Masters thesis of engineering mechanical properties and applications of additively manufactured thermoplastic polyurethane material

The mechanical properties of additively manufactured TPU materialcan be affected by various parameters, such as the build orientation, the mixratio of new and reused printing powder, and

Mechanical Properties and Applications of Additively Manufactured

Thermoplastic Polyurethane Material

A thesis submitted in fulfilment of the requirements for the degree of Master of Engineering

Tao Xu Bachelor of Engineering, Tongji University

School of Engineering College of Science, Technology, Engineering and Maths

RMIT University

August 2021

I certify that except where due acknowledgement has been made, the work is that of theauthor alone; the work has not been submitted previously, in whole or in part, to qualifyfor any other academic award; the content of the thesis is the result of work which has beencarried out since the official commencement date of the approved research program; anyeditorial work, paid or unpaid, carried out by a third party is acknowledged; and, ethicsprocedures and guidelines have been followed

Tao Xu

23 August 2021

First of all, I would like to give my heartfelt thanks to my senior supervisor, Professor

Yi Min (Mike) Xie, who provided me with the opportunity to study at RMIT and alwayshelped me with my research It has been a great privilege and joy to study under hisguidance and supervision I am particularly impressed by his sharp mind for detectingextremely detailed problems that exist in my research, which profoundly inspires me to be

a rigorous and responsible researcher Professor Xie is not only my academic leader butalso my life mentor I am sure that I will forever beneĄt from his direction and treasure itall of my life

My sincere and hearty thanks and appreciation also go to my associate supervisor, DrXiaoshan (Susanna) Lin After every dayŠs busy and exhausting work, she still devoted herconsiderate care and immense vigour in the supervision of my thesis writing, including thesuggestions on wording, the help in forming the structure, and the efforts to the reĄnement

of ideas in my thesis Without her keen insights and constant encouragement, the thesiswould not have been Ąnished

I would like to thank all my current and former colleagues I miss lunchtime withYulin Xiong, Hu Xu, Wei Li, Anbang Chen, Wenzheng Xu, Yunzhen he, Qi Cai, ZichengZhuang, Minghao Bi and Zhonggao Chen I wish to particularly thank Mr Dingwen Baofor giving me considerable care during my life in Melbourne

I am grateful for the support and company from my friends Zhouhan Jia, Hange Yang,Shangchi Yao, Rong Peng, Chenxing Zuo, Qi Lu, Jiajun Fan, Bo Fan, Mengge Cai, Zhi Li,

appreciation to my former teacher Mr Ercüment Görgül, who introduced me to ProfessorXie personally and offered me a lot of help and advice in my career.

I am thankful to Nanjing Ameba Engineering Structure Optimization Research tute for providing technical support I thank Mr Wei Shen for arranging the workspaceand experimental equipment and Mr Donghui Wang for helping me operate machines andconduct manufacturing

Insti-I thank my beloved girlfriend Yingqi Li, who has been with me and brought me less laughs and happiness We met in spring, however, when I saw her, a poem by Shake-speare appeared in my mind: "Shall I compare thee to a summerŠs day?"

count-I would like to express my appreciation for the funding for my research from the tralian Research Council (grant number: FL190100014)

Aus-Finally, but most importantly, I would like to thank my parents, who offer me tional support and encourage me to pursue my dreams I will keep your love in the quietestplace in my heart

1.1 Overview 1

1.1.1 Two Commonly Used Additive Manufacturing Technologies 2

1.1.2 Additively Manufactured TPU Material 5

1.2 Problem Statement 5

1.3 Objectives and Scope 7

1.4 Layout of Thesis 8

2 Literature Review 9 2.1 Chemical Properties of TPU 9

2.2 TPU Particles 11

2.3 Thermal Behaviour of TPU Material 12

2.4 Mechanical properties of TPU Material 13

3 Tensile Tests of TPU Dumb-Bell Specimens Printed with Various Parameters 16 3.1 Introduction 16

3.2 Materials and Testing Methods 17

3.2.1 Test Specimens 17

3.2.2 Processing Parameters 18

3.2.4 Scanning Electron Microscope 21

3.2.5 Particle-Size Distribution 22

3.2.6 Melt Flow Rate 23

3.2.7 Differential Scanning Calorimetry 24

3.3 Results and Discussion 25

3.3.1 Effect of Build Orientation 25

3.3.2 Effect of Post-processing 28

3.3.3 Effect of Powder Mix Ratio 30

3.4 Summary 35

4 A Novel Low-cost and Environmentally Friendly Method for Concrete Cast-ing UsCast-ing Additively Manufactured TPU Moulds 37 4.1 Introduction 37

