University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
The Extension of Selective Inhibition
Sintering (SIS) to High Temperature
Alloys

By:

Matthew R. Petros

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
1
Acknowledgements

It is a challenge to determine how to begin to acknowledge all of the people, relationships, and
circumstances that have influenced me as I completed my doctoral work in Engineering at the University
of Southern California.

First and foremost, I must thank my mother, Marlene Petros, for her truly unconditional love and
unwavering support throughout my educational journey. My mother has served as an inspiration to me
and my brother for our entire lives. She consistently emphasized the importance of education, and this
instilled in us a passion for the pursuit of knowledge. The example she set of hard work and self-
sacrifice for her children has been a beacon of light for which I will be forever thankful. I love you, Mom,
and thank you.

I would also like to thank my brother, and dearest friend, Christopher Petros. My brother has always
encouraged me to dream big and reach for nothing but the loftiest of goals. His encouragement and love
has been a source of indescribable empowerment. Chris, that “look” you give is really something else. I
am truly lucky to have a brother like you.

I would also like to thank my “doctoral brother” and dear friend Payman Torabi. Yoda, you motivate and
challenge me. I am excited to see what you will accomplish in your lifetime. I have no doubt that your
contributions, both scientifically and as a human being, will have a profound impact on this world.

2
My sincerest thanks also to my colleague and dear friend Jing Zhang. Jing, I will always remember the
time spent tinkering in my garage, attempting to come up with the next big thing. I have greatly
appreciated your friendship throughout this journey.

I would also like to thank a doctoral colleague and close friend, Babak Haji. Babak, it has been a pleasure
to navigate this doctoral journey together. You have a wonderful sense of humor and have truly been a
friend to me. I look forward to a lifetime of friendship with you.

I must also thank my undergraduate research assistants, Maria Yepremian and Trevor Nelson. The
completion of my educational journey was in large part due to the ability of these two bright minds to
perform. Thank you very much for your dedication, your commitment, and your friendship throughout.
Thank you to my committee members, Dr. Yong Chen and Dr. Steven Nutt. Your advice was instrumental
to my success as a research professional. I would like to especially thank Dr. Nutt for allowing access to
the world-class M.C Gill Composites Center for the purposes of my research.

Finally, I would like to thank my advisor and mentor, Dr. Behrokh Khoshnevis. Dr. Khoshnevis, you are a
true inspiration. Of the multitude of ways you have influenced my life, the most important to me was
your belief in the ability of any person to be creative. Your philosophy and teachings on creativity have
had a lasting impression on my day-to-day life, and I am thankful to you for that, sir.

3
Abstract
Selective Inhibition Sintering of metal alloys (SIS-metal) was previously proven effective in the additive
manufacture (AM) of low resolution bronze parts. Recent advancements in the use of a high-precision
inkjet print head represented an order of magnitude improvement in SIS-metal resolution. However, the
fabrication of complex three-dimensional metallic parts required new SIS-metal compatible, cross-
sectional image generation based on the part boundary profile. Multiple candidate layer-processing
approaches were identified and validated for basic geometries. These approaches were chosen from
previous research as well as preliminary investigations and were applied to a modified SIS-metal process
for validation. The validation criteria were based upon the amount of powder waste produced, the
ability to handle complex geometries, printing speed, extraction (post-processing) speed, and part
integrity. Results are discussed for implementation of the five candidate layer processing approaches in
the fabrication of basic shapes. An evaluation is presented for their use on more complex geometries.
Two approaches were then chosen for the construction of more complex geometries, the results of
which are presented.
These approaches were then extended to higher temperature, ferrous alloys. A new inhibiting agent,
Magnesium Nitrate, was identified and validated for use with a stainless steel alloy. Initial feasibility
experiments were conducted to verify the inhibition concept, and the results were used to fabricate a
stainless steel gear with the SIS-metal process. A set of experiments was conducted to realize higher
density values and hence improve the mechanical properties of as-built parts. Dog bone specimens were
tested in tension and a final tensile strength of 507 MPa (UTS) was achieved. A final sintered density of
87.47% was achieved in a SIS-metal fabricated stainless steel gear.

4
Table of Contents
Table of Contents .......................................................................................................................................... 4
List of Figures ................................................................................................................................................ 7
List of Tables ................................................................................................................................................. 9
1. Introduction ........................................................................................................................................ 10
1.1 Introduction to Additive Manufacturing ........................................................................................... 10
1.2 Advantages and Disadvantages of AM Processes ............................................................................. 11
1.3 Selective Inhibition Sintering - Metal ................................................................................................ 12
1.4 Advantages/Disadvantages of SIS-metal .......................................................................................... 13
1.5 SIS-metal Challenges ......................................................................................................................... 14
1.6 Purpose of the Study ......................................................................................................................... 17
1.7 Goal of the Study .............................................................................................................................. 17
1.8 Research Contribution ...................................................................................................................... 18
1.9 Organization of the Dissertation ....................................................................................................... 18
2. Background ......................................................................................................................................... 20
2.1 A Review of Metal AM Processes...................................................................................................... 20
2.2 A Review of Selective Inhibition Sintering Research ......................................................................... 22
2.3 Technology Overview Comparison of AM Processes ........................................................................ 28
2.4 Critique of Past Approaches .............................................................................................................. 29
3. Influence of Build Strategies ............................................................................................................... 31
3.1 Abstract ............................................................................................................................................. 31
3.2 Experimental Setup and Procedures................................................................................................. 31
3.2.1 Hardware ....................................................................................................................................... 31
3.2.2 Software ......................................................................................................................................... 34
3.2.3 Experimental Research Objectives ................................................................................................ 36
3.3 Experimental Plan ............................................................................................................................. 37
3.3.1 Hatching Scheme A (HSA) .............................................................................................................. 38
3.3.2 Hatching Scheme B (HSB)............................................................................................................... 39
3.3.3 Hatching Scheme C (HSC) ............................................................................................................... 41
3.4 Application to Complex Geometries ................................................................................................. 42
3.5 Mechanical Testing and Relative Density ......................................................................................... 44
3.6 Discussion .......................................................................................................................................... 46
5
4. Literature Review ................................................................................................................................ 47
4.1 Organic Material Burnout ................................................................................................................. 47
4.1.1 Introduction ............................................................................................................................... 47
4.1.2 Binder Burnout ........................................................................................................................... 47
4.1.3 Sucrose Decomposition ............................................................................................................. 53
4.2. Powder Metallurgy .......................................................................................................................... 54
4.2.1 Overview .................................................................................................................................... 54
4.2.2 Metal Powder Characteristics .................................................................................................... 55
4.2.3 Compaction ................................................................................................................................ 56
4.3 Introduction to Sintering ................................................................................................................... 56
4.3.1 Sintering Characterization.......................................................................................................... 58
4.3.2 Mechanical Testing in Powder Metallurgy ................................................................................ 59
4.3.3 Sintering of Copper Alloys .......................................................................................................... 61
4.3.4 Sintering of Ferrous Alloys ......................................................................................................... 62
4.4 Sintering Inhibition ............................................................................................................................ 66
5. Methodology ....................................................................................................................................... 70
5.1 Introduction ...................................................................................................................................... 70
5.2 Research Hypotheses ........................................................................................................................ 72
5.3 Research Plan & Procedure............................................................................................................... 72
5.4 Assumptions Derived from Literature .............................................................................................. 74
5.5 Analytical Identification of Candidate Oxides ................................................................................... 75
5.6 Analytical Identification of Candidate Precursors............................................................................. 77
5.7 Special Consideration of Deposition by Print Head (dilution talk) ................................................... 78
5.8 Stainless Steel Sintering .................................................................................................................... 78
5.9 Droplet Tests ..................................................................................................................................... 79
5.10 Part Fabrication ............................................................................................................................... 80
5.11 Improvement of Part Strength ........................................................................................................ 80
6. Results ................................................................................................................................................. 81
6.1 Stainless Steel Sintering .................................................................................................................... 81
6.1.1 Procedure ................................................................................................................................... 81
6.1.2 Results ........................................................................................................................................ 84
6.2 Saturated Droplet Testing ................................................................................................................. 85
6
6.2.1 Procedure ................................................................................................................................... 85
6.2.2 Results ........................................................................................................................................ 87
6.3 Diluted Droplet Testing ..................................................................................................................... 88
6.3.1 Procedure ................................................................................................................................... 88
6.3.2 Results ........................................................................................................................................ 91
6.4 Scanning Electron Microscopy .......................................................................................................... 92
6.4.1 Procedure ................................................................................................................................... 92
6.4.2 Results ........................................................................................................................................ 93
6.4 Part Fabrication ............................................................................................................................... 102
6.4.1 Procedure ................................................................................................................................. 103
6.4.2 Results ...................................................................................................................................... 104
6.5 Improvement of Part Strength ........................................................................................................ 105
6.5.1 Procedure ................................................................................................................................. 105
6.5.2 Results ...................................................................................................................................... 108
7. Conclusion and Future Work ............................................................................................................ 116
7.1 Conclusion ....................................................................................................................................... 116
7.2 Future Work .................................................................................................................................... 116
References ................................................................................................................................................ 118

7
List of Figures

Figure 1.1: The current SIS-metal process. ................................................................................................. 13
Figure 1.2: Fishbone diagram identifying potential areas for SIS-metal improvement.............................. 16
Figure 2.1: Typical DMLS setup. .................................................................................................................. 21
Figure 2.2: Typical 3DP setup. ..................................................................................................................... 22
Figure 2.3: The SIS-polymer process [25]. .................................................................................................. 23
Figure 2.4: The first implementation of an SIS-metal process [25]. ........................................................... 26
Figure 2.5: (a) SIS-metal result from Khoshnevis et al (b) Inhibited negative region [11]. ......................... 27
Figure 2.6: (a) SEM image of untreated bronze powder after sintering (b) SEM image of inhibited bronze
powder with inhibitor material covering the surface of bronze particles [16]. ......................................... 27
Figure 2.7: Comparison of various inhibitors [16]. ..................................................................................... 28
Figure 3.1: Metal SIS prototype machine. .................................................................................................. 32
Figure 3.2: Sintering profile used in experiments. ...................................................................................... 33
Figure 3.3: (a) SolidWorks model of simple shapes part file (b) MeshLab view of the associated STL file (c)
Sliced image of the STL file. ........................................................................................................................ 34
Figure 3.4: Illustration of two intersection points (red) created by each triangle intersecting a given z
plane. .......................................................................................................................................................... 35
Figure 3.5: (a) Digital model of a washer (b) Sliced image of a layer cross-section. Lighter ring region
represents part, dark regions represent the filled (printed) areas, and perimeter of ring represents the
original part boundary (c) Final sliced image. ............................................................................................. 36
Figure 3.6: (a) Basic shape design with 480 µm borders and a 2 mm thick printed perimeter (b) Results
immediately after sintering (c) Front of final parts after separation from the inhibited region. Dimensions
in cm (d) Back of final parts. ....................................................................................................................... 38
Figure 3.7: (a) Basic shape design with 200 µm shape borders and parting lines (b) Results of the HSD
experiment immediately after sintering (c) Final part results – front. Dimensions in cm (d) Final part
results - angled view. .................................................................................................................................. 40
Figure 3.8: (a) Results of inhibitor removal as a result of abrasive blasting (b) Manual grid removal with
the aid of pliers angled view (c) Manual grid removal with the aid of pliers – front. ................................ 40
Figure 3.9: (a) Three-dimensional cut-away view of HSE design (b) Top and bottom layer of HSE design
(c) Results of the HSE experiment immediately after sintering (d) Final part results - angled view. ......... 42
Figure 3.10: (a) Gear designed in SolidWorks (b) Final part result using HSB printing scheme (c) Albert
Einstein bust design (d) Final part result with added base, using HSC printing scheme. ........................... 43
Figure 3.11: (a) Möbius strip design (b) Final part result using HSC printing scheme. ............................... 44
Figure 3.12: (a) Micro-tensile dog bone design (b) Cross-section of build in first orientation . ................. 44
Figure 3.13: Stress-strain curves for dog bone specimens tested in tension. ............................................ 45
Figure 4.1: (a) Highly loaded binder powder system (b) surface volatiles escape first causing planar
binder-vapor interface. ............................................................................................................................... 48
Figure 4.2: Example of typical SIS-metal binder loading. Periphery of part is loaded with binder while the
core is an open pore network. .................................................................................................................... 48
8
Figure 4.3: Heating rate diagram from Pinwill. Shaded region indicates heating rates that lead to various
failure modes [30]. ...................................................................................................................................... 49
Figure 4.4: Modified apparatus for explicit monitoring and feedback of binder weight loss [39]. ............ 50
Figure 4.5: Minimum cycle time as a function of x dimension with fixed y, z dimensions [42]. ................ 52
Figure 4.6: Thermal decomposition of sucrose as observed with TGA equipment. ................................... 54
Figure 4.7: Bronze powder particle size distribution. ................................................................................. 55
Figure 4.8: Illustration of neck diameter and spherical particle diameter in sintering. ............................. 57
Figure 4.9: Changes in mechanical properties as a sintered material densifies. ........................................ 59
Figure 4.10: Sintered strength model fit to Yoozbashizadeh bronze sintering results............................... 60
Figure 4.11: Differential hardness between sintered and unsintered regions of bronze coupons. ........... 68
Figure 5.1: (left) Preliminary droplet test with Sucrose-based inhibitor solution in bronze where
mechanical separation by abrasive blasting was easily achievable. (right) In contrast, sintering inhibition
in stainless steel did not occur in the droplet test. ..................................................................................... 71
Figure 5.2: Graphical representation of proposed research. ..................................................................... 73
Figure 6.1: High vacuum sintering furnace utilized in experiments. .......................................................... 82
Figure 6.2: Generic sintering profile used in experiments. ......................................................................... 83
Figure 6.3: (left) The vacuum impregnation apparatus and (right) the support apparatus in accordance
with ASTM B962. ......................................................................................................................................... 84
Figure 6.4: Hold temperature vs relative density of the stainless steel sintered coupons. ....................... 85
Figure 6.5: (left) Pipette, (middle) pre-heat furnace, and (right) abrasive blasting equipment. ................ 87
Figure 6.6: Photograph of inhibited specimens after abrasive blasting. .................................................... 88
Figure 6.7: (left) The experimental design and (right) implementation of the diluted droplet testing. .... 89
Figure 6.8: Photographs of the seven dilution experiments in order of temperature from left to right. .. 91
Figure 6.9: Visual representation of the results of the dilution experiments. ........................................... 92
Figure 6.10: JEOL JSM 6610 SEM located at the USC CEMMA center. ....................................................... 93
Figure 6.11: Photograph of the surface of droplet (3, 2) after sintering. ................................................... 93
Figure 6.12: SEM micrograph of the MicroMelt 316L material in its original, unsintered condition. ........ 94
Figure 6.13: Electron micrograph of an uninhibited portion of the sintered sample. ............................... 95
Figure 6.14: SEM micrograph of 100% saturated droplet located in position (1, 1). ................................. 96
Figure 6.15: Magnified micrograph of the center of the 100% saturated droplet. .................................... 97
Figure 6.16: (left) The magnesium EDAX spectral peak highlighted in blue , and (right) the oxygen EDAX
spectral peak for the same region highlighted in green. ............................................................................ 97
Figure 6.17: EDAX spectral analysis of 100% saturated droplet. ................................................................ 98
Figure 6.18: Magnified micrograph of the center of the 100% saturated droplet after a thin layer of
material was removed with a hand tool. .................................................................................................. 100
Figure 6.19: (left) The magnesium EDAX spectral peak highlighted in green , and (right) the oxygen EDAX
spectral peak for the same region highlighted in red. .............................................................................. 100
Figure 6.20: Alternative view of 100% saturated droplet region after a thin layer of material was
removed. ................................................................................................................................................... 101
Figure 6.21: An SEM micrograph of a droplet region rich in magnesium oxide. The brightness was dialed
down for better visualization. ................................................................................................................... 102
9
Figure 6.22: (left) The gear designed for part fabrication, (center) the layer image representing layers 1-
30, and (right) the 32nd and final closing layer. ....................................................................................... 103
Figure 6.23: (left) Green part after blowing with low pressure air. .......................................................... 104
Figure 6.24: Alternate view of finished part after abrasive blasting. ....................................................... 104
Figure 6.25: Tensile specimen design. ...................................................................................................... 106
Figure 6.26: Fabricated dog bone specimen mold.................................................................................... 106
Figure 6.27: (left) Instron 5567 Testing System and (right) a dog bone speciment being tested with
extensiometer attached. ........................................................................................................................... 108
Figure 6.28: Shrinkage results from the second sintering pass. ............................................................... 109
Figure 6.29: Calculated corresponding densities for second sintering pass. ............................................ 110
Figure 6.30: ASTM E8 tensile test results for a hold temperature of 1200
o
C. ......................................... 111
Figure 6.31: ASTM E8 tensile test results for a hold temperature of 1250
o
C. ......................................... 111
Figure 6.32: ASTM E8 tensile test results for a hold temperature of 1300
o
C. ......................................... 112
Figure 6.33: ASTM E8 tensile test results for all dog bone specimens. .................................................... 112
Figure 6.34: SIS-metal fabricated gear after second sintering pass (scale in cm). ................................... 114
Figure 6.35: A visual representation of the difference in size between the two sintering passes. .......... 114
Figure 7.1: Diluted droplet test of H13 alloy at 1100
o
C with no observable inhibition. .......................... 117
List of Tables

