 shows a screen capture of the UI of the SIS-metal beta machine.
Slicing software, developed by Petros et al. [27], takes the STL model of the parts as input and
slices them into layers with predefined thicknesses. Each layer is then saved as an image file
which contains the boundary of the part in that layer in a specified color. Figure 5-4 illustrates a
STL model of a part along with the image of a layer generated by the software.

95

Figure 5-3: User interface software developed for SIS-metal beta machine
96

Figure 5-4: STL model of a part and one layer generated by SIS-metal slicing software
5.3 Calibration of the Machine Parameters
There are two sets of parameters that play key roles in the process; the composition of the
inhibitor fluid and the print settings. The fluid should be fine-tuned such that it is jettable and
inhibition still occurs. However, these properties of the fluid are at odds with each other. For
the greatest level of inhibition to occur, the fluid should contain a high level of solute. This
increases the viscosity of the fluid which reduces its jettability. Once the fluid is fine-tuned, the
print settings are important in calibrating the depth of penetration into the powder bed. Depth
of penetration determines the layer thickness, which in turn affects the resolution and surface
quality of the finished part.
5.3.1 Inhibitor
The key components of the inhibitor solution are water, sugar, and an industrial surfactant. The
concentration of sugar determines the degree of inhibition that can be achieved. The amount
of water and surfactant are important in the jettability of the solution as well as penetration of
97

the solution into the powder. As was previously mentioned, there is a trade-off between the
degree of inhibition and the jettability and the penetration of the solution. Therefore a series of
experiments was conducted to find the optimum composition of the inhibitor solution. The
properties of the fluid that play an important role in jettability are viscosity and surface tension.
Higher viscosities may result in poor jettability while lower viscosities will cause the inhibitor to
leak through the printhead. On the other hand, if the surface tension is too high, the droplets
will be too small and won’t penetrate enough into the powder layer. If viscosity is too low, the
droplets may spread too much causing excessive side penetration and consequently lower
precision. The characterization of the inhibitor fluid will be discussed in details in the next
chapter where a new inhibitor solution is developed to be used in SIS with optimum result. The
inhibitor used in the experiments of this chapter was based on the previous SIS research.
Previous SIS research provided some candidate solution compositions. The one with the best
inhibition result was chosen for the purpose of this research. This inhibitor solution contains
sucrose as the inhibitor precursor, water as the solvent and surfactants as the surface tension
adjustment agent[9].
Different solutions were prepared by changing the concentration level of sucrose in water and
the amount of surfactant used. These solutions were classified experimentally based on their
ease of penetration into the powder layer, the degree of inhibition, and their jettability. Several
different experiments were carried out to fine-tune the properties of the candidates in order to
make them jettable.
98

The viscosities of these solutions were also measured and the one with the closest viscosity to
that of original Epson Ink was chosen to be fine-tuned and used in the machine. For this
purpose, a VISCOlab 3000: Laboratory Viscometer” was used as the experimental setup (Figure
5-5).

Figure 5-5: VISCOlab 3000: Laboratory Viscometer
The results of the viscosity measurements for the original Epson ink and the final inhibitor
solution are shown in Table 5-2.
Table 5-2: Viscosity test results for magenta ink

Material Temperature(
o
C) Viscosity(cp) error ±
Epson Ink 30 2.5 4.1%
Inhibitor Solution 30 2.25 4.6%

99

5.3.2 Printer Settings
The Epson WF30 provides different options for the ink droplet sizes, quality, resolution, and the
number of nozzles engaged in the printing process. These factors, along with the number of
prints on each layer, determine the penetration depth into the powder bed, powder surface
quality after droplet deposition, and print speed.
Before calibrating the printer settings, the powder layer thickness needed to be adjusted. The
metal powder used in this research was a fully alloyed bronze with chemical composition
presented in Table 5-3. The powder particle size has a distribution of 98.7% -325 mesh, 1.3% -
200/+325 mesh as reported by the manufacturer. Thus, the layer thickness cannot theoretically
be less than the largest particle size, approximately 80 µm. For calibration of the layer
thickness, the current spreading mechanism in the machine was used to spread layers with
differing thicknesses. The minimum layer thickness that could be spread consistently was 120
microns. This layer thickness was used in the experiments for determining the optimum print
settings. It should be noted that the SIS process is applicable to smaller layer thicknesses. The
limiting factor in the current study was the particle size.
100

Table 5-3: Chemical Composition of the used bronze powder
Copper 90%
Tin 9.792%
Lead 0.025%
Zinc 0.04%
Iron 0.058%
Phosphorous 0.085%

A design of experiments approach was utilized to investigate the significant and optimum
values of the print parameters. The goal was to choose the best combination of factors that
gives a depth of penetration of slightly more than 120 microns, an acceptable surface quality,
and a satisfactory print speed. Each DoE consisted of a set of small squares printed on a thick
layer of loose powder. Each square was printed with a unique combination of the controllable
factors. The factors included: printed color which determines the number of nozzles engaged,
print quality which determines the droplet sizes, and number of passes which directly affects
the penetration depth as well as the surface quality. Three different levels were considered for
each factor. The factors and their corresponding levels can be seen in Table 5-4.
101

Table 5-4: Important factors and their levels used in the DOE
Levels
Factor Description 0 1 2
1 Color Cyan (C) Magenta (M) Both (B) (Blue)
2 # of Passes 3 4 5
3 Print Quality Text/Image (TI) Photo (P) Best Photo (BP)

The experiment was conducted by printing 27 small squares on a thick layer of powder each
with a unique combination of factors. Table 5-5 illustrates the factor combinations for all
squares.
102

Table 5-5: Factor combinations for the squares

After printing the squares with their assigned factors, the part was bulk sintered in the furnace.
The inhibited powder in the printed areas was then removed, resulting in the part shown in
Figure 5-6. As can be seen in the figure, the final part consists of small square depressions with
different depths of penetration. By quantitatively measuring the depth of each square and
qualitatively comparing the surface qualities, the settings that give the most penetration depth
as well as the best surface quality were determined.
Square # Color No. of
Passes
Print Quality Square # Color No. of
Passes
Print
Quality
1 C 5 P 15 B 4 P
2 B 5 TI 16 M 3 P
3 B 3 BP 17 B 3 TI
4 M 5 BP 18 C 3 TI
5 C 3 P 19 B 4 TI
6 M 3 TI 20 B 3 P
7 B 5 BP 21 C 3 BP
8 M 4 P 22 B 4 BP
9 C 4 P 23 B 5 P
10 M 3 BP 24 M 4 BP
11 M 5 P 25 C 4 TI
12 M 5 TI 26 C 5 BP
13 M 4 TI 27 C 4 BP
14 C 5 TI

103

Figure 5-6: The fabricated experimental part
Based on these experiments, it was concluded that printing with both colors simultaneously
(which equals to an RGB value of R0, G0, and B255) resulted in the greatest depth of
penetration. This results in a faster printing process due to less passes required for each layer.
Figure 5-7 illustrates the penetration depth vs. number of passes for printing in blue color
under different print qualities. As shown in the figure, “four prints” provides a depth of
penetration slightly over 120 microns as desired.
104

Figure 5-7: Penetration depth vs. number of passes under different print qualities
Factors such as speed and surface quality were also considered in choosing the most desirable
factors. Table 5-2 summarizes the comparison between different print qualities. Based on the
table, photo quality is set as the print quality factor of the machine.
Table 5-6: Comparison of different print qualities

5.4 Part Fabrication Process
Once the inhibitor solution and printer parameters were calibrated, the part fabrication process
needs to be determined. There are several steps in fabricating a part using SIS-metal as follows:
1) The STL model is sliced based on predefined slicing algorithms and build strategies.
Factors\print qualities Text/Image Photo Best Photo
Speed High Med Low
Surface quality Low High High
Penetration depth Low High Med

105

2) The slices are sent to the machine one by one to be inkjet printed on each powder
layer.
3) The part chamber is removed from the machine and the green part is prepared for
sintering.
4) The green part is sintered in an appropriate furnace with a specific sintering profile.
5) The sintered part is post-processed to reveal the final part.
Since there are many factors and parameters involved in the part fabrication process, and there
is not much research available on SIS, each step had to be studied by thorough and extensive
experimentations.
The first group of experiments were on slicing algorithms and build strategies. This was studied
by Petros et al. [27]. In his research, Petros studied the effect of five different build strategies or
so called hatching schemes (HSA, HSB, HSC, HSD, and HSE) on the fabrication of four simple
shapes shown in Figure 5-8.
106

Figure 5-8: 3D model of the simple shapes used for studying build strategies
5.4.1 Build Strategies
Hatching schemes directly determine how the sintered parts can be removed from the inhibited
area and what parts of the part boundaries should be inhibited. Hatching schemes are as
follows:
 HSA: In this method, the entire regions outside the parts are covered by inhibitor. The
idea is to make post processing and revealing of the final parts easier. The utilization of
this method in fabricating the mentioned simple shapes is illustrated in Figure 5-9. The
issue with HSA is that too much inhibitor is deposited which penetrates into the parts
and distorts the shapes. Material waste is another concern with using HSA.
107

Figure 5-9: Part fabrication using HSA [27]
 HSB: HSB includes printing only the boundary of the parts with a thick line. The result is
shown in Figure 5-10Figure 5-9. The fabricated parts are more accurate than the previous
method, however there are still shape distortion and dimensional variation involved.