4.2 Design Process 39

4.2.1 Topologically Optimized 2D cantilever 39

4.2.2 3D Free-form Column 44

4.3 Manufacturing Process 47

4.3.1 Mould Printing 47

4.3.2 Pre-processing of the TPU Moulds 47

4.3.3 Casting and Demoulding 52

4.4 Manufacturing WorkĆow 53

4.5 Discussion 55

5 Conclusions 58 Bibliography 61 Appendix A Python Program for 2D Topology Optimization 72 A.1 Boundary Conditions and Stiffness Matrix 72

A.2 Finite Element Solver 75

A.3 BESO Technique 77

A.4 Parameter Settings 83

List of Figures

1.1 Schematic of selective laser sintering 3

1.2 Schematic of fused deposition modelling 4

1.3 TPU materials for additive manufacturing: (a) TPU powder for SLS (im-age credit: SINTERIT); (b) TPU granules for FDM (im(im-age credit: Kingoda) and (c) TPU Ąlament for FDM (image credit: SUNLU) 6

2.1 Two-step synthesis of polyurethanes: (a) Synthesis of urethane prepoly-mer via reacting two diisocyanates with polyols (b) Synthesis of polyurethane via reacting the prepolymer with chain extenders 10

2.2 Particle shapes attainable by different powder generation processes (Schmid and Wegener [2016]) 11

2.3 Typical DSC-Thermogram with nature of Śsintering windowŠ as LS pro-cess temperature (Schmid and Wegener [2016]) 12

2.4 SEM image of an SLS part surface (Launhardt et al [2016]) 14

3.1 Shape of the dumb-bell test specimens 18

3.2 3D printed dumb-bell specimens 18

3.3 Three build orientations 20

3.4 Comparison of specimen surfaces: (a) specimen with post-processing and

LIST OF FIGURES

3.5 Tensile test setup 23

3.6 Illustration of sieve analysis 24

3.7 Stress-strain curves of specimens printed in three different directions 25

3.8 Average and standard deviation of the results: (a) Maximum tensile strength and (b) strain at break 26

3.9 Moduli of specimens printed in different orientations 27

3.10 Illustration of selective laser sintering (SLS) processing 28

3.11 Comparison of the average tensile strength ŤYŤ indicates specimens with post-processing, and ŤNŤ means specimens without post-processing 28

3.12 Comparison of the average strain at break ŤYŤ indicates specimens with post-processing, and ŤNŤ means specimens without post-processing 29

3.13 Comparison of moduli of specimens at different strains ŞYŤ indicates specimens with processing, and ŤNŤ means specimens without post-processing 31

3.14 Comparison of (a) maximum tensile strength and (b) strain at break 31

3.15 Microstructure of new TPU powder under different magniĄcations: (a) 500×; (b) 1000× and (c) 2000× 32

3.16 Microstructure of reused TPU powder under different magniĄcations: (a) 500×; (b) 1000× and (c) 2000× 33

3.17 Particle size distribution results for the new powder and the reused powder 34 3.18 Differential scanning calorimetry (DSC) traces of the new powder and the reused powder: (a) heating traces and (b) cooling traces 34

4.1 Boundary conditions of the cantilever 41

4.2 Topology optimization result 41

4.3 Smoothed shape of the topology optimization result 41

4.4 Extruded solid model 42

4.5 TPU Mould Design 42

4.6 Block design and cutting directions of the mould 43

4.7 Demolding sequence 43

4.8 Geometrical articulations of NerviŠs column 45

4.9 Mould design for the 3D column: white parts are the pieces of the mould, and the black part is NerviŠs column 46

LIST OF FIGURES

4.10 Six sections of the mould and the corresponding bases 46

4.11 Mould printing with FDM technique 48

4.12 Additively manufactured TPU mould of the 2D cantilever 48

4.13 Additively manufactured TPU mould of 3D column (top part) 49

4.14 Water-based putty on the surface of the mould 50

4.15 Machine oil (red) on the water-based putty 51

4.16 Pre-processing of the mould for 3D column (top part): (a) Cutting the mould into four pieces; (b) Appling TeĆon release paper on the internal surfaces; (c) Assembling the four pieces on a board 51

4.17 2D concrete cantilever 52

4.18 3D concrete column (top part) 53

4.19 Demoulding process of the 2D cantilever 54

4.20 After demoulding of the 2D cantilever: (a) TPU mould after demoulding; (b) 2D concrete cantilever; (c) Surface of the 2D concrete cantilever 55