Table 2.1: Qualitative comparison of commercial AM technologies. ......................................................... 29
Table 3.1: Chemical composition of bronze as reported by manufacturer. ............................................... 33
Table 3.2: Dogbone design dimensions. ..................................................................................................... 45
Table 4.1: Chemical Composition of Stainless Steel 316L as reported by the manufacturer. .................... 63
Table 5.1: Common metal oxides and their respective melting points. ..................................................... 76
Table 5.2: Potential chemical precursors to desired metal oxides. ............................................................ 77
Table 6.1: The measured masses and calculated relative density per Equation 4.5. ................................. 84
Table 6.2: Inhibition results. ....................................................................................................................... 88
Table 6.3: Droplet position with associated solution concentration. ......................................................... 90
Table 6.4: Run order for dilution experiments. .......................................................................................... 90
Table 6.5: Relative atomic quantities associated with the spectral analysis. ............................................. 99
Table 6.6: Shrinkage result summary from second sintering pass. .......................................................... 109
Table 6.7: Summary of mechanical testing results. .................................................................................. 113
Table 6.8: Measured diameter of re-sintered gear in all directions. ........................................................ 115

10
1. Introduction
1.1 Introduction to Additive Manufacturing
Traditional Manufacturing techniques such as casting, molding, forming, machining, and joining in metal
part fabrication are well established. The material properties of the end products are well understood,
and they are capable of targeting highly demanding applications. However, machining process
geometric constraints often result in parts with more material and weight than is needed to support the
design loads. Machining processes can subtract up to 98% of the original billet to achieve the desired
form [1]. These processes are generally slow with multiple manufacturing steps necessary for the
production of complex geometries. In addition, the processing time required to convert a given
geometry to a machine's tool path is not trivial. In many cases, the part design is constrained by the
limitations of the tool path. For example, a cutting bit such as an end mill must be able to physically
touch the work piece without the machine colliding with the part. This problem can be sometimes
alleviated by building fixtures specific to the work piece or even re-positioning the work piece multiple
times during a build.
These limitations created a need for a new branch of manufacturing methods that could quickly turn
digital models into functional physical parts. Multiple metal additive manufacturing (AM) technologies
have emerged over the last few decades. The American Society for Testing and Materials (ASTM)
describes AM as [2]:
"A process of joining materials to make objects from 3D model data, usually layer upon layer, as
opposed to subtractive manufacturing methodologies."
AM technologies are used to produce parts for various applications, including functional models,
fit/assembly, patterns for prototype tooling, and small to medium volume manufacturing [3]. Although
11
much research is needed to improve AM performance, there are numerous instances of the
implementation of AM technologies for high demand applications [4-6].
1.2 Advantages and Disadvantages of AM Processes
AM processes offer advantages over traditional manufacturing methods in a number of ways:
1. Greater geometric freedom,
2. Little to no tooling required,
3. Manufacturing lead times drastically reduced,
4. Costs largely independent of geometric complexity,
5. Minimal waste,
6. Low relative labor.
There are numerous AM technologies capable of fabricating metal and metal alloy parts. The various
commercially available AM processes are based on one of three core technologies: laser, electron beam,
or inkjet. While some of these technologies are high in performance with a wide variety of materials to
choose from, they also come at a high cost. Typical laser- and electron-beam based systems start from
$500k and rise to over $1M [7]. Although AM processes show promise for the production of end-use
metallic parts, there are currently many areas for improvement. The current disadvantages of metal AM
processes include:
1. Mechanical performance,
2. Poorly understood material properties under static load, fatigue, and damage tolerance,
3. Anisotropic material properties,
4. Dimensional accuracy, process speed, and surface finish,
5. High up-front machine- and material costs.
12
1.3 Selective Inhibition Sintering - Metal
Traditional research in powder sintering has mainly focused on enhancing and speeding the sintering
process. Similarly, the metal AM processes mentioned above selectively sinter or fuse particles within
each layer’s part cross-section. In contrast, the Selective Inhibition Sintering - Metal (SIS-metal) process,
originally ideated by Dr. Behrokh Khoshnevis, is based upon the retardation of sintering [8-10]. The core
concept behind the SIS process is the prevention of selected regions of each powder layer from
sintering; this is accomplished by treating the regions external to the part in each layer with a sintering
inhibitor. In this sense, the SIS process can be considered an inverse to traditional metal AM processes
[11].
In the current metal process, a commercial piezoelectric print head is utilized to deposit a liquid
chemical solution at the periphery of the part for each layer. Once all of the layers have been
completed, the entire part is removed from the machine and bulk sintered in a conventional sintering
furnace. The inhibitor deposited at the part's boundary decomposes into hard particles that retard the
sintering process. The particles in this region are prevented from fusing, allowing for removal of
inhibited boundary sections and revealing of the completed part. Figure 1.1 is an illustration of the SIS-
metal process.
13

Figure 1.1: The current SIS-metal process.
1.4 Advantages/Disadvantages of SIS-metal
The potential advantages of the SIS-metal process over existing metal AM processes are listed below:
 Cost Effective: When compared to using electron or laser beam generators, the use of an inkjet
print head represents orders of magnitude reduction in machine cost.
 High Speed: The use of an inkjet print head is proven to be much faster than the use of scanning
beams. SIS-metal proves to be faster as it only treats the periphery of parts as opposed to the
entire cross-section.
 Sacrificial Support Structures: The support structures in the SIS-metal process are constructed
of inhibited regions and thus are easily removed with the rest of the inhibited regions.
 Unadulterated Alloy: Since inhibitor solution is deposited only in regions of the powder bed that
are not included in the final part, the final part remains a pure metal alloy.
14
 Enhanced Sinterability: Since the final part is a pure metal alloy, there is no organic residue left
within the part. Thus, a higher sintered density is theoretically achievable.
 Shorter Burnout Schedule: Since only the periphery is treated with inhibitor solution, there is
ultimately less organic material to burn out prior to reaching the sintering temperatures.
Potential disadvantages of the SIS-metal process include:
 Removal of Inhibited Regions: Removal of inhibited regions necessitates an additional post-
processing step. Abrasive blasting can be cumbersome in the case of complex part geometries
such as internal passages.
 Finished Part Strength in a Single Sintering Cycle: There is a limit to the degree of sintering SIS-
metal parts can achieve in one sintering pass because it is possible to over-sinter inhibited
regions so they are no longer separable. Thus, a second sintering cycle must be employed to
achieve desired mechanical properties.
1.5 SIS-metal Challenges
The work of Khoshnevis and Yoozbashizadeh paved the way to new research areas in the SIS-metal
process with the first prototype SIS-metal machine. Previous research provided a benchmark for future
work in SIS-metal which led to a second prototype machine based on piezoelectric print head
technology. The second prototype machine was built and parts were fabricated, achieving orders of
magnitude better resolution as reported by Torabi et al. [12].
With these achievements the authors were able to identify six potential areas of improvement
necessary for the advancement of the SIS-metal process (Figure 1.2). The six research topics listed below
are common concerns in all AM processes.
For the purposes of this study, research topics 5-6 will be addressed.
15
1. Poor relative dimensional tolerances are of concern in the SIS-metal process. The proposed
process uses a bulk, loose powder sintering technique which results in greater shrinkages in
printed parts. Careful research into the factors that affect dimensional tolerances is needed
2. SIS-metal parts have been printed with internal cavities of a simple nature. However, printing a
part with complex internal passageways presents a problem for the SIS-metal process as there
may not be direct access to remove inhibited regions.
3. Surface Quality has seen drastic improvement with the implementation of a piezoelectric print
head. However, the surface quality of SIS-metal parts needs additional improvement.
4. Improper organic material burnout SIS-metal parts has led to defects such as, distortion,
cracks, and expanded pores (bloating). There is a need to focus on the burnout stage of the
sintering profile in order to minimize or eliminate such defects.
5. SIS-metal parts have shown a low relative density when compared to parts fabricated using
conventional methods. Thus, the mechanical properties of printed parts could be improved
through a focus on the sintering cycle.
6. The applications of the SIS-metal process are limited by the materials available. New materials
must be qualified in order to expand the potential applications of the technology.
16

Figure 1.2: Fishbone diagram identifying potential areas for SIS-metal improvement.
Preliminary results in a bronze material reported a relative density of 65.2-66.7% by volume when
measured using calipers and a standard scientific scale. A typically high sintered density for pre-alloyed
bronze materials is in the range of 85-90% [13]. Three non-standard dog bones were fabricated in one
sintering pass at low temperature and hold time. Preliminary micro tensile tests resulted in an average
ultimate tensile strength of 28.4 MPa with a standard deviation of 0.6 MPa. As a reference, ASTM B823
requires a minimum yield strength of 10 ksi [68.9 MPa] for the bronze material used in this research
[14].These parts are typically fabricated using conventional press and sinter powder metallurgy
processes.
While the improvement of material properties is desirable for the SIS-metal process, the applications of
a bronze base material are limited. Thus, it is highly desirable to qualify ferrous alloys for the process.
17
The 316 stainless steel alloy is one of the most commonly used metals in the world. Thus, the use of this
material with the SIS-metal process would broaden the potential applications by orders of magnitude.
1.6 Purpose of the Study
The ultimate goal of AM technologies is to either replace or work alongside traditional manufacturing
processes in order to fabricate industrial grade metallic parts. Industrial grade is defined here as a way
of describing metal parts capable of meeting the demands of a particular industrial application. In
general, AM processes have begun to be adopted by many applications across the automotive,
biomedical, aerospace, and general manufacturing industries. However, due to the disadvantages of
current metal AM technologies, there is an unmet need for a low-cost, high-performance AM solution.
The overall objective of SIS-metal research is to improve the process to potentially meet the demands of
industrial applications. As such, the process must gain the ability to print in advanced materials with
acceptable strength.
The purpose of the proposed research is to investigate the possibility and determine a method of
printing by SIS a net-shape part with maximized density, and hence mechanical properties from a 316
stainless steel material.
1.7 Goal of the Study
Building upon the previous work of Khoshnevis et al [15] and Yoozbashizadeh [16], the goal of the
proposed research is to further advance the SIS-metal process through an investigation of new
materials. The preliminary results presented focus on viable hatching schemes in the fabrication of
complex three-dimensional metallic parts. These hatching schemes are directly applicable to new
materials. The resulting parts were mechanically tested and an analysis was presented for a bronze alloy
material.
18
The proposed research will focus on using the SIS-metal process to fabricate 316 stainless steel parts.
This will be accomplished in three distinct stages.
1. The first stage of the effort will focus on candidate precursors to identify a new inhibitor.
Experiments will be conducted to determine the efficacy of a chosen inhibitor through
investigation of the sintering cycle and a characterization of the degree of inhibition.
2. In the second stage of the effort, the selected inhibitor solution will be diluted to investigate the
effects of inhibitor concentration on the degree of inhibition. The results will be put to practical
use in the fabrication of 2.5D extruded shapes using the SIS-metal process.
3. In stage three, the focus will be shifted to the characterization and improvement of part
strength through further investigation of the sintering cycle.
1.8 Research Contribution
This research is the first extensive research conducted in an effort to identify and validate a sintering
inhibitor for use with ferrous alloys in a metal additive manufacturing process. The methodology
presented is extensible in identifying inhibitor candidates for other metals and metal alloys in future
research.
1.9 Organization of the Dissertation
Chapter 1: This chapter includes a brief introduction to additive manufacturing and the SIS-metal
process, including the respective challenges.
Chapter 2: This chapter provides a more thorough review of additive manufacturing processes as well as
all previous research conducted in the field of SIS.
19
Chapter 3: This chapter presents the preliminary research conducted in SIS-metal associated with
improving upon the build strategies and modifying the necessary steps in the SIS-metal fabrication
process.
Chapter 4: This chapter reviews all fundamentals associated with powder metallurgy, sintering and the
inhibition of sintering.
Chapter 5: This chapter presents the analytical and experimental methodology used in this research.
Chapter 6: This chapter presents the experimental results.
Chapter 7: This chapter presents the research conclusion and identifies potential future work.
20
2. Background
2.1 A Review of Metal AM Processes
In general, metal AM processes work by adding material in thin layers from the bottom up, with each
layer representing a cross-section of the part [17]. There are various competing technologies in the AM
of metal parts such as Direct Metal Laser Sintering (DMLS), Selective Laser Sintering (SLS), Laser
Engineered Net Shaping (LENS), Electron Beam Melting (EBM) and Selective Laser Melting (SLM) [18].
There are four common steps to all of these AM technologies:
1) A digital model of a part is constructed using one of the various CAD packages (SolidWorks,
Alibre, Google SketchUP, etc.)
2) The model is imported into a slicing software which generates the profile and/or tool path
instructions for each layer of the part
3) A file containing the profile and tool path instructions is transferred to the AM machine, which
subsequently builds the part one layer at a time
4) The part is removed and post-processed to achieve desired aesthetic and/or mechanical
properties
Step 3 mechanically differentiates various the AM technologies. For example, DMLS is a powder AM
technology that utilizes high power CO 2 laser energy to selectively bond metal powder particles together
layer by layer. This process is an adaptation of the Selective Laser Sintering process developed by Carl
Deckard in 1989 [19]. The beam diameter and layer thickness of the EOS M250, a DMLS machine, was
reported to be 400 and 50 µm respectively. The build is typically performed in a nitrogen atmosphere
with a scan rate in the range of 100 – 275 mm/s. Sintering density, as in all powder metal processes is
strongly affected by the particle shape, size, distribution and chemical constituents of the powder
21
system. The DMLS process can produce parts near full density in materials such as steel, stainless steel,
titanium and various super alloys [20, 21]. A typical DMLS setup can be seen in Figure 2.1 [22].

Figure 2.1: Typical DMLS setup.
The setup and specifications of the EBM process is similar to laser processes except that an electron
beam is used to transfer energy to selected powder particles layer by layer. In addition, while laser
sintering takes place under a protective atmosphere, EBM takes place in a partial vacuum. In general,
EBM is able to deliver higher energies to the powder bed than typical laser processes. The EBM process
was developed at Chalmers, University of Technology, in Sweden and was commercialized by Arcam AB
[17].
In addition to beam based processes, there is one metal process that is based upon inkjet technology.
The Prometal process is an implementation of the binder jetting process developed by Sachs et al in
1992 [23]. In this process, also known as 3DP, an inkjet print head is used to selectively deposit a liquid
binder in the part cross section of each layer. Layer thicknesses are typically 100 to 150 µm . Once the
22
entire part has been built layer by layer, it is put into a sintering furnace at 350
o
F for 24 hours to harden
the binder/metal alloy system. Typical steel green density for this process is about 60% and is infused
with a bronze filler material to bring the part to full density. The final part is typically 60% stainless steel
and 40% bronze [18]. A typical 3DP setup can be seen in Figure 2.2.

Figure 2.2: Typical 3DP setup.
2.2 A Review of Selective Inhibition Sintering Research
In 2003, Behrokh Khoshnevis patented a process called Selective Inhibition Sintering, a method for
fabricating three-dimensional objects from a binderless powder [24]. Khoshnevis et al fabricated an SIS
machine capable of building polymeric parts (SIS-polymer) [8]. The machine utilized a single nozzle
solenoid valve in a three-axis machine. The SIS-polymer process steps can be seen below and are
illustrated in Figure 2.3.
Step 1: Spread thin powder layer
Step 2: Deposition of sintering liquid inhibitor via single printing nozzle
23
Step 3: Minimize radiation frame to reduce powder waste by masking powder outside of the
part area to prevent it from sintering
Step 4: Sinter layer by thermal radiation source such as a nichrome filament or
tubular halogen bulb.
 Repeat Steps 1-4 until entire part is complete

Figure 2.3: The SIS-polymer process [25].
Three polymers (polystyrene, polycarbonate, and Duraform®) used in SLS were evaluated in the SIS-
polymer development process.
The approximate sintering temperature and time for sintered layers of varying thicknesses were
determined experimentally. Several liquid inhibitor candidates were tested. Various concentrations of a
salt water and hydrogen peroxide mixture exhibited superior performance. These liquids were then
tested for compatibility with thermal and piezo inkjet printers. Proper inhibitor choices resulted in
sharp, well-defined edges.
Inhibition is a key factor in SIS in the ease of separation of parts with fine geometric details. Four
theories were developed to explain the inhibition process:
24
1. Macroscopic mechanical inhibition: powder particles at the point of droplet impact are
displaced to the point that particles are no longer close enough to adhere to each other during
the sintering process.
2. Microscopic mechanical inhibition: liquid inhibitor penetrates the powder layer filling voids and
gaps between particles. The inhibitor coats the particles and obstructs adhesion in the sintering
process.
3. Thermal inhibition: liquid inhibitor penetrated the powder layer filling voids and gaps and acts
as a heat absorber and cooling agent during the sintering process to prevent the inhibited
sections from reaching the proper sintering temperature.
4. Chemical inhibition: liquid inhibitor penetrates the powder layer filling voids and gaps between
particles. The inhibitor reacts with powder particles at their surface creating a new species
resistant to sintering.
Asiabanpour et al designed a set of experiments to determine the factors that most greatly affected the
mechanical strength of SIS-plastic parts [26]. Experiments were performed using saturated
potassium iodide solution as the inhibitor and polystyrene polymer powder for the base
material. Results showed that factors related to the heating and layer thickness most
significantly affected part strength.
Later in 2003, Asiabanpour et al again examined the SIS-polymer process [9]. This time, the study aimed
to use statistical analysis techniques in studying the effects of various factors on the surface quality and
dimensional accuracy of parts. Surface quality response was determined using the Shellabear method
where fabricated parts were compared to sample parts with varying surface qualities.
25
In 2004, Asiabanpour extended the SIS-polymer research to a wider array of polymer materials and
concluded that polystyrene served as the best choice of base material among candidates [10]. However,
it was observed that many other powders could potentially be used in the process.
Later in 2004, Asiabanpour and Khoshnevis developed a machine path generator for the SIS-polymer
process [27]. The machine path algorithm generated appropriate boundary and hatch paths. The
algorithm was successfully tested on various STL models including a model that was over 200 MB in size.
The generated paths were used to successfully fabricate parts in the research.
In 2005, Mojdeh under the guidance of Khoshnevis modified the existing SIS-polymer process for the
application of SIS to metal alloys for the first time [25]. The high temperatures and controlled
atmosphere required to sinter metallic powders dictated the use of bulk sintering instead of the
previous layer-by-layer sintering approach. Figure 2.4 illustrates the modified process for metals in
which compaction was implemented to achieve high densities.
1. An extrusion nozzle with fine orifice is used to deliver a ceramic sintering inhibitor to a selected
area. The inhibitor forms a wall between area inside and outside the layer profile
2. Powder is uniformly spread
3. An adhesive liquid is deposited outside of the part profile to act as a container for the part.
4. The layer is compressed
 Repeat Steps 1-4 until entire part is complete
5. After all layers are completed, the part is bulk sintered in a conventional sintering furnace
26

Figure 2.4: The first implementation of an SIS-metal process [25].
In 2012, SIS-metal research explored the potential for using a liquid delivery system to deposit inhibitor
into a metal powder bed. Khoshnevis et al utilized a single nozzle solenoid valve for deposition. The
process used a combination of raster and vector machine path strategies to accomplish builds. The
process steps are listed below. Figure 2.5 illustrates the resulting part builds.
In this implementation of the SIS-metal process, five steps were involved in the production of a metallic
part[11]:
1. A single layer of metal powder is spread over the build tank
2. A single nozzle solenoid valve deposits a liquid solution exterior to the periphery of the part
cross section
3. The layer is then heated to evaporate the liquid in the deposited regions
 Steps 1-3 are repeated for each layer, until the entire part is completed
4. The completed part is removed from the machine and bulk sintered in a conventional sintering
furnace. During this step, the deposited inhibitor decomposes into hard particles that retard the
sintering process, and thus prevent the deposited regions from fusing.
27
5. The part is removed from the furnace. The inhibited regions are subsequently removed via
abrasive blasting to reveal the completed part.