Figure 5-10: Part fabrication using HSB [27]
 HSC: In order to reduce the effects of HSB, HSC was developed similar to HSB with a
difference in border line thickness. In HSC, the thinnest line possible that can be
separated in SIS-metal process is used in part fabrication (Figure 5-11). The dimensional
108

accuracy is acceptable here and the shapes are fine. The issue with HSC is that it only
works for simple 2.5D shapes. 3D shapes cannot be separated using this method.

Figure 5-11: Part fabrication using HSC [27]
 HSD: HSD adds separation lines to HSC which help in the removal of the sintered
sections outside the part. As shown in Figure 5-12, this method results in accurate
shapes and 3D design can also be separated and fabricated. However, HSD adds another
step to post-processing and it also doesn’t help with the wasted material issue.

Figure 5-12: Part fabrication using HSD [27]
Figure shows a gear printed in SIS-metal beta machine using HSD.
109

Figure 5-13: Gear fabrication using HSD
 HSE: The final hatching scheme studied was HSE. In this method, a thin layer outside of
the part is completely covered with inhibitor. Here, the loose powder around the
printed part will be removed using air blower, prior to being sintered. Therefore, only
the part with a thin inhibited area around it, which acts as a sacrificial mold, will be put
into the furnace for sintering. This method is the most effective in fabrication of 3D
complex parts and generates the least amount of waste. The only issue with this
method is the fact that green parts require support structures for their overhanging
features. Figure 5-14 demonstrates HSE in part fabrication process.

Figure 5-14: Part fabrication using HSE [27]
110

Based on the results of these experiments, HSE was chosen as the standard build strategy for
SIS-metal. A novel slicing software was developed to implement HSE and to create a water-tight
mold around the part [27].
5.4.2 Support Structures
As mentioned earlier, in fabricating 3D parts with HSE, the overhanging features require
support. The use of support structures have been studied in the fabrication of two 3D models
with different features shown in Figure 5-15.

Figure 5-15: 3D models of support structure study test parts
Fabricating these parts without support structures has been a failure. Shows the failed
fabricated parts.
111

Figure 5-16: Failed parts due to lack of support structures
In order to overcome the issue, support structures shown in Figure 5-17 had been added to the
parts. These supports were generated by open source software used in the mask image
projection Stereolithography processes.

Figure 5-17: Test parts with support structures
The fabrication process using the generated supports was successful for the mentioned test
parts. However, the surface quality of the parts on locations where the supports were attached
to the parts was not acceptable. The traces of the supports were visible on the parts.
112

Experimenting with different types of support structures and methods led to the elimination of
the supports. Instead, the parts were supported by the addition of fine ceramic powders
around them. In this method, the parts will be fully supported by ceramic powder and the
quality of them will be intact. Figure 5-18 shows the application of this method on one the test
parts.

Figure 5-18: Test part covered in fine ceramic powder for support
This method generated the best results in the fabrication of complex 3D shapes. Figure 5-19
demonstrates successful fabrication of the test parts.

Figure 5-19: Successful fabrication of the test parts using ceramic powder support
113

5.5 Fabrication of Complex 3D Parts
Upon investigating the capabilities of the SIS-beta machine, several complex designs were
chosen to further demonstrate the process performance in fabricating high quality complex 3D
parts. Figure 5-20 shows the digital model of Einstein’s head beside the post-processed
fabricated part.

Figure 5-20: Digital model of Einstein's head beside the fabricated part in SIS-metal process
As the final result, 3D model of a jet engine provided by GE to the 3D printing community was
selected for fabrication. F shows the fabricated model.
114

Figure 5-21: Fabricated jet engine model (3D design provided by GE)
5.6 Benchmark Part Fabrication
After successful fabrication of 3D parts in SIS-metal, the capability of the machine in creating
small features was evaluated using two benchmark parts. The first part contained a simple
design to examine the minimum wall thickness and gap size that the machine can handle.
Figure 5-22 illustrates the design and the fabricated part used in this experiment. As can be
seen in the figure, the top portion of the design contains square holes with different
dimensions. On the left several lines with different thicknesses are printed that determine the
smallest gap size. Finally, the shapes on the right demonstrate the minimum wall thickness that
can be printed with the current SIS machine under the defined parameters.
115

Figure 5-22: Fabricated part for testing the resolution of the machine
The results of the measured features are demonstrated in Table 5-7.
Table 5-7: Feature size measurements
Feature Size in microns
Gap size 330
Square Hole Side 820
Wall Thickness 250

The second part was chosen among several suggested benchmark parts reported in literature. A
benchmark part contains various features that are designed to investigate accuracy and
capability of the considered AM machine or process in manufacturing complex geometries.
Moylan [42] reviewed common test artifacts and compared their features, field of use and
advantages. Based on his research, Moylan [43] introduced a standardized benchmark part
which was designed to include common features in other suggested test artifacts while
simultaneously highlighting both the capabilities and limitations of a machine or process. That
116

standard part served to quantify the accuracy and diagnose specific defects in an AM machine.
For the purpose of this research, the suggested design was modified to fit the build size of the
machine while maintaining all of the proposed features. The new design and the important
features that it encompasses are illustrated in Figure 5-23.

Figure 5-23: Modified benchmarking part design
Table 5-8 summarizes the characteristics of the machine that is investigated by each feature.
117

Table 5-8: Benchmarking part features

A digital CAD model was created using the SolidWorks software. The file was saved in STL
format and sliced using our internally developed slicing software. A rendering of the digital
model as well as the fabricated part can be seen in Figure 5-24. The fabricated benchmark part
showed very promising results as it exhibited good performance in fabrication of both vertical
and horizontal internal features, through holes, small gaps and extrusions, sharp corners and
interesting features throughout.
Feature Characteristics being investigated
Top Surface The flatness or warpage of a large surface
Center Hole Roundness and perpendicularity of the holes
Pins and holes aligned with x- and y- axes The straightness error of x- and y- axes
Staircases The capability of the machine in producing surfaces
parallel to the machine axes
Outer Edges The capability of the machine in producing surfaces
askew to the machine axes
Central Cylinders Concentricity of two cylinders
Ramp Stair-step effect
Fine features (rectangular bosses and holes) Establishing the minimum required separation of
features as well as the minimum size of rectangular
holes and bosses
Fine features (cylindrical bosses and holes) Establishing the minimum feature size achievable by the
machine
Lateral features Investigating the use of support structures

118

Figure 5-24: The benchmarking part (a) digital model (b) fabricated part
It was observed that the ability of the process to produce small extruded features was highly
dependent upon the extent of sintering of the part. In addition, the designed small gaps were
displaced during the spreading process and thus did not show in the final part results. This is
not a limitation of the SIS-Metal process, but of the powder spreading technique used. Future
research can fine-tune the calibrated parameters discussed here as well as newly identified
ones in order to improve the accuracy of the SIS-Metal process.
5.7 Surface Quality Measurements
There are several factors that affect the surface quality of parts such as sintering conditions,
powder material, spreading, and green part density. The surface quality is also dependent on
powder particle size, but no direct conclusion can be made solely based on that. In the SIS-
metal process, all the above-mentioned factors have been fine-tuned to achieve the best
results, so it is safe to consider them fixed for the purpose of surface quality measurement.
119

Therefore, it can be stated that the surface quality of the SIS-metal parts is mainly affected by
the behavior of the inhibitor deposited on the powder bed.
Cylindrical droplet penetrations in opposed to spherical penetrations create more uniform
surfaces. The spherical penetrations have different amounts of penetrations in z direction
which result in uneven surfaces with larger pores on the surface.
Saturation level and physical properties of the inhibitor also play important roles in surface
quality. Figure 5-25 visually compares the roughness of the bottom surface and top surface of a
test part. As can be seen, the top surface is visibly much rougher due to the deposition of
inhibitor fluid, whereas no inhibitor is deposited on the smooth bottom surface of the part.

