Investigation of Material and Mechanical Properties of Laser Beam Welded (LBW) Laser Powder Bed Fusion (LPBF) and Wrought Titanium Alloy Samples
A thesis submitted in fulfilment of the requirement for the degree of Master of Engineering
Ali Tamaddon
B. Sc. (Mechanical Engineering), Azad University, Tehran
P. E. (Institute of Engineers Australia), Canberra, ACT
School of Engineering
College of Science, Technology, Engineering and Mathematics
RMIT University
November 2021
Declaration I certify that except where due acknowledgement has been made, this research is that of the
author alone; the content of this research submission is the result of work which has been
carried out since the official commencement date of the approved research program; any
editorial work, paid or unpaid, carried out by a third party is acknowledged; and, ethics
procedures and guidelines have been followed.
In addition, I certify that this submission contains no material previously submitted for award
of any qualification at any other university or institution, unless approved for a joint-award
with another institution, and acknowledge that no part of this work will, in the future, be used
in a submission in my name, for any other qualification in any university or other tertiary
institution without the prior approval of the University, and where applicable, any partner
institution responsible for the joint-award of this degree.
I acknowledge that copyright of any published works contained within this thesis resides with
the copyright holder(s) of those works.
I give permission for the digital version of my research submission to be made available on
the web, via the University’s digital research repository, unless permission has been granted
by the University to restrict access for a period of time.
I acknowledge the support I have received for my research through the provision of an
Australian Government Research Training Program Scholarship.
Ali Tamaddon
Page | ii
Ali Tamaddon, 19/11/2021
Acknowledgements: This is to acknowledge the start of a new chapter in my life and a career in research.
I would like to express my sincere gratitude to Professor Sabu John and Distinguished
Professor Milan Brandt for their unequivocal support, guidance, and mentorship, which will
never be forgotten. Their invaluable knowledge and wisdom have been a beacon of reckoning
throughout these years. I would like to thank key RMIT staff that assisted me throughout this
research: Mr. Alan Jones of the Advanced Manufacturing Precinct (AMP) and Dr. Matthew
Field and Dr. Edwin Mayes from the RMIT Microscopy and Microanalysis Facility (RMMF) and
Dr. Wei Qian Song from the RMIT Bundoora East Material Testing Laboratories.
I would also like to thank my dearest son Mr. Daniel Tamaddon for his patience and
understanding during this period and Dr. Mohammad Mehdizadeh who encouraged me to
start this journey. A list of many good friends who supported me throughout this process
would be too long for this limited space and includes but is not limited to Dr. Joe Elambasseril
Ali Tamaddon
Page | iii
and Dr. Nabi Chowdhury to name a few.
Table of Contents Declaration .............................................................................................................................................. ii
Acknowledgements:............................................................................................................................... iii
Table of Contents ................................................................................................................................... iv
List of Tables ........................................................................................................................................ viii
List of Figures ......................................................................................................................................... ix
Abstract ................................................................................................................................................... 1
1.
Introduction .................................................................................................................................... 3
1.1.
Background ............................................................................................................................. 3
1.2. Motivation ............................................................................................................................... 5
1.3.
Project Objectives ................................................................................................................... 6
1.3.1. Aim .......................................................................................................................................... 6
1.3.2. Research Questions ................................................................................................................ 7
1.3.3. Research Scope ....................................................................................................................... 7
1.4.
Thesis Outline .......................................................................................................................... 8
1.5.
References .............................................................................................................................. 9
2. Literature Review .............................................................................................................................. 10
2.1
Chapter overview .................................................................................................................. 10
2.2
Overview of additive manufacturing (AM) ........................................................................... 10
2.3 Metal additive manufacturing .............................................................................................. 13
2.4
Laser powder bed fusion (LPBF) ............................................................................................ 15
2.4.1 Powder feedstock ................................................................................................................. 17
2.4.2 Laser fusion ........................................................................................................................... 18
2.4.3 Build shielding gas ................................................................................................................. 20
2.5
Laser beam welding (LBW) .................................................................................................... 23
2.5.1
Introduction to lasers ............................................................................................................ 23
2.5.2 Disk lasers.............................................................................................................................. 24
2.5.3 Laser welding ........................................................................................................................ 24
2.6 Welding considerations ........................................................................................................ 28
2.7 Microscopy ............................................................................................................................ 29
Research Questions .............................................................................................................. 30
2.8
References ............................................................................................................................ 31
2.9
3. Materials and Methodology ......................................................................................................... 35
Introduction .......................................................................................................................... 35
3.1
First Phase: Pilot Samples ..................................................................................................... 35
3.2
3.2.1 Design of experiments .......................................................................................................... 35
Ali Tamaddon
Page | iv
3.2.2 Material: ................................................................................................................................ 36
3.2.3
Fabrication of Laser Powder Bed Fusion (LPBF) parts .................................................. 38
3.2.4
Laser Beam Welding...................................................................................................... 41
3.2.5
Cross section and metallographic preparation ............................................................. 44
3.2.6
Microscopy (area, shape) .............................................................................................. 47
3.2.7 Weld porosity studies ................................................................................................... 47
3.2.8
EDS (chemical composition) .......................................................................................... 49
3.2.9
Microscopy (EBSD) ........................................................................................................ 50
3.2.10 Micro hardness ............................................................................................................. 52
3.3
Second Phase: Main Samples................................................................................................ 53
3.3.1 Design of experiments .......................................................................................................... 53
3.3.2 Fabrication of parts ............................................................................................................... 55
3.3.3 Pre-testing measurement and testing: ................................................................................. 59
3.3.4 Tensile test set up: ................................................................................................................ 59
3.3.5 Fatigue test set up: ............................................................................................................... 62
3.3.6 Post-test examinations: ........................................................................................................ 64
3.4
References ............................................................................................................................ 65
4. Results and Discussion .................................................................................................................. 66
4.1
Introduction .......................................................................................................................... 66
4.2
Pilot samples ......................................................................................................................... 66
4.2.1 Macroscopy ........................................................................................................................... 66
4.2.1.1 Weld appearance and comparison to AWI standards ...................................................... 66
4.2.2 Optical Microscopy ............................................................................................................... 68
4.2.2.1 Weld area measurement, shape, and penetration profile ............................................... 68
4.2.3 Scanning Electron Microscopy .............................................................................................. 73
4.2.3.1 EDS and chemical composition results ............................................................................. 73
4.2.3.2 Microstructure and EBSD results ...................................................................................... 75
4.2.4 Porosity analysis .................................................................................................................... 79
4.2.5 Hardness profile .................................................................................................................... 85
4.3 Main samples ........................................................................................................................ 89
4.3.1 Design of experiments .......................................................................................................... 89
4.3.2 Tensile tests .......................................................................................................................... 90
4.3.2.1
Tensile test setup .......................................................................................................... 90
4.3.2.2
Single piece wrought and LPBF tensile tests ................................................................. 90
4.3.2.3 Porosity analysis – Single piece wrought and LPBF ...................................................... 91
4.3.2.4 Welded wrought to LPBF assembly tensile tests ........................................................ 100
Ali Tamaddon
Page | v
4.3.2.4.1 Pre-test porosity analysis – Wrought to LPBF ............................................................. 100
4.3.2.4.2 Tensile tests – Wrought to LPBF ................................................................................. 103
4.3.2.4.3 Post-test CT scan – Wrought to LPBF .......................................................................... 106
4.3.2.5 Welded LPBF to LPBF assembly tensile tests .............................................................. 109
4.3.2.6
Tensile tests – LPBF to LPBF ........................................................................................ 110
4.3.2.7 Post-test CT scans – LPBF to LPBF ............................................................................... 111
4.3.3 Fatigue tests ........................................................................................................................ 116
4.3.3.1 Design of experiments ................................................................................................ 116
4.3.3.2
Single piece wrought fatigue tests .............................................................................. 117
4.3.3.3 Pre-test porosity analysis – single piece wrought....................................................... 118
4.3.3.4
Fatigue test – single piece wrought ............................................................................ 118
4.3.3.5 Post-test SEM – single piece wrought ......................................................................... 118
4.3.3.6
Single piece LPBF fatigue tests .................................................................................... 121
4.3.3.7 Pre-test porosity analysis – single piece LPBF ............................................................. 121
4.3.3.8
Fatigue test – single piece LPBF .................................................................................. 122
4.3.3.9 Post-test CT scan – single piece LPBF .......................................................................... 122
4.3.3.10 Post-test SEM – single piece LPBF ............................................................................... 123
4.3.3.11 Welded wrought to LPBF assembly fatigue tests ....................................................... 125
4.3.3.12 Pre-test porosity analysis – wrought to LPBF ............................................................. 125
4.3.3.13 Fatigue test – wrought to LPBF ................................................................................... 131
4.3.3.14 Post-test CT scan – wrought to LPBF .......................................................................... 136
4.3.3.15 Post-test SEM – wrought to LPBF ............................................................................... 138
4.3.3.16 Welded LPBF to LPBF assembly fatigue tests ............................................................. 150
4.3.3.17 Pre-test porosity analysis – LPBF to LPBF ................................................................... 150
4.3.3.18 Fatigue test – LPBF to LPBF ......................................................................................... 151
4.3.3.19 Post-test CT scan – LPBF to LPBF ................................................................................ 157
4.3.3.20 Post-test SEM – LPBF to LPBF ..................................................................................... 158
4.4
References .......................................................................................................................... 168
5. Conclusions ................................................................................................................................. 169
5.1
Chapter overview ................................................................................................................ 169
5.2 Welding parameters ........................................................................................................... 169
5.3
Tensile strength and fatigue performance ......................................................................... 170
5.4 Microstructural parameters ................................................................................................ 171
5.5
Future work ......................................................................................................................... 172
5.5.1 Dissimilar alloys ................................................................................................................... 172
5.5.2 Post weld processing........................................................................................................... 172
Ali Tamaddon
Page | vi
5.5.3 Build parameters effect on weldability ............................................................................... 172
5.5.4
Impact of geometry on weld performance ......................................................................... 172
5.5.5 Enhance statistical confidence of results ............................................................................ 172
Ali Tamaddon
Page | vii
List of Tables
Table 2-1. (Page 15) Cost comparison of titanium vs. steel and aluminium.
Table 3-1. (Page 36) Primary Design of Experiment for pilot sample weld parameters.
Table 3-2. (Page 37) Nominal chemical composition of Titanium alloy grade 5 used for this research, as per ASTM F2924-14.
Table 3-3. (Page 39) LPBF base material build parameters.
Table 3-4. (Page 53) Proposed design of experiment for tensile tests as per ASTM E8.
Table 3-5. (Page 54) Proposed design of experiment for fatigue tests as per ASTM E466.
Table 4-1. (Page 68) Surface colour in titanium welds; as per American Welding Society (AWS) G2.4M:2014 – guide for the fusion welding of titanium and titanium alloys.
Table 4-2. (Page 72) Summary of weld parameters, resulting energy density and welded joint fusion zone cross section area on pilot samples as measured using digital microscopy.
Table 4-3. (Page 75) Ti6Al4V raw material properties.
Table 4-4. (Page 77) Phase identification in fusion zone of pilot sample number 9.
Table 4-5. (Page 98) ASTM E8 tensile test results on wrought single piece samples.
Table 4-6. (Page 105) ASTM E8 tensile test results for LPBF to wrought welded samples.
Table 4-7. (Page 110) ASTM E8 tensile test results for LPBF-to-LPBF welded samples, note the close proximity of the UTS to breaking strength and relative short overall elongation prior to breakage.
Table 4-8. (Page 117) ASTM E466 fatigue test design of experiment (DOE).
Table 4-9. (Page 118) ASTM E466 fatigue test results for single piece wrought samples.
Table 4-10. (Page 122) ASTM E466 fatigue test results for single piece LPBF samples.
Table 4-11. (Page 131) ASTM E466 fatigue test results for LPBF to wrought welded assemblies.
Ali Tamaddon
Page | viii
Table 4-12. (Page 152) ASTM E466 fatigue test results for LPBF to LPBF welded assemblies.
List of Figures
Figure 2-1. (Page 11) Single step and multi-step AM process principles (Standard 2012).
Figure 2-2. (Page 12) Production of AM parts from independent service providers (in millions of dollars). (Wohlers Report 2021).
Figure 2-3. (Page 13) Experience of AM technologies per industry 2019 (%). Courtesy Ernest Young (EY_Global 2019b).
Figure 2-4. (Page 14) Overview of single-step AM processing principles for metallic materials courtesy ISO ASTM 52900 (Standard 2012).
Figure 2-5. (Page 16) Build rate comparison of small (blue), medium (red) and large (green) machines. Courtesy (Khorasani et al. 2020).
Figure 2-6 (Page 20) Summary of stress (S) versus cycles to failure (N) (S-N) data for PBF (laser), PBF (E-beam), and wire (DED) at R = 0.1. Metallic Materials Properties Development and Standardization (MMPDS) data for cast, wrought machined data are shown for comparison (Lewandowski & Seifi 2016).
Figure 2-7. (Page 22) Schematic representation of possible process by-products courtesy (Ladewig et al. 2016).
Figure 2-8. (Page 22) Schematic of defect induced by recoating during LPBF process.
Figure 2-9. (Page 23) Four major components of a laser (AWS 2010).
Figure 2-10. (Page 24) Principles for thin disk laser (AWS 2010).
Figure 2-11. (Page 25) Trump TruLaser Cell 7020 weld setup. Laser beam comes from above and nozzle feeds the inert shielding gas to the work area.
Figure 2-12. (Page 27) Tensile test results related to the porosity ratio in welding area, according to the average power variation (Akman et al. 2009).
Figure 2-13. (Page 27) Microhardness distribution of workpieces for different average power (Akman et al. 2009)
Figure 3-1. (Page 37) SEM image of the used Titanium grade 5 (Ti-6Al-4V) powder.
Figure 3-2. (Page 37) Titanium grade 5 alloy wrought sheets.
Figure 3-3. (Page 39) SLM Solutions ™ SLM 250, metal additive manufacturing machine.
Figure 3-4. (Page 40) 20mm x 20mm x 4mm LPBF fabricated Ti64 plates for preliminary welding experiments.
Ali Tamaddon
Page | ix
Figure 3-5. (Page 41) Trumpf TruLaser Cell 7020 welding machine and interface.
Figure 3-6. (Page 42) Custom jig to hold the square samples in compression for welding and diagram of weld direction used.
Figure 3-7. (Page 42) Welding set up inside the Trumpf Trulaser ™ Cell 7020 robotic laser welding machine chamber.
Figure 3-8. (Page 43) Trial runs on provisional samples, to verify weld parameters. Note the direction of welds marked on far-right sample in image. Start (S) to finish (F).
Figure 3-9. (Page 44) Work in progress of the pilot sample fabrication.
Figure 3-10. (Page 45) Struers Labotom-3 manual table-top lab size cutting machine.
Figure 3-11. (Page 46) Struers CitoPress phenolic resin mounting machine.
Figure 3-12. (Page 46) Struers RotoPol-21 and RotoForce-4 polishing machine.
Figure 3-13. (Page 47) Keyence VHX-5000 Digital optical microscope.
Figure 3-14. (Page 48) Phoenix v| Tome|x s CT scanner at RMIT Bundoora East campus.
Figure 3-15. (Page 49) Image of VG Studio 3.0 interface.
Figure 3-16. (Page 50) FEI Quanta 200.
Figure 3-17. (Page 52) FEI Nova NanoSEM 200.
Figure 3-18. (Page 52) Microhardness measurements performed on cross section.
Figure 3-19. (Page 55) Work in progress of main samples fabrication in the SLM® 250HL machine.
Figure 3-20. (left) (Page 56) Additively manufactured tensile testing half specimens as per ASTM E8 prior to welding, note elongated tabs the junction for welding, these are incorporated to accommodate any anomalies occurring at the start and finish of each weld. These tabs are machined off and result in a fully compliant geometry as per ASTM E8. (right) wrought titanium alloy half specimens alongside baseline samples, these are used to establish the experiment reference data necessary for this research.
Figure 3-21. (Page 56) Tensile testing sample geometry (courtesy ASTM E8 standard).
Figure 3-22. (Page 57) (left) Additively manufactured fatigue testing half specimens as per ASTM E466 prior to welding. Note the provision for welding tabs, similar to tensile testing samples. (right) wrought titanium alloy half specimens alongside baseline fatigue full samples.
Figure 3-23. (Page 57) Fatigue testing sample geometry (courtesy ASTM E466 standard).
Ali Tamaddon
Page | x
Figure 3-24. (Page 58) Custom jig to accommodate ASTM E8 and E466 autogenous welding in the Trumpf Trulaser ™ Cell 7020 robotic laser welding machine.
Figure 3-25. (Page 59) Machined Fatigue and Tensile samples post machining and ready for tests.
Figure 3-26. (Page 61) MTS Landmark 100kN Servohydraulic Test System, material testing machine. Material testing laboratory, RMIT University, Bundoora East, Victoria.
Figure 3-27. (Page 62) Close up of the tensile sample loaded onto the rectangular grips and the MTS extensometer attached and armed (note pin is removed and arm is aligned, visible via the opening in which the pin has been removed from).
Figure 3-28. (Page 63) Polycarbonate ballistic shield to protect against flying debris.
Figure 4-1. (Page 67) Top side of welded pilot samples.
Figure 4-2. (Page 67) Bottom side of welded pilot samples.
Figure 4-3. (Page 69 – 71) Combined cross section area of fusion and heat affected zones.
Figure 4-4. (Page 72) Scatterplot of combined fusion zone (FZ) and heat affected zone (HAZ) area versus welding energy density.
Figure 4-5. (Page 74) (Top) Cross section of welded joint as visible via SEM microscope. (Bottom) Energy dispersive spectroscopy (EDS) results across welded joint, sample number 8.
Figure 4-6. (Page 75) Energy dispersive spectroscopy (EDS) results of base metal (BM) at the additive manufactured side of sample number 8.
Figure 4-7. (Page 76) Electron backscatter diffraction images of fusion zone on sample 9. Top left shows the overall area under study, with the zoomed area highlighted in the top right image. Needle shaped titanium alloy grains captured middle left using the parameters above middle right. Phase identification bottom left and highlighted of Titanium hexagonal close packed (HCP) crystals bottom right.
Figure 4-8. (Page 77) Zoomed out phase identification of pilot sample number 9. (Right) Blue area is Ti hexagonal close packed (HCP) and yellow area is Titanium body centre cubic (BCC).
Figure 4-9. (Page 80) Porosity profile of pilot sample number 6, welded at 2000 Watts and 1200mm/min. Note that 1.28% of overall region of interest (ROI) consists of pores. 241 defects were detected at a mean diameter of 0.14mm, with a mean sphericity of 0.5171.
Figure 4-10. (Page 81) Porosity profile of pilot sample number 9, welded at 2200 Watts and 1200mm/min. Note that only 0.12% of overall region of interest (ROI) consists of pores. 99 defects were detected at a mean diameter of 0.12mm, with a mean sphericity of 0.4901.
Figure 4-11. (Page 82) There is direct relationship between porosity and mean pore size.