4.21 Six sections of NerviŠs column 56

4.22 Assembled NerviŠs column: (a) Assembled column; (b) Surface of the column 57

List of Tables

3.1 Dimensions of the dumb-bell specimens 17

3.2 Processing parameters used in this study 19

3.3 Eight sets of specimens printed with different processing parameters 21

4.1 Comparison of water-based putty and oil-based putty 50

Thermoplastic polyurethane (TPU) is a polymer material that has high ity, good biocompatibility and excellent abrasion resistance These propertiesopen a pathway to manufacturing functional TPU parts for applications in var-ious Ąelds, such as aerospace engineering, medical devices and sports equip-ment The mechanical properties of additively manufactured TPU materialcan be affected by various parameters, such as the build orientation, the mixratio of new and reused printing powder, and whether the printed part is post-processed The settings of printer parameters, i.e., printing speed, printingpath, and processing temperature, have been the focuses of previous studies.However, the inĆuences of other factors have not been systematically investi-gated, which would limit the application of TPU material In this project, themechanical properties of additively manufactured TPU material affected bythree different processing parameters are Ąrstly investigated, including buildorientation, mix ratios of the new and reused powder and post-processing.Then, additively manufactured TPU moulds are applied to cast complex con-crete structures

ductil-A series of tensile tests are conducted on TPU dumb-bell specimens It

is found that the mix ratio of the new and reused powder is the most criticalfactor in the mechanical properties of the printed TPU parts Compared toreused powder, new powder has plumper particle and properer thermal be-haviour that is more suitable for the additive manufacturing process Besides,TPU parts printed in Ćat and on-edge orientations show better tensile strengthand deformability than those printed in upright orientation In addition, post-processing is found to enhance the deformability of TPU parts by more than40%

Once the mechanical properties of printed TPU are characterised, a novel,low cost, and environmentally friendly method for concrete casting using printed

is developed to cast complex concrete structures A planar topologically mized cantilever and a free-form 3D column are cast using the proposed TPUmould Good surface quality is obtained for the cast concrete elements, andthe moulds can be reused many times after cleaning.

opti-The Ąndings of this thesis will provide clear guidelines for the selection ofprocessing parameters for additively manufactured TPU parts and the method

of casting complicated concrete structures with additively manufactured TPUmould

of complex geometries so as to realize the Şfree manufacturingŤ of parts (Agarwala et al.

[1995]; Kruth et al [2005]; Lamikiz et al [2007]; Calignano et al [2017]) AM hasbeen increasingly employed to manufacture functional parts in various areas, includingaerospace (Barroqueiro et al.[2019]), medical engineering (Berry et al.[1997];Liu et al

[2019]) and sports (Mărieş et al.[2008])

Three types of materials can be used in additive manufacturing: polymers, ceramicsand metals Thermoplastic polyurethane (TPU) is an additively manufacturable polymer

1.1 Overview

material that has high ductility, good hydrolysis resistance, excellent biocompatibility andexcellent abrasion resistance (Lu et al.[2003];Li et al.[2008];Ma and Yang[2008];Maand Yang [2008]; Aurilia et al.[2011]; Feng and Ye [2011];Mi et al [2013]; Lee et al

[2019]) It can be used in structures that require high ductility, such as energy-absorbingstructures (Bates et al [2016]) and wearable devices (Scarpello et al [2012]; Li et al

mate-Selective Laser Sintering

SLS uses a laser as its energy source and powder as the primary processing material.Figure1.1shows the schematic of SLS The platform on the left gradually rises, lifting thepowder to the printing height Then the roller pushes the powder into the printing chamberwhile the printing platform gradually lowers A portion of the powder is then irradiated

by a high-energy laser ray and sintered together, while the rest plays a role in supportingthe sintered part

The most prominent advantage of SLS technology is that no supporting structure isrequired The suspended structure of the printed part can be supported by the un-sintered

1.1 Overview

Figure 1.1: Schematic of selective laser sintering

powder Therefore, SLS can directly manufacture structures with complex geometries.High material utilization is another signiĄcant advantage of SLS technology Un-sinteredmaterials can be reused in subsequent production, so there is little material waste How-ever, the reuse of un-sintered materials will reduce the mechanical properties of the printedparts due to the inferior particle quality and thermal properties

Since the raw material of the SLS process is powdery, the sintered part always has

a grainy surface If a high-quality surface is required, then post-processing is needed toimprove the surface smoothness of the printed part

Fused Deposition Modelling

FDM is a method of heating and melting Ąlament and printing parts using the melted terial Figure1.2shows the schematic of FDM The heating nozzle moves in the XY planeaccording to the cross-sectional shape of the printed parts The thermoplastic Ąlament

ma-1.1 Overview

melted into a semi-liquid state in the nozzle The material is then extruded by the nozzleand selectively coated on the platform After the processing layer is formed, the nozzlerises to a certain height and then prints on the next layer