Figure 5-25: (a) Bottom surface quality (b) top surface quality of the SIS-Metal test part
In order to have a better measure for surface quality, the roughness of the two surfaces was
quantified by measuring the depths of the pores and pits of each surface of the test part at 10
different locations (Figure 5-26).
(a) (b)

120

Figure 5-26: Surface roughness experiment test part and the locations where measurements were taken
Figure 5-27 illustrates the measured values. The average values are 163.81 and 48.8 microns
respectively for the top and bottom surfaces. This measurement can serve as a metric to
quantify and compare surface qualities.

Figure 5-27: Top and bottom surface pore sizes
0
100
200
300
0 2 4 6 8 10 12
Hole Sizes [microns]
Measurement Location Along Y Axis
Top vs. Bottom Surface Roughness
Top Surface
Bottom Surface
1
2
3
4
5
6
7
8
9
10
121

5.8 Conclusion
This chapter has analyzed the performance of the SIS process using a high-resolution piezo-
electric printhead for the objective of developing a high-resolution SIS-Metal prototype
machine. The use of the printhead introduced several new parameters into the process which
were thoroughly studied and calibrated for the provided materials. The performance of the
developed SIS-beta machine in producing high resolution parts was evaluated and verified by
fabricating a standardized benchmark part. This experiment also investigated the capability of
the SIS-beta machine in creating various common features found in 3D printed parts with
complex geometries, which was further examined by fabrication of complex 3D parts.

122

6 Chapter Six: Application to high temperature alloys (Stainless Steel)
6.1 Introduction
The application of the Selective Inhibition Sintering process in the fabrication of high resolution
metallic parts with bronze alloy material had been thoroughly studied in previous chapters.
However, bronze, due to its mechanical and thermal properties, has limited applications in
industries such as aerospace, automotive, and dentistry which seem to be the target industries
for high-end metal AM technologies. If the SIS-metal process is to have an impact on such
target industries and compete with other Additive Manufacturing processes, the use of higher
temperature alloys such as stainless steel, titanium alloys, and cobalt-chrome is a necessity.
While titanium and cobalt-chrome are of interest to different AM technologies and have
specific applications in high-tech and bio-medical industries, processing them requires utilizing
cost prohibitive furnaces with extreme vacuum conditions which was out of the scope of this
research.
Stainless steel, on the other hand, has a broader range of applications and is also more
accessible and affordable in powder form. Therefore, stainless steel has been selected to be
used in SIS-metal as the case study in the development of a systemic approach for adapting
new materials in the process. It should be noted that the developed inhibitor and methodology
discussed in this chapter can theoretically be directly applied to the other high temperature
metals.
123

6.2 Applications of Stainless Steel in Industry
The most important characteristic of stainless steel that makes the material widely appropriate
for many different applications is its corrosion resistance property. The fact that different
grades of stainless steel can be hardened and shaped easily also contributes to the popularity
of them in industry. The following statements summarize a number of applications where
stainless steel has proven itself through many years of reliable service [44]:
 Cutlery and kitchenware: The most well-known application of stainless steels is for cutlery
and kitchenware. Some grades of stainless steel can be hardened, tempered, and shaped
into parts with sharp edges which together with the corrosion resistance property will make
them perfect for this application.
 Automotive: Stainless steels are used in automobiles structural parts as well as exhaust
systems and catalyst converters due to their mechanical and thermal characteristics.
 Power generation: Corrosion is a big concern in power generation industry especially at
elevated temperatures. This makes stainless steel alloys a good fit for the application.
 Food production: Stainless steel is widely used in food production and storage. The reason
behind this is the corrosiveness of the strong anti-bacteriological cleaning and rinsing
systems that are used in the processes.
 Architecture, building and construction: The low maintenance cost and anti-vandal
characteristics of stainless steels makes them suitable for public transport, ticket machines,
124

and street furniture. There is also a growing market for the use of stainless steel in
construction of buildings and bridges.
The aforementioned applications as well as numerous other ones make stainless steel a target
for different Additive Manufacturing processes. Therefore, the rest of this research will focus
on the development of a new inhibitor fluid for the SIS-metal beta machine in order to fabricate
high resolution stainless steel parts. Stainless steel 316L was chosen for this purpose as the
most common stainless steel alloy used in industry. The inhibitor development will be achieved
by following the guideline and methodology resulted from the previous work which includes
these steps: inhibitor candidate determination, fluid properties fine-tuning, jettability
experimentations, and powder-fluid interaction studies. Finally, the inhibitor will be tested by
fabricating high resolution stainless steel parts.
6.3 Inhibitor Candidate Determination
Petros [31] performed thorough research on the potential use of different metal oxides for the
inhibition of metal powder sintering. As mentioned in his study, since metal oxides are not
typically soluble in water, they are not directly used in the preparation of the inhibitor solution.
Rather, a molecular precursor is utilized which decomposes to the desired oxide during
sintering.
In order for inhibition to occur, the oxide material needs to be in solid form while the base
metal powder is being sintered. Therefore, the melting temperature of the final oxide is the
most important criterion in determining the possible candidates for inhibition. The other
125

factors to be considered are the solubility of the precursor in water and the number of oxide
molecules generated per desired unit of the inhibitor solution.
Thus, the considered criteria in determining the candidate oxides are as follows:
 A precursor should yield a high number of oxide particles after printing and
decomposition
o A candidate precursor should contain a high fraction of the final oxide
o A candidate precursor should be highly soluble
 The sintering temperature of the oxide should be higher than the sintering temperature
of the base material
Table 6-1 shows Common metal oxides and their respective melting points.
Table 6-1: Common metal oxides and their respective melting points [31]
Ceramic Oxide Melting Temperature High Melting Point Referenced in Literature
MgO 2852 YES YES
ZrO2 2715 YES YES
CaO 2572 YES YES
SrO 2531 YES NO
Al2O3 2072 YES YES
ZnO 1975 YES NO
BaO 1923 YES NO
TiO2 1843 YES YES
126

As for the initial requirement of high melting temperature of metal oxides, the ones with the
melting points of at least 30% greater than that of stainless steel 316L were chosen. This
resulted in the elimination of the oxides with melting points below 1820
o
C [31]. Further, the
candidates that haven’t been mentioned in literature as being effective inhibitors were also
ruled out resulting in the remaining candidates highlighted in green in the table.
The remaining candidates were compared with regards to the other assumptions of high
solubility in water and high number of oxides resulted from the decomposition of the
precursor.
For the identification of the precursors, nitrate and sulfate families were proposed to be the
candidates for inhibition precursors. These two families of molecules are commonly known to
have high solubility in water and decompose to metal oxides at high temperatures [31]. The
following table compares the four candidates regarding their solubility in water and the number
of oxide molecules they yield at high temperatures.
Table 6-2: Potential chemical precursors to desired metal oxides [31]
Chemical Name Formula Metal Atoms
per Molecule
Molar Mass
[g/mol]
Solubility in 100
mL H 2O [g]
Oxide Particle
Yield
Magnesium Nitrate Mg(NO 3) 2 1 256.4 125.0 2.9E+23
Aluminum Nitrate Al(NO 3) 3 1 213.0 73.9 2.1E+23
Aluminum Sulfate Al2(SO 4) 3 2 342.2 36.4 1.3E+23
Zirconium Sulfate Zr(SO 4) 2 1 285.4 52.5 1.1E+23
127

The decomposition of all these materials happens at a lower temperature than the sintering
temperature of 316L. Therefore, the metal oxides will be present during the sintering of the
base metal and inhibition will be likely to happen.
Based on this information, it seems logical to select magnesium nitrate as the main inhibitor
candidate for stainless steel 316L. The next step would be the experimental validation of the
candidate inhibitor material.
6.3.1 Concentration levels
In order to study the behavior of magnesium nitrate solution on 316L powder and to fine-tune
its jettability properties, different concentrations of the solute in water had been prepared.
100% concentration is obtained by fully saturating the solution and recording the amount of
salt needed. Other concentration levels are calculated based on that amount. Magnesium
nitrate salt with the formula Mg(NO 3) 2 is unstable in air and quickly forms the hexahydrate with
the formula Mg(NO 3) 2.6H 2O, therefore the most common magnesium nitrate in the market
comes in the hexahydrate form. The solubility of the hexahydrate form is different from the
anhydrous form (containing no water) and was not reported in the datasheet. The salt used in
this research is manufactured by the Avantor Performance Materials, Inc. under J.T.Baker®
brand.
128