Ali Tamaddon
Page | xi
Figure 4-12. (Page 83) There is direct relationship between laser welding power levels and the amount of porosity in the final resulting welded joint. This is shown with a probability value (P-value) of 0.686.
Figure 4-13. (Page 83) There is an inverse relationship between the welding laser beam travel speed and the resulting porosity, with a P-value of 0.796.
Figure 4-14. (Page 84) (left) Direction relationship between welding energy density and resulting porosity is not evident until (right) outlier datapoint (100 J/mm) are taken out, resulting in a clear demonstration of a relationship between the welding energy density (J/mm) and the resulting porosity (% volume) of the welded joints. Figure 4-15. (Page 84) (left) inverse relationship between laser beam welding power levels with a P-value of 0.448 and Figure 4-16 (right) show the inverse relationship between porosity sphericity to laser beam welding travel speeds with a P-value of 0.690.
Figure 4-16. (Page 85) Leco LM700AT micro-indentation hardness testing machine.
Figure 4-17. (Page 86) Microhardness profile across three different assemblies built using 1800W power level. The LPBF side is labelled as SLM.
Figure 4-18. (Page 86) Microhardness profile across three different assemblies built using 2000W power level. The LPBF side is labelled as SLM.
Figure 4-19. (Page 87) Microhardness profile across three different assemblies built using 2200W power level. The LPBF side is labelled as SLM.
Figure 4-20. (Page 87) The greatest variation in hardness happens on the HAZ area of LPBF. The LPBF side is labelled as SLM.
Figure 4-21. (Page 91) CT scan image of porosities. Note only pores with a probability greater or equal to 1 have been annotated.
Figure 4-22. (Page 92) CT scan data graphical representation of porosities along the length of the tensile test sample, note that the origin of the z position is around the middle of the sample. Circle diameters proportional to diameter of porosities.
Figure 4-23. (Page 93) Tensile test on single piece LPBF Ti6Al4V, note the fracture geometry.
Figure 4-24. (Page 94) Stress-strain diagram of single piece LPBF tensile sample, as per ASTM E8.
Figure 4-25. (Page 95) CT scan of single piece LPBF tensile test sample, not showing any significant porosities around the fractured surface, besides the single pore at 180 microns diameter (defect 7).
Ali Tamaddon
Page | xii
Figure 4-26. (Page 96) Fractured surface of LPBF single piece tensile sample. Note the flat surface of failure perpendicular to the axis of stress loading at the middle of the part. Top image at 38x magnification. Bottom left at 50x facing the upper and bottom right at 50x, facing the lower area of the face. Note dark points as porosities.
Figure 4-27. (Page 97) Three identical ASTM E8 wrought Ti6Al4V alloy samples were tested for tensile strength. The top 2 resulted in a cup and cone fracture surface and the last sample resulted in a single slanted 45 degrees fracture.
Figure 4-28. (Page 98) Stress-Strain diagram of wrought single piece tests as per ASTM E8.
Figure 4-29. (Page 99) Fractured surface of wrought single piece tensile sample. Note the 45 degrees angle between the surface of failure and axis of stress loading. Top image at 50x magnification, bottom left at 200x and bottom right at 400x, highlighting one of the points of failure.
Figure 4-30. (Page 100) Sample number 1, CT scan and porosity distribution histogram
Figure 4-31. (Page 101) Sample number 2, CT scan and porosity distribution histogram.
Figure 4-32. (Page 102) Sample number 3, CT scan and porosity distribution diagram.
Figure 4-33. (Page 103) Minitab combined analysis of porosity data for test samples 1, 2 and 3.
Figure 4-34. (Page 104) Three welded LPBF to wrought samples were tested up to failure. All failures were in the wrought side of the assembly and beyond the fusion zone.
Figure 4-35. (Page 105) Stress-Strain diagram of LPBF to wrought welded samples as per ASTM E8.
Figure 4-36. (Page 106) Sample number 1, post fracture CT scan data, shows that most of the porosities are in the fusion zone.
Figure 4-37. (Page 106) Sample number 2, post fracture CT scan data, again most of the porosities are concentrated at the fusion zone and away from the point of fracture.
Figure 4-38. (Page 107) Sample number 3, post fracture CT scan data, showing consistent results with all similar samples, with the porosities being concentrated at the fusion zone and away from the fracture area.
Figure 4-39. (Page 107) Surface of fractured damage. Left image shows entire face of damaged area in 50x magnification, right image shows a 100x magnification of cup and cone fracture detail.
Figure 4-40. (Page 108) Top left (200x) and top right (1600x) show details of the cup and cone failure (in ductile mode) at higher magnification.
Figure 4-41. (Page 109) Three LPBF to LPBF welded test samples.
Ali Tamaddon
Page | xiii
Figure 4-42. (Page 110) Stress-Strain diagram of LPBF-to-LPBF welded samples as per ASTM E8.
Figure 4-43. (Page 111) CT scan images of sample numbers 1 and 2 of the welded LPBF-LPBF tensile tests. Both images have annotations showing the location of the top 5 largest porosities by diameter.
Figure 4-44. (Page 112) Histograms showing the distribution of porosity diameter sizes in the samples number 1 (top) and 2 (bottom). Both show relatively similar pore sizes, with the mean sizes ranging between 260 and 227 microns.
Figure 4-45. (Page 113) Distribution of porosities relative to the z-axis, with pore sizes represented by diagram size. As evident, the majority of the pores are at the vicinity of the fusion zone (circa 400 to 600 voxels on the z-axis). Top is sample 1 and bottom is sample 2. Z datum point is at base metal and 600 microns is around FZ.
Figure 4-46. (Page 114) Large pores can be seen at the perimeter and on the surface of the fractured area of tensile test sample number 1 of the LPBF-to-LPBF welded assembly.
Figure 4-47. (Page 115) Large pores can be seen at the perimeter and on the surface of the fractured area of tensile test sample number 2 of the LPBF-to-LPBF welded assembly.
Figure 4-48. (Page 116) Fatigue test loading profile.
Figure 4-49. (Page 119) Single piece wrought samples subjected to 500MPa and R0.1 axial fatigue loading. Top image is the top view, and the bottom image is the side view. As evident from these images the failure mode is ductile, with a fracture plane at 45 degrees to the plane of loading.
Figure 4-50. (Page 120) Fatigue test results for single piece wrought samples as per ASTM E466 (Sample ID W1), which failed at 58,046 cycles, shows a complete ductile failure initiated at a material defect as evident in the highlighted images above.
Figure 4-51. (Page 121) Histogram of sample ID LL1 porosity analysis.
Figure 4-52. (Page 122) The largest 5 pores in sample ID LL1 were distributed along the entire length of the sample.
Figure 4-53. (Page 123) Distribution of pores in sample ID LL1.
Figure 4-54. (Page 124) LPBF single piece fatigue sample fractured surface. Note the multiple pores on the surface of the failure plane and edges of the fracture (bottom right).
Figure 4-55. (Page 126) CT scan images of samples (from top to bottom and left to right) LW10, LW11 and LW12 (alongside an image to show the orientation of the axis, note Z-axis is zero at the bottom and increases upwards)
Figure 4-56. (Page 127) Pore diameter to sphercity distribution graph (from top to bottom) LW10 and LW11.
Ali Tamaddon
Page | xiv
Figure 4-56. (continued) (Page 128) Pore diameter to sphercity distribution graph LW12.
Figure 4-57. (Page 128) Pore diameter, volume and Z-axis distribution of samples (from top to bottom) LW10, LW11 and LW12.
Figure 4-58. (Page 129) Pore diameter, volume and Z-axis distribution of samples (from top to bottom) LW11 and LW12.
Figure 4-59. (Page 130) Pore size distribution histogram (from top to bottom and left to right) LW10, LW11 and LW12. As can be seen in the summary graph the quantity of pores, their size and distribution are relatively consistent.
Figure 4-60. (Page 131) Welded LPBF to wrought fatigue behaviour is best represented by using an S-N curve, with the number of cycles to failure in logarithmic scale in the abscissa and the stress levels on the ordinate.
Figure 4-61. (Page 132) Failed welded LPBF to wrought samples subject to 200MPa (sample ID from top to bottom LW3, LW2 and LW1).
Figure 4-62. (Page 133) Failed welded LPBF to wrought samples subject to 200MPa (top 2, sample ID from top to bottom LW11 and LW12) and 600MPa (bottom sample LW7).
Figure 4-63. (Page 134) Failed welded LPBF to wrought samples subject to 400MPa (sample ID from top to bottom LW4, LW5 and LW6).
Figure 4-64. (Page 135) Failed welded LPBF to wrought samples subject to 500MPa (sample ID from top to bottom LW8, LW9 and LW10).
Figure 4-65. (Page 136) LW1 (left) and LW2 (right), both failed at the LPBF side of the welded assembly, but the largest pores were identified to be at the FZ as per CT scans.
Figure 4-66. (Page 136) LW3 (left) and LW4 (right), both failed at the LPBF side of the welded assembly, but the largest pores were identified to be at the FZ as per CT scans.
Figure 4-67. (Page 137) LW5 (left) and LW6 (right), both failed at the fusion zone (FZ) of the welded assembly and the largest pores were also identified to be at the FZ as per CT scans.
Figure 4-68. (Page 137) LW7 (left) and LW8 (right), both failed at the fusion zone (FZ) of the welded assembly and the largest pores were also identified to be at the FZ as per CT scans.
Figure 4-69. (Page 137) LW9 (left) failed at the LPBF side of the welded assembly, but the majority of the largest pores were at the FZ. LW10 (right), both failed at the LPBF side of the welded assembly, but the largest pores were identified to be at the FZ as per CT scans.
Figure 4-70. (Page 138) LW11 (left) and LW12 (right) both failed at the LPBF side of the welded assembly and as evident in the images, in both assemblies the porosities are concentrated at the FZ and HAZ.
Ali Tamaddon
Page | xv
Figure 4-71. (Page 139) Fracture surface on LPBF base-metal side of the welded assembly LW1, which failed under 200MPa loading, at 144,795 cycles and R=0.1 shows a brittle failure mode, with multiple points of crack initiations at the perimeter of the failure plane.
Figure 4-72. (Page 140) Fracture surface at the fusion zone of the welded assembly LW2, which failed under 200MPa axial stress, after 55,513 loading cycles and R=0.1 shows was porosities which acted as multiple points of crack initiations.
Figure 4-73. (Page 141) Fracture surface at the fusion zone of the welded assembly LW3, which failed at 200MPa axial stress, after 149,920 loading cycles and R=0.1. Note the porosities and enlarged image of one of the porosities showing crack initiation points at the surface of the pore.
Figure 4-74. (Page 142) Fracture surface on LPBF base-metal side of the welded assembly LW4, which failed under 400MPa loading, at 20,347 cycles and R=0.1. Note the pore at the centre of the fracture surface and multiple fracture initiation points at the perimeter of the failure zone.
Figure 4-75. (Page 143) Fracture surface on LPBF base-metal side of the welded assembly LW5, under 400MPa loading, at 16,949 cycles and R=0.1. Two major porosities can be clearly seen on these images.
Figure 4-76. (Page 144) Fracture surface on LPBF base-metal side of the welded assembly LW6, under 400MPa loading, at 23,352 cycles and R=0.1.
Figure 4-77. (Page 145) Fracture surface at fusion zone of welded assembly LW7, under the highest load applied during these tests of 600MPa. The part failed after only 60 cycles. Large porosities can be observed on the failure surface.
Figure 4-78. (Page 146) The contribution of surface roughness acting as crack initiation source can be seen by observing the perimeter of the fracture surface of LW9 which failed at 500MPa after 7,164 loading cycles
Figure 4-79. (Page 147) Large pores and their inherent surface cracks that have led to the failure of sample LW10 at 500MPa of stress after 9,496 loading cycles.
Figure 4-80. (Page 148) Surface roughness and inconsistencies have played a significant role in crack initiation on sample LW11 at 300MPa axial stress after 39,245 loading cycles. Figure 4-81. (Page 149) Surface conditions can also be observed to have had a significant effect on the failure mode of LW12 at 300MPa axial stress after 28,266 loading cycles. Crack initiation porosities highlighted in (bottom right) image.
Figure 4-82. (Page 150) Sample ID LL7, note the concentration of porosities at z-axis position around FZ and the HAZ.
Figure 4-83. (Page 151) Sample ID LL8, shows similar pattern for detectable porosities, with the concentration being at the FZ and HAZ.
Ali Tamaddon
Page | xvi
Figure 4-84. (Page 152) Welded LPBF to LPBF fatigue behaviour is best represented by using an S-N curve, with the number of cycles to failure in logarithmic scale in the abscissa and the stress levels on the ordinate.
Figure 4-85. (Page 153) Top and side view of LPBF to LPBF welded assemblies subjected to 200MPa axial loading cycles to failure. From top to bottom, sample IDs: LL1, LL2 and LL9
Figure 4-86. (Page 154) Top and side view of LPBF to LPBF welded assemblies subjected to 300MPa axial loading cycles to failure. Top LL8 and bottom LL7.
Figure 4-87. (Page 155) Top and side view of LPBF to LPBF welded assemblies subjected to 400MPa axial loading cycles to failure. Top LL4 and bottom LL3.
Figure 4-88. (Page 156) Top and side view of LPBF to LPBF welded samples, subject to 500MPa (top, LL5) and 100MPa (bottom, LL6). The sample subject to 100MPa didn’t fail after 1,000,001 cycles, which for the purpose of this study is considered the limit for infinite life.
Figure 4-89. (Page 157) (left to right and top to bottom): LL1, LL2, LL3, LL4, LL5, LL7. Note that sample ID LL6 did not fail.
Figure 4-90. (Page 158) (left) sample ID LL8 and (right) sample ID LL9
Figure 4-91. (Page 159) Sample LL1, subjected to 200MPa of axial load, survived 117,055 cycles. The perpendicular cut surface suggests the brittle nature of the fracture and the pores at the perimeter, are sources of crack initiation.
Figure 4-92. (Page 160) Sample LL2, subjected to 200MPa of axial loading, underwent 161,869 cycles prior to fracture5 as typical in LPBF to LPBF welded assemblies in this research, the failure mode is demonstrated to be brittle, with multiple points of fracture initiation at the perimeter of the welded assembly.
Figure 4-93. (Page 161) Sample LL3, subjected to 400MPa of axial stress loading, underwent 19,149 cycles prior to failure. Bottom right picture shows the starting point of a microcrack that contributed to the brittle fracture of the part.
Figure 4-94. (Page 162) Large pores pertaining to incompletely melted particles can be seen as dark patches in the images of sample LL4, which was subjected to 400MPa axial loading and maintained its integrity up to 16,544 cycles.
Figure 4-95. (Page 163) Sample LL5, which was subjected to the highest axial loading of this research at 500MPa, withstood 10,746 cycles before showing multiple failures, including a secondary fracture on the side of the body, as seen in the bottom two images.
Figure 4-96. (Page 164) LL6, close up image of the fusion zone (FZ) and heat affected zones (HAZ). Note that albeit surface anomalies, the welded assembly was able to maintain its integrity after 1,000,001 cycles of 100MPa axial loading.
Ali Tamaddon
Page | xvii
Figure 4-97. (Page 165) Sample LL7, which failed at the fusion zone at 300MPa axial loading after 4,094 cycles, shows a large cluster of porosities at the centre and perimeter of the fusion zone.
Figure 4-98. (Page 166) Sample LL8 which also failed at the fusion zone at 300MPa axial loading after only 2,264 cycles, has multiple large pores on the surface and perimeter of the fractured area, contributing to the failure of the part.
Ali Tamaddon
Page | xviii
Figure 4-99. (Page 167) Sample LL9, which failed at the fusion zone after 3,965 cycles at 200MPa axial force, also demonstrates large pores scattered all over the face of the fractured surface and perimeter.
Abstract Metals are one of the primary class of materials used for additive manufacturing. Minimal
material wastage and freedom in creating complex geometries have enabled this technology
to reap such benefits with more expensive metallic alloys, namely titanium alloys. Titanium
alloys are significantly more expensive than aluminium and steel alloys but have excellent
corrosion resistance and weight to strength ratios, making them appealing to highly corrosive
applications such as aerospace and medical devices. Despite the widespread adoption of
additive manufacturing technologies and increase in additive manufacturing machine size and
speed of fabrication, the limitations are still significantly considerable for the purpose of mass
production of parts at large scale. Laser powder bed fusion (LPBF), which is the most
promising of all metal additive manufacturing technologies has chamber sizes confined to
approximately one cubic meter or less; and has part build rates at 180 cm3/h or less. This is in
stark contrast to traditional manufacturing methods, which can produce large volumes at high
rates.
This research investigates a means of negating such shortcomings and reaping the benefits of
both LPBF additive manufacturing and traditional manufacturing methods by joining such
parts together using laser welding in the autogenous mode (no fillers required).
Based on the literature review conducted in this field and the gaps identified, a two-stage
methodology was utilised, as per following.
At the first stage of the research, 20x20x4mm pilot samples were made using wrought Ti6Al4V
titanium alloys and LPBF fabricated Ti6Al4V titanium alloys with the same geometry and
dimensions, then welded together using disc laser beam at various power levels and speed,
which were selected using existing literature. The welded joints were examined for porosity,
geometry of weld zones, microhardness and microstructure. Energy density and rate of
cooling resulting from the various welding parameters were the driving factors of the welded
joint results. The results of examinations were analysed, and laser beam power level of 2200
Watts and 1200 mm/ min were the most promising parameters.
At the second stage of the research more samples were fabricated, in a similar fashion to the
first stage, but had geometries and dimensions as per ASTM E8 and E466 standards. All
Abstract
Page | 1
samples in the second stage were welded using the most promising weld parameters found
during the first stage and subjected to tensile and fatigue tests. Each sample was checked for
porosity before and after each test and the results produced a profile of tensile and fatigue
characteristics of such assemblies.
It is demonstrated that the welding of LPBF to wrought samples, and LPBF to LPBF is a viable
solution to overcoming the limitations in build volume and production speed in LPBF
technologies. The governing factors in the success of such welds are laser energy density
Abstract
Page | 2
(cooling rates) and LPBF base metal initial build quality (porosities).