Figure 1.2: Schematic of fused deposition modelling

Compared with the SLS technology, the FDM technology uses a heated nozzle instead

of a laser, so the equipment cost is low Also, the Ąlament material can be effectivelyutilized with little waste, which signiĄcantly reduces the printing cost In addition, there

is no poisonous gas or chemical pollution during the printing process The installationand application of FDM equipment are relatively simple and very Ćexible It can be easilyassembled as a desktop-level device It can also be assembled onto a robotic arm to printlarger components The printing speed of FDM technology is fast, and nozzles of differentdiameters can be used to control the printing speed

However, more apparent grooves can be found on the surface of the prototype printed

1.2 Problem Statement

by FDM technology due to the layer-by-layer process The laminated structure also results

in low strength in the direction perpendicular to the cross-section (Chacón et al.[2017]).Besides, supporting structures are required for the hanging parts of the printed compo-nents, making post-processing more difficult

1.1.2 Additively Manufactured TPU Material

TPU materials in different forms can be used in different additive manufacturing niques Usually, TPU powder is used for SLS, and TPU granules and TPU Ąlament areemployed for FDM The TPU powder, TPU granules and TPU Ąlament are shown in Figure

tech-1.3

As shown in Figure1.3a, TPU powder used in SLS technology has a very small particlesize The diameter of the powder is usually between 20 µm and 120 µm The colour ofTPU powder is usually white TPU granules are transparent elliptical particles that have

a transparent white colour (Figure 1.3b) The size of the TPU granule particles can beselected according to actual needs TPU wire is usually coiled on a roll (Figure 1.3c),which is very easy to transport and is low in price Both TPU granules and Ąlament can beused for the FDM process The most signiĄcant advantage of granules is that the materialcan be replenished during the printing process The printing process will be interruptedwhen the Ąlament is used up, which may lead to an unsuccessful printing process

1.2 Problem Statement

Figure 1.3: TPU materials for additive manufacturing: (a) TPU powder for SLS

(image credit: SINTERIT); (b) TPU granules for FDM (image credit: Kingoda)

and (c) TPU Ąlament for FDM (image credit: SUNLU)

other factors will also inĆuence the mechanical properties of the printed TPU parts First

of all, due to the layer-by-layer process of SLS, the mechanical properties of the printedparts are anisotropic The TPU parts with different printing directions will have differentmechanical performances Secondly, post-processing of the printed part will change itscomposition, which will affect its mechanical properties In addition, in the SLS process,the powders used for supporting the printed parts are exposed to ambient temperature Thereuse of these powders will affect the mechanical properties of the printed parts Apartfrom the mechanical properties, the possibility of employing TPU material in civil engi-neering has not been studied previously The tasks of this thesis can be deĄned by threecore questions:

1 What is the inĆuence of print orientations and post-processing methods on the chanical properties of SLS manufactured TPU parts?

me-2 What is the difference between the new TPU powder and the reused TPU powder,and how do they affect the manufacturing process

1.3 Objectives and Scope

3 Is it possible to employ additively manufactured TPU moulds to fabricate complexconcrete structures?

The answers to these questions were provided by a series of experimental tests on thetensile tests of additively manufactured TPU specimens Then, two free-form concretemembers were fabricated using 3D printed TPU material

1.3 Objectives and Scope

The Ąrst objective of this project is to investigate the mechanical properties of SLS ufactured TPU material affected by three different processing parameters, including buildorientation, mix ratios of the new and reused powder and post-processing A series ofexperimental tests are conducted on TPU dumb-bell specimens manufactured with an in-dustrial SLS printer to achieve this goal The effects of the three processing parameters onthe mechanical properties of the 3D printed TPU parts are investigated Besides, the mi-crostructure, particle-size distribution, viscosity and thermal properties of new and reusedTPU powder are measured

man-The second objective is to propose a novel, low-cost, and environmentally friendlymethod for casting concrete structures with complex shapes using TPU moulds printedwith the FDM technique A topologically optimized 2D cantilever and a 3D free-formcolumn are then cast by using this method

It is worth mentioning that although this thesis mainly focuses on SLS technology, bothSLS and FDM techniques are employed in 3D printing TPU FDM technique is employed

in Chapter4 for casting concrete components because of its low cost, in addition to itsability to print larger components The print size capability of SLS is limited by the size

of the print chamber, while the nozzles used in FDM technology can be mounted on the