It was experimentally observed that 259.44 g of magnesium nitrate hexahydrate fully saturates
100 ml of water. Other concentration levels are achieved using fully saturated amount as
shown in Table 6-3.
Table 6-3: Magnesium Nitrate Concentration levels
Concentration level % Salt (g) Water (ml)
0 0.00 100
10 25.94 100
20 51.89 100
30 77.83 100
40 103.78 100
50 129.72 100
60 155.67 100
70 181.61 100
80 207.56 100
90 233.50 100
100 259.44 100
6.3.2 Droplet Experiments
In order to experimentally validate the inhibitor candidate, a set of experiments was designed
and conducted by Petros [31]. In those experiments, magnesium nitrate solutions were
prepared with different concentration levels as shown in Table 6-3. A droplet from each
solution was then deposited onto a thick layer of 316L powder. Next, the powder layer
129

containing the droplets was sintered at a specific temperature for a specified period of time.
The resulted sintered part was sand blasted to see how effective the inhibitions of the solutions
had been. In droplet experiments, if inhibition occurs by a solution, the corresponding droplet
would cause the powder in that region to be weakly bonded and therefore create a hole by
sandblasting. Otherwise, the powder would be sintered and inhibition would fail.
Droplet experiments were carried out for seven sintering temperatures of 1000
o
C, 1050
o
C,
1100
o
C, 1150
o
C, 1200
o
C, 1225
o
C, and 1250
o
C under vacuum. The position of and its
corresponding concentration level is shown in Table 6-4.
Table 6-4: Droplet positions and their corresponding concentration level [31]
Droplet Location with Solution Concentration
(1, 1) = 10% (1, 2) = 100%
(2, 1) = 60% (2, 2) = 40%
(3, 1) = 30% (3, 2) = 90%
(4, 1) = 50% (4, 2) = 70%
(5, 1) = 20% (5, 2) = 80%

The resulted parts with inhibited areas are presented in Figure 6-1. It was shown that all the
solutions above 30% saturation could fully inhibit sintering of stainless steel in temperatures
below 1225
o
C.
130

Figure 6-1: Droplet experiments with 10 different concentration levels [31]
6.4 Fluid Properties Fine-Tuning and Jettability
As mentioned earlier, surface tension and viscosity of the fluids are the determining factors in
their jettability. In the case of the commercial Epson printhead used in this research, in order to
get the best result and a stable jettability, the viscosity of the fluid needs to be in the range of
2-5 cP and the surface tension should be managed to stay between 30 and 50 dyn/cm.
These fluid properties also define the behavior of the droplets in touch with a porous medium.
Behaviors such as vertical and lateral penetration, penetration time, pore volumes saturation
level, and surface quality can be directly affected by the properties of the fluid and therefore
can be enhanced by the study of these properties.
The process of fine-tuning the viscosity and surface tension of the inhibitor fluid starts with
measuring the viscosity of solutions with different concentration levels. The solutions with
viscosities in the desired range are then chosen and their surface tensions are measured and
131

calibrated if needed. The calibration will be achieved by the addition of surfactant. The viscosity
of the resulting solutions are measured again and if they are still in the range, they will be taken
to the next step. Here, the jettability of the candidate solutions are experimentally tested by
running them through the actual printhead. In this step, the inhibitor fluid is printed on powder.
Different factors such as the printed surface quality, droplet sizes, jet stability of the fluid, and
fluid-powder interactions determine the final solution. The process flow is shown in Figure 6-2.

Figure 6-2: Inhibitor fluid preparation process
6.4.1 Viscosity
The viscosities of the solutions with the concentration levels previously shown in Table 6-3,
were measured using a DV1 Digital Viscometer Model LVDV-IP, manufactured by Brookfield
AMETEK (Figure 6-3).
•Chemical Study
•Inhibition Study
Candidate
Material
•Different
Saturation
Levels
Solution
Preparation
•Droplet
Experiments
Inhibition
•Candidates in the
desired range
Viscosity
Measurments
•Surfactant
addition
Surface Tension
Calibration
•Fluid Stability
•Print properties
Jettability
132

Figure 6-3: Brookfield AMETEK digital viscometer
The device is reported to be accurate within ±1.0% of the measurement range and repeatable
within ±0.2%. All Brookfield viscometers employ the principle of rotational viscometry, in which
the viscosity is measured by sensing the torque required to rotate a spindle at constant speed
while immersed in the target fluid through a calibrated spring. The torque, τ, is proportional to
the viscous drag (𝑐 𝑣 ) on the immersed spindle, and thus to the viscosity of the fluid, η as seen in
the following equation. The LV in the model type stands for low viscosity; thus, the viscometer
has a calibrated spring more elastic than most other models
𝜂 =
𝜏 𝑐 𝑣

Important parameters that may contribute to the error in viscosity measurements are
cleanliness of spindle (to reduce friction from equipment), and temperature of the fluid (to
maintain constant and uniform viscosity). In order to reduce the error in viscosity
133

measurements in this research, the spindle is carefully cleaned before each use. Additionally,
the temperature of the fluid is measured and maintained as close to room temperature as
possible. The temperature of the fluids were held in the range of 24.4±0.1
o
C.
Figure 6-4 shows the measured viscosities of the solution with concentration levels ranging
from 0 percent (pure water) to 100 percent (fully saturated) with increments of 10 percent. As
can be seen in the graph, the viscosity of the solution rises with the addition of more
magnesium nitrate.

Figure 6-4: Viscosity measurements for solutions with different concentration levels
From these measurements, inhibitor solutions with saturations ranging from 50% to 100% seem
to be in the desired range of 2-5cP for jettability. These will be considered as candidates for
further experimentations.
0
1
2
3
4
5
0 10 20 30 40 50 60 70 80 90 100 110
Viscosity (cP)
Saturation Level (%)
Viscosity Measurements
134

6.4.2 Surface Tension
Surface tension measurements were collected with the use of CSC-DuNouy Interfacial
Tensiometer 70545, manufactured by Central Scientific Company, Inc (Figure 6-5).

Figure 6-5: Surface tensiometer manufactured by Central Scientific Company, Inc (source: cscscientific.us)
The accuracy and repeatability is reported to be within ±0.05 dyne/cm. This tensiometer
employs the du Noüy ring method of measurement, which utilizes a platinum ring hanging
parallel to the surface of liquid. The ring is sunk into the liquid and gradually drawn apart from
the surface in the vertical direction. The surface tension of the liquid that is hanging by the ring
creates a force on it, which changes as the ring is drawn further. The maximum force, F,
required to raise the ring from the surface of the liquid is measured and related to the liquid’s
surface tension, γ. This relationship is shown in the following equation:
135

γ=F/4πR
where R is the center diameter of the ring [45].
CSC reports that the surface tension readings may differ from the true value by as much as 30
percent, although for most measurements the difference is less than 5 percent. Sources of this
error difference are the cleanliness of rings and vessels, in addition to the phenomena which
occurs while using the device. Briefly explained, the force measured as the surface of the liquid
is drawn, assumes it is only in the vertical direction. However, the surface film created at the
breaking point also has non-vertical components (Figure 6-6). Thus, the true value of the force
is less than that calculated using the mentioned equation.

Figure 6-6: Surface of film at breaking point.
Another factor affecting the error is the difference in pressure on top and bottom of the ring.
The pressure on top of the ring is atmospheric, whereas the pressure on the bottom is
136

atmospheric minus ρ∙g∙h, where ρ is the density of the liquid, g is the acceleration due to
gravity, and h is the surface of the liquid. This factor creates a larger force value than the
calculated one. Although these errors are opposite in sign, they do not generally compensate
each other.
It is also important to ensure that the temperature of the liquid is always uniform, because the
liquid’s surface tension is dependent on temperature. For the purpose of this research, the
temperature is kept in the range of 21.6±0.2
o
C. The error is also safely assumed to be uniform
for all solutions by randomly conducting the measurements on the same day and with the same
cleanliness conditions of the equipment. Figure 6-7 demonstrates the measurement results.