Figure 2.5: (a) SIS-metal result from Khoshnevis et al (b) Inhibited negative region [11].

In 2013, Yoozbashizadeh further investigated the fundamentals of SIS-metal based on the microscopic
mechanical inhibition principle. The inhibitors tested were ceramic salt solutions as well as a
carbohydrate solution with a bronze powder base material. It was determined that deposited inhibitor
material decomposed (Figure 2.6) into ceramic (in the case of salt) or carbon (in the case of
carbohydrate) that covered the surface of affected metal powder particles. This covering of particles
resulted in a retarding force on the surface transport mechanisms of the sintering process.

Figure 2.6: (a) SEM image of untreated bronze powder after sintering (b) SEM image of inhibited
bronze powder with inhibitor material covering the surface of bronze particles [16].
28
The degree of potency for various inhibitor choices was characterized by way of harness testing. Bronze
powder samples were first treated with individual inhibitors and then over-sintered. Hardness tests
were conducted on each sample and the experiment that resulted in the lowest hardness (sucrose) was
deemed the most suitable inhibitor (Figure 2.7).

Figure 2.7: Comparison of various inhibitors [16].
In addition, the research characterized the degree of sintering for the bronze alloy used in a pressure-
less sintering procedure. Various combinations of sintering temperatures and hold times were employed
under a vacuum environment. Bronze coupon specimens were successfully sintered to 90% relative
density and accompanied by a 7.9% shrinkage.
2.3 Technology Overview Comparison of AM Processes
As listed in Section 3.1, there are many competitive approaches in the AM of metal and metal alloy
parts. Investment into AM processes by both the academic and private sectors have greatly enhanced
the understanding as well as the capabilities of these processes. The level of performance of some AM
29
technologies, with the aid of post-processing efforts is slowly approaching that of conventional
machining techniques. A summary of the pros and cons of some commercialized technologies can be
seen in Table 2.1.
Table 2.1: Qualitative comparison of commercial AM technologies.

SIS-metal potentially offers the ability to fabricate complex metallic parts with good mechanical and
dimensional performance. In addition SIS-metal should be able to produce parts with highly complex
geometries in a wide range of materials.
2.4 Critique of Past Approaches
Previous SIS-metal research served to lay the groundwork for the process. Khoshnevis et al were
successful in fabricating copper alloy-based parts with two separate implementations of SIS-metal. In
the first implementation, the inhibitor material was extruded through a fine orifice nozzle. In the
second, the inhibitor was deposited via a carrier fluid through a pressure backed solenoid valve. The
following is a list of observations of key areas of the SIS-metal process that could be further improved
upon.
 Previous SIS research measured shrinkage and strengthening of metal alloy coupons outside of
the build process. Careful research should be conducted in measuring the mechanical
performance of “as-built” parts while simultaneously characterizing whether inhibited regions
are separable.
30
 Previous research measured shrinkage in one direction. The nature of SIS-metal material
isotropy in an “as-built” condition should be considered and dimensional tolerances should be
quantified.
 The effects of time and temperature factors on the degree of sintering of metal coupons were
well characterized by Yoozbashizadeh [16]. This study could potentially be extended by varying
these factors to determine whether inhibition effects are greater under specific factor settings
for a given degree of sintering.
 Previous research in SIS-metal investigated relatively bulky parts. High resolution metallic parts
with features in the sub-millimeter scale have yet to be fabricated.
 Research should be conducted to treat only the perimeter of parts. In previous work, there was
a great deal of waste in printing the entire negative part region.
 Previous research characterized the sintering profile and resulting mechanical properties for
"inhibited" and "uninhibited" coupons. However, research has yet to characterize a hold time
and/or temperature limit for successful inhibition.
 Previous research focused on a low temperature bronze alloy. Research should be conducted on
higher temperature, ferrous alloys.

31
3. Influence of Build Strategies
3.1 Abstract
Selective Inhibition Sintering of metal alloys (SIS-metal) was previously proven effective in the additive
manufacture (AM) of low resolution bronze parts. Recent advancements in the use of a high-precision
inkjet print head represented an order of magnitude improvement in SIS-metal resolution [12].
However, the fabrication of complex three-dimensional metallic parts required new SIS-metal
compatible, cross-sectional image generation based on the part boundary profile. As a preliminary
study, three candidate layer-processing approaches were identified and validated for basic geometries.
These approaches were chosen from previous research as well as preliminary investigations and were
applied to a modified SIS-metal process for validation. The validation criteria were based upon the
amount of powder waste produced, the ability to handle complex geometries, printing speed, extraction
(post-processing) speed, and part integrity. Results are discussed for implementation of the five
candidate layer processing approaches in the fabrication of basic shapes. A preliminary evaluation is
presented for their use on more complex geometries. Two approaches were then chosen for the
construction of more complex geometries, the results of which are presented.
3.2 Experimental Setup and Procedures
3.2.1 Hardware
The metal SIS beta prototype machine (Figure 3.1) was developed at the Center for Rapid Automated
Fabrication Technologies (CRAFT) and was utilized for the purposes of this research. The system consists
of a three-axis motion system which uses a commercial piezoelectric Epson Workforce 30 print head [28,
29].
32

Figure 3.1: Metal SIS prototype machine.

The inhibitor solution is deposited onto the powder bed by activating two color channels of the print head
simultaneously. These two channels collectively account for one hundred and twenty nozzles. Two colors
were used in order to maximize the amount of fluid deposited per pass of the print head. The inhibitor
chosen for this research was a solution of sucrose, water, and an industrial surfactant [12]. The sucrose is
believed to decompose into high temperature carbon particles that act as a sintering obstacle in affected
areas. Building upon the research of Khoshnevis et al. [11], the metal powder used was a fully alloyed
bronze with chemical composition shown in Error! Reference source not found.. The powder particle size
as a distribution of 98.7% -325mesh, 1.3% -200/+325mesh as reported by the manufacturer. The
manufacturer reported values for yield strength and theoretical density were 83 MPa and 8.77 g/cm
3
,
respectively.

33

Table 3.1: Chemical composition of bronze as reported by manufacturer.

Upon completion of inhibitor deposition, the green part was sintered in a programmable, high
temperature Neytech Centurion Qex porcelain furnace. Sintering took place in a vacuum environment.
The sintering profile based on previous research can be seen in Figure 3.2 [11]. A ramp of 25
o
C/min was
used up until the pre-heat temperature of 115
o
C. The part was held for three hours to ensure the sample
was dehydrated. The temperature was then ramped again at a rate of 25
o
C/min until the sintering
temperature of 780
o
C. Finally, the part was underwent passive cooling.

Figure 3.2: Sintering profile used in experiments.
After sintering, the part was ready for post-processing. Separation or removal of inhibited regions was
accomplished using abrasive blasting equipment. Sodium bicarbonate (45-90 µm nominal) and spherical
0
100
200
300
400
500
600
700
800
900
0 100 200 300 400 500
Temperature [
o
C]
Time [min]
Bronze Sintering Profile
34
glass bead (200 mesh) media were used at pressures between 30 – 60 psi for the removal of inhibited
regions.
3.2.2 Software
A simple Graphical User Interface (GUI) was developed in Visual C# for ease of implementation of
various hatching schemes. Parts were first designed and saved into STL format in SolidWorks CAD
software. MeshLab and Netfabb Basic, open-source software, were then used to view and manipulate
the STL files when necessary (Figure 3.3ab). Based on the aforementioned research of Asiabanpour et al.
and a literature review of other AM slicing approaches, the layered images (Figure 3.3c) were
constructed using a predetermined uniform layer thickness for a given binary STL file format. The layer
thickness was determined by choosing a thickness slightly greater than the minimum spreadable layer
thickness for the bronze powder used, 120 µm. Spreading thinner layers in with the current powder
system increases the risk of a failed spread, and ultimately a failed part build. It is conceivable, however,
that an adaptive slicing scheme could be implemented to speed the SIS-metal process. For example,
thicker layers could potentially be used simply by depositing more fluid per layer.

Figure 3.3: (a) SolidWorks model of simple shapes part file (b) MeshLab view of the associated STL file
(c) Sliced image of the STL file.
The cross-sectional profiles were obtained from the intersection of the object and regularly spaced planar
faces perpendicular to the z direction. The STL model triangles were first sorted by their respective vertex
(a) (b) (c)
35
z values in ascending order. For each layer, two intersections were calculated per triangle to define the
cross-sectional profile at a given z height (Figure 3.4).

Figure 3.4: Illustration of two intersection points (red) created by each triangle intersecting a given z
plane.
Within a given cross section, a series of connected intersecting points form loops that represent part or
hole boundaries. Loops in a given cross section were differentiated by a simple algorithm. Each triangle
for a given layer produces a pair of intersections, [x 1, y 1] and [x 2, y 2], which were stored in an irregular
sequence. The pairs were first sorted by checking against other pairs in the layer. The sorting process
started from the first pair of intersections in the array and searched for an adjacent pair. Once the
adjacent pair was found, the next adjacent pair was searched for. This was continued until the loop
closed upon itself. If there were points unaccounted for in the same layer, the process would begin
again to identify the next loop until all points were accounted for.
Each loop was then ranked using the ray casting method to distinguish between hole and part
boundaries; this process determines which side of the boundary will be marked for inhibitor deposition.
For example, if the loop represents a part boundary, the area external to the loop will be marked for
deposition. If the loop represents the boundary of a hole, the area internal to the loop will be marked
for deposition (Figure 3.5).
36

Figure 3.5: (a) Digital model of a washer (b) Sliced image of a layer cross-section. Lighter ring region
represents part, dark regions represent the filled (printed) areas, and perimeter of ring represents the
original part boundary (c) Final sliced image.
An even rank (0, 2, 4…) signifies part boundaries and an odd rank signifies hole boundaries. These
calculations resulted in the rendering of images representing the layered cross sections of a part (Figure
3.5c). The shaded regions in Figure 3.5c represent areas marked for deposition in the first printing
scheme implemented. Finally, a print command was sent to the Epson Workforce 30 to begin inhibitor
deposition.
3.2.3 Experimental Research Objectives
Improvements in the areas of resolution, part complexity, part accuracy, powder waste, and processing
time are necessary in order to qualify the modified SIS-metal process. Thus, the first objective was to build
high resolution, complex parts with the SIS-metal beta machine.
In order to build these parts, it was necessary to create a method for slicing STL files as well as to
determine a best practice for printing with respect to hatching schemes. Thus, multiple hatching schemes
were investigated, three of which are detailed in Section 3.3 Experimental Plan.
In addition, it was necessary to take a look at the possibility of in-furnace organic material burnout. It is
known from literature that organic material burnout is accomplished within the furnace cycle for a
(a) (b) (c)
37
multitude of materials and applications [30, 31]. Thus, the SIS-metal process was modified by eliminating
Step 3, the interlayer heating. Removal of excess liquid and organic material burnout would be
accomplished in the furnace for the purposes of this research. However, the consequences of the
modification of the process necessitated careful evaluation as preliminary parts exhibited overt warpage
and cracking.
Finally, it is of great interest to understand the mechanical properties of as-built SIS-metal parts. As such,
tensile tests were performed using a DEBEN MICROTEST 5kN micro-tensile apparatus. Tensile tests are
considered an industry standard for the examination of powder metallurgy parts [32].
It is believed that the results of the preliminary work are extensible to other materials. Thus, these
preliminary objectives will lay the ground work for the adaptation of the SIS-metal process to more
advanced materials such as ferrous-based alloys.
3.3 Experimental Plan
Three hatching schemes implemented are described below:
 Hatching Scheme A (HSA): HSA utilizes a scheme in which a thick border is printed at the
periphery of each part shape. In addition, a thick square boundary is printed to encompass all of
the parts.
 Hatching Scheme B (HSB): HSB represents a hatching scheme in which a thin border is printed at
the perimeter of the part; however, separation lines are also printed to aid in part removal.
 Hatching Scheme C (HSC): HSC represents a scheme in which a thin border is printed at the
perimeter of the part with no separation lines. Instead, the entire part is encompassed (closed)
by inhibited region.
38
3.3.1 Hatching Scheme A (HSA)
HSA called for a reduction in the area of inhibited region when compared to previous research due to
concern over the formation of cracks from the burnout stage. Thus, instead of printing the entire part
negative region as was done in previous research, a thinner boundary profile was designed (Figure 3.6a).
In HSA, an inhibited base was printed in the initial layer. To be clear, the entire first layer was printed
beneath the part. The experiment consisted of an extruded square, circle, pentagon, and triangle with
500 µm boundaries surrounding each shape. A 2 mm thick square perimeter was printed to provide a
protective bounding box for ease of transport of the green part. The same image was repeated over ten
layers to obtain the desired shape extrusions. Figure 3.6b illustrates the results immediately after
sintering, and Figure 3.6cd illustrates the results after removal of the inhibited regions.

Figure 3.6: (a) Basic shape design with 480 µm borders and a 2 mm thick printed perimeter (b) Results
immediately after sintering (c) Front of final parts after separation from the inhibited region.
Dimensions in cm (d) Back of final parts.
Upon removal from the furnace, the results were mixed. The triangle, circle, and pentagon shapes were
extracted. However, the square shape exhibited visible internal cracking. Cracks were observed at the
corners in and around the printed bounding box (Figure 3.6b). Slight differential shrinkage was observed
at the corners of the bounding box and the triangular extrusion, where the corners visibly pulled away
from adjacent edges.
(a) (b) (c) (d)
39
In addition, it seems that improper organic material burnout led to an eruption surrounding the large
border. The extracted basic shapes exhibited signs of excess fluid deposition as well. This was evidenced
during the inhibitor removal. While removing inhibited regions (via abrasive blasting) from the
underside of the parts, material from within the part itself was easily removed. This signifies that the
liquid inhibitor encroached into each extruded shape.
3.3.2 Hatching Scheme B (HSB)
It was decided to first take an approach of printing less fluid via thinner part boundaries. In addition, as
extruded shapes become more complex, it becomes difficult to access and remove inhibited regions. In
the case that every layered image is the same, access to inhibited regions begins to decline as the part
grows in height. For example, if the extruded shapes were 50 layers (6mm) in height, the deepest
inhibited regions are 6mm below the surface of the part. Thus, it was suggested that parting lines be
printed to aid in the removal of inhibited regions. In this way, the part could potentially be extracted by
manually breaking the part out at the parting lines.
HSB represents a hatching scheme with the addition of a disconnected grid of parting lines (Figure 3.7a).
The experiment consisted of an extruded square, circle, pentagon, and triangle with 200 µm boundaries
surrounding each shape. The parting lines were designed with a 200 µm thickness. In HSB, the all of the
layers are the same including the base layer. Figure 3.7b illustrates the results of the HSB experiment
immediately after sintering. Figure 3.7cd represent the final extracted part results after a final pass of
abrasive blasting.
40

Figure 3.7: (a) Basic shape design with 200 µm shape borders and parting lines (b) Results of the HSD
experiment immediately after sintering (c) Final part results – front. Dimensions in cm (d) Final part
results - angled view.
Upon removal from the furnace, the results of HSB showed there was no visible evidence of cracking or
deformation. However, the extraction process after sintering significantly increased the total processing
time in HSB. The part was abrasive blasted as in the previous hatching schemes, but the blaster was
unable to remove material from deep within the part (Figure 3.8a). Inhibited regions could be removed
up to a depth of between 1-2 mm. Subsequent to abrasive blasting, the remainder of the square grid
was separated manually with the use of pliers. Figure 3.8bc represent the part in the process of
extraction during manual grid removal.