Chapter 3 presents an experimental study on the effects of three processing eters on the mechanical properties of TPU parts manufactured by SLS, including buildorientation, mix ratios of the new powder and the reused powder, and post-processing.Furthermore, the microstructure and the thermal behaviour of TPU powder are charac-terized based on the test results obtained from scanning electron microscope (SEM) anddifferential scanning calorimetry (DSC).

param-Chapter4proposes a novel, low-cost, and environmentally friendly method for castingconcrete structures with additively manufactured TPU mould A topologically optimized2D cantilever and a 3D free-form column are manufactured using the new TPU mould Toachieve good concrete surface quality and easy demoulding, several demoulding methodsare also proposed in this chapter

Chapter5summarises the major conclusions and Ąndings of this thesis

Chapter 2

Literature Review

Thermoplastic polyurethane (TPU) is a material that is the bridge between rubber andplastics It is an additively manufacturable polymer material that has high ductility, goodbiocompatibility and excellent abrasion resistance These properties open a pathway tomanufacturing functional TPU parts for applications in various Ąelds such as aerospaceengineering, medical devices and sports equipment In this chapter, the chemical, thermaland physical properties of TPU material are introduced

2.1 Chemical Properties of TPU

TPU is composed of linear polymer chains with block structures (Coleman et al.[1990]).These chains contain relatively long low-polarity segments (soft segments) that alternatewith shorter but high-polarity segments (hard segments)

A common method of synthesizing TPU is via reacting polyols and diisocyanates,

2.1 Chemical Properties of TPU

2.1shows a two-step method of synthesizing polyurethane material proposed byDatta andWłoch[2016] The Ąrst step is to obtain urethane-prepolymer terminated with isocyanategroups by reacting diisocyanates and polyols (Figure2.1a) Polyurethane is then obtained

by reacting the semi-product from the Ąrst step with a low molecular weight chain extender(Figure2.1b) The purposes of reacting the prepolymer with low molecular weight chainextenders are to build polyurethane molecular weight and increase the block length of thehard segment

Figure 2.1: Two-step synthesis of polyurethanes: (a) Synthesis of urethane

prepolymer via reacting two diisocyanates with polyols (b) Synthesis of

polyurethane via reacting the prepolymer with chain extenders

The polarity of diisocyanates creates a strong attraction between them, leading to a highdegree of aggregation and order, forming crystalline or pseudo-crystalline regions in thesoft polyol matrix (Koerner et al [2008]) The crystalline or pseudo-crystalline regionsact as physical crosslinks, which cause a high level of elasticity of TPU, while polyolsimpact the polymer elongation properties (Kojio et al.[2020]) However, these "pseudo-crosslinks" will disappear under the heat effect, which accounts for the thermoplastic prop-erties of TPU Consequently, the classic injection moulding and extrusion methods are

[2012]) Cryogenic milled particles have the lowest cost but the poorest powder bility With the improvement of the production method in recent years, cryogenic milledparticles start to be used for SLS processing and have been studied by many researchers(Van den Eynde et al.[2015];Dadbakhsh et al.[2016];Verbelen et al.[2017]).

Ćowa-Figure 2.2: Particle shapes attainable by different powder generation processes

(Schmid and Wegener[2016])

2.3 Thermal Behaviour of TPU Material

Figure 2.3: Typical DSC-Thermogram with nature of Śsintering windowŠ as LS

process temperature (Schmid and Wegener[2016])

2.3 Thermal Behaviour of TPU Material

The thermal behaviour of TPU powder plays a decisive role in the additive ing process (Haponiuk et al.[1990]) Because of its thermoplasticity, TPU can be used

manufactur-in the SLS process Thermal properties of TPU can be measured by differential scannmanufactur-ingelectron microscope (Frick and Rochman[2004];Schmid and Wegener[2016]) A typicalDSC thermogram is shown in Figure2.3 In the SLS process, the laser beam is used to

selectively sinter the powder, that is, to heat the powder to the melting point (T m, heatingcurve, Figure 2.3) The heater in the SLS machine provides an appropriate processingtemperature (SLS sintering window, Figure2.3) to suppress the crystallization of the pow-der melted by the laser as long as possible so that the sintered layers could better adhere

2.4 Mechanical properties of TPU Material

to the next processing layer Then the sintered material cools to the crystallization point