1. Introduction 1.1. Background
Additive manufacturing has come a long way from the early days when it was regarded as a
rapid prototyping technology, used to create polymer components for visualisation and
demonstration purposes, to this day where it has become a viable production method for
end-use parts of various materials, including complex cellulous fibres, carbon fibre embedded
plastics and even metallic parts. There is a great amount of expectation of this technology,
due to the unique characteristics it provides, such as ability to create complex geometries,
freedom from the necessity of jigs and fixture and the time required to transform 3D CAD
data into actual parts. All which are done seamlessly with minimal need for post processing
and minimal wastage of raw material. This makes the technology extremely appealing in low-
volume, high-complexity scenarios. Many OEM (original equipment manufacturer)
organisations have started showing interest in such technologies and the market is growing
at a rapid pace and is not showing any signs of deceleration. (Brandt 2016)
Additive manufacturing has especially become extremely appealing in fabricating three-
dimensional metallic end-use parts, providing the possibility to fabricate novel shapes and
geometrical features never deemed possible prior to the birth of this technology. With the
proliferation of this technology the affordability factor is also becoming more favourable; to
such an extent that owning and operating a metal printing machine is no longer a capability
confined to only large OEMs but also available to small and medium enterprises (SMEs). This
change of dynamics has led to 3D printing becoming a changing force used to optimise the
supply chain and manufacturing today, filling in the gaps of small quantity, high complexity
and high value parts. With the increase of the quality of the end product and development of
3D-printing software applications for design and analysis, we are witnessing a large shift in
the application of additive manufacturing from prototyping to end-use products, with an
overall increase of 65% in applying such technologies in industries in 2019 compared to 2016;
and there are strong signs and economic model predictions that are showing these figures to
accelerate significantly by the end of 2022, with US companies alone shifting from a miniscule
share of 4% end use parts fabricated using additive manufacturing to a major share of 54% by
1 Introduction
Page | 3
end of 2022. (EY_Global 2019a)
The entire product development lifecycle requires to be re-engineered to harness the full
potential of this new technology. This includes but is not limited to all aspects of the supply
chain, from design, raw material to fabrication, testing, quality assurance and control and
recycling. (Huang et al. 2015)
Additively manufactured metal components have many benefits, such as complex geometry
not possible achieving using traditional subtractive methods, freedom from manufacturing
process restrictions, low lead times, minimum wastage of material, near-zero set up costs to
name a few. New concepts and parameters become important in this new technology, such
as the capital investment in machinery, customised computer aided design (CAD) tools;
machine feedstock; skillsets of operators, verification methods for final product and quality
assurance and control to safeguard client confidence in the product. The marketplace is now
much more aware of the potentials as well as the restrictions of this new technology and the
race has started to overcome those hurdles and gain a competitive advantage in this new-
found marketplace.
Metal additive manufacturing has evolved over the years and has seen the rise and fall of
various technologies and fabrication methods, such as stereolithography and sheet
lamination. The most common and highly popular method is the laser powder bed fused
(LPBF), formerly known as selective laser melted (SLM) technology, due to its advantages such
as moderate surface roughness, dimensional accuracy, high layer resolution and critical
design fabrication to name a few. LPBF is used to fabricate a large range of metal alloys such
Titanium (Ti), Nickle (Ni), Aluminium (Al) and Ferrous (Fe) based alloys. (Revilla et al. 2020)
Titanium alloys have a special place in the Aerospace industry, due to their high strength to
weight ratio; high mechanical strength in the range from cryogenic temperatures to
moderately elevated temperatures as well as high resistance to corrosion in the most
harshest environments (Putyrskii, Yakovlev & Nochovnaya 2018). Hence there has been a
motivation to utilise titanium alloys in fabricating additively manufactured end-use parts,
using various technologies. LPBF has proliferated in the past decade to be the driving force in
1 Introduction
Page | 4
industrial scale metal additive manufacturing (Curtis 2013).
1.2. Motivation
Engineering is the practical application of science that also has to deal with any limitations.
As for additive manufacturing of metals is concerned, the main draw backs that have hindered
the adoption of this technology at full scale, print size and speed of fabrication, make to the
top of the list. Current build chamber sizes are still restrictive, in comparison to subtractive
manufacturing techniques. Even with the rapid progress during the duration of this research
were their build volumes have increased dramatically from a mere 248x248x250mm (on SLM
Solutions SLM250) and their build methods have evolved from a single laser fabrication at a
rate of 20 cm3/ hour to a whopping build chamber volume of 600x600x600mm; using 12
concurrent lasers of 1000 Watts each delivering built parts at a rate of 1000cm3/ hour, i.e. 1
litre/ hour (on SLM Solutions NXG XII 600). Even such progress is not comparable to the rapid
production rates of mass subtractive and traditional manufacturing, thus, to overcome such
shortfalls, it is proposed in this research that additively manufactured parts be combined with
conventionally manufactured components; to reap the benefits of both methods and
overcome the limitations in size and speed of fabrication. Traditionally fabricated parts will
be especially economical in areas where additively manufactured components do not
demonstrate competitive advantage, such as parts that lack geometrical complexity, are bulky
and easily available in off the shelf formats, shapes, and cross sections such as in sheets,
standard beams and rods of various sizes. An example of such a synergic combination would
be an assembly in which long slender beams would be joined to complex geometrical
topologies. For instance, in the fabrication of an aircraft wing, the long slender beams would
be the wing spars and the complex geometries would be the wing root assembly and/ or
topologically optimised wing cross sections. The slender beams would benefit from
traditionally manufactured process such as extrusion and rolling and the complex geometry
would benefit from additive manufacturing. Joining two differently created titanium alloys of
same material and understanding the overall performance of the resulting assembly is the
main topic addressed in this research. Titanium alloys in wrought format have well
documented mechanical, weldability and fatigue characteristics; but the same material when
subject to production processes involving making the LPBF powder stock and eventual
fabrication of parts were layers upon layers of laser are melted and subjected to rapid and
extreme thermal cycles are yet to be fully investigated and documented. Laser beam welding
1 Introduction
Page | 5
of material that have undergone such extreme thermal cycles and combining them with
traditionally fabricated wrought parts will result in unique mechanical and fatigue
performance characteristics, that depend on the macroscopic characteristics, microstructure
of the LPBF and wrought parts as well as the parameters used in welding these parts together.
1.3. Project Objectives
(Qian et al. 2016)
This research investigates the material, mechanical properties, and fatigue performance of
titanium grade 5 (Ti-6Al-4V) laser powder bed fused (LPBF) additively manufactured parts
welded to wrought material of same chemical composition. This is achieved using Laser Beam
Welding (LBW), in autogenous mode, which means the two metals are fused using laser
energy and without any kind of filler material involved. The surface condition of each of the
welded faces requires to meet certain criteria. Since laser beam welding can be conducted
using various parameters, such as speeds, power settings, and beam geometry, the focus of
this research is to find the co-relationship between these weld parameters with the overall
mechanical and fatigue performance of the welded assembly. There is no prior literature on
the study the mechanical and fatigue performance of Ti-6Al-4V LPBF and wrought alloy
assemblies of same material using laser beam welds in autogenous mode. A detailed study of
macroscopic, microstructure, porosity, hardness, chemical composition, tensile strength and
fatigue performance are analysed in detail and related to the welding parameters that are
1.3.1. Aim
within the control of the technician, during the assembly process.
The aim of this research is to understand how laser beam welding parameters such as laser
power level and scanning speed can affect the overall performance of titanium alloy
assemblies made using laser powder bed fused (LPBF) and wrought parts welded together
using a disc laser beam weld method in autogenous mode. Failure modes are related to the
variations in the structure due to presence of porosities, non-uniformities, and other material
1 Introduction
Page | 6
features within the individual parts.
1.3.2. Research Questions The research questions addressed during this research are as follows:
1- What are the most predominant factors of laser beam welding on the overall
mechanical and fatigue performance of a welded SLM assembly?
2- What are the interaction effects of such parameters on the end result?
3- What microscopic parameters (phase, grain size and orientation) effect the overall
1.3.3. Research Scope
performance of the welded SLM assembly?
For practical purposes, the scope of the research is confined to the investigation of assemblies
fabricated using laser beam welding using disc lasers (1030nm) in autogenous mode with
variations in the following parameters only:
1) Three discreet laser power levels of 1800 Watts, 2000 Watts and 2200 Watts and
2) Three discreet scanning speeds of 800 mm/ min, 1000 mm/ min and 1200 mm/ min
The parameters used for the fabrication of the individual LPBF parts are deemed to be
constant (at laser power of 100 Watts, focal offset at 0mm, spot size at 80 microns, hatch
space at 120 microns, layer thickness at 30 microns and beam scanning speed at 375 mm/
min) and irrelevant to the scope of this research. Other laser beam welding parameters such
as focal offset distance (at 0mm) and the flow rate of the inert shielding gas (Argon at 20 litres
per minute) are also kept constant throughout the research.
The assemblies are scrutinised for the following:
1) Macroscopic appearance of the welded joint
2) Microscopic geometry of the fused area and heat affected zone of joint
3) Microstructure of base metal, heat affected zone and fusion zone
4) Chemical composition and any potential changes throughout the cross section of the
welded assembly
5) Hardness profile across the base metal, heat affected zone and fusion zone
6) Porosity at the fusion zone and heat affected zone
7) Tensile strength
1 Introduction
Page | 7
8) Fatigue performance
All such parameters were then correlated to the laser beam welding parameters and
Thesis Outline
1.4. The research is documented in five chapters. The outline of which are as follows:
conclusions are made.
Chapter one – Introduction: Describes the problem statement of the research. The industrial
background, motivation and research tasks are presented.
Chapter two – Literature review: Reviews the state of the art and literature available in the
field of laser beam welding (LBW), laser powder bed fusion technology (LPBF), welding of
titanium, electron back scattering (EBSD), energy dispersive spectroscopy (EDS), tensile and
fatigue testing.
Chapter three – Materials and methodology: Details the materials and goes into detail of the
methodology used to address the research questions, using a scientific and proven approach.
Chapter four – Results and discussion: Experiment results are studied and discussed in detail
against established theory and prior works and literature in the field.
Chapter five - Conclusion: Study results are used to draw a conclusion and verdict in response
to research questions. The stage is set for any further investigation into this research topic in
1 Introduction
Page | 8
the future.
1.5. References Brandt, M 2016, Laser Additive Manufacturing: Materials, Design, Technologies, and Applications, Woodhead Publishing series in electronic and optical materials, Elsevier Science & Technology, Cambridge.
Curtis, DI 2013, Microstructural and mechanical property characterization of Ti-6Al-4V produced by SLM and EBM.
EY_Global 2019a, How 3D printing brings transformational change,
Huang, Y, Leu, MC, Mazumder, J & Donmez, A 2015, 'Additive manufacturing: current state, future potential, gaps and needs, and recommendations', Journal of Manufacturing Science and Engineering, vol. 137, no. 1.
Putyrskii, SV, Yakovlev, AL & Nochovnaya, NA 2018, 'Benefits and Applications of High-Strength Titanium Alloys', Russian engineering research, vol. 38, no. 12, pp. 945-948.
Qian, M, Xu, W, Brandt, M & Tang, HP 2016, 'Additive manufacturing and postprocessing of Ti-6Al-4V for superior mechanical properties', MRS bulletin, vol. 41, no. 10, pp. 775-784.
Revilla, RI, Van Calster, M, Raes, M, Arroud, G, Andreatta, F, Pyl, L, Guillaume, P & De Graeve, I 2020, 'Microstructure and corrosion behavior of 316L stainless steel prepared using different additive manufacturing methods: A comparative study bringing insights into the impact of microstructure on their passivity', Corrosion science, vol. 176, p. 108914.
1 Introduction
Page | 9
2. Literature Review 2.1 Chapter overview Chapter 2 provides an overview of current literature and state-of-the-art of the various topics
related to this research. There has been significant focus on the application of titanium alloys
in metal additive manufacturing in the past decade, with research expanding all the way from
medical prosthetics applications to aerospace components. Metal additive manufacturing has
also advanced from the sidelines to mainstream end use part fabrication, increasing speed of
fabrication, build volume and affordability. This has resulted in greater accessibility and reach
and hence applications in more industries and new sectors. A variety of new opportunities
have arisen since the potential of this new technology has become more evident. This has
positively impacted the entire supply chain, with great benefits evident in the spare parts
business. Decreasing the dependence on overstocking of parts fabricated in large batches,
taking up space and causing long delays in supply due to the need for warehousing and
2.2 Overview of additive manufacturing (AM) As per ASTM F2792-12a, which made an attempt to provide standard terminology for additive
distribution (Kousa 2019).
manufacturing technologies, before being withdrawn in 2015, in favour of ISO ASTM 52900:
Additive Manufacturing (AM) is the process of joining material to make objects from 3D model
data, usually layer upon layer, as opposed to subtractive manufacturing methodologies
(ASTM_International 2013). Subtractive manufacturing methodologies include cutting,
drilling, machining, milling and other methods, in which the net shape is formed by removing
material from a bulk solid. In Additive Manufacturing the focus is creating a net shape by
adding material which come in various raw formats such as powder, gel, liquid or filaments
to name a few. The difference between rapid prototyping and additive manufacturing is the
difference in the objectives which are served. Where rapid prototyping products are used for
a design often iterative for form, fit or function testing or combination thereof
(ASTM_International 2013). Whereas additive manufacturing (AM) focuses on making end
2 Literature Review
Page | 10
use parts and not prototypes.
Rapid prototyping was a consequence of the speedy adoption and advancement of computer
aided design (CAD) in the 1980’s which brought about the need for prototyping designs in real
physical models and parts. Rapid prototyping is one of the earlier forms of additive
manufacturing. CAD data is converted to a stereolithography (STL) file, in which three-
dimensional data created in CAD is approximated by triangles and sliced containing the
information of each layer that is going to be printed. The thickness of each layer as well as the
resolution of the build depends on many factors, including raw material, machinery used and
heat and material flow limitations.
Categorisation of additive manufacturing technologies in the third decade of the twenty first
century has started to look like Figure 2-1 (Standard 2012). With the main single step additive
manufacturing processes being categorised as metallic, polymer, ceramic and composite. This
research has focused on metallic single-step additive manufactured titanium Ti6Al4V alloys
joined via autogenous disc laser beam welding to other AM Ti64 parts or wrought parts.
2 Literature Review
Page | 11
Figure 2-1. Single step and multi-step AM process principles (Standard 2012)
Additive manufacturing is now expanded to more industries than ever before and now serves
not just engineers in aerospace and manufacturing, but also medical practitioners in the
health sector, architects in the building industry and scientists and students in academia and
research. Rapid prototyping is also a very useful tool for market researchers to gauge public
and target market opinion on a new product.
Additive manufacturing has grown significantly since the early days of its inception in the form
of Selective Laser Sintering in the 1980s, until it became a rapid prototyping tool in the 1990s
and then as a rapid direct manufacturing tool in the 2000s. (Rayna & Striukova 2016)
Additive manufacturing has always rapid growth and even has been relatively immune to
massive market forces such as the global pandemic in the past year, with a market size
increase of 7.5% to USD 12.8 billion in 2020. Yet, this was significantly much less compared to
the average rate of 27.4% over the last 10 years, as seen in Figure 2-2.
Figure 2-2. Production of AM parts from independent service providers (in millions of US dollars). (Wohler’s Report 2021)
A thorough study of 404 firms providing 3D printing services to the European market has
revealed the major impact of additive manufacturing on supply chains across various
industries. Impacts on reducing setup costs, decreasing costs per unit for small production
runs and transferring the risk of production to 3D printing service providers. Inventory and
2 Literature Review
Page | 12
logistics costs are also reduced, due to the decrease for the need of transportation and
warehousing. Supply chains also enjoy greater flexibility with the ability to absorb variations
in production volume. (Rogers, Baricz & Pawar 2016)
Experience in using additive manufacturing technologies is increasing substantially with 78%
of the aerospace industry having used this technology in their operations and other industries
following suit, as seen in Figure 2-3.
2.3 Metal additive manufacturing Metals are one of the primary class of materials used for additive manufacturing. In 2010, the
Figure 2-3. Experience of AM technologies per industry 2019 (%). Courtesy Ernest Young (EY_Global 2019b)
results of a US Navy workshop were published with the title “direct digital manufacturing
(DDM) of metallic components: affordable, durable and structurally efficient aircraft” which
laid out a vision of on demand part availability for aircraft. The technology at the time was
still facing major challenges in design, qualification and certification, maintenance and repair
and the science and technology of DDM (Frazier 2010). ISO ASTM 52900 categorises metallic
2 Literature Review
Page | 13
additive manufacturing process as per Figure 2-4.
Figure 2-4. Overview of single-step AM processing principles for metallic materials courtesy ISO ASTM 52900 (Standard 2012)
One of the main reasons that titanium alloys are not as widely used in aerospace engineering
is the high cost of such material. Additive manufacturing provides many benefits, one such
benefit is the minimal material wastage and hence improving the cost effectiveness of such
alloys for high demanding applications in highly corrosive environments. Table 2-1 provides a
cost comparison of titanium to other common engineering material such as steel and
aluminium. The high cost of titanium is due to the high extraction and processing costs
associated with this mineral (Dutta & Froes 2017).
Hence titanium and namely Ti-6Al-4V which due to its high strength to weight ratio and
anticorrosive characteristics is one of the most common titanium alloys becomes an attractive
2 Literature Review
Page | 14
option as a metal additive manufacturing feedstock.
Material ($/ pound) Steel
Aluminium
Titanium
0.02 0.10 0.15
0.01 1.10 1.15
0.22 5.44 9.07
Format Ore Metal Ingot Sheet
0.30 to 0.60
1.00 to 5.00
15.00 to 50.00
2.4 Laser powder bed fusion (LPBF) Laser powder bed fusion (LPBF) technology employs, as the naming suggests, a bed of powder
Table 2-1 Cost comparison of titanium vs. steel and aluminium (Dutta & Froes 2017) Cost of titanium - Comparison
that is selectively melted layer upon layer by high power laser, then is lowered and covered
by another layer of powder which in turn is melted again, as per the slices of the part design
pattern. The melted layers fuse together at their interface due to the energy of the scanning
laser. A variety of material can be utilised in this technology. Other forms of additive
manufacturing such sheet lamination do not enjoy the same level of freedom and material
savings. In sheet lamination thin layers are cut and fused together forming the final shape.
Excessive material is produced as the result which would need to be reprocessed again and
turned into sheets. Directed energy deposition methods, use a filament wire or metal powder
that is fed through a nozzle and is melted layer upon layer using high energy laser or electron
beams. The resolution of such processes is yet to be competitive with the precision of LPBF
technology. Laser powder bed fusion (LPBF) technologies have enjoyed special attention from
industry and have gained popularity. This research focuses on this technology and attempts
to address these objectives (Nichols 2021).
LPBF has the lowest build rate compared to other metal additive manufacturing processes
such as directed energy deposition (DED), electron beam powder bed fusion, binder jetting
and sheet lamination. Yet, it has superior surface quality and the highest dimensional
accuracy. The most significant parameters in LPBF technology are laser power, scan speed,
2 Literature Review
Page | 15
layer thickness, hatch space, build size, and powder handling system (Khorasani et al. 2020).
Figure 2-5. Build rate comparison of small (blue), medium (red) and large (green) machines. Courtesy (Khorasani et al. 2020).
Part quality and build rates are the two main areas of recent advancements in LPBF
technology by the manufacturers of such machinery. This is characterised by an increase in
the number of concurrent lasers, maximum wattage and build chamber size (Figure 2-5).
Ti6Al4V parts fabricated using LPBF and electron beam melting (EBM) technologies have been
demonstrated to have relatively similar fatigue properties, albeit differences in
microstructure, both methods being sensitive to the main three types of defects, surface
defects, unmelted zones and small internal defects. Machining improves the mechanical
properties of the parts by removing the surface defects (Chastand et al. 2018).
But for the purpose of this research, additional processes post-fabrication is not considered,
2 Literature Review
Page | 16
since that would contradict the benefits of direct digital additive manufacturing.