Figure 3.8: (a) Results of inhibitor removal as a result of abrasive blasting (b) Manual grid removal
with the aid of pliers angled view (c) Manual grid removal with the aid of pliers – front.
(a) (b) (c) (d)
(a) (b) (c)
41
3.3.3 Hatching Scheme C (HSC)
As part complexity evolved from extruded shapes into truly three-dimensional shapes, a new hatching
scheme and build method was desirable to promote part extraction. For example, in the case of delicate
thin features, the use of pliers to manually remove grid sections may break the features. In addition, as
features become thinner and more abundant, the use of additional parting lines may no longer be
practical. In the case of an abundance of small features, the amount of inhibited section increases as the
number of small features increases until the scheme is no longer distinguishable from earlier hatching
schemes in previous research [16].
An additional motivation for an alternative build method is the conservation of base material. Sintering
the entire negative region along with the part creates a great deal of wasted base material that can no
longer be recycled for subsequent builds.
Thus, HSC represents a scheme in which the designed part is entirely enclosed by inhibited region.
Figure 3.9a illustrates a three-dimensional view of the extruded square shape design. The lightly shaded
transparent section in the figure are areas in which inhibitor solution were deposited and not yet
removed, while the inner brown region represents the final part. The first and last sliced images are a
filled triangle, circle, square, and pentagon respectively (Figure 3.9b). All other sliced images are similar
to HSA (Figure 3.7a); however, the 2mm thick border has been removed.
In implementing this hatching scheme, two additional steps were added to the SIS-metal process. After
removal from the machine, the green part was heated for fifteen minutes at 100
o
C in order to improve
the green strength of the organic material. The part was then removed from the furnace and placed into
a powder recycling station where all of the excess powder in the part negative region was blown away.
Figure 3.9c illustrates the results of the HSC experiment immediately after sintering. Figure 3.9d
illustrates the final part results.
42

Figure 3.9: (a) Three-dimensional cut-away view of HSE design (b) Top and bottom layer of HSE design
(c) Results of the HSE experiment immediately after sintering (d) Final part results - angled view.
Upon removal from the furnace, there was a slight discoloration in the inhibited regions of all parts as
can be seen in Figure 3.8c. This may have been due to a change in inhibitor solution used for this
experiment as a second batch was needed to conduct these experiments. The parts were abrasive
blasted from various angles around each part to remove the inhibited regions. Part extraction proved
much easier in this regard due to additional access to inhibited regions as well as the lack of need for
additional tools. The integrity of all part shapes was acceptable. However, the quality of top surfaces
had degraded slightly due to the addition of an inhibited layer above the top surface of the part.
3.4 Application to Complex Geometries
A gear, a human bust and a crescent wrench were designed to implement two of the printing schemes,
HSB and HSC. The gear was designed in SolidWorks using the design library toolbox and can be seen in
Figure 3.10a. The outer diameter of the gear was designed to be 48.8mm with a face width (height) of
6mm. The result of building the part using HSB can be seen in Figure 3.10b.
The bust of Albert Einstein was designed to be printed with the HSC hatching scheme. The original STL
file for the bust was downloaded from the MakerBot Thingiverse website [33]. The file was healed,
modified, and scaled down using Netfabb (Figure 3.10c). The total height of the bust was designed to be
(a) (b) (c) (d)
43
23mm, which represented the largest height that could be accommodated in the laboratory furnace
without interfering with the thermocouple. The resulting part can be seen in Figure 3.10d.

Figure 3.10: (a) Gear designed in SolidWorks (b) Final part result using HSB printing scheme (c) Albert
Einstein bust design (d) Final part result with added base, using HSC printing scheme.
Finally, a Möbius strip was designed to be printed with HSB. The Möbius strip was downloaded from the
GRABCAD community [34]. It was healed, modified and scaled down in Netfabb (Figure 3.11a). The total
height of the part was designed to be 15 mm and the outside diameter at the widest point was 75 mm.
The final result can be seen in Figure 3.11b.
(a) (b)
(c) (d)
44

Figure 3.11: (a) Möbius strip design (b) Final part result using HSC printing scheme.
3.5 Mechanical Testing and Relative Density
Three dog bone specimens were printed using HSA in a single batch. The design of the dog bone
experiments can be seen in Figure 3.12a, and a cross-sectional image of the print can be seen in Figure
3.12b. The part dimensions, as designed, can be seen in .
, and the associated stress-strain curves for the specimens can be seen in Figure 3.13.

Figure 3.12: (a) Micro-tensile dog bone design (b) Cross-section of build in first orientation .

(a) (b)
45
Table 3.2: Dogbone design dimensions.

Figure 3.13: Stress-strain curves for dog bone specimens tested in tension.
A square prism test specimen was designed and built with the SIS-metal process in order to characterize
the relative density. The dimensions were directly measured with calipers, and the specimen was
46
weighed to determine the relative density. The resulting relative density of the test specimen was found
to be between 5.72 - 5.85 g/cm^3, 65.2-66.7% of the manufacturer reported material density.
3.6 Discussion
As a result of the preliminary experiments, it was understood that printing less inhibitor solution
(thinner borders) around the part periphery led to desirable burnout behavior. However, it is also
understood that taller parts and parts with supports will necessitate larger amounts of deposited
inhibitor. In general, a low ramp rate throughout the burnout period will be necessary for future
research.
In addition, while the inhibitor solution used produces adequate results, it is desirable to search out a
new inhibitor for use with ferrous alloys. Previous research by Yoozbashizadeh [16] reported the
ineffectiveness of a carbohydrate-based inhibitor solution for a low alloy steel material. Thus, an
extensive review of inhibitor candidates will be necessary for the application to ferrous alloys.
Finally, it is understood that there is a limit to the degree of sintering achievable with the SIS-metal
process in one sintering pass. This limit is based on the over-sintering of inhibited regions, which makes
it impossible to mechanically separate inhibited regions from the part regions. This phenomenon has
been reported by Yoozbashizadeh [16] as well as observed in this preliminary work. Thus, parts that
undergo one sintering pass are porous with poor corresponding mechanical properties. Based on the
tensile and density data in Section 3.5 Mechanical Testing and Relative Density, further research must
be performed to improve the strength of as-built SIS-metal parts. The proposed research will include a
characterization of the sintering limit for a specific choice of inhibitor, as well as a post-sintering cycle to
improve mechanical properties.

47
4. Literature Review
4.1 Organic Material Burnout
4.1.1 Introduction
The majority of literature that focuses on the burnout of organic binders is based in the injection
molding or tape casting processes. In general, binder must be removed from PM parts prior to
densification. During debinding, the organic material is heated thermally, melted into a liquid, and
decomposed into a vapor. Burnout models consider a multitude of factors such as pyrolysis of the
material, heat transfer, multiphase fluid flow (liquid and gas), vapor diffusion, capillary forces,
convection and liquid bonding forces to name a few. Incomplete binder removal and uncontrolled
thermal decomposition rates can lead to part defects such as deformation, distortion, cracks, and
expanded pores. Binder thermolysis behavior is affected by the thermochemistry of the binder, binder
concentration, size and configuration of the part, heating rate and furnace atmosphere [35]. It is
important to note that most models for binder burnout behavior are developed empirically for a specific
binder-powder bed system.
4.1.2 Binder Burnout
Strijbos reported two distinct cases of burnout processes in the literature [36]. In the first case, the
amount of binder is sufficiently low (<10 vol%) that the open porosity allows volatile materials to escape
from the interior to the compact surface. In the second case, high binder loading (15-50 vol%) creates
closed pore compacts where the void space is filled with binder (Figure 4.1). This high loading creates a
situation in which open pores are created only as volatiles are removed starting from the green body
surface. Alternatively, a nonplanar binder-vapor interface may result.
48

Figure 4.1: (a) Highly loaded binder powder system (b) surface volatiles escape first causing planar
binder-vapor interface.
Although the organic material in SIS-metal is only deposited at the periphery of the part, the loading is
considered to be high in the deposited regions. Figure 4.2 illustrates a 3D-view of the organic material
loading for an extruded square shape. It is assumed that the organic residue completely fills the voids
between particles in the deposited regions, while empty pore space characterizes the internal part
volume.

Figure 4.2: Example of typical SIS-metal binder loading. Periphery of part is loaded with binder while
the core is an open pore network.
Mistler reported that if the degradation and removal of organic materials reaches a critical point, the
volatiles will eventually be liberated through the formation of a crack at a weak point (or point of high
stress concentration) in the tape casting process. The use of water-based binder systems has been
documented in literature as being more prone to crack formation than their solvent-based counterparts
(a) (b)
Untreated
empty pores
Deposited
regions closed
pores
49
[37]. Binder burnout is of particular interest in the tape casting process. Careful attention must be paid
to the burnout schedule (ramp rate, hold time) to prevent crack defects prior to sintering.
Pinwill et al. developed temperature heating rate diagrams for the removal of a polypropylene-based
binder for powder injection molding [30]. In this study, researchers heated samples at various heating
rates (2, 4, 8, 16, 32, 64, 128
o
C/hr) in static air and flowing nitrogen to a predetermined temperature.
The specimen was removed and inspected visually by the naked eye as well as with low power optical
microscopy when necessary to determine if cracks were present. The experiment was carried out at 20
o
C lower if cracks were present and 20
o
C higher otherwise. The results were plotted and heating rate
diagrams were produced for a particular binder powder system (Figure 4.3).

Figure 4.3: Heating rate diagram from Pinwill. Shaded region indicates heating rates that lead to
various failure modes [30].
The Pinwill diagram coupled with TGA traces suggest that a slow heating rate is necessary at the onset of
binder decomposition throughout the aggressive decomposition temperature range. This has been
verified by other literature as well.
More than one study has been conducted on controlling the rate of decomposition through gravimetric
process control [38, 39]. A special debinding furnace with pressure and gas flow control was designed
50
for this approach (Figure 4.4). The temperature was controlled according to the rate of weight loss in a
staggered method of increase.

Figure 4.4: Modified apparatus for explicit monitoring and feedback of binder weight loss [39].
Other researchers have shown the ability to intrinsically control the rate of decomposition through
selecting blends with weight loss over wider temperature ranges [40].
Expanded pores and blistering in PM parts are a result of bubble formation and growth in the powder
compact. PM parts greater than 10-15 mm in thickness often show bloating or cracking after pyrolysis.
Low molecular weight constituents generally evaporate early at lower temperatures without thermal
degradation. The higher temperature material degrades next and diffuse to the surface and evaporate.
Matar el al studied the early stages of pyrolysis, before continuous porosity is developed, to determine a
critical heating rate above which defects are produced. A bubble is assumed to form when the
concentration of remaining binder constituents decomposition is greater than atmospheric. It was
shown that the lower the heating rate, the lower the vapor pressure was inside the part [41]. Very low
heating rates, on the order of 10
-3
K/s were necessary for defect free parts.
51
Lombardo and Feng developed an analytical equation for the minimum time for binder removal from
green bodies when a threshold pressure is specified [42]. Equations 4.1 – 4.2 give insight to the
significant parameters and physics behind binder burnout.
(4.1)
𝑡 ∗
=
𝐺 𝑇 𝑠 𝑘 𝑆 2
𝑃 𝑡 2
− 1
× {𝑙𝑛
1 − 𝜖 𝑠 1 − 𝜖 𝑠 − 𝜖 𝑏 ,𝑜 − 2 [
1
1 − 𝜖 𝑠 − 𝜖 𝑏 ,𝑜 −
1
1 − 𝜖 𝑠 ] +
1
2
[
1
(1 − 𝜖 𝑠 − 𝜖 𝑏 ,𝑜 )
2
−
1
(1 − 𝜖 𝑠 )
2
]}
Where:
𝑡 ∗
= minimum burnout time
𝑇 𝑠 = starting temperature
𝑘 = Kozeny-Carman parameter
𝑆 = surface area per unit volume
𝑃 𝑡 = threshold pressure
𝜖 𝑠 = solid particle volume fraction
𝜖 𝑏 ,𝑜 = binder volume fraction

And G is defined by Equation 4.2,
(4.2)
𝐺 = 0.8365
𝜇 𝜌 𝑏 2𝜌 𝑜 2
𝑅𝑀 𝑇 𝑜 2
𝐿 𝑥 2
𝐿 𝑦 2
𝐿 𝑧 2
𝐿 𝑥 2
𝐿 𝑦 2
+ 𝐿 𝑥 2
𝐿 𝑧 2
+ 𝐿 𝑦 2
𝐿 𝑧 2

Where:
52
𝜇 = gas viscosity
𝜌 𝑏 = binder density
𝜌 𝑜 = initial ambient atmospheric density
𝑅 = gas constant
𝑇 𝑜 = initial ambient temperature
𝐿 𝑥 ,𝑦 ,𝑧 = dimension in each direction
It can easily be seen from Equations 4.1 and 4.2 that the solution is heavily dependent on the part
geometry and threshold pressure specified. Figure 4.5 was a result of the research that best summarizes
this conclusion. The figure illustrates the dependence of minimum time as a function of 𝐿 𝑥 for fixed
vales of 𝐿 𝑦 ,𝑧 . The idea of a “controlling resistance” can be visualized from the figure in that the amount
of time required for burnout is driven by the smallest dimension.

Figure 4.5: Minimum cycle time as a function of x dimension with fixed y, z dimensions [42].
This bears great importance in SIS-metal in determining the cycle time. Since only the perimeter of the
part contains organic material, the thickness of the perimeters printed, whether they are top, bottom,
53
or side surfaces are the limiting factors, regardless of the orientation. Thus if the sacrificial mold
surrounding the part is of less than 1mm thickness, it could be hypothesized that the burnout time is of
the same order of magnitude for a 1mm thick part.
Another important finding for SIS-metal is that since the threshold pressure value is low (𝑃 𝑡 =1.1 in
Lombardo and Feng’s work), a vacuum environment would make the burnout of organic materials
impractical if not impossible. This has already been confirmed with preliminary experiments. The part
should instead be burnt out in a protective and/or reducing atmosphere close to atmospheric pressure
(1 atm).
4.1.3 Sucrose Decomposition
Typically, the thermolysis behavior of a binder sample under specific atmospheric conditions is
characterized using a thermogravimetric analyzer (TGA). A slow ramp rate (between 1 – 10
o
C) is
employed with a less than 20 mg mass of binder. The current research, in the absence of TGA
equipment, has canvassed the literature in search of similar TGA curves. Suitable TGA curves for the
sucrose have been found to give insight to a starting point for the proposed research. The TGA curves
found indicate that sucrose decomposition predominantly occurs in the range of 167.9 – 196.1
o
C, which
indicates that a slow heating rate is required beginning in this temperature range [43]. The wide range
of decomposition temperature makes sucrose a good candidate for binder burnout. Gaussling showed
with a 10
o
C/min ramp rate under nitrogen atmosphere, the most aggressive range of decomposition
temperature was 226.5 – 324.6
o
C. From ambient temperature to 500
o
C, a 78.55% loss of mass was
recorded (Figure 4.6) [44]. This corresponds to a complete decomposition of sucrose with only carbon
residue remaining.
54

Figure 4.6: Thermal decomposition of sucrose as observed with TGA equipment.
In addition, the melting or decomposition of sucrose has been thoroughly studied in literature. Rhoos et
al reported the beginning of a liquid phase at 160
o
C from a ramp rate of 5
o
C/min and confirmed the
thermodynamic melting point of 185
o
C [45]. Lee et al showed that, depending on the ramp rate used (1,
10, 25
o
C), different melting temperatures were observed for a specific sample ranging from 184 - 206
o
C
[46]. The results of a study performed by Hurtta et al with a ramp rate of 1 and 10
o
C/min showed an
initial melting temperature of 184.5 and 189.2
o
C respectively.
4.2. Powder Metallurgy
4.2.1 Overview
Powder metallurgy (PM) is considered one of the most diverse manufacturing approaches among the
various traditional metalworking technologies. PM provides the ability to fabricate high quality, highly
complex parts to tight tolerances economically. Typically, powder metal parts are fabricated by first
shaping or compacting the powder base material followed by thermal bonding of the particles by
sintering. The many industries and applications where PM has made an impact include: abrasives,
55
agriculture, aerospace, automotive, chemical, electrical, electronics, hardware, industrial,
manufacturing, medical/dental, nuclear, ordnance, and petrochemical.
4.2.2 Metal Powder Characteristics
Understanding the nature of the powder base material is crucial in understanding a powder metallurgy
process. Fundamental particle characteristics include: particle size and distribution, particle shape and
variation with particle size, surface area, interparticle friction, flow and packing, internal particle
structure, composition, homogeneity, and contamination. Typical powders used in metal AM processes
are spherical, prealloyed/homogeneous materials in the subsieve (<44 µm) size range. Spherical
particles are standard in AM due to a high flowability. Figure 4.7 illustrates a microscopic image of the
bronze powder utilized in the current research.