(T c, cooling curve, Figure2.3) and solidiĄes

2.4 Mechanical properties of TPU Material

In the past, the effect of TPU powder on the mechanical properties of printed parts hasbeen studied by many researchers Dadbakhsh et al.[2016] investigated the inĆuences ofthe size and shape of TPU powder on SLS processability and mechanical properties ofprinted TPU parts The effects of powder bed temperature, laser energy density, powderparticle size and powder composition on the properties of laser-sintered TPU were stud-ied by Gan et al [2019] The powder Ćow, rheology of melt, shrinkage and hardeningbehaviour involved in TPU sintering were examined byVerbelen et al.[2017] through theanalysis of four different TPU grades A systematic evaluation of the physical and thermalproperties of TPU powder for SLS was presented byYuan et al.[2017] In all these studies,the mechanical properties of additively manufactured TPU parts were found to be highlydependent on the powder quality, especially the size and thermal behaviour of particles.Build orientation also has a signiĄcant inĆuence on the additively manufactured partssince it directly affects the mechanical properties, build time and cost of printed parts(Zhang et al [2017]) Barba et al [2020] studied the effect of build orientation on themechanical properties of additively manufactured metal parts Chacón et al.[2017] char-acterized the effect of build orientation on polylactide (PLA) samples However, the effect

of build orientation on the additively manufactured TPU parts has not been studied ously

previ-Figure2.4 illustrates a Scanning Electron Microscope (SEM) image of the surface of

an SLS part It can be seen from the Ągure that the surface roughness of the

manufac-2.4 Mechanical properties of TPU Material

tured components is signiĄcantly high since the powder is used as the printing material inSLS (Ramos and Bourell[2002]) Different techniques have been proposed in the past toimprove the surface quality of the SLS part Ramos and Bourell[2002] proposed a highlaser power polishing technique for smoothing the surface of the printed part By usingthis technique, the grainy SLS part surface could be melted by CO2 and Nd: YAG lasers

at high scanning speed Galantucci et al.[2009] investigated the effect of chemical ment using acetone, ester and chloride solvents on additively manufactured material Thepost-processed part has a signiĄcantly improved surface quality The chemical treatmenthas been widely employed in industries since it does not require human intervention, and

treat-it signiĄcantly improves the surface Ąnish at the expense of negligible part size changes(Kumbhar and Mulay[2018]) So far, very few studies have been reported on the effect ofpost-processing on the mechanical properties of printed parts However, it shall be notedthat the composition of the printed part may be changed by chemical treatment, leading

to the changes of mechanical properties During the SLS process, the un-sintered powder

Figure 2.4: SEM image of an SLS part surface (Launhardt et al.[2016])

is exposed to Ąeld temperature and serves as a support for the printed part In order tosave material, after removing the printed part, the un-sintered powder is generally reused

2.4 Mechanical properties of TPU Material

as printing material (Kumar [2003]) Dadbakhsh et al [2016] investigated the effect ofreused polyamide powder on the microstructure and mechanical behaviour of SLS parts

It is found that the new particles have a relatively spherical shape The reused particleshave similar sizes, but apparent cracks can be observed on the surface Besides, the SLSparts printed using the new powder show higher tensile strength and shear strength com-pared with parts printed with mixed and reused powder Plummer et al [2012] studiedthe recyclability of TPU powder used in SLS No signiĄcant trend was found in the TPUparticle size, and the thermal properties were unchanged

3.2 Materials and Testing Methods

3.2 Materials and Testing Methods

A series of experimental tests are conducted on TPU dumb-bell specimens manufacturedwith an industrial SLS printer The specimens are printed with three different processingparameters, i.e build orientation, post-processing method and mix ratio of new and reusedpowder Besides, the microstructure images of TPU particles are obtained from a scanningelectron microscope (SEM) Particle-size distribution and viscosity are measured by sieveanalysis and differential melt Ćow rate (MFR), respectively The thermal behaviour of TPUpowder is characterized based on the test results from differential scanning calorimetry

3.2.1 Test Specimens

In this study, all specimens were printed using an industrial SLS machine (Farsoon-HS403,Changsha, China) with the TPU powder type WANFAB-PU95AN According to ISO37standard: Rubber, vulcanized or thermoplasticŮDetermination of tensile stress-strain prop-erties (International Organization for Standardization[2017]), the outline of the dumb-bellpieces for tensile tests is shown in Figure3.1 The dimensions of the dumb-bell are sum-marized in Table3.1 Figure3.2shows the 3D printed dumb-bell specimens

Table 3.1: Dimensions of the dumb-bell specimens

Dimension value (mm)