Improvements in each of the process parameters involved in the powder bed fusion process
results in an overall improvement in part quality, decrease lead time and increase the safety
2.4.1 Powder feedstock Additive layer manufacturing (ALM) processes such as laser powder bed fusion (LPBF), aim to
and user experience of the machines.
produce fully dense parts, hence metallic powder processing gains great importance. Recent
studies have showed a direct relationship between the particle size distribution of a metallic
powder material to the density, surface quality and mechanical properties of the part
fabricated using LPBF technology. Although process parameters can be adjusted accordingly
to each particle size to enable a densely built part using various particles sizes, but as-built
surface roughness will always be greater with larger particles sizes (Spierings, Herres & Levy
2011).
Laser powder bed fusion (LPBF) part characteristics are well documented to be governed by
the laser power, scanning speed, focal length, layer thickness and powder material to name
a few. The effects of particle size distribution on the mechanical properties of the end result
parts has been investigated in detail (Mori et al. 2018). Particle size has little effect on
elongation but is inversely proportional to the tensile strength of the produced part, at
constant energy densities. One contributing factor to this is the increased oxygen content
during the sieving process of powders and processing of LPBF parts, from the oxidised film on
surface of the particles. Another factor is the grain size decrease due to repeated heat cycles.
Therefore it is essential that stockfeed particle sizes be controlled prior to the LPBF process.
(Mori et al. 2018)
The chemistry of the surface of Ti6Al4V components fabricated using additive manufacturing
and no secondary process and also Ti6Al4V fabricated using AM, but post-processed using
polishing methods, showed some variations in the thickness of the oxide layer thickness, and
a higher aluminium content on the surface than as-built AM Ti6Al4V (Vaithilingam, Prina, et
al. 2016), due to this finding, it was decided to measure the changes in chemistry across the
2 Literature Review
Page | 17
various areas in the assembly using electron dispersion spectroscopy (EDS) in this project.
Various types of titanium alloys have been investigated for additive manufacturing purposes,
alloys such as Ti–24Nb–4Zr–8Sn (Zhang, LC et al. 2018), Ti-Si, Ti-8.5Cu, Ti-5Cu, Ti6.5Cu, Ti-7Cu,
Ti10Cu (Zhang, D et al. 2019) Ti-Ta-Zr, Ti-Nb-Zr and Ti-Fe-Ta (Trevisan et al. 2018), Ti-Cr, Ti-
Nb, Ti-Mo, Ti-Ta (Nagase et al. 2019) but Ti6Al4V and pure titanium are the most common
material used for additive manufacturing and are considered the most common types due to
their excellent strength to weight ratio and corrosion resistance characteristics.
Significant research has been performed to study the application of Ti6Al4V as a superior
material of choice of metal additive manufacturing. (Qian et al. 2016)
Titanium Ti-6Al-4V equilibrium microstructure contains alpha (α) and beta (β) phases which
respectively have hexagonal close packed (HCP), and body centred cubic (BCC) lattice
structures. The α is stable at room temperature and provides the strength of the alloy,
whereas the β phase metastable and contributes to the ductility of the alloy. Ti-6Al-4V under
rapid cooling conditions, produces a brittle HCP martensitic phase (α’) which is not favourable
2.4.2 Laser fusion Multiple experiments have been conducted to establish the relationship between the main
to the overall mechanical properties of the part (Galindo-Fernandez et al. 2018).
parameters in laser powder bed additive manufactured parts such as laser power levels, layer
thickness, scanning velocity, hatch spacing, energy density and focal offset distance (FOD) and
the mechanical properties of the resulting parts. In this section we will look into the literature
available for studying these effects under constant feedstock and build shielding gas
conditions. One of such studies concluded that LPBF fabricated Ti6Al4V can achieve yield
strengths over 1300MPa, but tensile elongations to be noticeably below the minimum
threshold of 10% suggested for critical structural applications. The mechanical properties of
LPBF Ti6Al4V are defined by the phases, morphology, length scales, size and orientation of its
crystal grains (Xu et al. 2015). Laser power levels used and scanning speed used in fabricating
LPBF Ti6Al4V samples was noted during the study and their effects documented and
correlated to the microstructure and overall behaviour of the samples. For a layer thickness
of 60µm and constant energy density, a focal offset distance (FOD) of zero to 2mm was proven
2 Literature Review
Page | 18
to be favourable in terms of resulting in decrease of the alpha-prime phase and hence
improved ductility o the samples. Reducing the energy density, which is governed by equation
2-1, results in preserving alpha-prime martensite phase. Thus to achieve an ultrafine lamellar
(alpha + beta) structure, it is necessary to maintain energy densities at the desired level. (Xu
et al. 2015)
𝐸 = 𝑃/𝑣𝑡ℎ (Eq. 2-1)
Where:
E: Energy density (J/mm3)
P: Power level in Watts
v: Scanning velocity (mm/s)
t: Layer thickness (mm)
h: Hatch spacing (mm)
Literature reviews have also been conducted across several publications to compare the
fatigue and mechanical properties of parts fabricated using laser powder bed fusion (LPBF)
and electron beam fusion (EBF) methods. Figure 2-6, graphically summarises one such review,
which emphasises the need for further study on low and high cycle fatigue of LPBF AM parts.
Fatigue behaviour of Ti-6Al-4V tested at R=0.1 compared to cast and wrought, the difference
in performance was assumed to result from process-induced defects; with machining, and
heat treatment each having positive impact on the fatigue behaviour of resultant parts
2 Literature Review
Page | 19
(Lewandowski & Seifi 2016).
and wire (DED) at R = 0.1. Metallic Materials Properties Development and Standardization (MMPDS)
data for cast, wrought machined data are shown for comparison (Lewandowski & Seifi 2016).
2.4.3 Build shielding gas Shielding gas flow in the laser powder bed fusion (LPBF) additive manufacturing process plays
Figure 2-6 Summary of stress (S) versus cycles to failure (N) (S-N) data for PBF (laser), PBF (E-beam),
a crucial factor in the stability of the laser melting and the consequent layer by layer build-up
of material. This is not just to avoid oxidization during the melting process, but also influences
the state of the powder bed during the melting process, removing spatter and fumes away
from the build area, just in time for the molten area to solidify. Shielding gas are usually
recirculated in the build chamber. The shielding gas used in this research flowed directly on
the build area from above (SLM Solutions, SLM 250). The two main by-products of the laser
powder bed fusion process are welding plume and reposits and spatter and ejected powder
2 Literature Review
Page | 20
(Figure 2-7). Welding plumes consist of metal vapour and plasma plume. Metal vapour is a
result of the focused high energy laser beam taking some of the metal elements in the molten
pool of the alloy to their evaporation point and ejecting them out of the molten pool (Ladewig
et al. 2016). This was one of the reasons that an electron dispersive spectroscopy (EDS)
analysis was performed across the welded samples in this research project to verify if such a
mechanism has come into play and created any major changes in the chemical composition
in the individual heat affected zones (HAZ) or fusion zone (FZ). The plasma plumes are
generated by the ionisation of the gas between the energy (laser) source and the interaction
zone (Beck, Berger & Hugel 1995). The second by-product of the laser powder bed fusion
process is spatter and ejected powder which upon analysis of spatter formation during laser
melting it has been found that the size of such formations are significantly larger in size than
the used powder, are spherical and have chemical compositions identical to the original raw
powder, all which suggests that spatters are a result of instabilities in the molten pool rather
than being the result of vaporisation (Simonelli et al. 2015). Laser powder bed fusion process
by-products also have secondary effects on the process, namely beam attenuation and
reposition of ejected powder and spatter. Beam attenuation can be minimised by changing
the focal point position of the laser beam and minimise the chances of a “splashy process”
(Grünberger & Domröse 2015). Redeposition of spatter material could potentially lead to
major variations in layer thickness and decrease in the accuracy of build. This is usually in the
form of increased layer thickness, caused by large spherical particles from the spatter and
redeposits, which can then be torn off the base material by the re-coater and leave behind a
hold in the part, as demonstrated schematically in Figure 2-8 (Gong et al. 2013).
This research has utilised an argon shielding gas at maximum flow rate of the welding
2 Literature Review
Page | 21
machine, which was 15 litres per minute at 1.5 bars of pressure (SLM_Solutions 2016).
Figure 2-7. Schematic representation of possible process by-products courtesy (Ladewig et al.
2016)
2 Literature Review
Page | 22
Figure 2-8. Schematic of defect induced by recoating during LPBF process. (Gong et al. 2013)
2.5 Laser beam welding (LBW) 2.5.1 Introduction to lasers Laser is an acronym for Light Amplification by Simulated Emission of Radiation, a process first
proven in 1960. The laser process generates a coherent optical beam with a constant
wavelength, phase and amplitude that at any particular time has a predictable relationship to
that of any other time (AWS 2010). Laser can be produced at various wavelengths, but only a
few have industrial applications. Figure 2-9 depicts the major elements required to generate
laser. These include (1) a liaising gain medium, (2) a means of exciting the gain medium, (3)
an optical resonator, and (4) a heat sink. Gain mediums can be either solid, liquid or gaseous.
The optical cavity provides a feedback system for the system to achieve sufficient
amplification to overcome internal losses.
Material processing lasers vary in power levels and beam delivery mode. Wavelengths vary
from less than 0.2 to over 10 micrometres. Beam delivery modes also vary from pulsed to
continuous ranging from less than 1 nanoseconds up to a full continuous beam. Peak powers
reaching multiple megawatts (MW). Continuous beams average up to several kilowatts (kW).
Laser beams are delivered to the work area either through an optical cable or directly through
the atmosphere (AWS 2010).
Two of the most common industrial lasers are disc (Ytterbium doped Yttrium Aluminium
garnet – Yb:YAG) and fibre (Neodymium doped Yttrium Aluminium Garnet – Nd:YAG).
2 Literature Review
Page | 23
Figure 2-9. Four major components of a laser (AWS 2010)
2.5.2 Disk lasers Disk laser is characterised by a combination of a heat sink and a laser output that are
positioned on the opposite sides of a thin layer of active gain medium. Disk lasers are not
always circular, despite their naming since other geometries have also been used and
demonstrated as being feasible and effective. Disk lasers use a diode pumped Yb:YAG
(Ytterbium: Yttrium Aluminium Garnet) medium, which is excited via diodes at the front
surface and cooled at the rear surface using a cooling block which has a highly reflective
surface and acts as a mirror in the laser resonator. Figure 10 shows the principles of how a
thin laser disk works.
2.5.3 Laser welding Focused laser beams are one of the highest power dense sources of energy available to
Figure 2-10. Principles for thin disk laser (Antognini et al. 2009)
industry today. Industrial laser beams are so highly dense that at all materials will evaporate
if the energy can be absorbed. Therefore, when welding with laser, a hole is created due to
evaporation of the material it is subject to. This “hole” is then moved through the material,
creating a narrow trail of molten material in a parallel sided fusion zone (FZ) that collapse and
create a sealed joint. The resulting joint is called a “keyhole” weld. The concentration of the
energy of the weld is so dense that it only affects the joint and has minimal impact on the
surrounding areas. Albeit the great efficiency of laser welding, the surrounding areas are still
2 Literature Review
Page | 24
affected by the thermal gradient and are deemed as the “heat affected zone - HAZ”. The
greater the efficiency the smaller the heat affected zone (Steen & Mazumder 2010). A typical
laser welding setup is depicted in Figure 2-11.
Figure 2-11. Trump TruLaser Cell 7020 weld setup. Laser beam comes from above and nozzle feeds the inert shielding gas to the work area. Prior studies have demonstrated that weld geometry is a key factor in ruling the mechanical
properties and fatigue life of welded joints. Underfill is observed on both the top and bottom
of all butt welds. These are not defects but rather a feature and inherent characteristic of
such welds. The complex shape of weld beads is due to the interaction of many physical
phenomena, namely volume contraction surface tension, gravity, vapour pressure and phase
transformation. All of which result in complex geometries that govern the mechanical and
2 Literature Review
Page | 25
fatigue behaviour of such assemblies (Boccarusso et al. 2017).
Butt joint welding of 2mm thick titanium Ti6Al4V alloy sheets using disk laser welding were
carried out using Trumpf TruDisk 3302, at a wavelength of 1.03 micron, at Silesian University
of Poland, were laser welding parameters on weld quality and mechanical properties of test
joints were studied. The weld was performed in autogenous mode (no filler material used),
and the shielding gas was high purity argon. The study demonstrated the energy density
required to obtain full penetration keyhole welds to be 50 J/mm for a 2mm thick material.
(Aleksander 2012)
The effects of laser power and penetration depth for titanium alloys have been well
documented. It has been established that the energy density needs to be adjusted to
accommodate the thickness of the alloy. Too much energy will result in excessive evaporation
of the metal and greater thermal gradients in the heat affected zone, this has been studied
for both continuous as well as pulsed Nd:YAG laser welding of titanium alloys (Akman et al.
2009), in this study the effect of energy density on wrought titanium alloys on final assembly
hardness profile as well as porosity has been established (Figures 2-12 and 2-13). In pulsed
laser welding the peak power is the most important parameter while determining the
penetration depth which is equal to pulse energy per pulse duration (Akman et al. 2009). The
average powers used ranged from 340 Watts to 550 Watt, for a 3mm thick Ti6Al4V wrought
2 Literature Review
Page | 26
titanium alloy sheet.
Figure 2-12. Tensile test results related to the porosity ratio in welding area, according to the
average power variation (Akman et al. 2009)
Figure 2-13. Microhardness distribution of workpieces for different average power. Origin of
2 Literature Review
Page | 27
distance is measured from symmetrical midline of the welds (Akman et al. 2009)
The effects of welding travel speed on the quality of the weld have also been investigated on
wrought 2mm thick sheets using 2.5kW power levels and speeds varying from 3m/min to
10.5m/min (Cao & Jahazi 2009). Several conclusions were drawn as follows:
The FZ width decrease as the speed is increased. The HAZ was found to have a mixture
of primary alpha and alpha prime microstructures
The underfill depth decreases with the increased welding speed, with microporosities
up to 100 microns appearing in the joint area
FZ did not show any signs of solidification cracks, but the HAZ showed some
microcracks
Laser welded joints have similar or higher strength compared to the Base Metal (BM),
but have lesser ductility that then BM. Loss of ductility is contributed to the presence
2.6 Welding considerations Butt welds contain defects that occur during the fabrication process. For calculation purposes
of micropores and aluminium oxides
the probability of having a crack at the weld is almost always more than zero. Welds are
regarded as regions of weakness due to the presence of opportunities for stress
concentration, poor material properties, variations of material property in the fusion zone
and heat affected zones. Fracture mechanics is an approach used to determine propagation
life, while strain life approach is used to determine initiation life. Therefore total life is the
sum of these lives (Al-Mukhtar et al. 2011).
The fatigue assessment of welds has advanced in recent decades and is now dependant on
numerical methods such as finite element (FE) or occasionally boundary element method
(BEM) (Poutiainen & Marquis 2006). Fatigue assessment of welded structures can be based
on strains, stress or stress intensity factors and categorised into global and local methods.
Initial cracks used in the fatigue analyses are often in the range of 0.05-0.2mm (Poutiainen &
Marquis 2006), and guidelines for the application of fracture mechanics are well established
and the application of this method to welded structures is well documented and successful
2 Literature Review
Page | 28
(Maddox 1991).
2.7 Microscopy During the laser beam welding process, the joint regions are subject to severe thermal
gradients, varying from the melt pool to the base material, resulting in variations and non-
homogeneity in the microstructure of various regions of the welded assembly. In the wrought
side of the welded assembly the base metal is composed of dark beta-phase in a dominating
alpha-phase matrix, which is a typical annealed structure for alpha-beta titanium alloys. The
beta-phase is distributed at the boundary of the elongated alpha-grains. At the fusion zone
(FZ), during the solidification, body centred cubic (BCC) beta dendritic grains grow in the
direction of the heat flow. When the temperature reaches around 980 degrees Celsius (beta-
transus), diffusionless transformation [A diffusionless transformation is a phase change that
occurs without the long-range diffusion of atoms but rather by some form of cooperative,
homogeneous movement of many atoms that results in a change in crystal structure (Porter
& Easterling 2009)] of beta to alpha-prime martensitic structures occur only when the cooling
rate is sufficient at approximately faster than the critical cooling rate of 410 degree Celsius
per second; resulting in a full martensitic alpha-prime structure. The alpha-prime martensitic
plates have an acicular morphology. Heat affected zones have more complex microstructure
due to the temperature gradients present, dictated by the cooling rate. The highest cooling
rate surely takes place in the heat affected zone (HAZ) adjacent to the fusion zone. HAZ consist
of a mixture of alpha and alpha-prime (Boccarusso et al. 2017).
Laser welding of Ti6Al4V alloys results in changes to the mechanical properties such as
ductility and tensile strength, resulting from changes to the microstructure. Microcracks are
observed due to the effect of incomplete homologous wetting and residual stresses produced
by the large solidification undercooling of the molten pool (Facchini et al. 2010).
Grain size and amount of martensite formation affect the tensile strength of any alloy and
both of these factors are in turn related to the cooling rate of the welding process. The greater
the welding heat input and the lower the cooling rate, the lower the tensile strength of the
welded joint. Higher cooling rates during welding result in higher tensile strength and also
2 Literature Review
Page | 29
larger relative elongation and hence higher weld ductility (Yung, WK et al. 1997).
2.8 Research Questions There is a large volume of
literature on microscopic, microstructural and general
characterisation of laser beam powder bed fused (LPBF) parts additively manufactured
metals, namely titanium and steel (Slotwinski et al. 2014), (Lewandowski & Seifi 2016;
Slotwinski, Garboczi & Hebenstreit 2014) with limited focus on the effects of post additive
manufacturing remelting effects due to welding (Vaithilingam, Goodridge, et al. 2016). Hence
for the objective of this research in making larger additively manufactured parts it was found
that it was necessary to address the following research questions:
1. What are the most critical factors in autogenous laser beam welding that influence
the overall performance of a welded LPBF and wrought assembly?
2. What is the correlation of these factors to the performance of the assembly?
3. What microscopic parameters effect the overall performance of the welded LPBF
2 Literature Review
Page | 30
assembly, particularly in strength and fatigue performance?
2.9 References Akman, E, Demir, A, Canel, T & Sınmazçelik, T 2009, 'Laser welding of Ti6Al4V titanium alloys', Journal of Materials Processing Technology, vol. 209, no. 8, pp. 3705-3713.
Al-Mukhtar, AM, Biermann, H, Hübner, P & Henkel, S 2011, 'The Effect of Weld Profile and Geometries of Butt Weld Joints on Fatigue Life Under Cyclic Tensile Loading', Journal of Materials Engineering and Performance, vol. 20, no. 8, pp. 1385-1391.
Aleksander, APL 2012, 'Laser welding of titanium alloy Ti6Al4V using a disk laser', MTM J, vol. 7, pp. 53-56.
Antognini, A, Schuhmann, K, Amaro, FD, Biraben, F, Dax, A, Giesen, A, Graf, T, Hansch, TW, Indelicato, P & Julien, L 2009, 'Thin-disk Yb: YAG oscillator-amplifier laser, ASE, and effective Yb: YAG lifetime', IEEE Journal of Quantum Electronics, vol. 45, no. 8, pp. 993-1005.