Figure 6-7: Surface tension measurements for solutions with different concentration levels
0
10
20
30
40
50
60
70
80
90
0 10 20 30 40 50 60 70 80 90 100
Surface Tension (dyn/cm)
Saturation Level (%)
Surface Tension Measurements
137

The following conclusions can be drawn from the graph:
 The graph shows that as the concentration level increases, the surface tension of the
fluid decreases.
 The most significant drop in surface tension occurs in the beginning of the graph where
the solution is only 10% saturated. From this point on, increasing the saturation is not as
effective as the first 10%.
 All of the solutions above 30% concentration level have surface tensions in the range of
approximately 50-60 dyn/cm which doesn’t make any of them preferable to the others
regarding the jettability of the solution. Therefore, the viscosity criterion is still the
dominant factor.
6.4.3 Droplet Penetration Time
Droplet penetration time can serve as an important measure to qualify the penetration of
solutions into the powder. Faster penetration is preferred, but at the same time, the
penetration of the droplets shouldn’t be too fast in order to allow for adjacent droplets to
attach to each other to produce a more uniform surface. On the other hand, too slow of a
penetration will result in incomplete deposition before the spreading of the next layer.
The experimental setup for measuring the penetration time of the solutions consists of a
powder container filled with stainless steel powder, a timer, iPhone 6 slo-mo video recording,
and a micropipette. The micropipette used here is single channel variable droplet size
manufactured by SciLogex (Figure 6-8). This micropipette can generate droplets with volumes
138

between 10-100ul. For the purpose of penetration time measurements, a droplet size of 15ul
has been chosen.

Figure 6-8: SciLogex single channel variable micropipette
Figure 6-9 shows the experimental setup at the instance when a droplet is being deposited on
the powder bed.

Figure 6-9: Experimental setup for droplet penetration time measurements
139

The experiment is randomized for different concentration levels and the measurements are
repeated three times. Figure 6-10 presents the average penetration time of the droplets from
solutions with concentration levels.

Figure 6-10: Droplet penetration time vs. concentration level
As the graph shows, as the concentration level increases, the droplets have a harder time
penetrating into the powder. This was expected due to the increase in the viscosity of the
solution.
The printed droplets need to penetrate into the powder before the next layer is spread. Three
seconds is a reasonable time for letting the droplets penetrate while the machine is preparing
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Penetration time (s)
Saturation Levels
Droplet Penetration Time
140

for the next layer spread. Therefore, setting a goal of three seconds results in ruling out 90%
and 100% saturated solutions.
6.4.4 Droplet Shapes and Sizes
By combining the results of viscosity measurements, surface tension measurements, and
droplet penetration time measurements, 50% to 80% saturated solutions have been chosen for
further experimentations. These experiments include the determination of droplet shapes and
droplet sizes.
Droplets jetted out of the printhead are in picoliter size range and are hard to visualize. Due to
the fast rate of deposition, high speed cameras would be required to capture the motion of the
droplets which significantly increases the cost of experiments.
In this research a different approach is adopted to see and compare the shape of the droplets
from different saturated solutions. In this method, 50%, 60%, 70%, and 80% solutions are put
into different color channels of the printhead. The y axis of the machine is held fixed, forcing
the printhead to move only along the x axis. Then, a single line with a specific color and a
thickness of 80 microns is sent to the printer. It was experimentally discovered that printing an
80-micron line will make the printhead print the line using all the nozzles in that specific color.
For instance, if the y axis is fixed, printing an 80-micron line with a single color (cyan, magenta,
or yellow) will result in 60 rows of thin lines printed beside each other. This is because the color
channels in the printhead have 60 nozzles each positioned close to each other in y direction.
141

In order to make the droplets visible, the lines are printed on a black piece of ABS plastic. Figure
6-11 illustrates the microscopic view of the printed lines on the mentioned plate for 50%-80%
solutions.

Figure 6-11: Inkjet-printed droplets for inhibitor solutions with different precursor amounts and concentration
levels: (a) 50% (b) 60% (c) 70% and (d) 80%
From this image, it can be inferred that 50% solution droplets are more circular and have less
shape variations.
In the next experiment, the size of the droplets are measured. For this purpose, 1 ml of each
solution is put into a 1-ml syringe which replaces a specific ink cartridge on the printhead. The
142

strategy was to print an exaggerated amount of each solution to increase the precision and
reduce the error of the measurements.
A 500x500 pixel square shape was designed to be printed with a resolution of 635 DPI. This
translates into a 20x20 mm square being printed 100 times onto the powder bed. Figure 6-12
shows the top view of the powder tank with the square printed on it. As it can be seen in the
image, the edges of the square are not straight lines. This happens when the printed inhibitor,
saturates all the pores in the powder layer, and then it starts penetrating side-wise. This
wouldn’t happen with 100 prints if a thicker layer of powder had been chosen.

Figure 6-12: 500x500 px square printed for droplet size measurements
143

This square was printed 100 times for each solution and the total amount of solution consumed
(𝑣 𝑇 ) was recorded in [ml]. The printhead deposits one droplet for each pixel. Therefore, droplet
size (𝑣 𝐷 ) in [pl] can be calculated as follows:
𝑣 𝐷 =𝑣 𝑇 /100/500/500∗10
9

Also the amount of fluid deposition (𝑣 𝑚𝑚
) in [nl] for one pass of the printhead for 1 mm
2
will
be:
𝑣 𝑚𝑚
=𝑣 𝑇 /100/20/20×10
6

Table 6-5 shows the amounts of solutions deposited and the corresponding droplet sizes for
each solution.
Table 6-5: Droplet size measurements for different concentration levels
Concentration
level (%)
Amount Deposited After
100 Prints (ml)
Droplet
Volume (pl)
Droplet Diameter
[micons]
Deposition
per mm
2
[nl]
50 0.60 24.0 35.8 15
60 0.58 23.2 35.4 14.5
70 0.51 20.4 33.9 12.75
80 0.44 17.6 32.3 11

It is shown that as the concentration levels increase, the generated droplets get smaller. While
the difference in the droplet sizes is not significant, a solution with larger corresponding
144

droplets is preferred. It simply translates into more solution deposited on each pass of the
printhead and increases the speed of part fabrication.
Droplet sizes and shapes are important factors in determining the final concentration level.
However these experiments didn’t highlight a significant difference between selected solutions.
6.4.5 Surface quality
Print surface quality is the most important parameter in selecting the best concentration level
for the process. Surface qualities of the prints are visually compared with each other. In order
to amplify the effect, each solution is printed 10 times on a thick powder layer. Each printed
surface is then observed under microscope and compared with the rest.
Figure 6-13 presents the printed surface qualities of 50%-80% solutions under microscope. As it
can be seen in the image, the 50% solution provided a more uniform surface. Large cracks are
visible in the surfaces with higher concentration levels. As explained in the third chapter, this
phenomenon is due to the higher viscosity of the solutions. In fluids with higher viscosity, as the
droplets penetrate into the pores of the powder bed, they have a tendency to pull powder
particles inward. In certain regions, the droplets attach with each other and create the opening
and cracks on the surface of the powder.
145

Figure 6-13: Printed surface quality of the solutions with different concentration levels:
(a) 50% (b) 60% (c) 70% and (d) 80%
Based on this experiment, the composition of the solution with 50% saturation is selected as
the final composition of the inhibitor fluid to be fine-tuned and used in the SIS-metal process.
Although the surface quality of the 50% solution is acceptable, however, it’s not ideal and still
shows small cracks and non-uniformity. The next step includes adding surfactant to the solution
and tailoring its surface tension in order to optimize the depth of penetration and surface
quality.
146

6.4.6 Diluted Surfactant
Different industrial surfactants have been studied in this research. The one with the best result
was DeMULS DLN-2314 surfactant from Deforest Enterprises, Inc. This surfactant is
biodegradable and is mostly used in making emulsions. Table 6-6 summarizes some
characteristics of DeMULS DLN-2314 surfactant.
Table 6-6: DeMULS DLN-2314 characterisitcs (source: DeForest website)
Appearance Clear to hazy amber liquid
pH (1% in DW) 7.0 - 9.0
Specific Gravity @ 25
o
C 0.99 gm/ml
Brookfield Viscosity@ 21
o
C 100 cPs
Solubility @ 10% by wt.
Soluble in water, alcohols, glycols,
Dispersible or insoluble in aromatic
solvents & oils
% Actives 93% Minimum

DeMULS DLN-2314 is soluble in water which means it can be used in the prepared inhibitor
fluid. Also its pH level is not acidic and is close to neutral state, therefore it will not be corrosive
to the materials used in the printhead. The minimum activity number states that at least 93% of
the surfactant consists of surfactant particles and the rest includes filler, such as water,
preservative, glycerin, and so on.
The typical applications of this surfactant in industry includes: degreasers, engine cleaners,
paint strippers, floor strippers, all-purpose cleaners, and spot removers.
147