Figure 4.7: Bronze powder particle size distribution.
There are various densities associated with powder metal materials and are a function of inter-particle
friction. Inter-particle friction is characterized by the surface area, surface roughness and chemistry of
the particles. As the surface area increases so does the resistance to flow. Therefore, spherical particles
are most suited to being spread layer by layer for AM processes. Apparent density, or bulk density, is the
density of powder in the loose state without agitation. Tap density is the highest density that can be
achieved by vibration of a powder without the addition of an external pressure. Typical apparent and
tap densities for spherical particles are 60% and 64%, respectively. Finally, the theoretical density
56
describes the density of the material when no porosity is present. These three definitions of density play
important roles in the characterization of a powder metallurgy process and the final mechanical
behavior of the parts produced.
4.2.3 Compaction
One of the main uses of metal powders is the fabrication of complex shapes using a compaction process.
Compressibility, or compactibility, is of significant importance in traditional press and sinter applications;
however, compressibility is generally not of concern in AM processes. Powder AM processes generally
bond loose powder particles together in a solid (ExOne, SIS, SLS) or liquid phase (EBM, DMLS) sintering
process. Past SIS research did consider the addition of a compaction stage to the SIS process. However,
the current research is focused on loose powder sintering [25].
4.3 Introduction to Sintering
Randall German describes sintering as “a thermal process which creates inter-particle welds, improving
upon properties observed in the green state.” It is the process by which particles bond at high
temperatures below the melting point in solid-state atomic transport events, but it can also occur in a
liquid phase as well. Excess surface energy is the driving force for sintering. Particles sinter by atomic
motions that decrease or eliminate the surface energy of the powder. Particles with high surface have
more surface energy and sinter faster. Thus, smaller particles sinter faster. Important factors in sintering
include particle size, applied pressure, formation of a liquid phase, sintering time, heating rate, and
process atmosphere.
A typical measure of degree of sintering is the neck size ratio (X/D). This is defined as the neck diameter
divided by the particle diameter (Figure 4.8).
57

Figure 4.8: Illustration of neck diameter and spherical particle diameter in sintering.
There are three characteristic stages of sintering. Depending on the target final density and properties
of the SIS-metal material, the mechanism, or stage, of sintering can be generally known. The initial stage
of sintering is characterized by rapid neck growth, grain sizes that are no bigger than the size of
particles, and low shrinkage (<3%). This corresponds to a density below 70% of theoretical [47]. The
model in Equation 4.3 for initial stage sintering is valid for an X/D ratio of less than 0.3 and focus on
isothermal neck growth. It is also important to note that neck growth can occur without accompanied
shrinkage.
(𝑋 𝐷 ⁄ )
𝑛 = 𝐵𝑡 /𝐷 𝑚 (4.3)
Where t is the isothermal sintering time, and B is a collection of material and geometric constants. B, m,
and n depend on the mechanism of mass transport. Initial stage sintering shows a high sensitivity to
small changes in temperature and particle size. Changes in sintering time, on the other hand, have a
smaller relative effect in comparison.
Intermediate stage sintering pore structure becomes smoother and has a cylindrical shape as compact
properties develop. Intermediate stage sintering is characterized by grain growth, shrinkage, and some
pore isolation in the latter portion. The relative density range in the intermediate stage is generally 70-
92%. Pore isolation occurs at roughly a 92% relative density, or 92% of the material’s full density value.
58
Isolated pores indicate the final stage of sintering where densification is slowed. Densification of a
sintered compact can be characterized by Equation 4.4.
𝜌 𝑠 = 𝜌 𝑔 /(1 − ∆𝐿 /𝐿 𝑜 )
3
(4.4)
Where 𝜌 𝑠 is the sintered density, 𝜌 𝑔 is the green, or apparent density and ∆𝐿 /𝐿 𝑜 is the shrinkage.
Shrinkage is easily measured with micrometers or calipers using samples heated to various
temperatures and hold times. In general, high sintering temperatures contribute to higher shrinkage
rates and part warping.
High sintering temperatures indicate liquid phase sintering of a powder compact. Liquid phase sintering
enhances densification at the expense of shape loss (slumping). High temperature strengths during the
sintering process are important in the context of part warpage and cracking [48]. Slumping of
unsupported overhanging features in AM can result in cracking or collapse of the final part.
4.3.1 Sintering Characterization
Densification, or degree of sintering, greatly increases the strength of sintered metals. Both in situ and
post mortem data give insights in the quantification of sintering. However, it is more common for
mechanical properties such as tensile strength, radial crush strength, hardness and density to be the
focal point of sintering studies. The sintering response parameters are generally measured with respect
to independently adjusted parameters such as peak temperature or hold time. Both factors affect the
sintered density; however, temperature is reported to have a more dramatic effect [49]. Figure 4.9
illustrates the general case of mechanical property development for metals during densification.
59

Figure 4.9: Changes in mechanical properties as a sintered material densifies.
4.3.2 Mechanical Testing in Powder Metallurgy
Tensile strength, hardness, radial crush strength and transverse rupture strength are common post
mortem responses measured in PM processes. Mechanical tensile testing of dog bone specimens is a
standard practice in PM strength characterization. At low sintered densities, below approximately 80%
of theoretical, the size of the inter-particle sinter bonds is the main determination of strength. At higher
densities, the particle boundaries and pore structure become primary factors. It is important to note
that hardness shares a generally linear relationship with sintered density.
Various literature refers to using the Archimedes principle for determining the relative density of a PM
specimen [32, 50] of complex shape. A summary of ASTM B962 is as follows. The experimental tools
needed are an analytical balance with 0.01 g sensitivity, a transparent water container with a water level
rise of less than 0.10 in, a test specimen support for weighing in water, oil (20-65cP) for impregnation of
specimen, a vacuum impregnation apparatus and thermometer. The sintered density is calculated using
Equation 4.5.
𝑆𝑖𝑛𝑡𝑒𝑟𝑒𝑑 𝐷𝑒𝑛𝑠𝑖𝑡𝑦 , 𝐷 𝑠 =
𝐴 𝜌 𝑤 𝐵 −(𝐶 −𝐸 )
(4.5)
Where:
60
A = Mass of the specimen, g
B = mass of oil-impregnated specimen, g
C = mass of oil-impregnated specimen and support immersed in water, g
E = mass of specimen support, g
𝜌 𝑤 = density of water, g/cm
3
.
Empirical observations have led to simple models of strength as a function of fractional density for
shrinkages less than 6% under specific sintering conditions (Equation 4.6).
𝑆𝑖𝑛 𝑡𝑒𝑟𝑒𝑑 𝑆𝑡𝑟𝑒𝑛𝑔𝑡 ℎ, 𝜎 = 𝜎 𝑜 𝑓 𝑚 (4.6)
Where 𝜎 𝑜 represents the theoretical strength, f is the fractional density, and m is an empirical constant.
Applying Equation 4.6 to the research of Yoozbashizadeh seems to show agreement (Figure 4.10). The fit
is assumed to be roughly valid for values of fractional density between 0.6 and 0.883, the latter
correlating to a shrinkage of 5.9%.

Figure 4.10: Sintered strength model fit to Yoozbashizadeh bronze sintering results.
61
4.3.3 Sintering of Copper Alloys
As the previous research and preliminary work were focused in SIS-metal for the bronze process, special
consideration is given to literature associated with the sintering of bronze in order to gain a deeper
understanding of how to apply knowledge of printing in bronze to printing in stainless steel.
Sintered bronze can be produced from mixing copper powder and tin powder or from a pre-alloyed tin
bronze powder. Pre-alloyed bronze composition is nominally 90Cu-10Sn, with complementary
constituents of graphite, lead, iron, and/or phosphorous depending on the specifics of the material
grade. Self-lubricating bearings or bushings are produced at relative densities of 5.8 to 7.2 g/cm
3
, which
corresponds to roughly 75 - 93% fractional density. The basic manufacturing procedure consists of
compacting a powder shape and then sintering to achieve a homogenous alpha bronze structure.
Additionally, a sizing operation is employed to ensure dimensional accuracy and surface integrity [13].
Both premixed and pre-alloyed bronze powders are used in the production of bronze components.
However, the increased liquid phase sintering makes premixed bronze alloys a poor choice for use in SIS-
metal.
According to literature, bronze alloys are sintered between 760 - 870
o
C [13, 51, 52]. More typically,
bronze is sintered from 815 - 860
o
C for 15 - 30 minutes. It is well known that higher sintering
temperatures give rise to a more rapid shrinkage than do longer sintering times [13]. Another
disadvantage of liquid phase sintering is compact slumping, or shape distortion which is a result of too
much liquid being formed during sintering [53].
Densification in sintering depends on several variables, but principal factors include [13]:
 Sintering temperature
 Sintering time
62
 Particle size
 Green density
Bronze sintering atmospheres are typically protective and reducing. Reduction of oxides that surround
particles promote the rate of sintering. However, reducing atmospheres may not be necessary in the
application of SIS-metal due to a lower sintering rate being preferable [13]. Control of sintered
dimensions is achieved by manipulating sintering time and/or temperature.
Filters made from sintered bronze have porosity between 30-50%. Typical particle sizes range from 325-
1000 µm spherical particles depending on the pore size being targeted. Sintering is generally performed
in a loose powder sintering process under hydrogen or cracked ammonia atmosphere. The typical
sintering temperature and hold time are 785
o
C and 60 minutes, respectively. The addition of 0.35%
phosphorous reduces the solidus temperature of tin bronze to about 700
o
C, which promotes liquid
phase sintering at lower temperatures [54].
Various studies have reported the sintered strength of sintered bronze compacts. Cu-10Sn spherical
bronze particles with 26 µm nominal diameter was reported to reach approximately 120 MPa when held
for 60 minutes at 800
o
C. German noted that sintering for this material typically begins at 400
o
C.
Yoozbashizadeh reported similarly sized Cu-10Sn bronze coupons sintered for 60 minutes to have a yield
strength of 103 MPa.
ASTM B823 sets a specification for copper based PM structural parts [14]. According to the standard,
bronze alloy parts with 10% tin content should have a minimum yield strength of 10 ksi (68.9 MPa).
4.3.4 Sintering of Ferrous Alloys
As the current research is focused in SIS-metal for 316L stainless steel, special consideration is given to
literature associated with the sintering of stainless steels. Stainless steels are a class of ferrous alloys
63
which are distinguished by their superior resistance to corrosion and oxidation at elevated temperatures
[55]. Sintered SS-316L can be premixed or pre-alloyed. Pre-alloyed powder SS-316L composition is
predominantly nickel and chromium, with smaller constituents of carbon, molybdenum, manganese,
silicon, sulfur, phosphorus, and nitrogen. The addition of molybdenum improves general corrosion, and
provides higher creep and tensile strength at elevated temperatures. Alloy 316L resists atmospheric
corrosion, in addition to moderately oxidizing and reducing environments [55, 56]. Both premixed and
pre-alloyed stainless steel powders are used in the production of stainless steel components. However,
the increased liquid phase sintering makes premixed stainless steel alloys a poor choice for use in SIS-
metal. The metal powder used in this study is pre-alloyed SS-316L with chemical composition as
reported by the manufacturer shown in Table 4.1 [57].
Table 4.1: Chemical Composition of Stainless Steel 316L as reported by the manufacturer.

Stainless steels are typically sintered in one of three different atmospheres: dissociated ammonia (75
vol% and 25 vol% nitrogen), nitrogen-base (3 to 10 vol% hydrogen with the remaining volume of
nitrogen) or vacuum. Sintering stainless steel in dissociated ammonia requires a dew point of -45 to -
50
o
C to prevent oxidation. Davis et al found that lower dew points are required at lower temperatures if
a reducing environment is to be maintained [58]. Since nitrogen content is proportional to the square
root of the positive pressure of nitrogen, sintering in an atmosphere of 90% or more nitrogen results in
64
almost twice the amount of nitrogen as that obtained in dissociated ammonia. Slow cooling in this
environment results in more nitrogen pickup due to increased nitrogen solubility down to a temperature
of 1095
o
C and reduced nitrogen solubility coupled with chromium nitride precipitation below that
temperature [58]. Sintering in vacuum is an alternative to the two methods presented and is most
practical when considering CRAFT laboratories equipment. A key consideration is partial pressure
sintering; it is important to consider the vapor pressure of alloying elements such as chromium. If the
furnace pressure falls below elemental chromium’s vapor pressure, the element will evaporate and
corrosion resistance will be highly reduced in the final part. To prevent this from happening, the vacuum
vessel may be backfilled with an appropriate gas, such as nitrogen or argon, to a partial pressure above
the vapor pressure of the different elements within the alloy [58].

Various studies have reported the sintered strength of sintered stainless steel compacts. Samal et al
found that conventionally processed P/M stainless steels may reach sintered densities of 6.6 to 7.3
g/cm
3
, which corresponds to 84 - 93% fractional density [55]. ASTM B783-13 sets a specification for
ferrous powder metallurgy structural parts and reports that P/M SS-316L parts have relative densities of
6.6 to 6.9 g/cm
3
, which corresponds to roughly 83 - 86% fractional density [59]. 6.6 g/cm
3
corresponds
to a yield strength of 15 ksi (100 MPa) and a hardness of 20 HRB, while 6.9 g/cm
3
corresponds to a yield
strength of 22 ksi (150 MPa) and hardness of 45 HRB [59].

ASTM-B783-13 is accommodating in that it allows for re-sintering of the part to achieve a target density.
This may be useful in attempting to create structural stainless steel alloy parts with the SIS-metal
process. For the purposes of this research, a target density of 83% will be set as the goal for obtaining
structural stainless steel parts per ASTM B783-13.

65
Densification in sintering depends on several variables, but principal factors include [13]:
 Sintering temperature
 Sintering time
 Particle size
 Green density

According to literature, stainless steel is sintered between 1120 - 1315
o
C from 0 - 180 minutes
depending on the desired mechanical properties [55, 58]. It is well known that higher sintering
temperatures give rise to a more rapid shrinkage when compared to longer sintering times [55].
Sintering stainless steels are known for displaying greater shrinkage during sintering than P/M irons.
Davis et al found that sintering in either hydrogen or vacuum results in greater shrinkage than nitrogen-
base atmospheres [58]. However, sintering in vacuum with appropriate partial pressure conditions is
desirable due to energy conservation and ecological considerations.

In terms of market distribution, sintered stainless steels have applications in the automotive industry,
which constitutes the largest volume followed by hardware and tools, filters, appliances, office
machines and miscellaneous uses. Grade SS-316L is the most widely used in the market. For example,
P/M SS-316L is used to fabricate steel ejector pads (used for a refrigerator icemaker), which are pressed
to a density of 6.6 g/cm
3
and have a hardness of 65 HRB [58]. There are major applications for P/M
stainless steels for porous mediums, such as liquid retention in bearings, filtering, and sound
attenuation in telephones, microphones and hearing aids. The method for fabricating stainless steel
filters involves first mixing loose powder with resin and pouring it in a mold. The mixture is lightly
pressed at an appropriate temperature to cure the resin. The resin decomposes during sintering and the
porous sheet is then re-pressed and re-sintered [58].
66
4.4 Sintering Inhibition
The inhibition or retarding effects of second phase particles has extensively been studied in literature.
Tikkanen [60] and Tikkanen and Ylasaari [61] found that mixing MgO or CaO in a cobalt or nickel
powders significantly retarded the rate of densification. Johnson concluded that these particles pin the
grain boundaries and create a drag force that retards the sintering process in its earlier stages [62, 63]. If
force exerted by the moving grain boundary interface is low enough that the particles are dragged by
the surface, these particles along with the surface are less mobile. This is true in the case of the initial
stage sintering models.
Ashby et al confirmed the inhibition effects of dispersed particles in metal powders and wires[64]. In
general, a dispersion of particles is said to inhibit sintering. Inhibition effects are greater as the
dispersion particle size decreases and the volume fraction increases. The particles were said to have an
effect on the efficiency of the sources and sinks of matter flow. A crude measure of the inhibition effects
of dispersed particles was given by Equation 4.7.
𝛽 =
𝑎 √𝑓 𝑑 (4.7)
Where a is the metal powder particle radius, f is the volume fraction of dispersed particles, and d is the
dispersed particle diameter. It was concluded that 𝛽 < 1 signified negligible inhibition while 10 < 𝛽 < 100
signified marked inhibition. For 𝛽 > 100, inhibition is very strong and only small necks may form.
Densification is almost completely suppressed in this case.
The study concluded that when inhibition is strong, sintering can only proceed when a “threshold driving
force” is exceeded. The driving force comes in the form of an external pressure in cases where inhibition
is significant. It is shown that particle dispersions affect all stages of sintering, but they particularly
67
retard the densification stages of sintering. Densifying mechanisms tend to be more strongly inhibited
then those that do not densify.
Maak et al. observed that a dispersion of SbO particles greatly suppressed the densification mechanisms
during sintering of Ag. The researchers found that the specimens were no longer controlled by diffusion
but rather interface reaction. This was found to be related to the oxide particle size and spacing [65].
Ness mixed copper powder 2-4 µm in size into aluminum nitrate or magnesium nitrate solution. An
induced chemical reaction left aluminum oxide or magnesium oxide with a 30 nm particle size and
spheroid morphology, respectively. These oxides covered the surface of the copper particles. Compacts
with oxide concentrations between 0.01 wt. % and 0.1 wt. % by weight were pressed and sintered.
Inhibition was found to be the result of an interface reaction mechanism due to particle drag. The extent
of densification was found to decrease as the amount of alumina increased [63].
In 2000, Lu and Hwang investigated the effect of mixing a dispersion of alumina particles within a
carbonyl iron powder compact and sintering. The researchers found through dilatometry that alumina
impeded the early stage sintering of iron in the α phase, but that it improved densification in the γ phase
at high temperatures. They concluded that the presence of alumina played two roles. The first was that
the physical presence of alumina blocked the diffusion of iron atoms which led to the inhibition of
sintering. The second role was that their grain boundary pinning effect helps densification. Finally, the
researchers found that as the alumina content increased, the sintered density of the compact decreased
[66].
In 2010, Razavi-Tousi et al. investigated the sintering behavior of Al-Al 2O 3 composite powders with
varying volume fractions of dispersed oxide particles. He concluded that the smaller, nanometeric
alumina particles had a more hindering effect on sintering than the submicron particles. The hindering
effect was observed to be more aggravated by higher volume fractions of the oxide [67].
68
This literature review serves to guide the current SIS-metal research by suggesting that inhibition is
more likely to occur in the early stage of sintering, or lower temperatures. This is in agreement for
example with the work of Shukla et al who concluded that copper compacts with a fine disbursement of
Cr 2Nb precipitates were strongly inhibited at low temperatures and pressures [68]. However, the same
precipitate was shown in previous studies to enhance mechanical properties at high temperatures. This
literature review seems to indicate that lower temperatures are most likely preferable for ease of
separation for inhibited sections. This is in agreement with preliminary experiments as well.
The work of Yoozbashizadeh observed that higher temperatures and longer hold times at peak
temperatures contributed to larger differential mechanical properties such as hardness (Figure 4.11)
[16]. This is useful in SIS due to the fact that the difference in degree of sintering between these two
regions serves as the pretext for the process. Removal of inhibited region is much easier if the
mechanical properties are significantly lower than uninhibited regions. However, there is a limit to
which inhibited regions can be sintered and no longer removed by abrasive blasting which was not
covered by the research.