ASTM_International 2013, F2792-12a-Standard Terminology for Additive Manufacturing Technologies.
AWS, AWS- 2010, Recommended Practices for Laser Beam Welding, Cutting, and Allied Processes,
Beck, M, Berger, P & Hugel, H 1995, 'The effect of plasma formation on beam focusing in deep penetration welding with CO2 lasers', Journal of Physics D: Applied Physics, vol. 28, no. 12, p. 2430.
Boccarusso, L, Arleo, G, Astarita, A, Bernardo, F, De Fazio, P, Durante, M, Minutolo, FMC, Sepe, R & Squillace, A 2017, 'A new approach to study the influence of the weld bead morphology on the fatigue behaviour of Ti–6Al–4V laser beam-welded butt joints', The International Journal of Advanced Manufacturing Technology, vol. 88, no. 1-4, pp. 75-88.
Cao, X & Jahazi, M 2009, 'Effect of welding speed on butt joint quality of Ti–6Al–4V alloy welded using a high-power Nd: YAG laser', Optics and Lasers in Engineering, vol. 47, no. 11, pp. 1231-1241.
Chastand, V, Quaegebeur, P, Maia, W & Charkaluk, E 2018, 'Comparative study of fatigue properties of Ti-6Al-4V specimens built by electron beam melting (EBM) and selective laser melting (SLM)', Materials Characterization, vol. 143, pp. 76-81.
Dutta, B & Froes, FS 2017, 'The additive manufacturing (AM) of titanium alloys', Metal Powder Report, vol. 72, no. 2, pp. 96-106.
EY_Global 2019b, How 3DP is moving from hype to game changer,
2 Literature Review
Page | 31
Facchini, L, Magalini, E, Robotti, P, Molinari, A, Höges, S & Wissenbach, K 2010, 'Ductility of a Ti-6Al- 4V alloy produced by selective laser melting of prealloyed powders', Rapid Prototyping Journal, vol.
Frazier, WE 'Direct digital manufacturing of metallic components: vision and roadmap', pp. 9-11.
Galindo-Fernandez, M, Mumtaz, K, Rivera-Diaz-del-Castillo, P, Galindo-Nava, E & Ghadbeigi, H 2018, 'A microstructure sensitive model for deformation of Ti-6Al-4V describing Cast-and-Wrought and Additive Manufacturing morphologies', Materials & Design, vol. 160, pp. 350-362.
Gong, H, Rafi, K, Starr, T & Stucker, B 'The effects of processing parameters on defect regularity in Ti- 6Al-4V parts fabricated by selective laser melting and electron beam melting', pp. 424-439.
Grünberger, T & Domröse, R 2015, 'Direct Metal Laser Sintering: Identification of process phenomena by optical in-process monitoring', Laser technik journal, vol. 12, no. 1, pp. 45-48.
Khorasani, A, Gibson, I, Veetil, JK & Ghasemi, AH 2020, 'A review of technological improvements in laser-based powder bed fusion of metal printers', International Journal of Advanced Manufacturing Technology, vol. 108.
Kousa, J 2019, 'How to digitalize spare parts business', Masters of International Business thesis, Lathi university of applied sciences.
Ladewig, A, Schlick, G, Fisser, M, Schulze, V & Glatzel, U 2016, 'Influence of the shielding gas flow on the removal of process by-products in the selective laser melting process', Additive Manufacturing, vol. 10, pp. 1-9.
Lewandowski, JJ & Seifi, M 2016, 'Metal additive manufacturing: a review of mechanical properties', Annual Review of Materials Research, vol. 46, pp. 151-186.
Maddox, SJ 1991, Fatigue strength of welded structures, Woodhead publishing.
Mori, S, DOI, K, Maruyama, S, HANAMI, K, KITAGAKI, H, TERAUCHI, S & MATSUSHITA, T 2018, 'Effect of Particle Size on the Quality Characteristics of Pure Titanium Fabricated Using Metal Additive Manufacturing', Journal of the Japan Society of Powder and Powder Metallurgy, vol. 65, no. 7, pp. 431-435.
Nagase, T, Hori, T, Todai, M, Sun, S-H & Nakano, T 2019, 'Additive manufacturing of dense components in beta‑titanium alloys with crystallographic texture from a mixture of pure metallic element powders', Materials & Design, vol. 173, p. 107771.
Nichols, MR 2021, 'The pros & cons of powder bed fusion', Metal Powder Report, vol. 76, no. 2, pp. 90-91.
2 Literature Review
Page | 32
Porter, DA & Easterling, KE 2009, Phase transformations in metals and alloys (revised reprint), CRC press.
Poutiainen, I & Marquis, G 2006, 'A fatigue assessment method based on weld stress', International journal of fatigue, vol. 28, no. 9, pp. 1037-1046.
Qian, M, Xu, W, Brandt, M & Tang, HP 2016, 'Additive manufacturing and postprocessing of Ti-6Al-4V for superior mechanical properties', MRS bulletin, vol. 41, no. 10, pp. 775-784.
Rayna, T & Striukova, L 2016, 'From rapid prototyping to home fabrication: How 3D printing is changing business model innovation', Technological Forecasting and Social Change, vol. 102, pp. 214-224.
Rogers, H, Baricz, N & Pawar, KS 2016, '3D printing services: classification, supply chain implications and research agenda', International Journal of Physical Distribution & Logistics Management, vol. 46, no. 10, pp. 886-907.
Simonelli, M, Aboulkhair, N, Maskery, I, Tuck, C, Ashcroft, I, Everitt, N, Wildman, R & Hague, R 'Aspects of the process and material relationships in the selective laser melting of aluminium alloys', Springer, pp. 397-404.
SLM_Solutions 2016,
Slotwinski, JA, Garboczi, EJ & Hebenstreit, KM 2014, 'Porosity measurements and analysis for metal additive manufacturing process control', Journal of research of the National Institute of Standards and Technology, vol. 119, p. 494.
Slotwinski, JA, Garboczi, EJ, Stutzman, PE, Ferraris, CF, Watson, SS & Peltz, MA 2014, 'Characterization of metal powders used for additive manufacturing', Journal of research of the National Institute of Standards and Technology, vol. 119, p. 460.
Spierings, AB, Herres, N & Levy, G 2011, 'Influence of the particle size distribution on surface quality and mechanical properties in AM steel parts', Rapid Prototyping Journal, vol.
Standard, A 2012, 'ISO/ASTM 52900: 2015 Additive manufacturing-General principles-terminology', ASTM F2792-10e1, vol.
Steen, WM & Mazumder, J 2010, Laser material processing, springer science & business media.
Trevisan, F, Calignano, F, Aversa, A, Marchese, G, Lombardi, M, Biamino, S, Ugues, D & Manfredi, D 2018, 'Additive manufacturing of titanium alloys in the biomedical field: processes, properties and applications', Journal of applied biomaterials & functional materials, vol. 16, no. 2, pp. 57-67.
2 Literature Review
Page | 33
Vaithilingam, J, Goodridge, RD, Hague, RJ, Christie, SD & Edmondson, S 2016, 'The effect of laser remelting on the surface chemistry of Ti6al4V components fabricated by selective laser melting', Journal of Materials Processing Technology, vol. 232, pp. 1-8.
Vaithilingam, J, Prina, E, Goodridge, RD, Hague, RJ, Edmondson, S, Rose, FR & Christie, SD 2016, 'Surface chemistry of Ti6Al4V components fabricated using selective laser melting for biomedical applications', Materials Science and Engineering: C, vol. 67, pp. 294-303.
Xu, W, Brandt, M, Sun, S, Elambasseril, J, Liu, Q, Latham, K, Xia, K & Qian, M 2015, 'Additive manufacturing of strong and ductile Ti–6Al–4V by selective laser melting via in situ martensite decomposition', Acta Materialia, vol. 85, pp. 74-84.
Yung, WK, Ralph, B, Lee, WB & Fenn, R 1997, 'An investigation into welding parameters affecting the tensile properties of titanium welds', Journal of Materials Processing Technology, vol. 63, no. 1-3, pp. 759-764.
Zhang, D, Qiu, D, Gibson, MA, Zheng, Y, Fraser, HL, StJohn, DH & Easton, MA 2019, 'Additive manufacturing of ultrafine-grained high-strength titanium alloys', Nature, vol. 576, no. 7785, pp. 91- 95.
Zhang, LC, Liu, Y, Li, S & Hao, Y 2018, 'Additive manufacturing of titanium alloys by electron beam melting: a review', Advanced Engineering Materials, vol. 20, no. 5, p. 1700842.
2 Literature Review
Page | 34
3. Materials and Methodology 3.1 Introduction This chapter describes the experimental procedures, including planning and design of
experiment, materials, methods and machinery used. The steps taken to fabricate test
samples, their preparation, testing and analysis are provided in detail. Observations,
conclusions and underlying theory are addressed in subsequent chapters. The experiments
were performed in two phases:
The first phase consisted of planning, fabrication, preparation, and analysis of pilot samples
that scrutinised macroscopic, microscopic, microstructural, hardness and porosity effects of
various weld parameters. The objective of the first was to assist with choosing the most
promising welding parameters that would eventually be used for the next phase. In the first
phase square shape samples were fabricated and did not follow any standard shape profiles,
since they were not subject to any strain or fatigue loading and only provided a means to
measure weld parameter effects on porosity, microstructure, and hardness.
The second phase used the welding parameters concluded in the first phase and fabricated
standard samples as per American Standard for Testing of Material (ASTM), ASTM E8 and
E466. These standards define a uniform method for quantifying tensile and fatigue behaviour
of metallic parts. Prior to the tests, the porosity profiles of each sample were measured and
recorded. Failure modes were analysed in this phase and correlated to the properties of the
joints and records of porosity to find the root cause of each failure mode. Microscopic
characteristics, such as microstructure and hardness profiles were already studied during
3.2 First Phase: Pilot Samples 3.2.1 Design of experiments To minimise the number of test samples required to deduct the optimum weld parameters, a
phase one.
Taguchi design of experiment (DOE) methodology was utilised to plan for the multiple welding
scenarios of the various parts. The main control parameters (factors) for autogenous laser
3 Materials and Methodology
Page | 35
beam welding were deemed to be laser output power (P: in Watts) and laser beam travel
speed (S: in millimetres per minute); with laser focal point (set at 0mm) and shielding (set at
20 litres per minute) considered to be constant.
Each factor was chosen to have 3 levels, based on common industry values and within the
operating range of the machinery available. Power values at levels were defined to be at 1800,
2000 and 2200 Watts and Scanning Speeds at 800, 1000 and 1200 millimetres per minute.
The combination was derived by an L9(3^2) Taguchi Array as per Table 3-1 below.
Table 3-1 Primary Design of Experiment for pilot sample weld parameters
Sample Power (Watts) Speed (mm/ minute)
1 2 3 4 5 6 7 8 9 1800 1800 1800 2000 2000 2000 2200 2200 2200 800 1000 1200 800 1000 1200 800 1000 1200
The conclusion of this Taguchi array is that a minimum of 9 samples were required to be built
and tested. An additional 3 were added to the list as contingency, bringing the total to 12.
Therefore a total of 24 square samples were printed to be welded in pairs for this phase of
3.2.2 Material: Titanium grade 5, Ti-6Al-4V (Table 3-2) in spherical granular powder form in the range 20 to
the experiments.
45 microns in diameter (Figure 3-1) was utilised in the fabrication of both the pilot and main
additively manufactured samples. The powder was manufactured by electrode inert gas
atomisation (EIGA) by Eckart TLS in Germany.
Wrought titanium alloy of the same grade (grade 5, Ti-6Al-4V) sourced via Titanium
International, a division of CEM International Pty Ltd, was used in fabricating the second half
of each test assembly specimen. The wrought titanium alloy utilised was in sheet form of 4
3 Materials and Methodology
Page | 36
mm thickness.
Figure 3-1 SEM image of the used Titanium grade 5 (Ti-6Al-4V) powder.
Figure 3-2 Titanium grade 5 alloy wrought sheets.
Table 3-2 Nominal chemical composition of Titanium alloy grade 5 used for this research, as per ASTM F2924-14.
5.5-6.75 3.5-4.5
3 Materials and Methodology
Page | 37
V Element Al wt.% Fe 0.4 O 0.2 C 0.08 N 0.05 H 0.015 Ti Balance
3.2.3 Fabrication of Laser Powder Bed Fusion (LPBF) parts
The machine used to fabricate the additive manufactured samples was an SLM Solutions ™
SLM® 250HL. This is a commercial generative selective laser melting system that produces
homogeneous metal components using fine metal powders with material densities up to
99.9%. The machine used is installed at RMIT’s advanced manufacturing precinct, in Carlton,
Victoria. This machine has a preheated building platform that can maintain temperatures up
to 200o Celsius throughout the build process to minimise the residual stresses in the
manufactured parts. This building platform can be moved up to 350mm in the Z direction. The
machine is equipped with a bi-directional loader, a safe filter system and optimised protective
gas feed. The software used to control the machine is SLM® Auto-FabCAM, which handles the
production data processing and functional strategy. Build chamber dimensions are
248x248x250 mm. Laser power available is provided by a YLR-fibre laser outputting up to 400
Watts. Nominal build speed rate is estimated at 20 cubic centimetres per hour. Laser beam
focus diameter is variable between 80 to 250 microns at a scanning speed of 20 metres per
second. Practical layer thickness varies between 20 to 75 microns. Parts can be built with a
wall thickness of down to 150 microns. The machine can utilise Argon or Nitrogen to create
an inert build environment to avoid oxidisation during the build process, this is fed at a rate
of 1.5 litres per minute. After the build is complete the inert gas can be purged at 100 litres
per minute. The machine requires a steady supply of compressed air delivered at 15 litres per
3 Materials and Methodology
Page | 38
minute at a pressure of 1.5 bars. (SLM_Solutions 2016)
Figure 3-3 SLM Solutions ™ SLM 250, metal additive manufacturing machine
Prior studies on additive manufactured titanium alloys that resulted in part densities of up to
99.98% suggest a set of build parameters (Xu et al. 2015) to be used and entered into the
SLM® Auto-FabCAM interface alongside the design of the squares cubes (Table 3-3). The
resulting parts manufactured were extracted from the machine and showed a dense and
uniform structure (Figure 3-4).
Table 3-3 LPBF base material build parameters
Raw material
Titanium Ti-6Al-4V Grade 5, 20 to 45 microns diameter spherical powder bed 100 Watts 120 microns
3 Materials and Methodology
Page | 39
Laser power Hatch spacing Laser focal point Zero Layer thickness Laser spot size Scanning speed 30 microns 80 microns 375 mm per second
Figure 3-4 20mm x 20mm x 4mm LPBF fabricated Ti64 plates for preliminary welding experiments
Once the fabrication process was complete, the chamber was evacuated from the inert gas
and square samples were removed from the powder bed, and then cleaned using a fine brush.
Adjacent faces of each square that were to be welded to their wrought counterparts were
machined down from the initial surface roughness of 20 microns down to a surface flatness
of just 1 micron. This is to ensure maximum surface contact to enable a full penetration
3 Materials and Methodology
Page | 40
autogenous welding.
3.2.4 Laser Beam Welding
The machine used to butt weld the laser powder bed fused (LPBF) titanium alloy samples to
the wrought titanium alloy samples was the Trumpf TruLaser Cell 7020 (Figure 3-5)
Figure 3-5 Trumpf TruLaser Cell 7020 welding machine and interface.
Custom jigs were fabricated and kept the two squares in compression inside the Trumpf
Trulaser ™ Cell 7020 robotic laser welding machine (Figure 3-5). This machine utilises a solid-
state disc laser, operating at 1030nm wavelength, and an inert gas weld shielding nozzle that
covers the welded area with a continuous flow of argon gas at 20 litres per minute, the welds
3 Materials and Methodology
Page | 41
were carried out using a 2mm spot size and focal distance of 223mm (Figure 3-6 and 3-7).
LPBF
Wrought
Weld direction
Figure 3-6. Custom jig to hold the square samples in compression for welding and diagram of weld direction used.
Figure 3-7. Welding set up inside the Trumpf Trulaser ™ Cell 7020 robotic laser welding
3 Materials and Methodology
Page | 42
machine chamber.
Prior to welding the 9 pilot samples, test runs were conducted on the spare samples
fabricated, to verify that the weld parameters were providing acceptable results and were
free from any major defects, such as undercuts, lack of penetration and oxidisation. These
samples were scrutinised against the American Welding Society’s standard AWS G2.4:2014.
(Figure 3-8).
Figure 3-8 Trial runs on provisional samples, to verify weld parameters. Note the direction of
welds marked on far-right sample in image. Start (S) to finish (F)
Once the parameters were validated to produce full penetration butt welds and were
sufficiently acceptable by visual inspection (Figure 3-9), the main pilot samples were loaded
3 Materials and Methodology
Page | 43
for welding on the jig and welded as per the original design of experiments table.
Figure 3-9 Work in progress of the pilot sample fabrication
Upon completion of the welding procedure, the samples were photographed and
documented for the records. The differences in the samples are examined and discussed in
Chapter 4.
3.2.5 Cross section and metallographic preparation
To study the cross section of the welds; the samples needed to be cut across the middle of
the joint area and polished. Special considerations needed to be made for cutting, grinding
and polishing of titanium alloy samples. During cutting of titanium alloys, large burrs can occur
3 Materials and Methodology
Page | 44
due to overheating. During grinding and polishing, deformities can easily occur leading to
scratches. These need to be considered in metallographic cutting, grinding and polishing.
Therefore, cutting with regular aluminium oxide cut-off wheels is very ineffective, as heat
damage in the samples can easily occur. A good alternative is to use silicon carbide with plenty
of water lubrication and cooling during the process (Figure 3-10). The welded assemblies were
cut perpendicular to the joint and to a size that would enable mounting with phenolic resin.
This was achieved using Struers Hard - 54A25 cutting wheels at 2845 rpm. All parts were then
subsequently, washed, and dried. As a result, two sets of samples were created, each from
one half of the original sample.
Figure 3-10 Struers Labotom-3 manual table-top lab size cutting machine
To facilitate microscopic analysis of the weld, it is necessary to mount the cross section into
conductive phenolic resin. This is to provide a conductive platform which could be mounted
onto scanning electron microscopes, optical microscopes and hardness testing machines to
name a few. This was achieved by placing the sample face down from the section of interest
and moulding in conductive phenolic resin using hot compression at 180 degrees Celsius for
7 minutes under 1200N force (Figure 3-11). The resulting resin cylinders were marked for
3 Materials and Methodology
Page | 45
identification and traceability on the side using a handheld engraving tool.
Figure 3-11 Struers CitoPress phenolic resin mounting machine.
The surface of the resin cylinders containing the cross sections were then placed in a Struers
RotoPol-21, polishing machine and polished with coarse P-240 grit sandpaper co-rotating at
300 rpm under 40N of force provided by a Struers RotoForce-4 machine for 5 minutes; using
water as lubricant, followed by a polishing treatment using fine OP-S 0.04-micron discs
counter-rotating at 150rpm under 30N of force for another 6 minutes (Figure 3-12).