For surface tension calibration of the 50% saturated inhibitor solution, different amounts of the
surfactant needs to be added to it. For this purpose, a 1% by volume diluted surfactant in water
is prepared for two reasons. First, DeMULS DLN-2314 is a strong surfactant and small amounts
of it easily saturate the solution and minimizes the surface tension. Second, its viscosity is so
high that larger amounts of it may increase the viscosity of the solution.
In this experiment, 45 g of 50% inhibitor solution was used. Different volumes of diluted
surfactant were added to it and its surface tension was measured. Figure 6-14 presents the
surface tension measurements versus the amount of diluted surfactant added.
The amount of diluted surfactant added starts from 10µl and goes all the way up to 1 ml. It is
shown that the solution is saturated with surfactant after 300 µl and the surface tension
doesn’t decrease any further.
It was also experimentally observed that solutions with surface tensions below 45 dyn/cm are
not stable in the printhead. These solutions leak due to low surface tensions and the droplets
cannot be stopped. The solution that worked the best was detected to be the one with 50ul of
diluted surfactant in the graph below which has a surface tension of 48.6 dyn/cm.
148

Figure 6-14: Surface tension measurements versus the amount of diluted surfactant added to 45g of 50%
saturated inhibitor solution
After finalizing the inhibitor solutions and before starting part fabrication, Powder-related
parameters need to be characterized. These parameters include Layer thickness, powder bed
density and porosity, vertical and side penetration depth, and inhibition effectiveness.

0
10
20
30
40
50
60
70
0 100 200 300 400 500 600 700 800 900 1000
Surface Tension (dyn/cm)
Diluted Surfactant (µl)
Surface Tension Calibration
149

6.5 Powder-related Parameters
6.5.1 Stainless Steel 316L
The metal powder used in this research is 316L stainless steel powder with spherical particles
purchased from Carpenter Powder Products. This material is called “Micro-Melt 316L” with part
no. 2310103-0010. The chemical composition of the powder is shown in Table 6-7.
Table 6-7: Chemical composition of 316L powder
Fe Bal.
C ≤0.03%
Cr 17%
Ni 10.5%
Mo 2.5%
Si ≤1.0%
Mn ≤2.0%
Cu ≤1.0%

The particles have a maximum size of 44 µm (325- mesh) and a minimum size of 16 µm. A
microscopic SEM image of the particles is shown in Figure 6-15.

Figure 6-15: 316L stainless steel powder used in SIS-metal research
150

6.5.2 Layer Thickness
Comparing the particles sizes with the provided scale in Figure 6-15 (50 µm), it can be seen that
the majority of particles are much smaller than 50 microns. This decreases the flowability of
powder and make spreading more difficult. Different layer thicknesses were tried with the SIS-
metal machine. Among them, 120 µm was the smallest thickness that produced smooth
surfaces with an acceptable repeatability.
6.5.3 Powder Bed Density
Determining powder bed density provides valuable information about the process. By knowing
void spaces in the powder, the amount of inhibitor solution that needs to be deposited can be
precisely calculated. This specifies the number of passes that the printhead needs to make for
printing each layer.
For measuring the bed density, a 50x50 mm square is printed for 25 layers each containing 120
microns of powder. This produces a 50x50x3 mm rectangular prism as shown in Figure 6-16.
Using a density of 8 g/cc for 316L stainless steel, a fully dense part with the mentioned
dimensions should weight 60 g. However, the printed part weighed 31.37 g. This means that
the printed part is
31.37
60
×100=52.28% dense and 47.72% of it includes porosity and void
spaces.
151

Figure 6-16: 50x50x3 mm rectangular prism printed to determine bed density
6.5.4 Print Saturation Level
Number of print passes for each layer to fill all the pores and voids of the powder bed in the
selected areas with inhibitor can be determined as follows.
Figure 6-17 shows a 1 mm
2
shape element of a layer.

Figure 6-17: A 1x1x0.12 mm shape element of a powder layer
This element has a porosity of 47.725%. Therefore, the void space is calculated as:
𝑣𝑜𝑖𝑑 𝑠𝑝𝑎𝑐𝑒 =1×1×0.12×0.47.725= 0.05727𝑚𝑚
3
=57.27 𝑛𝑙
152

From previous experiment (Table 6-5), it is known that the printhead deposits 15 nl of inhibitor
in one pass of the print for 1 mm
2
of surface. Therefore, the number of passes required to fill all
the voids in the powder bed is equal to:
𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑝𝑟𝑖𝑛𝑡 𝑝𝑎𝑠𝑠𝑒𝑠 =
57.27
15
=3.818≈4
This means four passes of print will provide enough inhibitor to penetrate into the powder and
fill all the voids and also attach to the previous layer. This translates into 100% print saturation
level meaning that the deposited fluid equals the void space. In some cases, saturation of more
than 100% might be required to ensure that layers properly attach to each other. In this
research, it was observed that 4 passes is sufficient for part fabrication.
6.5.5 Sintered Density
The density of the parts was studied by Petros [31]. In his study, he reports a density of 57.01%
for sintering with a temperature of 1100
o
C and a hold time of 30 minutes at that temperature.
It was shown by Petros [31] that using a second sintering pass, the density of the final parts can
get to 91.46% which is more than the density of 85% defined by ASTM B783-13 as a standard
density for parts with structural applications [46].
6.6 Part Fabrication
Less complex 2.5D designs are chosen to be fabricated for evaluating the process in order to
avoid complex features and make debugging easier by de-aggregating the problem. These
complex features which will be added to the next steps of part fabrication experiments include
153

small feature, inclined surfaces, thin walls, overhanging features, internal channels, and
bottom-facing surfaces.
The first part to be printed was a gear. The cross section of the part and its dimensions are
shown in the Figure 6-18. The part consists of 25 layer with a total thickness of 3mm. The
thickness of the boundary lines is chosen to be 0.6 mm which provides enough inhibitor for
easier removal of the sacrificial mold around the part and also makes the green part strong
enough to be handled before sintering. The line thickness for small features inside the part that
don’t require much strength can be as thin as 100 microns to allow for the fabrication of
delicate details.

Figure 6-18: Cross section design of the first SIS-metal stainless steel part
The first printed gear was sintered at 1050
o
C for 30 minutes under vacuum. The inhibition was
fine and it was relatively easy to blast away inhibited regions using a micro sandblasting
equipment. However, the part itself wasn’t completely sintered which resulted in the loss of
154

material on edges and surfaces of the part during sandblasting. As it can be seen in Figure 6-19,
the top surface of the part has poor quality and edges are not sharp.

Figure 6-19: (Left) Stainless steel gear- first trial - (Right) Microscopic image of the surface
The same part design with the same dimensions and machine parameters was printed for the
second time. This time, the green part was sintered at 1100
o
C (instead of 1050
o
C) for 30
minutes under the same vacuum condition. The inhibition was still quite acceptable. The higher
temperature resulted in an increase in sintered part strength. Figure 6-20 shows the second
trial of stainless steel gear. It can be seen that this time, the surface quality is good and the
edges are straight and smooth.

Figure 6-20: (Left) Stainless steel gear- second trial - (Right) Microscopic image of the surface
155

This result presented the capability of the SIS-metal process and SIS-metal beta machine in
fabricating parts with 316L stainless steel alloy which is one of the most common high
temperature alloys used in industry. However, this capability needs to be further investigated
by printing 3D parts with complex geometries and fine details. Figure 6-21 shows the 3D model
and the dimensions of the part considered for fabricating with SIS-metal beta machine.

Figure 6-21: 3D complex part model to be fabricated by SIS-metal (source: GE)
This part is a simplified version of a third-stage compressor in a jet engine provided by GE for
prototyping purposes. Figure 6-22 shows the printed part after the sintering stage. This part
includes inhibited regions around the part which will be removed by sand-blasting.
156

Figure 6-22: Sintered compressor part with inhibited regions around it
The final part is revealed after removing the inhibited regions. Figure 6-23 presents the final 3D
part fabricated with stainless steel.