Figure 4.11: Differential hardness between sintered and unsintered regions of bronze coupons.
69
There is a trade off in loose powder sintering between final sintered density and in situ transverse
rupture strength (TRS). The higher the sintering temperature is the greater the final sintered density of
the specimen. However, this results in a lower in situ TRS. In situ transverse rupture strength (TRS) for
loose powder bronze sintering was reported in literature as well. TRS seems to be maximized at
sintering temperatures below 800
o
C [69]. This is significant in a loose powder AM process such as SIS-
metal because it attests to the strength of the powder compact throughout the sintering process.
Greater TRS could mean less support structures necessary as well as a lower probably of failure due to
part collapse.
One can also extend the TRS tradeoff to playing an important role in binder burnout as well. In the
process of thermally decomposing the deposited organic materials in the SIS process, the material first
softens and melts. If this is performed at too high of a ramp rate, the entire part could conceivably
collapse. This has been seen in preliminary experiments as well.

70
5. Methodology
5.1 Introduction
Previous research showed that the carbon degradation products of sucrose displayed the greatest
degree of inhibition out of a multitude of ceramic- and organic-based inhibitors tested for a bronze base
material [16]. Preliminary experiments utilizing carbohydrate, or carbon-based inhibitors resulted in
ineffective inhibition when used with ferrous alloys.
The experiments involved first filling a ceramic crucible with 316L SS powder and compacting through
light vibration. A 15 µL droplet of the sucrose solution previously used for bronze inhibition was
deposited into the 316L powder. This sample was placed into the furnace and sintered at 1100
o
C for 30
minutes. Upon removal from the furnace, the sample was abrasive blasted with no evidence of
inhibition (Figure 5.1). In other words, it was not possible to achieve mechanical separation of the
inhibited region from the surrounding part region.
It is believed that a lack of inhibition was observed as the result of a mechanism similar to the
carburization process for iron and steel materials. In this process, iron or steel is heated in the presence
of carbon in order to harden the surface of the part. Similarly, in the case of the preliminary experiment,
the sucrose inhibitor degraded to carbon in the furnace and the sample was heated. It is believed that
the presence of carbon in this experiment had the reverse of the intended effect and instead actually
hardened the affected region.
71

Figure 5.1: (left) Preliminary droplet test with Sucrose-based inhibitor solution in bronze where
mechanical separation by abrasive blasting was easily achievable. (right) In contrast, sintering
inhibition in stainless steel did not occur in the droplet test.
This result necessitated further research on candidate ceramic precursors to identify and test a new
inhibitor. It was believed that a different choice of inhibitor would exhibit a large enough degree of
inhibition in order to achieve mechanical separation of inhibited regions for a 316 stainless steel base
material. Thus, a number of ceramic precursors were examined through a review of literature and
droplet tests to determine which candidate would exhibit the greatest degree of inhibition. Inhibition
was gauged through abrasive blasting and visual observation.
It should be noted here that producing parts with maximum fractional density and hence mechanical
properties is desirable. However, there is a limit to the degree of sintering in SIS-metal based on the
degree of inhibition realized in printed regions. If the printed regions are sintered too much, they will
not be separable from the final part.

72
5.2 Research Hypotheses

The following research hypotheses are to be addressed in the current work:
Hypothesis 1: There exists a sintering inhibitor solution for a 316L SS powder that, when the treated
area is sintered under specific parameters, the inhibitor produces a condition of inhibition in which the
inhibited area is mechanically separable from uninhibited material.
Candidate precursors will need to be identified for use in an aqueous solution since metallic oxides are
typically no soluble in water. Experiments will be conducted to determine the efficacy of a chosen
inhibitor through investigation of the sintering cycle and a characterization of the degree of inhibition.
Hypothesis 2: It is possible to achieve an SIS-metal part density of at least 85% with a carefully planned
sintering strategy.
5.3 Research Plan & Procedure
The current research is focused on six key objectives:
1. Identify metal oxide dispersants with high melting points for potential use as inhibitors.
2. Refine list with the most likely inhibitor candidates through an analytical review of their
respective chemical precursors.
3. Determine the sintering conditions for loose powder sintering of the 316L SS base material.
4. Experimentally verify inhibition through the sintering of droplet tests at various temperatures
and concentration levels and subsequent abrasive blasting.
5. Use the newly discovered and verified inhibitor solution in printing rudimentary 2.5D shape with
the existing SIS beta prototype machine.
6. Improve part strength and characterize shrinkage.
Please refer to Figure 5.2 for a graphical representation of the proposed research.
73

Figure 5.2: Graphical representation of proposed research.
74
5.4 Assumptions Derived from Literature
From the literature review (4.4 Sintering Inhibition), the following assumptions were made to guide and
focus the search for a likely 316L SS inhibitor:
 Oxides inhibit or retard the sintering of metals and metal alloys
 The larger the number of oxide particles, the greater the degree of inhibition
 The smaller the oxide particle size, the greater the degree of inhibition
 The lower the temperature, the higher the efficacy of the inhibiting oxide
 An even dispersion and hence careful mixing of the dispersion is extremely important in drawing
conclusions with respect to inhibition
In the SIS-metal process, oxides are not used directly in the preparation of an inhibitor solution. Rather,
the molecular precursor to a desired oxide is utilized. This is due to metal oxides or other inhibitors not
typically being soluble in water. For example, carbon is not soluble in water; thus, Yoozbashizadeh [16]
used a highly soluble carbohydrate (sucrose) in order to obtain carbon through pyrolysis in the furnace.
The carbon degradation product served to actively inhibit or retard the sintering process in the
fabrication of bronze alloy parts.
The intended focus of the current research is on the use of oxides for the inhibition of stainless steel.
Similar to the work of Yoozbashizadeh, the author of this work has focused on identifying the best
candidate precursors for use in the SIS-metal process. This is a direct result of oxides not being soluble in
water. In addition, it is assumed that the sintering temperature of the final oxide is of great importance
in determining likely candidates for inhibition. The use of oxides as inhibitors is predicated upon the fact
that the oxides will not sinter at the temperature of the base material, allowing for ease of separation.
Thus, two additional assumptions in the current research were as follows:
75
 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 degradation product, or
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
5.5 Analytical Identification of Candidate Oxides
The first step in developing the SIS-metal process for stainless steel was to identify potential metal oxide
dispersants. Table 5.1 is a list of typical metal oxides and their respective melting temperatures, ordered
from highest melting point to lowest.
The initial requirement for a metal oxide to be a viable candidate for inhibition was that the melting
point of the material must be sufficiently high. According to the British Stainless Steel Association, the
melting point range for 316L is between 1375 – 1400
o
C. It is interesting to note that the 316L alloy
melting temperature is far lower than that of pure iron at 1535
o
C [70]. The first step in filtering the
inhibition candidates was to accept only the candidates whose melting points were at least 30% greater
than that of 316L. Thus, all candidates with melting points below 1820
o
C were eliminated.
The next step in the filtering process involved a review of the metal oxides mentioned in literature that
had previously shown evidence of retarding the sintering of various metals and metal alloys. The oxides
highlighted in yellow represent materials that had a sufficiently high melting point but were not
referenced in literature.
76
Oxides highlighted in green represent the materials with sufficiently high melting points and were found
effective in literature. Thus, five candidate oxides were identified for further investigation: Al 2O 3, MgO,
CaO, ZrO 2, and TiO 2.
Table 5.1: Common metal oxides and their respective melting points.

While this table consists of a comprehensive list of high temperature ceramic melting points, the
sintering temperatures of these ceramic materials were determined to be less significant and thus not
included. The most important scientific reasoning for this is that melting temperature is standard and
not dependent on various factors in the manner that sintering temperature is. The sintering
temperature of any material is incredibly dependent on the particle size, shape, regularity of size
distribution, purity of the material, chosen additives, and the sintering atmosphere. For example, while
the sintering temperature used for copper is known to be between 750
o
C – 1000
o
C, it is well-reported
77
in literature that the sintering temperature of nanoparticle copper was less than 250
o
C in a reducing
atmosphere[71].
5.6 Analytical Identification of Candidate Precursors
The next step in the identification and filtering of inhibitor candidate was to examine if the chemical
precursors would likely decompose to the desired metal oxide at high temperature. This refinement was
based on two conditions. The first condition was that the precursor must be highly soluble in water. The
logic behind this condition was that a higher solubility would lead to a higher yield due to higher metal
salt content. The second was that the product yield of oxide from the chemical decomposition be high;
this refers to the product that results from the thermal degradation of precursor within the furnace.
These two conditions taken in conjunction result in the highest possible amount of oxide.
A cursory review of chemical precursors resulted in the identification of nitrates and sulfates to be
highly likely candidates for inhibition precursors. These two families of molecules are generally found to
be highly soluble and generally decompose to metal oxides at high temperature. All nitrates in Group (2)
undergo thermal degradation to give the metal oxide, nitrogen dioxide and oxygen.
A list of the chemical precursors to aforementioned oxides can be seen in Table 5.2.
Table 5.2: Potential chemical precursors to desired metal oxides.

78
The thermal decomposition temperatures of all of these compounds were found to be lower than the
sintering temperature of the 316L base material. Some nitrates and sulfates were eliminated from
consideration due to unavailability of information or instability of the molecule at room temperature.
From previous research, it was known that aluminum sulfate decomposes to aluminum oxide at high
temperature [16]. Magnesium Nitrate was analytically determined to have a 40.5% higher yield of metal
oxide than aluminum nitrate by number of particles. Thus, due to the high oxide yield of the magnesium
nitrate precursor and the high temperature nature of magnesium oxide, magnesium nitrate was
selected as the most likely inhibitor candidate.
5.7 Special Consideration of Deposition by Print Head (dilution talk)
From previous work with the Epson Workforce 30 printer platform [12], it was known by this author that
the viscosity necessary for a fluid to be jettable is roughly between 2-5 cP. The surface tension is not as
significant in printing a solution but should generally be less than approximately 50 dynes. As the scope
of this work was to identify a proper inhibitor solution, it was unknown what the viscosity of the final
inhibitor solution would be. The objective was first to find a solution that would properly inhibit.
However, it was of great interest to the present research to have the ability to print a part in the SIS-
metal beta prototype. Thus, it was not enough for an inhibitor solution to be efficacious at full
saturation; it had to also be efficacious at a diluted saturation level to ensure jettability from the print
head.
5.8 Stainless Steel Sintering
Qualifying a new material for the SIS-metal process involved first sintering of the base material. From
literature as well as experience, it was well understood that a dispersion of oxides was only effective in
the early stages of sintering. Thus, the premise behind the choice of sintering profile and hold
temperature was the notion that the lower the sintering temperature, the more likely inhibition would
79
result. Once an understanding of the degree of sintering was reached, a likely sintering profile and hold
temperature could be chosen for droplet testing.
5.9 Droplet Tests
Once the degree of sintering for the stainless steel material was reached, droplet tests were conducted
at low sintering temperatures. The solution used for these initial droplet tests was saturated with solute
in order to maximize the probability of successful inhibition. The goal was to sinter a stainless steel
coupon such that the affected droplet region could be removed with a pressurized abrasive media (e.g.,
sand blasting) while leaving the unaffected regions intact.
Once an inhibitor candidate was validated, further experiments were necessary to ensure effective
inhibition at a diluted solution content. This was done to ensure an efficacious solution would be
printable. As such, various droplet tests were conducted in this research with various levels of dilution to
observe the minimum saturation level of a solution that can inhibit the 316L powder.
It was also important to gain a deeper understanding of the physical mechanism behind the inhibition
process. As such, it was necessary to review various sintered samples under microscope to verify the
final decomposed inhibition product was indeed a metal oxide, as well as to make new observations
regarding inhibition. This could be done using elemental analysis in addition to electron microscopy by
way of X-ray spectrometry. The principal behind the evaluation of resulting decomposition products is to
observe the ratio of elements in a given region. For example, if a small enough region is observed, and
the ratio of Mg to O is approximately 1:1 for that region, then it is highly likely that the material present
is MgO.
Based on previous experiments with bronze, it was expected that the density of a sintered coupon with
successful inhibition would be low, on the order of 55-65%. While this relative density would not be
sufficient for industrial applications, it would be sufficient for extracting a fabricated part of net shape.
80
5.10 Part Fabrication
From the result of droplet testing, various ad-hoc inhibitor solutions were created in order to print net-
shape parts with the SIS-metal beta prototype machine. Similar to preliminary experiments with the
bronze 5807c material, rudimentary two-dimensional shape extrusions were designed and printed as a
final verification for the use of the newly discovered inhibitor solution with the SIS-metal process. More
complex 2.5D shapes were later designed to gain a better understanding of process capabilities.
While part fabrication was not a primary focus of this research, it served to create a bench mark for as-
built material properties. This helped to create a basis of comparison for improving part strength.
5.11 Improvement of Part Strength
As a result of the combination of droplet tests, the sintering of the stainless steel base material and part
fabrication, the material properties of as-built parts was well understood. A target of 85% density was
set for the SIS-metal process based on ASTM B783-13. Thus, if only a low relative density (less than 85%)
was achievable in a single sintering pass, a second sintering pass would be utilized to achieve the desired
density. Final density and tensile strength of SIS-metal parts was measured and reported.

81
6. Results
6.1 Stainless Steel Sintering
6.1.1 Procedure
The material used in this research was purchased from Carpenter Powder Products headquartered in
Bridgeville, Pennsylvania. A spherical 316L stainless steel alloy, “Micro-Melt 316L” (Part No. 2310103-
0010), was chosen for the purposes of this research. The mesh size was 325- with a minimum particle
size of 16 µm.
The base material was sintered in a 10 kW, high vacuum sintering furnace purchased from Nanyang
Xinyu Furnace Co., Ltd. and can be seen in Figure 6.1. The furnace is equipped with a mechanical
vacuum pump as well as a diffusion pump. The maximum achievable vacuum reported by the
manufacturer is 10
-4
Pa. The furnace utilizes MoSi 2 heating elements with a maximum working
temperature of 1600
o
C. A 30-step programmable PID controller was utilized in order to achieve the
desired sintering profiles.
82

Figure 6.1: High vacuum sintering furnace utilized in experiments.
The sintering profile was designed so that the sintering hold temperature was the only factor varied.
Five hold temperatures were initially chosen at equal intervals from 1000
o
C up to 1400
o
C in order to
evaluate the corresponding degree of sintering. Once the densities of those coupons were
characterized, additional hold temperatures of interest were tested, specifically those hold
temperatures in the range where densification is prominent. The sintering profile strategy was as
follows: (1) 5
o
C/min ramp to 200
o
C, (2) 10
o
C/min ramp to 1000
o
C, (3) 5
o
C/min ramp to Hold
Temperature, (4) 30 min hold, (5) -25
o
C/min ramp to 700
o
C, (6) Passive cooling. An illustration of the
sintering profile for the 1400
o
C hold temperature can be seen in Figure 6.2 as an example.
83

Figure 6.2: Generic sintering profile used in experiments.
Cylindrical alumina crucibles purchased from AdValue Technology were used in these experiments. The
crucible inner dimensions, or pre-sintered compact dimensions, were nominally 10.5 mm in height and
23 mm in diameter.
The resulting density of the sintered coupons was measured per ASTM B962, using the Archimedes
principle. The procedure was summarized in Section 4.3.2. A Tree HRB103 electronic scale with a 0.001g
sensitivity was utilized for measuring sample mass. The samples were immersed in a standard 250 mL
graduated beaker. A vacuum impregnation apparatus (Figure 6.3-left), consisting of a Welch vacuum
pump (Model 2565B-50) and a desiccator, was used to impregnate the specimens with Mobil Almo 525
Air Tool Oil. A simple support was constructed with the apparatus as seen in Figure 6.3-right.
84

Figure 6.3: (left) The vacuum impregnation apparatus and (right) the support apparatus in accordance
with ASTM B962.
6.1.2 Results
The sintered density was calculated using Equation 4.5. The value of 0.998 g/cm
3
was used for the
density of water at a measured temperature of 22.4
o
C. The manufacturer reported density of 8.0 g/cm
3

was used for that of the stainless steel base material. Table 6.1 displays the calculated relative density
values for the sintered coupons. Figure 6.4 is a visual representation of the hold temperature versus the
calculated relative density for the sintered coupons.
Table 6.1: The measured masses and calculated relative density per Equation 4.5.

85

Figure 6.4: Hold temperature vs relative density of the stainless steel sintered coupons.
As evidenced in Figure 6.4, it was apparent that densification began at approximately 1200
o
C. Thus, two
additional hold temperatures (1225
o
C, 1250
o
C) were later added to the experiments in order to further
flush out the relationship between sintering temperature and relative density. This information served
as the basis for the droplet tests in the next section. It should be noted that at 1400
o
C, the sample was
melted.
6.2 Saturated Droplet Testing
With an understanding of the loose powder sintering behavior for the stainless steel base material in
question, it was time to determine whether or not a sufficient level of inhibition could be achieved such
that an inhibited region could be mechanically separated from an uninhibited region.
6.2.1 Procedure
From previous research [16, 72], it was understood that there is a limit to the density achievable while
maintaining the ability to mechanically separate inhibited regions through abrasive blasting. Thus, a set
86
of experiments was designed in order to first validate that an inhibitor would work for the base material
in question and determine at which temperature the inhibitor was no longer effective.

The first step was to create a saturated inhibitor solution using magnesium nitrate. Five hundred grams
of Magnesium Nitrate, 6-Hydrate, chemical formula Mg(NO 3) 2
.
H 2O, (CAS No. 13446-18-9) was purchased
from VWR for the purpose of testing its viability as an inhibitor. The solubility of Magnesium Nitrate was
reported by the manufacturer to be 125 g/100 mL. However, for the purposes of saturation, excess
magnesium nitrate crystals were added to 20 mL of distilled water to ensure saturation.