A Overall length (minimum) 115

B Width of ends 25 ± 1

C Length of the narrow portion 33 ± 2

D Width of the narrow portion 6.2 ± 0.2

E Transition radius outside 14 ± 1

F Transition radius inside 25 ± 2

Test length 25 ± 0.5Thickness 2 ± 0.2

3.2 Materials and Testing Methods

Figure 3.1: Shape of the dumb-bell test specimens

Figure 3.2: 3D printed dumb-bell specimens

3.2.2 Processing Parameters

The processing parameters investigated in this study include build orientation, post-processingand mix ratio, which are listed in Table 3.2 In this study, the SLS machine settings aredetermined according to the recommendations of the TPU powder manufacturer, and theyare consistent for all samples The laser power is 55 W The scanning speed and the scan-

3.2 Materials and Testing Methods

ning space are set at 15.2 mm/s and 0.1 mm, respectively

Table 3.2: Processing parameters used in this study

Parameters value

Build orientation Flat, on-edge, uprightPost-processing method Chemical treatment, nonePowder mix ratio (new

powder: reused powder) 3 : 7; 5 : 5; 10 : 0

Build Orientation

Build orientation refers to the orientation of a printed part on the print platform Theprinted part may show an anisotropic property due to the layer-by-layer printing process.Thus it is essential to analyse the effect of build orientation on the mechanical properties

of 3D printed TPU parts In this study, three build orientations are assessed: Ćat, on-edgeand upright (Figure 3.3) The Ćat specimens are printed along the thickness direction,whereas the on-edge and upright specimens are printed along the width and length direc-tions, respectively

Post-processing Method

During post-processing, the chemical solvent would react with the surface of the printedpart, which may change the composition of the part Therefore, it is necessary to study theeffect of post-processing on the mechanical properties of 3D printed TPU parts In thisstudy, the chemical treatment is performed by immersing the printed part into an amidesolvent for 3 min Figure3.4shows the surfaces of the specimens with and without post-processing As can be seen, the specimen with post-processing has a white and smoothsurface, whereas the specimen without post-processing has a yellowish-white and grainy

3.2 Materials and Testing Methods

Figure 3.3: Three build orientations

Powder Mix Ratio

After SLS processing, the un-sintered powder is often recycled and reused As TPU ticles may stick together during the SLS process, the large particles are Ąltered out fromthe un-sintered powder using a sieve with a mesh size of 0.178 mm Also, the temperature

par-in the SLS machpar-ine may affect the thermoplastic properties of the un-spar-intered powder Inthis study, three mix ratios commonly utilized in the industry are investigated: 30% newpowder and 70% reused powder, 50% new powder and 50% reused powder, and 100%new powder A total of 8 sets of specimens are printed with ten specimens in each set.Details of the test specimens are summarized in Table3.3

3.2 Materials and Testing Methods

Table 3.3: Eight sets of specimens printed with different processing parameters

Set Number Build Orientation Post-Processing

Method

Powder Mix Ratio (New Powder: Reused Powder)

1 Flat Chemical treatment 3 : 7

2 On-edge Chemical treatment 3 : 7

3 Upright Chemical treatment 3 : 7

3.2.3 Tensile Test Setup

A 10 kN universal testing machine with a 250 N load cell is used for tensile tests Theelongation of the length of the specimen is measured by an extensometer During the test,

a small prestress (about 2 N) is applied to the specimen to avoid bending The tensile load

is applied under displacement control at a constant rate of 100 mm/min (Figure3.5)

3.2.4 Scanning Electron Microscope

Scanning electron microscopes (SEM) use a focused electron beam to scan the surfaces

of tiny structures to provide detailed images of their microstructures According to theliterature (Cheng et al.[2005];Schmid and Wegener [2016]), to achieve an almost free-Ćowing behaviour during the SLS process, the ideal shape of TPU powder is plump andspherical It is believed that a better powder shape could improve the quality of the printedparts In this study, a scanning electron microscope (SEM, Tescan-MIRA 3, Kohoutovice,Czech Republic) is used for shape analysis of the new powder and the reused powder

3.2 Materials and Testing Methods

Figure 3.4: Comparison of specimen surfaces: (a) specimen with

post-processing and (b) specimen without post-post-processing

3.2.5 Particle-Size Distribution

Particle-size distribution (PSD) is an index indicating the proportion of particle sizes in asample particle group The PSD of TPU powder is very important because the narrower theparticle size distribution, the better the uniformity of the powder and the better the quality

of the printed part (Plummer et al.[2012]) One of the most commonly used methods formeasuring PSD is sieve analysis (or gradation test), which contains a series of sieves withdecreasing mesh sizes The particles pass through the sieves, and the weight of materialmaintained at each level of sieving is measured (Figure3.6) In this study, a sieve analysis(Mastersizer-2000, Malvern, UK) is carried out to obtain quantitative information aboutthe difference in the particle size distribution between the new powder and the reusedpowder