Figure 6-23: Final fabricated 3D stainless steel part
6.7 Systematic approach for incorporating new materials in the SIS-metal process
Figure 6-24 illustrates the proposed flowchart for incorporating new materials in the SIS-metal
process. In this approach, the process starts with the chemical study of inhibition process and
identification of potential candidate materials. The important criteria for ranking these
materials are their solubility in water and the number of generated oxide particles during the
157

sintering process. Among these candidates, one with better ranking is chosen for further
experimentations.
The main purpose of this approach is fine-tuning the viscosity and surface tension of the
inhibitor fluid to match the printhead desired range. The viscosity is adjusted by preparing
different concentrations of the inhibitor solution. The performance of these concentrations in
regards with inhibition is tested through droplet experiments.
The concentrations that perform acceptable inhibition are then chosen for viscosity
measurements. The next step involves surface tension adjustment of the solutions in the
desired range of viscosity. This is done by the addition of surfactant to them. Iterations need to
be performed for achieving the final solution with the desired properties.
The final solution is used in print experiments on powder using the printhead. If the print
parameters are satisfying, the inhibitor solution will be used for part fabrication experiments.
This process, summarized in Figure 6-24, has been utilized in this research for incorporating
stainless steel into the SIS process and was validated by the successful fabrication of high
resolution stainless steel parts.
158

Figure 6-24: Systematic approach for incorporating new materials in the SIS-metal process
159

7 Chapter Seven: Conclusion and Future Work
7.1 Conclusion
The following objectives were defined in the current study:
 Understanding the droplet penetration into the powder bed phenomenon
 Enhancing resolution, surface quality, and dimensional accuracy of SIS-metal parts
 Extending the application of SIS-metal to high resolution stainless steel parts
 Providing a systematic approach to incorporate new materials in the SIS-metal process
These objectives have been met to a satisfactory extent in two phases of the research. In the
first phase, SIS-metal beta machine was designed and built to be used as the experimental
setup throughout the research. This machine, utilizing a high resolution commercial inkjet
printhead, is capable of depositing droplets as small as 6 pL into the powder bed. For the first
application, the machine parameters and inhibitor fluid properties were calibrated for the
fabrication of bronze parts. The resulted parts show that the SIS-metal beta machine has
significantly increased the resolution and quality of the parts with minimum feature sizes and
surface quality comparable with those of the parts fabricated with the expensive commercial
metal AM machines.
In the second phase of the research, stainless steel was chosen as the second target material
for the SIS-metal process. A great deal of research was dedicated to the study of fluid
properties and their effects on the jettability of the fluid from the printhead and droplet
160

penetration into the porous powder beds. This research created a knowledge base that was
used in fine-tuning and optimizing the inhibitor fluid properties used in the fabrication of
stainless steel parts. The second objective was met by developing the inhibitor solution with
the desired properties and printing high resolution stainless steel parts.
The preparation and calibration of the inhibitor solution for stainless steel resulted in the
creation of a systematic approach for adaptation of new materials in the SIS-metal process
which was the third objective of this research. The provided step-by-step guideline paves the
road for future research on targeting materials with specific applications in certain industries.
7.2 Contributions
Contributions of this research to the SIS process include:
 Obtaining an understanding of the behavior of the droplets in contact with porous
powder beds. The study describes how changing fluid properties will affect its behavior
when deposited on powder.
 Researching different inkjet technologies and identifying the most effective one for the
SIS process. The developed machine while being very cost-effective is capable of
fabricating 3D complex parts with resolution comparable with that of the cost-
prohibitive commercial metal AM machines.
 Creating a systematic procedure for configuring the print conditions.
161

 Introducing the use of surfactants to the inhibitor fluids in order to adjust their surface
tension. The introduced surfactants are stable in the solution and are compatible with
the inkjet printhead used in this research.
 Systematic engineering of inhibitor characteristics such as choice of material, extent of
saturation, and surfactant selection.
 A thorough study of the process which led to the identification of important parameters
that affect the part fabrication process. These parameters can be used in the future
research on SIS in order to further improve the results.
 A massive improvement in the surface quality and dimensional accuracy of SIS-metal
parts which resulted from the development of the SIS-metal beta machine and the fine-
tuning of the inhibitor solution properties.
 The development of the inhibitor solution for stainless steel material which is
compatible with the printhead and with properties optimized for obtaining acceptable
surface quality and resolution.
 The fabrication of the first 3D stainless steel part with the SIS-metal process.
 Developing a systematic approach for incorporating new materials into the SIS-metal
process. Variations of this approach can also be applied to any other processes that deal
with inkjet technologies and inks or printing fluids.
7.3 Proposed Future Research Directions
Future research can focus on either of these two categories:
162

 In-depth study of the factors affecting SIS-metal process
 Adaptation of new materials into the process
7.3.1 In-depth study of the factors affecting SIS-metal process
There are several important factors in SIS-metal that if studied, can result in parts with better
mechanical, dimensional, and aesthetic properties. The following two factors can be studied
next:
 Shrinkage: Understanding and modeling shrinkage in the SIS-metal process will play a
very important role in the density and dimensional properties of the parts. In a powder-
based technology, studying relationship of the shrinkage with powder parameters such
as particle sizes and shapes is a necessity.
 Powder particle sizes: Several researches can be conducted on the effect of particle
sizes on different process parameters such as layer thickness, green part density,
shrinkage, porosity, and dimensional accuracy of the parts. Numerous studies can be
found in literature on mixing powders with different size distributions in order to
increase the density of the parts (bimodal powders). These studies can be adapted to
SIS-metal and their effects on the inhibition process can be studied.
7.3.2 Adaptation of new materials into the process
In order to extend the application of SIS-metal to different popular industries in AM, other high-
temperature alloys need to be adapted in the process.
163

Different industries have different standards for the material and mechanical properties of the
parts. For instance, in aerospace industries which are investing heavily on the application of AM
processes in fabricating their complex parts, the weight of the parts as well as their strength
and tolerance to higher temperatures are determining factors. Therefore, titanium alloys are
very popular in aerospace industries. Medical and dental applications are other major targets
for AM metal processes. Cobalt-chrome alloys are vastly used in these industries due to their
biocompatibility with blood and soft tissues. Thus, future studies on SIS-metal should also focus
on the adaptation of titanium and cobalt alloys. These studies have to include the engineering
of appropriate inhibitor solutions and also studying the sintering behavior of these materials.

164

8 References
1. Coykendall, J., et al., 3D opportunity in aerospace and defense: Additive manufacturing
takes flight. A Deloitte series on additive manufacturing, 2014. 1.
2. Khoshnevis, B., et al., SIS–a new SFF method based on powder sintering. Rapid
Prototyping Journal, 2003. 9(1): p. 30-36.
3. Torabi, P., M. Petros, and B. Khoshnevis, Selective Inhibition Sintering: The Process for
Consumer Metal Additive Manufacturing. 3D Printing and Additive Manufacturing, 2014.
1(3): p. 152-155.
4. Asiabanpour, B., et al. Advancements in the SIS process. in Proceedings from the 14th
SFF Symposium, Austin, Texas. 2003.
5. Khoshnevis, B., M. Yoozbashizadeh, and Y. Chen, Metallic part fabrication using selective
inhibition sintering (SIS). Rapid Prototyping Journal, 2012. 18(2): p. 144-153.
6. Gibson, I., D.W. Rosen, and B. Stucker, Additive manufacturing technologies. 2010:
Springer.
7. Asiabanpour, B., K. Palmer, and B. Khoshnevis, An experimental study of surface quality
and dimensional accuracy for selective inhibition of sintering. Rapid Prototyping Journal,
2004. 10(3): p. 181-192.
8. Mojdeh, M., Metallic parts fabrication using the SIS process. 2005.
165

9. Yoozbashizadeh, M., Metallic part fabrication with selective inhibition sintering (SIS)
based on microscopic mechanical inhibition. 2012: University of Southern California.
10. Epson Micro Piezo Print Head Technology. Epson Exceed Your Vision, 2010.
11. Hapgood, K.P., et al., Drop penetration into porous powder beds. Journal of Colloid and
Interface Science, 2002. 253(2): p. 353-366.
12. Denesuk, M., et al., Capillary penetration of liquid droplets into porous materials. Journal
of colloid and interface science, 1993. 158(1): p. 114-120.
13. Nguyen, T., W. Shen, and K. Hapgood, Drop penetration time in heterogeneous powder
beds. Chemical Engineering Science, 2009. 64(24): p. 5210-5221.
14. Litster, J., B. Ennis, and L. Lian, The science and engineering of granulation processes.
Vol. 15. 2004: Springer.
15. Popovich, L.L., D.L. Feke, and I. Manas-Zloczower, Influence of physical and interfacial
characteristics on the wetting and spreading of fluids on powders. Powder Technology,
1999. 104(1): p. 68-74.
16. Ayala, R., E. Casassa, and G. Parfitt, A study of the applicability of the capillary rise of
aqueous solutions in the measurement of contact angles in powder systems. Powder
technology, 1987. 51(1): p. 3-14.
17. Holman, R.K., et al., Spreading and infiltration of inkjet-printed polymer solution droplets
on a porous substrate. Journal of colloid and interface science, 2002. 249(2): p. 432-440.
166