The cylindrical alumina crucibles utilized for the sintering of stainless steel coupons in the previous
section were also utilized for the droplet tests. The crucibles were filled with powder and vibrated to
improve packing density. A Scilogex MicroPette (Figure 6.5-left) with a full scale range of 10 – 100 µL
was utilized to deposit 15 µL droplets onto each sample. Two separate droplets were deposited side-by-
side for each experiment to ensure valid results. The order of experiments was randomized, and there
were two replicates for each of the five temperature settings, or factor levels.
After deposition of the inhibitor droplets onto the green samples, the green parts were pre-heated for
60 minutes in a Neytech Centurion Qex porcelain furnace (Figure 6.5-middle) at 110
o
C. This was done in
an ambient environment to dry the excess water in the droplet region while mitigating the risk of
sample rupture due to pressure build up from rapid vaporization. The green part was then removed
from the furnace and carefully transported to the high-temperature vacuum furnace for sintering. The
five original hold temperatures from 1000
o
C – 1400
o
C were utilized for these experiments. The
sintering profile design from section 6.1 was utilized for consistency.
87
Upon completion of the sintering cycle, the coupons were removed from the furnace and placed into an
abrasive blasting chamber. The MB10 model from MicronBlaster (Figure 6.5-right) was used with glass
bead media. Pressures were increased slowly from 30-60 psi as necessary for inhibited region
separation. The distance of the blasting tip from the specimen was also decreased from 4 inches down
to 1 inch in order to intensify blasting.

Figure 6.5: (left) Pipette, (middle) pre-heat furnace, and (right) abrasive blasting equipment.
6.2.2 Results
Upon completion of abrasive blasting, it was discovered that the samples sintered at a hold temperature
of 1000
o
C- 1200
o
C were indeed sufficiently inhibited. At pressures below 40 psi, the inhibited droplet
region for the 1000
o
C and 1100
o
C samples were easily separated. For 1200
o
C, the pressure was raised
to the maximum 60 psi in order to achieve separation. Table 6.2 summarizes the results of the findings
with a column containing a subjective assessment of ease of removal. The ease of removal was ranked
from 1 – 10, with “1” representing separation at 1000
o
C and “10” representing no discernable inhibition
or an inability to mechanically separate. In the case of the 1300
o
C hold temperature, the pressure was
raised to 60 psi and the distance was reduced to 1 inch. A very low degree of inhibition was discerned,
however the uninhibited region showed signs of wear as well. This leads to the conclusion that if a part
cannot be sintered at 1300
o
C without being damaged during abrasive blasting. Figure 6.6 is a
photograph of the specimens, post-abrasive blasting, lined up in order of ascending hold temperature
from left to right.
88
Table 6.2: Inhibition results.

Figure 6.6: Photograph of inhibited specimens after abrasive blasting.
6.3 Diluted Droplet Testing
In the SIS-metal process, it is important to understand the level of concentration needed to achieve
separable inhibition. This was a special consideration given to an inhibitor candidate with regards to its
jettability from a print head as a fully concentrated solution was too viscous to be jettable. However,
designing the final tailored inhibitor solution was out of the scope of the current research. Without
knowing the actual viscosity and surface tension of the final designed solution, this research was to
determine the levels of concentration that proved effective through the inhibition of stainless steel for
the purposes of future research.
6.3.1 Procedure
From the research assumptions, it was understood that the greater number of inhibiting particles would
increase the likelihood of inhibition. However, since the fully saturated solution would not be printable,
89
a separate set of experiments was designed in order to determine the settings of concentration level
and temperature where mechanical separation was still achievable.
The first step was to create separate inhibitor solutions for ten different levels of magnesium nitrate
concentration. The levels of concentration were chosen between 10-100% at equal intervals. The
approach was to create an experiment in which all levels of concentration could be tested in a single
sintering pass for a given hold temperature.

The physical design of the experiment can be seen in Figure 6.7. For each experimental run, there were
two columns and five rows, for a total of ten droplets. Stainless steel powder was initially spread on an
82mm diameter alumina disc between two 0.125 inch thick spacers. This allowed to spread loose
powder in a rectangular shape with approximate dimensions 1 x 3 x
1
/ 8 inch. The droplet spacing was
designed to be 0.5 inches apart, but practically was difficult to maintain. The Scilogex MicroPette was
again used in these experiments to dispense 15 µL of a given solution to the appropriate location. An
example of the design in practice can be seen in Figure 6.7. The droplet order was also randomized to
reduce bias (Table 6.3).

Figure 6.7: (left) The experimental design and (right) implementation of the diluted droplet testing.
90
Table 6.3: Droplet position with associated solution concentration.

After droplet deposition, the samples were moved into a furnace pre-heat at 110
o
C for 60 minutes to
remove excess fluid. This was again done in an ambient environment to dry the excess water in the
droplet region while mitigating the risk of sample rupture due to pressure build up from rapid
vaporization. The green part was then removed from the furnace and carefully transported to the high-
temperature vacuum furnace for sintering. The hold temperatures investigated were between 1000-
1250
o
C at 50
o
C intervals, with the addition of a 1225
o
C hold temperature. The experimental order was
randomized to reduce bias. The experimental runs can be seen in Table 6.4.
Table 6.4: Run order for dilution experiments.

After sintering, all of the resultant samples were abrasive blasted and results were observed to
characterize the degree of separation.
91
6.3.2 Results
Upon completion of abrasive blasting, it was observed that all sintered samples showed some degree of
inhibition at a concentration level of 90%-100%. However, the samples sintered at 1250
o
C were
abrasive blasted at maximum pressure (60 psi) to achieve separation and the uninhibited surrounding
region was damaged. At a concentration of 10%, it was difficult to achieve separation at the lowest
temperature setting of 1000
o
C (Rank 8). Photographs of the results after abrasive blasting can be seen
in Figure 6.8 for all seven hold temperature settings. Various degrees of cracking and ruptures were
visually evident throughout all of the sintered samples. This was believed to be due to inconsistencies in
the pre-heating furnace when removing the excess fluid.

Figure 6.8: Photographs of the seven dilution experiments in order of temperature from left to right.
Figure 6.9 is a chromatic visual representation of the level of separation achievable for the different
levels of concentration versus the sintering hold temperature. The areas in green illustrate regions of
the graph where separation was easily achievable, and the areas in red illustrate the regions where
separation was not achievable. Areas in yellow are the regions that were not explicitly tested. More
92
specifically, an area located between green and red areas are colored yellow because it is unclear where
in the region separation is achievable and where it is not.

Figure 6.9: Visual representation of the results of the dilution experiments.
6.4 Scanning Electron Microscopy
6.4.1 Procedure
The dilution experiment sintered at 1100
o
C was repeated for the purpose of visual observation. Upon
removal from the furnace, the samples were visually observed and then transferred to scanning electron
microscopy equipment at the Center for Electron Microscopy and Microanalysis (CEMMA). The
instrument used was a JEOL JSM 6610 equipped with an EDAX energy dispersive X-ray spectrometer
(EDS) with a Saphire Si(Li) detecting unit, 10 mm
2
crystal and genesis software (Figure 6.10).
93

Figure 6.10: JEOL JSM 6610 SEM located at the USC CEMMA center.
6.4.2 Results
Upon removal from the furnace, the surface of the droplet regions were noticed to be white in color.
Figure 6.11 is a photograph of the 90% saturated solution in position (3, 2). The color indicated that the
surface was likely composed of magnesium oxide, but further observations and elemental analysis were
necessary in order to confirm the indications.

Figure 6.11: Photograph of the surface of droplet (3, 2) after sintering.
Figure 6.12 was an electron micrograph of the MicroMelt 316L material in its original, unsintered
condition. While the material was designated by the manufacturer to be of spherical morphology, it was
clear from the micrograph that this was not entirely accurate. It seemed that the majority of the
material was of spherical or rounded morphology, but there were many instances of cylindrical,
94
irregular, and spongy morphologies as well. In addition, it was found that some of the particles were
already aggregated and bonded together in some capacity.

Figure 6.12: SEM micrograph of the MicroMelt 316L material in its original, unsintered condition.
Figure 6.13 is an electron micrograph of an uninhibited portion of the sintered sample. The sample was
sintered at 1100
o
C for 30 minutes and was expected to have a density of 57.0%, based on the previous
sintering results.
95

Figure 6.13: Electron micrograph of an uninhibited portion of the sintered sample.
The X/D ratio for the uninhibited region was observed to be 0.32. It should be noted that this is a
difficult measurement to make in practice as it is highly dependent upon the size of the particles chosen.
The reported measurement was based on the particles located within the red outlined region of Figure
6.13.
Figure 6.14 is an SEM micrograph of the 100% saturated droplet located in position (1, 2). The droplet
diameter was measured to be 3.55 mm. The bright white regions were initially believed and later
confirmed to be magnesium oxide. Figure 6.15 is a magnified micrograph of the center of the same
droplet region. Figure 6.16 is the EDAX spectral analysis for the same region focused on the bright white,
web like substance in the center of the micrograph. The highlighted regions in blue (left) and green
(right) clearly indicate the location of magnesium and oxygen spectral peaks, respectively.
96

Figure 6.14: SEM micrograph of 100% saturated droplet located in position (1, 1).
97

Figure 6.15: Magnified micrograph of the center of the 100% saturated droplet.

Figure 6.16: (left) The magnesium EDAX spectral peak highlighted in blue , and (right) the oxygen
EDAX spectral peak for the same region highlighted in green.
98
Figure 6.17 illustrates the entire associated EDAX spectrum with the peaks identified by elemental symbol.

Table 6.5 illustrates the relative atomic quantities by Wt% and by At%. The blue box was drawn to
highlight the relative quantities of magnesium and oxygen. From the At%, it can be seen that it is highly
likely magnesium oxide (MgO) is present. The difference in atomic percentages between magnesium
and oxygen was slight and was believed to be attributable to measurement errors during spectral
acquisition.

Figure 6.17: EDAX spectral analysis of 100% saturated droplet.

99

Table 6.5: Relative atomic quantities associated with the spectral analysis.

The same droplet region was next scratched with a metal tool to remove a thin layer of material and
observed again under SEM. Figure 6.18 is a magnified SEM micrograph of the region. Figure 6.19
indicates the brighter regions of the image to be the presence of magnesium oxide, it is interesting to
note that magnesium and oxygen are present throughout the entire region of the micrograph, indicating
a high coverage of the metallic particles being not just in the bright regions.

100

Figure 6.18: Magnified micrograph of the center of the 100% saturated droplet after a thin layer of
material was removed with a hand tool.

Figure 6.19: (left) The magnesium EDAX spectral peak highlighted in green , and (right) the oxygen
EDAX spectral peak for the same region highlighted in red.
101
Figure 6.20 illustrates another view of the same droplet region under less magnification. The image
shows magnesium oxide sparsely scattered across the area being examined.

Figure 6.20: Alternative view of 100% saturated droplet region after a thin layer of material was
removed.
Due to the nature of scanning electron microscopy, the densely covered magnesium oxide regions were
too bright for a clear image when compared to the surrounding stainless steel material (see Figure 6.15).
102
Thus, a separate micrograph was taken (Figure 6.21) with the brightness dialed down in order to better
view the magnesium oxide structure in the droplet region.

Figure 6.21: An SEM micrograph of a droplet region rich in magnesium oxide. The brightness was
dialed down for better visualization.
6.4 Part Fabrication
The purpose of discovering a new base material/inhibitor system was to ultimately produce parts by
way of the selective inhibition sintering process. Previous SIS-metal research served as a viable proof of
concept for the creation of high resolution bronze alloy parts, however mechanical properties were
lacking. This necessitated the exploration of new material systems. The open design of the SIS-metal
beta machine allowed for ease of testing of these new systems. Thus, the magnesium nitrate inhibitor
solution was implemented with the SIS-metal beta machine in an attempt to fabricate a stainless steel
part.
103
6.4.1 Procedure
From the result of the droplet testing, magnesium nitrate based inhibitor solutions were created in
order to print net-shape parts with the SIS-metal beta prototype machine. It was known that a sintering
temperature of 1100
o
C provided adequate part strength to resist wear from abrasive blasting and at the
same time allowed for freedom in the choice of concentration between 20-90% saturation. In other
words, all of the concentration levels tested between 20-90% were adequately separable after sintering.
The jettability of a solution is dependent upon many factors, including the solution’s viscosity and
surface tension. In addition, the use of surfactants and a great deal of tailoring of the formulation is
needed to ensure stability of the solution. Based on the work of Torabi et al [73], a solution with 50%
concentration of magnesium nitrate was used for part fabrication. The solution was placed into the
refillable ink cartridges of the SIS-metal beta machine.
The printing process involved the use of 120 um layers. Two color channels were employed similar to
preliminary experiments with the bronze 5807c base material. HSE was utilized for the fabrication of a
two-dimensional gear similar to the gear designed and used in the preliminary experiments. This gear
was designed to be 48.8 mm in diameter and 3.6 mm in height (Figure 6.22-left). The layer image order
can be seen in (Figure 6.22-center/right).

Figure 6.22: (left) The gear designed for part fabrication, (center) the layer image representing layers
1-30, and (right) the 32nd and final closing layer.
104
6.4.2 Results
After the green part was removed from the platform, the peripheral powder was blown away carefully
using low pressure air (Figure 6.23). The green part was then pre-heated at 115
o
C for 4 hours to ensure
the vaporization of excess fluid prior to sintering. It was then sintered at a hold temperature of 1100
o
C
for 30 minutes and abrasive blasted to remove inhibited regions. Figure 6.24 represents the finished
part results after abrasive blasting.

Figure 6.23: (left) Green part after blowing with low pressure air.

Figure 6.24: Alternate view of finished part after abrasive blasting.
The fabricated gear in Figure 6.24 represented the first part ever fabricated with a ferrous material using
the SIS-metal process.
105
6.5 Improvement of Part Strength
While part fabrication was not a primary focus of this research, it was in line with the overall purpose
and served to confirm a bench mark for as-built material properties. This helped to create a basis of
comparison for improving part strength. Based on the previous sintering results (Table 6.1), the relative
density of a part sintered for 30 minutes at 1100
o
C was 57.01%.
6.5.1 Procedure
Based on ASTM B783-13 [59], a target density of 85% was established for the SIS-metal process.
However, only a low relative density (63.79% at 1225
o
C) was achievable in a single sintering pass while
maintaining the ability to mechanically separate the inhibited regions from the uninhibited regions.
Thus, a second sintering pass was utilized to achieve the desired density.
For this purpose, a tensile dog bone specimen was designed based on ASTM E8 [74]. It should be noted
that while the specimen as a green compact fit within the length requirement of ASTM E8, the final
sintered length of the specimens did not due to shrinkage during sintering. This was due to the size
106
limitation of the furnace used. The final dog bone length was designed as large as possible while being
able to fit into the furnace hot zone. Figure 6.25 describes the detailed design of the specimen mold.

Figure 6.25: Tensile specimen design.
The mold was comprised of alumina silicate material, purchased from McMaster-Carr (Part # 8479K11)
and machined on a CNC milling machine. The final mold can be seen in Figure 6.26.

Figure 6.26: Fabricated dog bone specimen mold.
107
The mold cavity was filled with loose powder and lightly vibrated for compaction. The strategy for
sintering the tensile coupons was to mimic the SIS-metal fabrication process as closely as possible. In
this regard, the green part was transferred to the furnace and sintered at 1100
o
C for 30 minutes,
simulating the first sintering pass for the SIS-metal built part.
The dog bone was then removed from the mold, measured and placed back into the furnace for a
second sintering pass.
The strategy for the second sintering pass was to use the lowest practical sintering temperature to
achieve the desired relative density of 85%. According to Equation 4.4, this relative density correlated to
a linear shrinkage of 15.3 mm or 13.5%. Practicality was solely based on sintering duration, the upper
limit of which was set at 8 hours of hold time. This was done in an effort to thwart unwanted warpage in
the part while maintaining the ability to produce a part within a 24 hour period. Thus, the goal for
improving strength was to determine the sintering temperature at which less than 8 hours of hold time
produced a part with greater than 15.3 mm linear shrinkage in the long dimension.
The baseline for the second sintering pass was a hold temperature of 1200
o
C for 4 hours. The
subsequent specimen would be sintered at the same hold temperature for 8 hours. If the goal of 85%
was not reached, the sintering temperature was increased by 50
o
C. The process was repeated until the
linear shrinkage goal of 15.3 mm was reached.
Once the shrinkage goal was reached, all of the dog bone specimens were tensile tested on an Instron
5567 Material Testing System (Figure 6.27-left) with a test speed of 0.5mm/min. A secondary
extensiometer (Instron Model: 2630-109) was attached to the specimen for greater accuracy (Figure
6.27-right).
108

Figure 6.27: (left) Instron 5567 Testing System and (right) a dog bone specimen being tested with
extensiometer attached.
Finally, the gear produced in the previous section was re-sintered with the newly determined secondary
sintering profile.
6.5.2 Results
The average linear shrinkage after the first sintering pass was found to be 0.92% with a standard
deviation of 0.019%. This correlated to a sintered relative density of 52.69% according to Equation 4.4.
Table 6.6 illustrates the shrinkage results after the second sintering pass. It can be seen from the table
that a sintering temperature of 1300
o
C and hold time of 8 hours resulted in a calculated relative density
of 85.23%, according to Equation 4.4. Figure 6.28 is an alternative representation of the shrinkage
results. Figure 6.29 represents the corresponding sintered densities from the same set of experiments.
109
Table 6.6: Shrinkage result summary from second sintering pass.

Figure 6.28: Shrinkage results from the second sintering pass.
110

Figure 6.29: Calculated corresponding densities for second sintering pass.
The resulting sintered dog bones were tensile tested, with the results shown in Figure 6.30 – Figure 6.33.
For comparison, wrought 316L stainless steel has a density of 7.95 g/cm
3
, yield strength of 205 MPa, and
ultimate tensile strength of 515 MPa according to the manufacturer. The density result represented the
minimum standard of structural powder metallurgy parts according to ASTM B783-13 at 85.23%. The
strength improvement culminated in a part with a UTS of 507 MPa. There was a small amount of
warpage in the resulting dog bones which was believed to be due to slight temperature differences in
the furnace chamber. This was evident in the 1300
o
C specimen stress-strain curves shown in Figure 6.32
and made it difficult to ascertain an accurate 0.02% offset yield strength.
111

Figure 6.30: ASTM E8 tensile test results for a hold temperature of 1200
o
C.