3.2 Materials and Testing Methods

Figure 3.5: Tensile test setup

3.2.6 Melt Flow Rate

Melt Flow Rate (MFR) is a parameter that indicates the ease of the Ćow of the meltedthermoplastics It determines the viscosity and quality of the thermoplastics To obtainMFR, the melted TPU material would Ćow through a speciĄc capillary at a prescribedtemperature and pressure, and the weight of the material Ćowing through the capillarywithin 10 min is measured (melt Ćow indexer: XNR-400A, Qingdao, China) According toASTM standard 1238 (Standard test method for melt Ćow rates of thermoplastics) (ASTMInternational[2020]), the test conditions for TPU are 230 ℃/2.16 kg

3.2 Materials and Testing Methods

Figure 3.6: Illustration of sieve analysis

3.2.7 Differential Scanning Calorimetry

TPU can be used in the SLS process because of its thermoplasticity The thermal ties of the TPU powder directly determine the ease of the sintering process Differentialscanning calorimetry (DSC, Mettler Toledo-Q2000, Columbus, USA) could be used tomeasure the thermal properties of the TPU powder In this study, the thermal properties

proper-of both new powder and reused powder are measured by DSC The tests are carried outunder a nitrogen atmosphere to simulate the environment of SLS processing The powder

is heated from 30 ℃at a rate of 50 ℃/min and then cooled from 240 to Ű70 ℃at a rate of

10 ℃/min

3.3 Results and Discussion

3.3 Results and Discussion

3.3.1 Effect of Build Orientation

The stress-strain relationships obtained from the tensile tests for sets 1, 2 and 3 are pared in Figure3.7

com-Figure 3.7: Stress-strain curves of specimens printed in three different

direc-tions

As can be seen in Figure3.7, all stress-strain curves demonstrate a non-linear behaviour

of SLS printed TPU The average maximum tensile strengths of specimens printed in Ćatand on-edge orientations are similar (6.3 MPa and 6.7 MPa, respectively) The specimensprinted in on-edge orientation have a larger strain capacity with an average strain at break

of 125.3%, whereas it is 110.5% for specimens printed in Ćat orientation Compared to Ćat

3.3 Results and Discussion

and on-edge orientations, the specimens printed in upright orientation show a signiĄcantdrop in both tensile strength and strain capacity The average maximum tensile strength

of the specimens printed in upright orientation is only 3.7 MPa, and the strain at break

is 58.4% Figure3.8presents the average and standard deviation of the maximum tensilestrength and strain at break obtained from the experimental tests It is worth noting thatthe sample printed in Ćat orientation contains the least number of layers, resulting to thelarger variations in the test results than the others

Figure 3.8: Average and standard deviation of the results: (a) Maximum tensile

strength and (b) strain at break

Since TPU is highly non-linear even at a low strain level, the moduli of printed TPUspecimens at different strains are calculated and shown in Figure3.9 It can be seen that themoduli of specimens printed in Ćat and on-edge orientations are very close to each otherwith slightly larger values for the specimen printed in on-edge orientation when the strain

is at extremely small and extremely large levels The modulus of specimens printed inupright orientation is signiĄcantly lower than those printed in Ćat orientation and on-edgeorientation The similarity of Ćat and on-edge directions is due to the fact that their moduliare determined by the mechanical properties of each layer, while the smaller modulus of

3.3 Results and Discussion

upright orientation is due to the smaller bonding force between the layers

Figure 3.9: Moduli of specimens printed in different orientations

The tensile test results are consistent with the studies conducted by Chacón et al

[2017], where the specimens were printed using the fused deposition modelling (FDM)technique The lower tensile strength, stiffness and deformability obtained for the spec-imens printed in upright orientation are attributed to the weak bond between the printedlayers In this study, the laser melts the powder in the processing layer, as illustrated inFigure3.10 As heat is conducted from the processing layer to the sintered layer, a smallarea of the sintered part is re-melted and bonded with the molten material Thus, the in-terlayer bonding strength is much higher than the bonding strength between the layers

In the tensile test, the specimens printed in upright orientation are pulled in a directionperpendicular to the layers, leading to premature bond failure

3.3 Results and Discussion

Figure 3.10: Illustration of selective laser sintering (SLS) processing

3.3.2 Effect of Post-processing

The average tensile test results of the specimens without post-processing (sets 4, 5 and 6)and those with post-processing (sets 1, 2 and 3) are compared in Figure3.11and Figure

3.12

Figure 3.11: Comparison of the average tensile strength ŤYŤ indicates

speci-mens with post-processing, and ŤNŤ means specispeci-mens without post-processing