3 Materials and Methodology
Page | 46
Figure 3-12 Struers RotoPol-21 and RotoForce-4 polishing machine.
3.2.6 Microscopy (area, shape)
After finalising the polishing process, the samples were etched using a phosphoric acid
solution. This was to pronounce the crystal structure. The cross section of welded regions was
measured using a Keyence VHX-5000 digital optical microscope (Figure 3-13). The
magnification was set at 20x (Digital Microscope VHX-5000).
Figure 3-13 Keyence VHX-5000 Digital optical microscope
Using the digital optical microscope linear and area dimensions of each distinct zone in the
cross section of the welds were measured and correlated to the weld parameters. Results of
the measurements are analysed and discussed in chapter 4.
3.2.7 Weld porosity studies
The next step to study the characteristics of the welds was to establish the relationship
between the weld parameters and their associated porosity profiles. Porosities act as stress
3 Materials and Methodology
Page | 47
concentrators and are often responsible for weak sites in structures, particularly in a welded
joint. A GE Phoenix V Tome X nano CT (Computed Tomography) scanning machine (Figure 3-
14) was used to take computed tomography images at resolutions up to 4+L /100microns as
per VDI 2630 guidelines (Richardson). The machine utilised is based at RMIT Bundoora East
campus, Victoria.
Figure 3-14 Phoenix v| Tome|x s CT scanner at RMIT Bundoora East campus
The settings used during the CT scan were as follows:
X-ray settings at 200 kilovolts and 50 micro-Amps current; resulting in a power rating
of 10 Watts, with standard focusing.
Detector settings at scan timing: 200ms; averaging 3 images; image skip at 1.
Resulting in a resolution of 13.2 microns
The results of the scans were later exported and analysed using Volume Graphics VGStudio
3.0 software on a dedicated Windows Based computer at the laboratory. Due to the large
volume of data, the computer had to have a minimum of 256GB of RAM to be able to handle
such large amount of data.
The settings for detecting porosities were as follows:
Algorithm: VDGefX; analysis mode: Void; probability threshold: 1
Minimum size at 8 voxels or an equivalent of 30 microns, which is the LPBF build
3 Materials and Methodology
Page | 48
resolution.
Figure 3-15 Image of VG Studio 3.0 interface
The results of the CT scans and their correlation with the weld parameters are discussed in
detail in Chapter 4
3.2.8 EDS (chemical composition)
The next phase of the study investigated the uniformity of the chemical composition between
the various regions (base metal, heat affected zone and fusion zone) of the welded joint
assembly. This was done at the RMMF (RMIT Microscopy and Microanalysis Facility), based in
the RMIT city campus. Using Oxford XMax20 energy dispersive spectroscopy (EDS) detector
capabilities of the FEI Quanta™ 200 scanning electron microscope and accompanying Oxford
Instruments AZTEC software (Figure 3-16). This composition examination is necessary to
determine if any variation in the chemical composition across the aforementioned regions of
3 Materials and Methodology
Page | 49
the weld can be attributed to the subsequent strength and fatigue performance of the weld.
Figure 3-16 FEI Quanta 200
EDS also known as energy dispersive X-ray spectroscopy (EDX) is a qualitative and quantitative
X-ray microanalysis technique and provides detailed information on the chemical
composition of samples. The methodology is such that a single electron beam is focused on
the surface of the sample, making the electrons penetrate the sample and interact with the
constituent atoms of the material. As a result of this interaction two types of X-rays are
generated: Bremsstrahlung X-rays which are also known as background X-rays and
Characteristic X-rays. These rays are detected by a specialised Energy Dispersive sensor which
in turn enables the software attached to the microscope display the signals as a spectrum, or
histogram, of intensity (count of X-rays or rate of X-rays) vs. energy.
3.2.9 Microscopy (EBSD)
One of the most crucial microanalyses on the samples, was undertaken to verify the
microstructure of welded assemblies across the various regions. Grain structure and changes
thereof throughout the welded joint from base metal, heat affected zone and fusion zone was
scrutinised using electron back scatter diffraction (EBSD) methodology using FEI Nova
3 Materials and Methodology
Page | 50
NanoSEM ™ 200 equipment and Oxford AZTEC software (Figure 3-17). These facilities are
available at RMIT’s Microscopy and Microanalysis Facilities (RMMF), which is one of largest
multi-disciplinary microscopy and microanalysis facilities in Australia, located at RMIT
University City campus, build 12, level 7.
Electron BackScatter Diffraction (EBSD) is a technique in Scanning Electron Microscopy (SEM),
which provides crystallographic data on the microstructure of the material in study. Target
materials are tilted in relation to the electron beam emission, resulting in diffraction of
electrons bouncing off the target material; these diffracted electrons create a pattern on a
fluorescent screen which is unique to the crystal structure and orientation of grains of that
area. These patterns are used to identify the crystal orientation, phases of each area and
boundaries of the grains and local crystalline perfections.
SEM using the EBSD technique, scan samples in a grid and establish grain morphology,
orientation and boundaries. Based on this data it can be established what the prevalent
crystal orientation is in the sample. Hence, the EBSD technique is a robust method to study
microstructures in a quantitative manner.
Typical microstructural information available via EBSD is as follows:
Visual maps and images of spatial information at resolutions of 10’s of nanometres
Grain size and phase distribution visualisation of microstructures
Visual maps of local texture and local crystal misorientations at resolutions from
millimetres up to 20 nanometres
Global crystallographic texture
Determination of local strains
3 Materials and Methodology
Page | 51
Accurate identification of minor phases
3.2.10 Micro hardness
Figure 3-17 FEI Nova NanoSEM 200
Microhardness was measured across the welded joint from additively manufactured base
metal to wrought base metal crossing both heat-affected zones and the fusion zone. A total
of 52 datum points were measured in 500-micron intervals and in three rows of 14, each 500
microns apart and symmetric to the centreline (Figure 3-18). This was performed using a LECO
LM700AT micro-hardness tester, applying a load of 300gF for 15 seconds, as per standard
ASTM procedure E384.
3 Materials and Methodology
Page | 52
Figure 3-18 Microhardness measurements performed on cross section
3.3 Second Phase: Main Samples
3.3.1 Design of experiments
The main experiment investigates the effects of the laser beam weld parameters on the
tensile strength and fatigue performance using standard sample shapes. These are the main
parameters required to judge the suitability of any LPBF-Wrought assembly part for end-use
applications. Since these characteristics are defined by the processes and conditions that
were used to create the parts and assemblies. Such processes in manufacturing are deemed
as “Special Processes” and are of vital importance. No quality control inspection can verify
the suitability of any resulting product without judging the manufacturing process used to
fabricate them.
Tensile tests were performed by subjecting the welded samples to tensile loadings until
failure and recording the stress and strain values throughout the loading to establish a stress
to strain diagram. Each test was performed at least two times to increase the confidence level
in the data obtained. The following tensile samples were fabricated and tested (Table 3-4).
Table 3-4 Design of experiment for tensile tests as per ASTM E8
Sample Description Test
1 2 3 4 5 6 7 8 Wrought single piece Wrought single piece LPBF single piece LPBF single piece Wrought and LPBF welded Wrought and LPBF welded LPBF and LPBF welded LPBF and LPBF welded Destructive tensile test Destructive tensile test Destructive tensile test Destructive tensile test Destructive tensile test Destructive tensile test Destructive tensile test Destructive tensile test
Fatigue tests were performed, using an MTS Landmark 100kN Servohydraulic Test System, by
subjecting the samples to tensile-compression loading cycles using various stress loadings,
until failure and the number of cycles to failure were recorded as an S/N curve. This curve
provides the overall fatigue performance profile of the assembly based on the various
loadings and can be extended to predict the fatigue performance of the assembly beyond the
3 Materials and Methodology
Page | 53
loadings exerted during the experiments.
Based on the results of the tensile tests, the fatigue loading settings for each combination of
LPBF-Wrought and single piece baseline samples were selected, resulting in the following
loading scenarios (Table 3-5):
Table 3-5 Proposed design of experiment for fatigue tests as per ASTM E466
Sample Description Test
1, 2, 3 4, 5, 6 7, 8, 9 10, 11 12, 13 14, 15 16, 17 18, 19 Wrought and LPBF welded Wrought and LPBF welded Wrought and LPBF welded LPBF and LPBF welded LPBF and LPBF welded LPBF and LPBF welded LPBF single piece Wrought single piece 200 MPa 400 MPa 600 MPa 200 MPa 400 MPa 600 MPa 400 MPa 500 MPa
The weld parameters to be used for the various tensile and fatigue sample assemblies had to
be uniform, hence based on the experimentation and scrutiny on the pilot samples, all
samples were welded using 2200 Watts at 1200 mm/ min which had provided the best results
3 Materials and Methodology
Page | 54
in terms of porosity, hardness and microstructure.
3.3.2 Fabrication of parts
The main experiment uses standard sample shapes as per ASTM E8 and E466 standards and
subjects them to tensile and fatigue loading and then examines their behaviour and
correlation with the welded assembly parameters in detail. The fabrication method for the
main samples is identical to the pilot samples explained in Section 3.2, with the only
difference being the geometry, where in the pilot samples the shape of the samples was
chosen as a simple square shape, because no loading conditions were explored (Figure 3-19).
Due to the main samples needing to have final shapes as per standards ASTM E8 and E466,
additional tabs were incorporated on the welded joint side of each half. These tabs allowed
the main section of the welded assemblies to be free from edge effects (effects of weld start
and finish on the parts (Figures 3-20, 3-21, 3-22 and 3-23)).
3 Materials and Methodology
Page | 55
Figure 3-19 Main sample fabrication in the SLM® 250HL machine
Figure 3-20. (left) Additively manufactured tensile testing half specimens as per ASTM E8 prior to welding, note elongated tabs the junction for welding, these are incorporated to accommodate any anomalies occurring at the start and finish of each weld (edge effects). These tabs are machined off and result in a fully compliant geometry as per ASTM E8. (right) wrought titanium alloy half specimens alongside baseline samples. These are used to establish the reference data necessary for this research.
3 Materials and Methodology
Page | 56
Figure 3-21. Tensile sample geometry according to ASTM E8 standard.
Figure 3-22. (left) Additively manufactured fatigue testing half specimens as per ASTM E466 prior to welding. Note the provision for welding tabs, similar to tensile testing samples. (right) wrought titanium alloy half specimens alongside baseline fatigue full samples.
Figure 3-23. Fatigue sample geometry according to ASTM E466 standard. (units are in mm).
Welding of the two sections into one single part was done using the same TRUMPF
experimental set up defined in the design of experimentation (Section 3.3.1). A custom jig
was fabricated to keep the two segments in compression during welding. The adjacent faces
of the two parts were also machined down to 1 micron surface roughness to ensure full
3 Materials and Methodology
Page | 57
autogenous welding (Figure 3-24).
Figure 3-24. Custom jig to accommodate ASTM E8 and E466 autogenous welding in the Trumpf Trulaser ™ Cell 7020 robotic laser welding machine.
Post welding the tabs were removed and overall shapes as per ASTM E8 and E466 were
3 Materials and Methodology
Page | 58
restored using machining (Figure 3-25) at the AMP facilities at RMIT, Carlton, Victoria.
3.3.3 Pre-testing measurement and testing:
Figure 3-25. Machined Fatigue and Tensile samples post machining and ready for testing.
All samples were measured for porosity prior to undergoing any tensile or fatigue loading
conditions. This was undertaken to ensure future analysis of failures be correlated to any
potential flaws in the joint. The GE Phoenix V Tome X S CT scanner based at RMIT Bundoora
East campus, Victoria, was used for this purpose.
The mode of scanning and porosity detection were identical to prior porosity detection
3.3.4 Tensile test set up:
activities on the pilot samples, i.e., minimum 30 microns diameter (equivalent to 8 voxels).
One of the main objectives of the tensile tests was to establish how much linear tensile force
can be exerted on the samples prior to their failure and how does the part behave under such
3 Materials and Methodology
Page | 59
loading. The yield strength was calculated by measuring when the relationship between stress
and strain cease to be linear. This includes understanding how much elastic behaviour does
the welded assembly demonstrate and what elongation profiles does it follow prior to failure.
An MTS Landmark 100kN Test Machine, based at Bundoora East campus, was utilised for the
fatigue tests. Samples were securely loaded using a rectangular chuck grip and an MTS
extensometer connected to measure the elongation and therefore strain of the part
throughout the loading cycle until failure (Figure 3-26 and 3-27). Once loaded, a
polycarbonate projectile barrier is placed in front of the machine and the run command is
given to the interface to commence the loading.
The samples were subjected to gradual tensile loads in steps of 100 Newtons. The data was
captured at a rate of 2 data points per second. A failsafe provision was included in the test
method to stop the machinery once the deformations exceeded 10mm.
Once the sample breaks, the displacement of the chucks exceeds the 3mm threshold and
triggers the failsafe mechanism of the machine; forcing a system shut down. Once it is safe to
do so, the machine is powered back up again, and the chucks are opened and the part
3 Materials and Methodology
Page | 60
released for further investigation, and made ready for the next round of experiments.
Figure 3-26 MTS Landmark 100kN Servohydraulic Test System, material testing machine. Material testing laboratory, RMIT University, Bundoora East, Victoria.
3 Materials and Methodology
Page | 61
3.3.5 Fatigue test set up:
Figure 3-27 Close up of the tensile sample loaded onto the rectangular grips and the MTS extensometer attached and armed (note pin is removed and arm is aligned, visible via the opening in which the pin has been removed from).
The fatigue performance of the welded assembly is one of the main characteristics that were
to be studied and understood in this research. To perform fatigue tests, the same MTS
Landmark 100kN machine was used.
The test methodology consisted of imposing a cyclic linear tension and compression to the
sample and measuring the number of cycles to failure. The results were graphed in what is
known as an S/N curve, where the abscissa represents the number of cycles in logarithmic
scale and the ordinate is the stress applied during the fatigue test. The tension and
compression loadings were not uniform, with the tension to compression ratio being 10 to 1.
The frequency of loadings was set at 30Hz. This frequency was chosen based on prior
literature and to accommodate the variations in build quality such as base material porosity,
inclusions, etc. that can weaken the parts beyond the nominal involved in this study.
A failsafe provision was set for the test to avoid any damage to the sample or machinery which
would come into effect were the machine detects displacements greater than 3mm. A
3 Materials and Methodology
Page | 62
polycarbonate ballistics shield was also put in place to prevent any flying debris (Figure 3-28).
Figure 3-28 Polycarbonate ballistics shield to protect against flying debris
Fatigue test settings and loadings were applied as per the design of experiment stated in Table
4 of Section 3.3.1. with the higher loadings tested first to ensure the welded parts are able to
withstand designed loadings, then lesser loadings were tested gradually to produce the S/N
3 Materials and Methodology
Page | 63
curve.
3.3.6 Post-test examinations:
After the completion of all tensile and fatigue tests the location of the fracture was imaged
using scanning electron microscopy (SEM) and computed tomography (CT) scanning. The
main objective of these examinations is to find the root cause of each failure mode and
correlate that to the original build and weld parameters.
The main areas examined were regions of discontinuity including porosities, inclusions and
3 Materials and Methodology
Page | 64
metallurgical changes across the joint.
3.4 References
Digital Microscope VHX-5000, www.keyence.com,
Richardson, H, Phoenix V|tome|X Series - 2D X-Ray /3D CT /3D Metrology,
SLM_Solutions 2016,
Xu, W, Brandt, M, Sun, S, Elambasseril, J, Liu, Q, Latham, K, Xia, K & Qian, M 2015, 'Additive manufacturing of strong and ductile Ti–6Al–4V by selective laser melting via in situ martensite decomposition', Acta Materialia, vol. 85, pp. 74-84.
3 Materials and Methodology
Page | 65
4. Results and Discussion 4.1 Introduction The results of the experiments performed at each stage of this research are detailed and
discussed within this chapter. The methods and equipment used to achieve these results were
4.2 Pilot samples The pilot samples were fabricated to examine the effects of various weld parameters that
detailed in chapter 3. The final conclusions are summarised in chapter 5.
were chosen in the design of experiments. Three levels of laser power were chosen, and three
levels of laser welding scanning speed were also shortlisted. The resulting 9 samples were
scrutinised for discolouration, weld geometry, changes to chemical composition,
4.2.1 Macroscopy 4.2.1.1 Weld appearance and comparison to AWI standards Upon completion of the welds as per the design of experiments, the welded joints were
microstructure, porosity and variations in hardness.
photographed using a Keyence VHX-5000 digital microscope (Figures 4-1 and 4-2). The
resulting images were compared to the American Welding Society, welding guideline, AWS
G2.4M:2014 – guide for the fusion welding of titanium and titanium alloys, (AWS 2014) (Figure
4-3) none of the welds had experienced any major contamination. This was with the exception
of some minor surface contamination that was verified to have not affected the integrity of
the welds, via subsequent examination of the weld cross section. In some cases, the welded
surfaces showed straw colour in the heat affected zone adjacent to the fusion zone, which is
an indication of mild surface contamination, and as per the guideline, is acceptable. The
recommendations for such cases are surface treatment by brushing and correction of
4 Results and Discussion
Page | 66
shielding, hence for the purpose of this research, it is deemed as acceptable.
Figure 4-1 Top side of welded pilot samples
4 Results and Discussion
Page | 67
Figure 4-2 Bottom side of welded pilot samples
Table 4-1 Surface colour in titanium welds; as per American Welding Society (AWS) G2.4M:2014 – guide for the fusion welding of titanium and titanium alloys.
Colour Silver Light straw to dark straw
Purple Action None Remove surface oxide by brushing, correct shielding Remove weld bead by grinding correct shielding
Dark blue to light blue Remove weld bead
Yellow to grey to white Interpretation Correct shielding Slight surface contamination, acceptable Moderate surface contamination, unacceptable Heavy contamination, unacceptable Very heavy contamination, unacceptable
4.2.2 Optical Microscopy 4.2.2.1 Weld area measurement, shape, and penetration profile
Remove weld bead and 1/32 in [0.8mm] to ¼ in [6mm] of the material beneath
One of the factors that affect the performance of a welded joint is the overall shape of the
fusion zone of the weld. Undercut geometry has a significant effect on fatigue performance,
due to the stress concentration phenomena. This has been studied (Ninh Nguyen & Wahab
1995) and modelled by many scholars (Liinalampi, Remes & Romanoff 2019).
To study the geometry of the welded joint, metallographic specimens were prepared, using
the methodology described in Chapter 3. The surfaces of the metallographic samples were
polished and etched in preparation for microscopic investigation.
Using the Keyence VHX-5000 mentioned earlier, the shape and area of the three distinct,
fusion zone (FZ), heat affected zone (HAZ) were captured and the area of each, distinguished
from the base metal (BM) and measured using the software available on the digital
4 Results and Discussion
Page | 68
microscope (Figure 4-3). Sample number 9 had one of the most uniform cross sections.