Figure 6.31: ASTM E8 tensile test results for a hold temperature of 1250
o
C.
112

Figure 6.32: ASTM E8 tensile test results for a hold temperature of 1300
o
C.

Figure 6.33: ASTM E8 tensile test results for all dog bone specimens.
113
A summary of the results of mechanical testing can be seen in Table 6.7.
Table 6.7: Summary of mechanical testing results.

Finally, the gear fabricated in the previous section was re-sintered at 1300
o
C for 8 hours and can be
seen in Figure 6.34. The linear shrinkage when measuring the diameter of the gear across all directions
was 17.4%, with a standard deviation of 0.003%. This shrinkage was greater than expected and was
believed to be due to the difference of gear placement and proximity to the heating elements within the
furnace. The measured relative density result was 87.47%, representing the first structural powder
metallurgy part fabricated with the SIS-metal process in accordance with ASTM B783-13. Figure 6.35 is a
visual representation of the final relative size of the gear (green) when compared to the original sintered
size (red).
114

Figure 6.34: SIS-metal fabricated gear after second sintering pass (scale in cm).

Figure 6.35: A visual representation of the difference in size between the two sintering passes.

115
Upon visual inspection, the re-sintered gear seemed to have maintained its shape after 17.4% linear
shrinkage. Table 6.8 illustrates the caliper-measured diameter of the gear across all five directions.
Table 6.8: Measured diameter of re-sintered gear in all directions.

116
7. Conclusion and Future Work
7.1 Conclusion
The scope of this research began with the identification of candidate ceramic inhibitors and their
precursors. Magnesium oxide and its precursor of magnesium nitrate were discovered to be effective
inhibitors in this regard. The research moved through an investigation of the effects of diluting the
successful inhibiting agent, observing that for low temperatures (1000
o
C – 1200
o
C), low concentrations
(>20%) of magnesium nitrate in water proved to inhibit sintered samples enough for mechanical
separation. Finally, sintering temperatures and hold times were investigated in order to improve
mechanical properties to a point where, according to ASTM B783-13, the sintered coupons could be
classified as structural powder metal parts. A relative density of 87.47% was achieved in an SIS-metal
fabricated part.
The proposed research was successful in establishing a method of printing by SIS a net-shape part with
high relative density, and hence mechanical properties from a 316 stainless steel powder material.
7.2 Future Work
While the current research proved successful in fabricating structural grade metallic parts, there is a
great deal to be learned about the inhibition concept as well as its application to alternate materials.
The following were identified as potential areas of future research in the SIS-metal process:
 While this research was able to produce a 2.5D net shape part with the SIS-metal process, a
great deal of research is needed to print more complex three-dimensional parts. This research
should be focused upon identifying additives that could serve as better binders during part
printing. The current inhibitor solution lacks green strength for part handling which causes total
part collapse.
117
 There are many industry applications interested in more exotic materials such as nickel- and
titanium- based alloys. Research should be conducted to determine whether the current
magnesium nitrate solution would serve as a good inhibitor for these materials as well.
 While the inhibition of stainless steel proved successful, it was noticed in a separate set of
experiments by the author that the identified inhibitor was not effective with all ferrous alloys.
Inhibition was attempted on a commercial grade tool steel from Carpenter powder products
(Type H13) with no observable inhibition (Figure 7.1). The effects of chemical composition of
ferrous alloys should be studied with respect to inhibition.

Figure 7.1: Various droplet tests on H13 alloy sintered at 1100
o
C with no observable inhibition.

 It was noticed by the author that in the case of using other inhibition precursors, the opposite of
the intended effect of inhibition could potentially have occurred. In other words, instead of
inhibiting sintering, the droplet may have actually enhanced sintering. This was evidenced by
the region surrounding the droplet mechanically separating during abrasive blasting before the
droplet region itself.

118
References
1. Allen, J., An Investigation into the Comparative Costs of Additive Manufacture vs. Machine from
Solid for Aero Engine Parts. 2006, DTIC Document.
2. Standard, A., F2792. 2012. Standard Terminology for Additive Manufacturing Technologies.
ASTM F2792-10e1.
3. Hopkinson, N. and P. Dickens, Analysis of rapid manufacturing - using layer manufacturing
processes for production. PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART
C-JOURNAL OF MECHANICAL ENGINEERING SCIENCE, 2003. 217(1): p. 31-39.
4. Tomlin, M. and J. Meyer. Topology optimization of an additive layer manufactured (ALM)
aerospace part. in The 7th Altair CAE Technology Conference, Gaydon, UK, 10th May. 2011.
5. Dehoff, R., et al., Case study: additive manufacturing of aerospace brackets. Advanced Materials
& Processes, 2013. 171: p. 19+.
6. Edwards, P., A. O'Conner, and M. Ramulu, Electron Beam Additive Manufacturing of Titanium
Components: Properties and Performance. Journal of Manufacturing Science and Engineering,
2013. 135(6): p. 061016-061016.
7. Grimm, T. 3D Printer Landscape: The View from 30,000 Feet. 2012 November 11, 2014];
Available from: http://www.engineering.com/3DPrinting/3DPrintingArticles/ArticleID/3916/3D-
Printer-Landscape-The-View-from-30000-Feet.aspx.
8. Khoshnevis, B., et al., SIS–a new SFF method based on powder sintering. Rapid Prototyping
Journal, 2003. 9(1): p. 30-36.
9. 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.
10. Asiabanpour, B., An experimental study of factors affecting the selective inhibition of sintering
process. 2003, University of Southern California: United States -- California. p. 168-168 p.
11. Khoshnevis, B., M. Yoozbashizadeh, and Y. Chen, Metallic part fabrication using selective
inhibition sintering (SIS). Rapid Prototyping Journal, 2012. 18(2): p. 144-153.
12. 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.
13. Davis, J.R., Copper and copper alloys. 2001: ASM international.
14. ASTMB823, Standar Specification for Materials for Copper Base Powder Metallurgy (PM)
Structural Parts. 2009, ASTM International: West Conshohocken, PA.
15. Khoshnevis, B., M. Mojdeh, and K. Palmer, Selective Inhibition Sintering for Fabrication of
Metallic Parts. IIE Annual Conference. Proceedings, 2004: p. 1-1.
16. Yoozbashizadeh, M., Metallic part fabrication with selective inhibition sintering (SIS) based on
microscopic mechanical inhibition. 2012, University of Southern California: Ann Arbor. p. 187.
17. Gibson, I., D.W. Rosen, and B. Stucker, Additive manufacturing technologies. 2010: Springer.
18. Wong, K.V. and A. Hernandez, A review of additive manufacturing. ISRN Mechanical Engineering,
2012. 2012.
19. Deckard, C.R., Method and apparatus for producing parts by selective sintering. 1989, Google
Patents.
20. EOSINT M 270, EOS, Editor. p. 1-2.
21. EOS Titanium Ti64 for EOSINT M 270 Systems (Titanium Version), EOS, Editor. p. 1-3.
22. Simchi, A., F. Petzoldt, and H. Pohl, On the development of direct metal laser sintering for rapid
tooling. Journal of Materials Processing Technology, 2003. 141(3): p. 319-328.
23. Sachs, E., et al., Three dimensional printing: rapid tooling and prototypes directly from a CAD
model. Journal of Manufacturing Science and Engineering, 1992. 114(4): p. 481-488.
119
24. Khoshnevis, B., Selective inhibition of bonding of power particles for layered fabrication of 3-D
objects. 2003, Google Patents.
25. Mojdeh, M., Metallic parts fabrication using the SIS process. 2005, University of Southern
California: Ann Arbor. p. 135.
26. Asiabanpour, B., K. Palmer, and B. Khoshnevis, Performance Factors in the Selective Inhibition of
Sintering Process. IIE Annual Conference. Proceedings, 2003: p. 1-7.
27. Asiabanpour, B. and B. Khoshnevis, Machine path generation for the SIS process. Robotics and
Computer-Integrated Manufacturing, 2004. 20(3): p. 167-175.
28. Ink Jet Printers Technical Brief. Epson America Inc. p. 1-3.
29. Inc, E.A., Epson Micro Piezo
TM
Print Head Technology. p. 1-7.
30. Pinwill, I., M. Edirisinghe, and M. Bevis, Development of temperature-heating rate diagrams for
the pyrolytic removal of binder used for powder injection moulding. Journal of materials science,
1992. 27(16): p. 4381-4388.
31. Tseng, W.J. and C.-K. Hsu, Cracking defect and porosity evolution during thermal debinding in
ceramic injection moldings. Ceramics international, 1999. 25(5): p. 461-466.
32. German, R., Powder metallurgy science, 1994. Metal Powder Industries Federation, Princeton,
NJ: p. 290-291.
33. 3DWP, Albert Einstein Bust 3D Scan. 2014, Thingiverse.
34. Frei, Mobius Chain Bracelet Assembly. GRABCAD.
35. Moon, J., et al., Ink ‐Jet Printing of Binders for Ceramic Components. Journal of the American
Ceramic Society, 2002. 85(4): p. 755-762.
36. Strijbos, S., Burning-out of a carbonaceous residue from a porous body. Chemical Engineering
Science, 1973. 28(1): p. 205-213.
37. Mistler, R.E., Tape casting: the basic process for meeting the needs of the electronics industry.
Am. Ceram. Soc. Bull., 1990. 69: p. 1022-1026.
38. Johnsson, A. Minimization of the Extraction Time for Injection-Moulded Ceramics. in Proc. Br.
Ceram. Soc. 1983.
39. Trunec, M. and J. Cihlář, Thermal debinding of injection moulded ceramics. Journal of the
European Ceramic Society, 1997. 17(2): p. 203-209.
40. Frank, B.D., Moulding ceramic composition. 1966, Google Patents.
41. Matar, S., et al., The effect of porosity development on the removal of organic vehicle from
ceramic or metal moldings. Journal of materials research, 1993. 8(03): p. 617-625.
42. Lombardo, S.J. and Z. Feng, Analytic method for the minimum time for binder removal from
three-dimensional porous green bodies. Journal of materials research, 2003. 18(11): p. 2717-
2723.
43. Hurtta, M., I. Pitkänen, and J. Knuutinen, Melting behaviour of D-sucrose, D-glucose and D-
fructose. Carbohydrate research, 2004. 339(13): p. 2267-2273.
44. Gaussling, T. Thermal Decomposition of Sucrose. 2008 December 15, 2014].
45. Roos, Y.H., et al., Comment on the Melting and Decomposition of Sugars. Journal of agricultural
and food chemistry, 2012. 60(41): p. 10359-10362.
46. Lee, J.W., L.C. Thomas, and S.J. Schmidt, Effects of heating conditions on the glass transition
parameters of amorphous sucrose produced by melt-quenching. Journal of agricultural and food
chemistry, 2011. 59(7): p. 3311-3319.
47. Fang, Z.Z., Sintering of advanced materials. 2010: Elsevier.
48. Gaal, P.S. and D.E. Apostolescu, Thermal Conductivity 24/Thermal Expansion 12. Vol. 24. 1999:
CRC Press.
49. German, R., Sintering: From Empirical Observations to Scientific Principles. 2014: Butterworth-
Heinemann.
120
50. ASTMB962, Standard Test Methods for Density of Compacted or Sintered Powder Metallurgy
(PM) Products Using Archimedes’ Principle. ASTM International: West Conshohocken, PA.
51. Trinks, W., Industrial furnaces. Vol. 1. 2003: John Wiley & Sons.
52. Chandler, H., Heat treater's guide: practices and procedures for nonferrous alloys. 1996: ASM
international.
53. German, R.M., Liquid phase sintering. 1985: Springer.
54. Neĭkov, O.D., Handbook of non-ferrous metal powders: technologies and applications. 2009,
Amsterdam ;Boston ;London: Elsevier.
55. Klar, E., and Prasan K. Samal, Powder metallurgy stainless steels: processing, microstructures,
and properties. 2007: ASM International.
56. Schade, C.T., Schaberl, J. W., & Lawley, A., Stainless steel AISI grades for PM applications.
International journal of powder metallurgy, 2008. 44(3): p. 57-67.
57. Near Net Shapes. 2012, Carpenter Powder Products. p. Near Net Shape Manufacturing Powders.
58. Davis, J.R., Stainless Steels. 1994: ASM International.
59. ASTMB783-13, Standard Specification for Materials for Ferrous Powder Metallurgy (PM)
Structural Parts. 2009, ASTM International: West Conshohocken, PA.
60. TIKKANEN, M.H., B.-O. ROSELL, and Ö. WIBERG, SINTERING OF METAL POWDER COMPACTS
CONTAINING CERAMIC OXIDES. Powder Metallurgy, 1962. 5(10): p. 49-60.
61. Tikkanen, M. and S. Ylasaari, ROLE OF DISLOCATIONS IN THE DENSIFICATION PROCESS. Physics
Sinter, 1969. 1(1).
62. Johnson, D.L., Particle drag retardation of surface smoothing and sintering. Journal of Materials
Science, 1976. 11(12): p. 2312-2318.
63. Ness, E., Sintering Inhibition of Copper Compacts With a Dispersed Oxide. Diss. Abstr. Int., 1989.
49(9): p. 179.
64. Ashby, M., et al., The influence of a dispersion of particles on the sintering of metal powders and
wires. Progress in Materials Science, 1980. 25(1): p. 1-34.
65. Maak, P., G. Weatherly, and K. Aust, The sintering behaviour of internally oxidized Ag Sb
powder. Materials Science and Engineering, 1980. 44(1): p. 139-146.
66. Lu, Y. and K. Hwang, Improved densification of carbonyl iron compacts by the addition of fine
alumina powders. Metallurgical and Materials Transactions A, 2000. 31(6): p. 1645-1652.
67. Razavi-Tousi, S., R. Yazdani-Rad, and S. Manafi, Effect of volume fraction and particle size of
alumina reinforcement on compaction and densification behavior of Al–Al 2 O 3
nanocomposites. Materials Science and Engineering: A, 2011. 528(3): p. 1105-1110.
68. Shukla, A.K., et al., Densification behavior and mechanical properties of Cu–Cr–Nb alloy powders.
Materials Science and Engineering: A, 2012. 551(0): p. 241-248.
69. Xu, X., W. Yi, and R.M. German, Densification and strength evolution in solid-state sintering Part
I Experimental investigation. Journal of materials science, 2002. 37(3): p. 567-575.
70. Melting temperature ranges for stainless steels 03-22-2016].
71. Jang, S., et al., Sintering of inkjet printed copper nanoparticles for flexible electronics. Scripta
Materialia, 2010. 62(5): p. 258-261.
72. 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.
73. Torabi, P., Enhancing the surface quality and dimensional accuracy of SIS-Metal, in Daniel J.
Epstein Department of Industrial and Systems Engineering. 2016, University of Southern
California.
74. ASTME8/E8M-15A, Standard Test Methods for Tension Testing of Metallic Materials. 2015,
ASTM International: West Conshohocken, PA.

Abstract (if available)

Abstract Selective Inhibition Sintering of metal alloys (SIS-metal) was previously proven effective in the additive manufacture (AM) of low resolution bronze parts. Recent advancements in the use of a high-precision inkjet print head represented an order of magnitude improvement in SIS-metal resolution. However, the fabrication of complex three-dimensional metallic parts required new SIS-metal compatible, cross-sectional image generation based on the part boundary profile. Multiple candidate layer-processing approaches were identified and validated for basic geometries. These approaches were chosen from previous research as well as preliminary investigations and were applied to a modified SIS-metal process for validation. The validation criteria were based upon the amount of powder waste produced, the ability to handle complex geometries, printing speed, extraction (post-processing) speed, and part integrity. Results are discussed for implementation of the five candidate layer processing approaches in the fabrication of basic shapes. An evaluation is presented for their use on more complex geometries. Two approaches were then chosen for the construction of more complex geometries, the results of which are presented. ❧ These approaches were then extended to higher temperature, ferrous alloys. A new inhibiting agent, Magnesium Nitrate, was identified and validated for use with a stainless steel alloy. Initial feasibility experiments were conducted to verify the inhibition concept, and the results were used to fabricate a stainless steel gear with the SIS-metal process. A set of experiments was conducted to realize higher density values and hence improve the mechanical properties of as-built parts. Dog bone specimens were tested in tension and a final tensile strength of 507 MPa (UTS) was achieved. A final sintered density of 87.47% was achieved in a SIS-metal fabricated stainless steel gear.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

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

PDF

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

PDF

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

PDF

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

PDF

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

PDF

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

PDF

Scalable polymerization additive manufacturing: principle and optimization

PDF

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

PDF

Experimental and numerical techniques to characterize structural properties of fresh concrete

PDF

Analyses of strength of layered structures fabricated by Contour Crafting

PDF

Hybrid vat photopolymerization: methods and systems

PDF

Contour crafting construction with sulfur concrete

PDF

Fabrication-aware machine learning for accuracy control in additive manufacturing

PDF

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

PDF

Multi-scale biomimetic structure fabrication based on immersed surface accumulation

PDF

Machine learning-driven deformation prediction and compensation for additive manufacturing

PDF

High temperature latent heat thermal energy storage to augment solar thermal propulsion for microsatellites

Asset Metadata

Creator Petros, Matthew R. (author)

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

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Industrial and Systems Engineering

Publication Date 07/26/2016

Defense Date 05/10/2016

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

Tag 3D printing,additive manufacturing,OAI-PMH Harvest,powder metallurgy,selective inhibition sintering,sintering,sis

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Khoshnevis, Behrokh (committee chair), Chen, Yong (committee member), Nutt, Stephen (committee member)

Creator Email mattpetros@gmail.com,petros@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-279478

Unique identifier UC11280610

Identifier etd-PetrosMatt-4619.pdf (filename),usctheses-c40-279478 (legacy record id)

Legacy Identifier etd-PetrosMatt-4619-1.pdf

Dmrecord 279478

Document Type Dissertation

Format application/pdf (imt)

Rights Petros, Matthew R.

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