18. Middleman, S., Modeling axisymmetric flows: dynamics of films, jets, and drops. 1995:
Academic Press.
19. Chua, K.W., et al., Time scale analysis for fluidized bed melt granulation-II: Binder
spreading rate. Chemical engineering science, 2011. 66(3): p. 327-335.
20. Heinzl, J. and C. Hertz, Ink-jet printing. Advances in electronics and electron physics,
1985. 65: p. 91-171.
21. Lu, K. and W.T. Reynolds, 3DP process for fine mesh structure printing. Powder
technology, 2008. 187(1): p. 11-18.
22. Moon, J., et al., Ink‐Jet Printing of Binders for Ceramic Components. Journal of the
American Ceramic Society, 2002. 85(4): p. 755-762.
23. de Gans, B.J., P.C. Duineveld, and U.S. Schubert, Inkjet printing of polymers: state of the
art and future developments. Advanced materials, 2004. 16(3): p. 203-213.
24. Lee, M., J.C. Dunn, and B.M. Wu, Scaffold fabrication by indirect three-dimensional
printing. Biomaterials, 2005. 26(20): p. 4281-4289.
25. Onuh, S. and K. Hon, Optimising build parameters for improved surface finish in
stereolithography. International Journal of Machine Tools and Manufacture, 1998.
38(4): p. 329-342.
26. Lu, K., M. Hiser, and W. Wu, Effect of particle size on three dimensional printed mesh
structures. Powder Technology, 2009. 192(2): p. 178-183.
167

27. Petros, M., P. Torabi, and B. Khoshnevis, The influence of build strategies in selective
inhibition sintering (SIS). The International Journal of Advanced Manufacturing
Technology, 2015: p. 1-11.
28. Hutchings, I.M. and G.D. Martin, Inkjet technology for digital fabrication. 2012: John
Wiley & Sons.
29. Tekin, E., P.J. Smith, and U.S. Schubert, Inkjet printing as a deposition and patterning
tool for polymers and inorganic particles. Soft Matter, 2008. 4(4): p. 703-713.
30. Fukano, T., Ink jet printer. 2004, Google Patents.
31. Petros, M., The Extension of Selective Inhibition Sintering to High Temperature Alloys.
2016, University of Southern California.
32. Holmberg, K., et al., Surfactants and polymers in aqueous solution. Journal of Synthetic
Lubrication, 2004. 20(4): p. 367-370.
33. Bredt, J.F., Binder stability and powder/binder interaction in three dimensional printing.
1995, Massachusetts Institute of Technology.
34. Ahsan, M.N., A.J. Pinkerton, and L. Ali, A comparison of laser additive manufacturing
using gas and plasma-atomized Ti-6A1-4V powders. 2012.
35. Upadhyaya, G., Powder metallurgy technology. 1997: Cambridge Int Science Publishing.
36. Standard Terminology of Powder Metallurgy, in ASTM standards, 2-5: B243 –13 (2004).
168

37. Eisenmann, M., Porous powder metallurgy technology. ASM International, ASM
Handbook, 1998. 7: p. 1031-1042.
38. Liu, B., et al. Investigation the effect of particle size distribution on processing
parameters optimisation in Selective Laser Melting process. in International solid
freeform fabrication symposium: an additive manufacturing conference. University of
Texas at Austin, Austin. 2011.
39. Torabi, P., M. Petros, and B. Khoshnevis, Enhancing the resolution of selective inhibition
sintering (sis) for metallic part fabrication. Rapid Prototyping Journal, 2015. 21(2): p.
186-192.
40. Ink Jet Printers Technical Brief. Epson 2007.
41. Kwan, J.G., Design of electronics for a high-resolution, multi-material, and modular 3D
printer. 2013, Massachusetts Institute of Technology.
42. Byun, H.-S. and K.H. Lee, Design of a new test part for benchmarking the accuracy and
surface finish of rapid prototyping processes, in Computational Science and Its
Applications—ICCSA 2003. 2003, Springer. p. 731-740.
43. Moylan, S., et al. Proposal for a standardized test artifact for additive manufacturing
machines and processes. in Proceedings of the 2012 Annual International Solid Freeform
Fabrication Symposium. 2012.
169

44. Nilsson, J.-O. Can mankind survive without stainless steels? December 2014; Available
from: http://www.stainless-steel-world.net/pdf/history_of_stainless_steel.pdf.
45. Butt, H.-J., K. Graf, and M. Kappl, Physics and chemistry of interfaces. 2006: John Wiley &
Sons.
46. ASTMB783-13, Standard Specification for Materials for Ferrous Powder Metallurgy (PM)
Structural Parts, in 2009, ASTM International: West Conshohocken, PA.

Abstract (if available)

Abstract Selective Inhibition Sintering (SIS) is a disruptive platform Additive Manufacturing (AM) process, capable of printing parts from polymer, metal and ceramic base materials. The principle idea behind the SIS process is the prevention of selected segments in each powder layer from sintering. This research is focused on the enhancement of surface quality and dimensional accuracy of the SIS-metal parts as well as incorporating new materials into the process. These are achieved in two phases of the research. ❧ In the first phase, an SIS-metal machine, called SIS-metal beta machine, is developed which incorporates a high resolution commercial piezo-electric printhead. A systematic procedure is proposed and implemented to experimentally configure and optimize print parameters. The SIS-metal beta machine is successfully utilized in fabrication of high resolution bronze parts with surface qualities and dimensional accuracies comparable to those of commercial metal AM machines. The presented results suggest a massive improvement in the properties of the fabricated parts. ❧ The second phase of this research is dedicated to the adaptation of high temperature metals in the SIS-metal process. For this purpose, a systematic step-by-step approach is developed for preparing new inhibitor solutions and fine-tuning their fluid properties for optimum jettability from the printhead and optimized droplet penetrations into the porous powder beds. This approach is validated by fabricating high resolution stainless steel parts.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The extension of selective inhibition sintering (SIS) to high temperature alloys

PDF

Metallic part fabrication wiht selective inhibition sintering (SIS) based on microscopic mechanical inhibition

PDF

Energy control and material deposition methods for fast fabrication with high surface quality in additive manufacturing using photo-polymerization

PDF

Selective separation shaping: an additive manufacturing method for metals and ceramics

PDF

3D printing of polymeric parts using Selective Separation Shaping (SSS)

PDF

Selective Separation Shaping (SSS): large scale cementitious fabrication potentials

PDF

Multi-scale biomimetic structure fabrication based on immersed surface accumulation

PDF

Slurry based stereolithography: a solid freeform fabrication method of ceramics and composites

PDF

Hybrid vat photopolymerization processes for viscous photocurable and non-photocurable materials

PDF

Contour crafting construction with sulfur concrete

PDF

Hybrid vat photopolymerization: methods and systems

PDF

Motion-assisted vat photopolymerization: an approach to high-resolution additive manufacturing

PDF

Contour Crafting process planning and optimization

PDF

Experimental and numerical techniques to characterize structural properties of fresh concrete

PDF

Scalable polymerization additive manufacturing: principle and optimization

PDF

Deformation control for mask image projection based stereolithography process

PDF

Evaluation of interlayer bonding strength in contour crafted structures: experimental and statistical approach

PDF

Fabrication-aware machine learning for accuracy control in additive manufacturing

PDF

Real-time flow control of viscous fluids using 3D image processing

PDF

Machine learning-driven deformation prediction and compensation for additive manufacturing

Asset Metadata

Creator Torabi, Payman (author)

Core Title Enhancing the surface quality and dimensional accuracy of SIS-metal parts with application to high temperature alloys

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Industrial and Systems Engineering

Publication Date 07/27/2016

Defense Date 06/17/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag 3D printing,additive manufacturing,inhibition,inkjet technologies,OAI-PMH Harvest,powder metallurgy,sintering

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Khoshnevis, Behrokh (committee chair), Chen, Yong (committee member), Goo, Edward (committee member)

Creator Email paymantorabi2007@gmail.com,ptorabi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-282301

Unique identifier UC11280386

Identifier etd-TorabiPaym-4644.pdf (filename),usctheses-c40-282301 (legacy record id)

Legacy Identifier etd-TorabiPaym-4644.pdf

Dmrecord 282301

Document Type Dissertation

Format application/pdf (imt)

Rights Torabi, Payman

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