Area
Parameters Dimensions Sample #1 1800 Watts @800 mm/ min
Sample #2 1800 Watts @1000 mm/ min
Sample #3 1800 Watts @1200 mm/ min
Sample #4 2000 Watts @800 mm/ min
Figure 4-3. Combined cross section area of fusion and heat affected zones (continued in
4 Results and Discussion
Page | 69
next two pages).
Area
Parameters Dimensions Sample #5 2000 Watts @1000 mm/ min
Sample #6 2000 Watts @ 1200 mm/ min
800
Sample #7 2200 Watts @ mm/ min
Sample #8 2200 Watts @ 1000 mm/ min
Figure 4-3. (continued) Combined cross section area of fusion and heat affected zones.
4 Results and Discussion
Page | 70
Area
Parameters Dimensions Sample #9 2200 Watts @ 1200 mm/ min
Figure 4-3. (continued) Combined cross section area of fusion and heat affected zones.
(cid:3017)
Considering that the energy density for welding is calculated using the following formula (Equation 4-2):
(cid:3049)
𝐸 = Eq. (4-2)
Where:
𝑃: Laser power
𝑣: Scanning speed
Using Minitab® the results of the area measurements were compared to the weld energy
density figures, as per Equation 4-2. Pearson’s correlation coefficient measures the degree of
linear relationship between two variables. The correlation coefficient assumes a value
between -1 to +1.
(cid:3289) (cid:3284)(cid:3127)(cid:3117)
For the values 𝑥 and 𝑦 :
∑ ((cid:3051)(cid:3284)(cid:2879)(cid:3051)̅)((cid:3052)(cid:3284)(cid:2879)(cid:3052)(cid:3364)) ((cid:3041)(cid:2879)(cid:2869))(cid:3046)(cid:3299)(cid:3046)(cid:3300)
𝜌 = Eq. (4-3)
Where:
𝑥̅ Sample mean for the first variable
Standard deviation for the first variable 𝑠(cid:3051)
𝑦(cid:3364) Sample mean for the second variable
Standard deviation for the second variable 𝑠(cid:3052)
4 Results and Discussion
Page | 71
𝑛 Column length
Figure 4-4. Scatterplot of combined fusion zone (FZ) area and heat affected zone (HAZ) area (in mm2) versus welding energy density (J/mm).
Upon populating the Pearson’s correlation coefficient as per Table 4-2, which summarises the
fusion zone (FZ) cross section area to the energy density, a coefficient of 0.997 was concluded
that testifies to the strong positive correlation between those two. Hence, an increase in
energy density will also lead into an increase in fusion zone cross section, therefore it would
be prudent to deduce that it will also increase the volume of the overall fusion zone at the
welded joint (Figure 4-4).
4 Results and Discussion
Page | 72
Table 4-2 Summary of weld parameters, resulting energy density and welded joint fusion zone cross section area on pilot samples as measured using digital microscopy. Energy (E=P/v) Sample 135 1 108 2 90 3 150 4 120 5 100 6 165 7 132 8 110 9 Area 18.166 14.148 11.684 20.614 16.205 13.347 21.942 17.929 14.941 Power 1800 1800 1800 2000 2000 2000 2200 2200 2200 Speed 800 1000 1200 800 1000 1200 800 1000 1200
4.2.3 Scanning Electron Microscopy
At the conclusion of the optical study of the cross section of the samples, the focus shifted
towards the study of the samples using SEM (scanning electron microscopy) technology.
4.2.3.1 EDS and chemical composition results The first stage of the study was to establish the chemical composition of the samples across
the welded joint, to establish if subjecting the assemblies to the weld energies have resulted
into any changes in the chemical composition of the fusion zone (FZ) or heat affect zones
(HAZ). The methodology used is detailed in Section 3.2.8 of the Material and Methodology
Chapter.
The FEI Quanta 200 and Oxford XMax20 energy dispersive spectroscopy (EDS) detector
4 Results and Discussion
Page | 73
provided the following graph (Figure 4-5) using the Oxford Instruments AZTEC software.
Figure 4-5. (Top) Cross section of welded joint as visible via SEM microscope. (Bottom) Energy dispersive spectroscopy (EDS) results across welded joint, sample number 8.
The resulting data matched the EDS study results performed on the base metal (BM) section
4 Results and Discussion
Page | 74
of the welded joints (Figure 4-6).
Figure 4-6. Energy dispersive spectroscopy (EDS) results of base metal (BM) at the additive manufactured side of sample number 8.
The values derived by the EDS analysis are in line with the initial material properties provided
in the certificate of compliance (CoC) of the raw material (Table 4-3).
Ti Table 4-3 Ti6Al4V raw material properties Grade 5 Bal. V 3.5-4.5 Al 5.5- Fe ≤0.40 C ≤0.08 N ≤0.05 H ≤0.015 O ≤0.20
6.75
4.2.3.2 Microstructure and EBSD results The grain structure of additive manufactured and wrought parts and the changes to each,
within and around the molten pool that eventually creates the fusion zone (FZ) and the
surrounding heat affected zones (HAZ) was of great interest for this research, since these
factors affect the mechanical properties and fatigue performance of the assembly. The best
tool available to capture microstructure properties of the parts in question, in this research
was the electron backscatter diffraction (EBSD) technique in scanning electron microscopy
(SEM). This technique was carried on pilot sample 9, using the parameters shown and
4 Results and Discussion
Page | 75
resulting images as per Figure 4-7.
20.00 kV
Accelerating Voltage
69.93°
Specimen Tilt (degrees)
73.51%
Hit Rate
4.00 Hz
Speed of Acquisition
Figure 4-7. Electron backscatter diffraction images of fusion zone on sample 9. Top left shows the overall area under study, with the zoomed area highlighted in the top right image. Needle shaped titanium alloy grains captured middle left using the parameters above middle right. Phase identification bottom left and highlighted of Titanium hexagonal close packed (HCP) crystals bottom right. Blue indicates Ti-Hex (alpha) and Red indicates Ti-Cubic (beta).
4 Results and Discussion
Page | 76
The dominant phase in the fusion zone was identified as hexagonal close packed (HCP/ alpha)
grain structure, as evident in the phase colour image bottom row of Figure 4-7 and Figure 4-
8, as well as the report generated by the Oxford Instruments AZTEC software (Table 4-4).
Table 4-4. Phase identification in fusion zone of pilot sample number 9.
Phase Phase Phase Mean Standard Min band Max band
name fraction count band deviation contrast contrast
(%) contrast band
contrast
Titanium 0.02 3 73.67 8.38 65.00 85.00
cubic
Ti-HEX 73.49 10546 108.85 23.10 37.00 160.00
Zero 26.49 3802 70.75 17.39 0.00 138.00
solution
4 Results and Discussion
Page | 77
Figure 4-8. Zoomed out phase identification of pilot sample number 9. (Right) Blue area is Ti hexagonal close packed (HCP) and yellow area is Titanium body centre cubic (BCC).
Grain size and amount of martensite (α or hexagonal close packed (HCP) crystal) formation
affect the tensile strength of the alloy. (Yung, WKC et al. 1997). Martensite grains are higher
in strength but lower in ductility.
Titanium Ti6Al4V alloy is made of Titanium, Aluminium and Vanadium. Aluminium is an alpha-
phase stabiliser, and hence increases the beta-transus temperature, so that alpha phases can
stay stable at higher temperatures, the alpha phase is responsible for the ductility of the
material. Vanadium on the other hand is a beta-phase stabiliser, which tends to prevent the
formation of martensite alpha-prime phases within the alloy. Alpha-prime phases have
superior tensile strength but are brittle in nature.
Titanium-base alloys such as Ti6Al4V have reversible allotropic transformation properties in
which crystal structures change from an alpha structure (hexagonal close packed) to beta
(body centred cubic) when the temperature exceeds a certain level called beta-transus
(approximately 995 degrees Celsius). Additive manufacturing Ti6Al4V feedstock have a crystal
structure comprised of both phases (alpha + beta), in which the alpha phase forms a matrix,
and the inter-granular beta phase is distributed at the boundaries of this matrix. With the
cyclic thermal heating and high cooling rates inherent to LPBF additive manufacturing
processes the beta phase transforms to martensite alpha-prime (hexagonal close packed) and
the alpha grains are elongated in the direction of the heat transfer. Alpha-prime phases
compete with the stable alpha phase, which provides the material’s strength, and the
4 Results and Discussion
Page | 78
metastable beta phase provides the ductility.
4.2.4 Porosity analysis Weld porosities are one of the most common areas for stress concentration and act as crack
initiation sources. The more non-spherical the porosity the higher the stress concentration.
To establish the porosity profile of each pilot sample and to find their correlation with weld
parameters, all samples were scanned using the computed tomography (CT) scanning
machine at the RMIT University Bundoora East campus.
The welded joint of each sample was CT-scanned and then post-processed to include a
consistent region of interest (ROI) throughout all samples. The ROI was established using the
software capabilities of the VG Studio 3.0 software.
The parameters used during the post-processing of the CT data, were set so that any porosity
above 30 microns were to be detected and made part of the statistics. All data were entered
into Minitab ® for statistical analysis and histogram plot generation. Figures 4-9 and 4-10 show
the CT scan and post processing of the porosities as well as detail the porosity distribution in
a histogram for the pilot samples with the most and least porosities. The most promising
profile was found to be in sample number 9, with only 99 pores with a mean size of
0.1328mm. The largest pore to be found was only 0.2mm in diameter and the pores were
distributed evenly with a standard deviation of 0.0382mm Figure 4-10 (bottom).
The histograms plotted on Figures 4-9 and 4-10, show that sample number 6, which has the
highest porosity, in terms of percentage of total volume, has a high concentration of pore
diameters within the 0.14mm to 0.3mm and a few pores at the larger diameters of 0.9mm
and 1.8mm, with a standard deviation of 0.1416; whereas sample number 9, which has the
lowest porosity has a more uniform pore size distribution between up to 0.21mm and a
4 Results and Discussion
Page | 79
standard deviation of 0.03824.
4 Results and Discussion
Page | 80
Figure 4-9. Porosity profile of pilot sample number 6, welded at 2000 Watts and 1200mm/min. Note that 1.28% of overall region of interest (ROI) consists of pores. 241 defects were detected at a mean diameter of 0.14mm, with a mean sphericity of 0.5171.
Figure 4-10. Porosity profile of pilot sample number 9, welded at 2200 Watts and 1200mm/min. Note that only 0.12% of overall region of interest (ROI) consists of pores. 99 defects were detected at a mean diameter of 0.12mm, with a mean sphericity of 0.4901.
4 Results and Discussion
Page | 81
The relationship between the porosity volume (in percentages of overall volume of the region
of interest – ROI) to the mean pore size was investigated using a Pearson correlation equation
(Equation 4-3, referred to earlier in this chapter) with the following conclusion that all samples
with the average pore size of 0.142mm demonstrate a direct relationship between porosity
Scatterplot of Porosity % vs Mean pore size
1.4
1.2
1.0
%
0.8
0.6
y t i s o r o P
0.4
0.2
0.0
0.00
0.05
0.10
0.20
0.25
0.30
0.15 Mean pore size (mm)
Figure 4-11. There is direct relationship between porosity and mean pore size
and mean pore size. The P-value for this correlation is 0.237 (Figure 4-11).
The relationship between porosity volume generated in samples to welding parameters was
investigated. A Pearson correlation equation (Equation 4-3) was utilised again. This was done
in three steps:
Step 1: Correlation of porosity volume to laser welding power (Figure 4-12).
Step 2: Correlation of porosity volume to laser welding travel speed (Figure 4-13).
Step 3: Correlation of porosity volume to energy density (Figure 4-14).
As predicted, there was a direct relationship between porosity and welding power levels (ρ =
4 Results and Discussion
Page | 82
0.157), and inverse relationship between porosity and welding travel speeds (ρ = -0.101) and
an overall direct relationship between porosity and energy density of the welding process (ρ
= 0.175).
Figure 4-12. There is direct relationship between laser welding power levels (Watts) and the amount of porosity (% volume) in the final resulting welded joint. This is shown with a probability value (P-value) of 0.686.
4 Results and Discussion
Page | 83
Figure 4-13. There is an inverse relationship between the welding laser beam travel speed (mm/ min) and the resulting porosity (% volume), with a P-value of 0.796.
Scatterplot of Porosity % vs Energy (E=P/v)
0.8
0.7
0.6
0.5
0.4
% y t i s o r o P
0.3
0.2
0.1
90
100
110
120
130
140
150
160
170
Energy (E=P/v)
Figure 4-14. (left) Direction relationship between welding energy density and resulting porosity is not evident until (right) outlier datapoint (100 J/mm) are taken out, resulting in a clear demonstration of a relationship between the welding energy density (J/mm) and the resulting porosity (% volume) of the welded joints. Since sphericity (defined as a unitless measure where the surface area of a sphere of the same
volume as that of the pore is divided by the actual surface area of the pore) of the pores is
also one of the factors affecting the fatigue performance (due to stress concentration) of
welded joints, this too was scrutinised using the Pearson correlation equation (Equation 4-3).
As demonstrated in the scatterplot diagrams of Figures 4-15 and 4-16. There is an inverse
relationship between the resulting sphericity of pores and both laser beam welding power
levels and welding speed levels.
4 Results and Discussion
Page | 84
Figure 4-15. (left) inverse relationship between laser beam welding power levels with a P- value of 0.448 and Figure 4-16 (right) show the inverse relationship between porosity sphericity to laser beam welding travel speeds with a P-value of 0.690.
4.2.5 Hardness profile Variations in hardness across the various regions in a welded joint also affect the overall
mechanical and fatigue performance of the assembly (Leuders et al. 2015).
A Leco LM700AT micro-hardness testing machine (Figure 4-18) was utilised to establish the
profile of the Vickers hardness values across the various regions such as the base metal (BM),
heat affected zone (HAZ) and fusion zone (FZ). Each location was measured three times across
intervals detailed in Chapter 3.
4 Results and Discussion
Page | 85
Figure 4-16. Leco LM700AT micro-indentation hardness testing machine.
The profile of the average hardness values vs. location within each assembly showed a great
variation, which is plotted in the diagrams on Figure 4-17 through to Figure 4-19. As is evident
in all three diagrams the hardness of the LPBF side of the assembly is greater than the wrought
side, which testifies to the fact that additively manufactured parts have greater hardness due
to the cyclic heating and cooling cycles they go through during the build process resulting in
Figure 4-17. Microhardness profile across three different assemblies built using 1800W power level.
Figure 4-18. Microhardness profile across three different assemblies built using 2000W power level.
4 Results and Discussion
Page | 86
changes in their microstructure and formation of harder crystal structure.
Figure 4-19. Microhardness profile across three different assemblies built using 2200W power level.
Throughout the 9 pilot samples studied, the average Vickers microhardness of wrought and
Laser powder bed fused (LPBF) Ti6Al4V is measured at 300 HV and 390 HV, respectively
Figure 4-20. The greatest variation in hardness happens on the HAZ area of LPBF. Each graph represents the power level listed below the diagram. 90 J/mm2 is highlighted in red.
4 Results and Discussion
Page | 87
(Figure 4-20).
The heating and cooling effects during welding process have an impact on the softening or
hardening of the various region of each welded assembly, due to the changes they make in
the microstructure of each region, resulting in strength mismatch across the cross section. By
analysing the variations in hardness across various samples, it was demonstrated that sample
number 9 had the least amount of sudden variation in hardness across the various cross
4 Results and Discussion
Page | 88
sections. The greatest variations observed were at the LPBF side of the HAZ.
4.3 Main samples After scrutinising the various results from the tests performed on the pilot samples, it was
concluded the parameters leading to the most uniform geometrical shape fusion zone cross
section (Section 4.2.2), least amount of porosity (Section 4.2.4), and most uniform distribution
of hardness (Section 4.2.5), are the ones that were used to make pilot sample number 9; i.e.
2200 Watts at a laser welding scanning speed of 1200 mm/ min. Hence for the main portion
of the research which focuses on the tensile strength and fatigue performance of the
4.3.1 Design of experiments Based on the lessons learnt from the pilot samples, main samples needed to be fabricated as
autogenously welded Titanium alloys, all parts were fabricated using these parameters.
per standard shapes (ASTM E8 and ASTM E466) and assessed via a non-destructive testing
method (i.e., CT scanning), tested in such a quantity and fashion that would be acceptable to
capture any testing variations and studied further after the tests concluded using CT scanning
and Scanning Electron Microscopy. Hence the following design of experiment was proposed:
Tensile tests:
o Single piece wrought – to establish baseline
o Single piece Laser Powder Bed Fused (LPBF) – to establish baseline
o Welded wrought to LPBF
o Welded LPBF to LPBF parts
Fatigue tests
o Single piece wrought (baseline)
o Single piece LPBF (baseline)
o Welded wrought to LPBF parts (at various stress levels, defined at the conclusion
of tensile tests)
o Welded LPBF to LPBF parts
The nature of all tests was destructive, and the failure areas were to be scrutinised using
scanning electron microscopy (SEM) and computed tomography (CT) scanning, to establish a
4 Results and Discussion
Page | 89
clear understanding of the mode of failure and root cause of destruction.
4.3.2 Tensile tests 4.3.2.1 Tensile test setup The first step in establishing the mechanical performance of the welded assemblies is to
establish their tensile strength, hence tensile tests were carried out at RMIT Bundoora East
campus using MTS Landmark Servohydraulic test system and MTS extensometer model
634.12F-24 (Serial Number 10413718A) controlled via the MTS FlexTest 60 application
software. Material tests were carried out on ASTM E8 standard specimens using the following
test parameters:
Axial displacement.
Data logging for time, axial force, axial displacement (elongation) and axial strain.
Data acquisition at 0.5 second intervals.
Ramp in tension rate at 0.00625 mm/s.
By plotting the engineering stress (σ, calculated by dividing axial force by smallest cross
section of test specimen) to engineering strain (ε) readings from the test apparatus, a stress-
strain curve was generated. Subjecting the specimen to linear forces and elongating the
sample (strain) results in generation of stress in the part. More strain results in more stress.
Releasing the strain, will result in the stress to be released, up to the point in which the
relationship between the stress and strain ceases to be linear, the part is regarded to be
elastic and deformations reversible. At the point in which the relationship between the strain
and stress cease to be linear, the part will be at its yield strength. In ductile material, further
application of strain beyond the yield strength will result in decrease in stress, then eventually
this will continue to result in more stress again, up to the maximum value of ultimate strength.
Further strain beyond the ultimate strength of the material will result in a decrease in stress
up to the point of rupture, where the curve will end. Brittle material lacks two distinct points
for yield strength and ultimate tensile strength (UTS), and both coincide at a singular value,
corresponding to the UTS.
4.3.2.2 Single piece wrought and LPBF tensile tests To establish a baseline understanding of the stress-strain characteristics of the constituent
4 Results and Discussion
Page | 90
material making the wrought and additive manufacture welded assembly, separate tests
were carried out on individual single piece wrought and single piece additively manufactured
samples.
4.3.2.3 Porosity analysis – Single piece wrought and LPBF Prior to the tensile tests on the parts, each were subjected to computed tomography (CT)
scanning. None of the scans showed any detectable porosity defects in the wrought samples.
The LPBF single piece CT scan image (Figure 4-21) and porosity distribution and size graph
(Figure 4-22), demonstrated that most of the porosities are concentrated at the ends of the
part and away from the mid-section, with the median diameter of the 19 pores detected being
240 microns.
4 Results and Discussion
Page | 91
Figure 4-21. CT scan image of porosities. Note only pores with a probability (as per CT Scan data post-processing analysis) greater or equal to 1 have been annotated here.
Porosity Diameter
6000000
5000000
4000000
3
m ) µ (
3000000
l
e m u o V
2000000
1000000
0
0
200
400
600
800
1000
1200
Position z axis (voxels); 1 voxel = 25.967 microns
Porosity Diameter
Figure 4-22. CT scan data graphical representation of porosities along the length of the tensile test sample, note that the origin of the z position is the middle of the sample. Circle diameters proportional to diameter of porosities. Tensile tests – LPBF single piece The single piece laser powder bed fused (LPBF) sample demonstrated a brittle failure
behaviour, as shown in Figure 4-23, the plane of failure is almost perpendicular to the axis of
4 Results and Discussion
Page | 92
force application, which is consistent with brittle fracture.
Figure 4-23. Tensile test on single piece LPBF Ti6Al4V. Note the fracture geometry.
The LPBF sample brittle failure mode was also confirmed by the way it behaved under axial
loading. The Yield Strength (YS) and Ultimate Tensile Strength (UTS) of the sample were a
4 Results and Discussion
Page | 93
singular point at 877.28 MPa and the breakage point was at 1232.27 MPa (Figure 4-24).
ASTM E8 tests on non-welded single piece LPBF sample
1400
1200
1000
800
) a P M
600
( s s e r t S
400
200
0
0
2
4
8
10
12
6
Strain (%)
Post-test CT scan – LPBF single piece After the tensile tests were concluded, the broken parts were CT scanned to identify the
Figure 4-24. Stress-strain diagram of single piece LPBF tensile sample as per ASTM E8.
location of porosities within the vicinity of the fractured zone. The single piece LPBF tensile
test sample didn’t show any noticeable concentration of porosities within the fractured zone.
Only three pores were identified as having a greater probability equal or greater than 1. Only
one of these pores was at a proximity to the fractured surface (defect 7, Figure 4-25). We will
investigate the porosities on the surface of the fracture via the scanning electron microscopy
4 Results and Discussion
Page | 94
at the section.
Post-test SEM – LPBF single piece The LPBF single piece sample was studied using the Quanta 200 scanning electron microscope
Figure 4-25. CT scan of single piece LPBF tensile test sample, not showing any significant porosities around the fractured surface, besides the single pore at 180 microns diameter (defect 7).
to image the fracture surface and investigate the failure mode and root cause of the rupture
(Figure 4-26). As is evident in the images, the brittle failure is evident by the flat surface at the
mid-section of the part. Multiple porosities are evident on the surface of the fracture and also
4 Results and Discussion
Page | 95
on the perimeter of the ruptured area (see Figure 4-26 bottom-right).
4 Results and Discussion
Page | 96
Figure 4-26. Fractured surface of LPBF single piece tensile sample. Note the flat surface of failure perpendicular to the axis of stress loading at the middle of the part. Top image at 38x magnification. Bottom left at 50x facing the upper and bottom right at 50x, facing the lower area of the face. Note dark points as porosities.
Tensile tests – Wrought single piece To gain statistical confidence in the results, three additional single piece wrought samples
were tested to destruction. The point of rupture shows the dominance of sheer stress forces
at the 45 degrees plane, resulting in a cup and cone (Figure 4-27 – top two samples) or slanted
45 degrees (Figure 4-27 – bottom sample) fracture shape. This cup and cone fracture is
consistent with a ductile fracture failure zone.
4 Results and Discussion
Page | 97
Figure 4-27. Three identical ASTM E8 wrought Ti6Al4V alloy samples were tested for tensile strength. The top 2 resulted in a cup and cone fracture surface and the last sample resulted in a single slanted 45 degrees fracture.
All three test samples demonstrated elastic behaviour. The following values for Yield Strength
(YS), Ultimate Tensile Strength (UTS), and Breaking Strength and averages were derived:
TW1
858.93 MPa
Table 4-5 ASTM E8 tensile test results on wrought single piece samples Yield Strength Breaking Strength Max Elongation
825.77 MPa
TW2
890.99 MPa
952.12 MPa
837.12 MPa
3.157
TW3
862.53 MPa
970.47 MPa
897.87 MPa
3.259
Mean
870.82 MPa
944.90 MPa
853.59 MPa
3.266
Ultimate Tensile Strength 912.12 MPa (mm) 3.383
The ductile behaviour was also evident in the three stress-strain diagrams as depicted in
ASTM E8 tests on Wrought Single Piece
1200
1000
800
) a P M
TW1
600
TW2
( s s e r t S
TW3
400
Mean
200
0
0
2
4
6
8
10
12
Strain (%)
Figure 4-28.
Post-test SEM Images were taken from the fracture surface of the wrought tensile samples using scanning
Figure 4-28. Stress-Strain diagram of wrought single piece tests as per ASTM E8.
electron microscopy (SEM). Multiple points of rupture initiation were evident (Figure 4-29),
4 Results and Discussion
Page | 98
all showing signs of failure on the plane of maximum sheer (45 degrees).
4 Results and Discussion
Page | 99
Figure 4-29. Fractured surface of wrought single piece tensile sample. Note the 45 degrees angle between the surface of failure and axis of stress loading. Top image at 50x magnification, bottom left at 800x and bottom right at 400x, highlighting one of the points of failure.
4.3.2.4 Welded wrought to LPBF assembly tensile tests 4.3.2.4.1 Pre-test porosity analysis – Wrought to LPBF To study the tensile strength of samples made of welded wrought and LPBF Ti6Al4V and also
gain statistical confidence in the results, three specimens were fabricated and machined to
comply with the ASTM E8 standard profile. Then the assemblies were CT scanned using the
equipment and methods detailed in Section 3.2.7. The focus of all scans was mainly the
welded region, with the 5 largest pores being highlighted (Figures 4-30, 4-31 and 4-32). The
porosity data was analysed using Minitab (Figure 4-33). The majority of the pores are found
to be concentrated at the trailing and lower end (in reference to the direction of travel of the
Histogram of Dia (microns) Normal
14
Mean 302.1 StDev 198.6 34 N
12
10
8
6
y c n e u q e r F
4
2
0
0
200
400
600
800
1000
Dia (microns)
welding laser beam) of the fusion zone of the welded assemblies.
4 Results and Discussion
Page | 100
Figure 4-30. Sample number 1, CT scan and porosity distribution histogram
Histogram of Dia (microns) Normal
20
Mean 225.9 StDev 108.1 46 N
15
10
y c n e u q e r F
5
0
0
100
200
500
600
700
300 400 Dia (microns)
4 Results and Discussion
Page | 101
Figure 4-31. Sample number 2, CT scan and porosity distribution histogram.
Histogram of Dia (microns) Normal
14
Mean 205.8 StDev 66.52 43 N
12
10
8
6
y c n e u q e r F
4
2
0
80
160
240
320
400
Dia (microns)
4 Results and Discussion
Page | 102
Figure 4-32. Sample number 3, CT scan and porosity distribution diagram.
N
lamro
700.0
600.0
elbairaV 1 elpma S 2 elpmaS 3 elpmaS
500.0
NveDtSnaeM 436.8911.20 3 3425.668.502 641.8019.522
400.0
y t i s n e D
300.0
200.0
100.0
000.0
0
002
006
008
000 1
004 D
ata
Figure 4-33. Minitab combined analysis of porosity data for test samples 1, 2 and 3.
4.3.2.4.2 Tensile tests – Wrought to LPBF To gain statistical confidence in the results, three welded LPBF to wrought samples were
tested to destruction. The area of fractures was imaged using CT scanning and scanning
electron microscope (SEM) methods, to capture the root cause and mode of failure. Upon
completion of all three tests, it was obvious that all failures were in the wrought side of the
assembly and the failure mode was ductile, resulting in a cup and cone fracture surface at a
45 degrees angle relative to the axis of loading. The locations of the failures were all beyond
the fusion zone (FZ) and at the border areas of the heat affected zone (HAZ). See Figure 4-34.
Therefore, it can be concluded that such welded assemblies made of wrought and LPBF, are
not compromised at the welded joint, thus making such assemblies viable for industrial
applications, as per the objectives of this research. The performance of the welded joint will
depend on the characteristics (porosity or embrittlement) of the heat affected zone (HAZ),
4 Results and Discussion
Page | 103
which needs to be monitored and controlled as per welding the relevant procedures.
Figure 4-34. Three welded LPBF to wrought samples were tested up to failure. All failures were in the wrought side of the assembly and beyond the fusion zone.
4 Results and Discussion
Page | 104
The force (and calculated stress) and strain readings were plotted in a stress-strain diagram, which
also demonstrated characteristics of a ductile behaviour (Figure 4-35), i.e., large strain and distinct
ultimate and breaking strengths, as well as cup and cone failure planes at 45 degrees to the plane of
tension.
ASTM E8 tests on welded wrought to LPBF assemblies
1200
1000
800
) a P M
600
( s s e r t S
TLW1 TLW2 TLW3 Mean
400
200
0
0
1
2
3
4
5
6
7
Strain (%)
The yield strength, ultimate tensile strength and breaking strength values of the three samples are
summarised in Table 4-6.
Figure 4-35. Stress-Strain diagram of LPBF to wrought welded samples as per ASTM E8.
Table 4-6 ASTM E8 tensile test results for LPBF to wrought welded samples
TLW1
848.88 MPa
Yield Strength Breaking Strength Max Elongation
822.08 MPa
TLW2
847.27 MPa
891.32 MPa
826.01 MPa
1.832
TLW3
912.67 MPa
971.12 MPa
894.04 MPa
2.162
Mean
855.63 MPa
923.46 MPa
847.38 MPa
2.027
4 Results and Discussion
Page | 105
Ultimate Tensile Strength 907.93 MPa (mm) 2.086
4.3.2.4.3 Post-test CT scan – Wrought to LPBF Post-test CT scanning of all three samples showed that the failures are not related to the
porosities picked up during the CT scans prior to the test, since all fractures occurred in the
ductile portion of the samples. Therefore, the strength of the welded joints was defined by
the strength of the wrought section, which was more ductile due to the microstructure in the
area, rather than its porosity profile.
Figure 4-36. Sample number 1, post fracture CT scan data, shows that most of the porosities are in the fusion zone.
4 Results and Discussion
Page | 106
Figure 4-37. Sample number 2, post fracture CT scan data, again most of the porosities are concentrated at the fusion zone and away from the point of fracture.
Post-test SEM – Wrought to LPBF Using the Quanta FEI scanning electron microscope (SEM) at RMIT University’s city campus
Figure 4-38. Sample number 3, post fracture CT scan data, showing consistent results with all similar samples, with the porosities being concentrated at the fusion zone and away from the fracture area.
RMMF facilities, detailed images were taken from the location of fracture on the welded
wrought and LPBF assembly. The fracture had occurred in the wrought base metal area,
Figure 4-39. Surface of fractured damage. Left image shows entire face of damaged area in 50x magnification, right image shows a 100x magnification of cup and cone fracture detail.
4 Results and Discussion
Page | 107
clearly showing a cup and cone failure characteristic to ductile failures (Figure 4-39 and 4-40).
Figure 4-40. Top left (200x) and top right (1600x) show details of the cup and cone failure (in ductile mode) at higher magnification.
4 Results and Discussion
Page | 108
Ductile failures result in a fractured surface at the plane of maximum sheer stress at 45
degrees angle to the plane of applied stresses. There was no sign of other fractures or
initiation of any cracks beyond the point of failure.
4.3.2.5 Welded LPBF to LPBF assembly tensile tests The third testing scenario was the study of welded assemblies made up of two LPBF parts. All
parts were welded and then machined to match the profile prescribed in the ASTM E8
standard. The same equipment and method were used as previously discussed tensile tests,
resulting in three failed samples, with failures mainly around the fusion zone (Figure 4-41).
This was mainly due to the presence of large porosities and material inconsistencies.
4 Results and Discussion
Page | 109
Figure 4-41. Three LPBF to LPBF welded test samples.
4.3.2.6 Tensile tests – LPBF to LPBF LPBF to LPBF welded assemblies were tested to destruction and the results showed a brittle
fracture and a failure mode behaviour (Figure 4-42) similar to the single piece LPBF test
ASTM E8 tests on welded LPBF to LPBF assemblies
1200
1000
800
) a P M
TLL1
600
TLL2
( s s e r t S
Mean
400
200
0
0
2
4
6
8
10
12
Strain (%)
sample (Figure 4-41) with an average UTS of 1032MPa.
Figure 4-42. Stress-Strain diagram of LPBF-to-LPBF welded samples as per ASTM E8. The brittle failure behaviour is more predominant when observing the close proximity of the
Ultimate Tensile Strength (UTS) to the point of breakage (Table 4-7).
Table 4-7 ASTM E8 tensile test results for LPBF-to-LPBF welded samples, note the close proximity of the UTS to breaking strength and relative short overall elongation prior to breakage. Breaking Strength Max Elongation Yield Strength
TLL1
835.40 MPa
1002.90 MPa
TLL2
892.55 MPa
1055.27 MPa
1019.80 MPa
1.879
Mean
863.98 MPa
1032.04 MPa
1011.35 MPa
1.796
4 Results and Discussion
Page | 110
Ultimate Tensile Strength 1008.81 MPa (mm) 1.714
cause of failure (Figures 4-43). Most porosities were at a mean diameter size of 259.6 microns (Figure
4-44) and concentrated at the fusion zone (FZ), contributing significantly to the failure (Figure 4-45).
4.3.2.7 Post-test CT scans – LPBF to LPBF The failed LPBF to LPBF tensile test samples were CT scanned to establish the failure mode and root
4 Results and Discussion
Page | 111
Figure 4-43. CT scan images of sample numbers 1 and 2 of the welded LPBF-LPBF tensile tests. Both images have annotations showing the location of the top 5 largest porosities by diameter.
Histogram of Dia (microns) Normal
14
Mean 259.6 StDev 187.6 28 N
12
10
8
6
y c n e u q e r F
4
2
0 -200
0
200
400
600
800
1000
Dia (microns)
Histogram of Dia (microns) Normal
9
Mean 227 StDev 150.9 20 N
8
7
6
5
4
y c n e u q e r F
3
2
1
0
0
200
600
400 Dia (microns)
4 Results and Discussion
Page | 112
Figure 4-44. Histograms showing the distribution of porosity diameter sizes in the samples number 1 (top) and 2 (bottom). Both show relatively similar pore sizes, with the mean sizes ranging between 260 and 227 microns.
Porosity Diameter
120000000
100000000
3
80000000
m ) µ (
60000000
l
40000000
e m u o V
20000000
0
100
200
300
400
500
600
0
Position z axis (voxels; 1 vox = 25.967microns)
Porosity Diameter
Porosity Diameter
25000000
20000000
3
15000000
m ) µ (
l
10000000
e m u o V
5000000
0
0
100
200
300
400
500
600
700
800
Position z axis (voxels; 1 vox = 25.967microns)
Porosity Diameter
4 Results and Discussion
Page | 113
Figure 4-45. Distribution of porosities relative to the z-axis, with pore sizes represented by diagram size. As evident, the majority of the pores are at the vicinity of the fusion zone (circa 400 to 600 voxels on the z-axis). Top is sample 1 and bottom is sample 2. Z datum point is at base metal and 600 microns is around FZ.
Post-test SEM imaging – LPBF to LPBF Detailed images taken from the location of fracture on the welded LPBF-to-LPBF assembly.
Relatively large pores can be seen at the surface of fracture as highlighted in the images of
the two samples (Figures 4-46 and 4-47).
4 Results and Discussion
Page | 114
Figure 4-46. Large pores can be seen at the perimeter and on the surface of the fractured area of tensile test sample number 1 of the LPBF-to-LPBF welded assembly
Figure 4-47. Large pores can be seen at the perimeter and on the surface of the fractured area of tensile test sample number 2 of the LPBF-to-LPBF welded assembly
The two images result in the following conclusions:
Laser powder bed fusion (LPBF) samples were prone to failures initiated by porosity
The failure mode in LPBF samples were brittle, initiated at porosities, resulting in
perpendicular surfaces of failure.
4 Results and Discussion
Page | 115
Majority of the pores observed were at the perimeter of the samples produced
4.3.3 Fatigue tests 4.3.3.1 Design of experiments To establish the fatigue behaviour characteristics of titanium alloys in various welded
configurations (Wrought-to-LPBF and LPBF-to-LPBF), samples were manufactured, welded,
machined and prepared as per ASTM E466 standards in the following scenarios:
1. Baseline single piece wrought
2. Baseline single piece laser powder bed fused (LPBF)
3. LPBF to wrought laser beam welded
4. LPBF to LPBF laser beam welded
All welds were conducted using the same parameters as the ones used for tensile test
samples, i.e., 2200 Watts of power and at 1200 mm/ min scanning speed.
As per ASTM E466 samples were subjected to frequent consecutive (30 Hz) axial 100% tension
and 10% compression loadings, also known as an R= -0.1 loading ratio (Figure 4-48). The
loading was calculated based on the cross section of the samples, which in this case was 32
mm2 and the stress levels of each test criteria.
4 Results and Discussion
Page | 116
Figure 4-48. Fatigue test loading profile.
The test cases considered for the fatigue tests were planned as per table below (Table 4-8).
Table 4-8 ASTM E466 fatigue test design of experiment (DOE)
Identification Material Loading (MPa) Loading R
Wrought single piece Wrought single piece Wrought single piece LPBF single piece LPBF single piece W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded W to LPBF welded LPBF to LPBF welded LPBF to LPBF welded LPBF to LPBF welded LPBF to LPBF welded LPBF to LPBF welded LPBF to LPBF welded LPBF to LPBF welded LPBF to LPBF welded LPBF to LPBF welded Frequency (Hz) 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 value -0.1 -0.1 -1 -0.1 -1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1
W1 500 W2 500 W3 400 L1 500 L2 400 LW1 200 LW2 200 LW3 200 LW4 400 LW5 400 LW6 400 LW7 600 LW8 500 LW9 500 LW10 500 LW11 300 LW12 300 LL1 200 LL2 200 LL3 400 LL4 400 LL5 500 LL6 100 LL7 300 LL8 300 200 LL9 The loadings were derived based on the tensile test results. Following the literature available for similar tests, as per Chapter 2.
4.3.3.2 Single piece wrought fatigue tests Single piece non-welded test samples were fabricated, tested and analysed to establish the
baseline on the inherent mechanical behaviour characteristics of Titanium Ti6Al4V alloy in
wrought form. Not having gone through any additive manufacturing process or other heat-
treatment, provided the opportunity to see how such alloys behave under their natural state
4 Results and Discussion
Page | 117
in wrought form.
4.3.3.3 Pre-test porosity analysis – single piece wrought Similar to tensile tests, fatigue test samples were also CT scanned prior to testing. CT Scan
data record any pre-existing porosities that would potentially affect the fatigue test results
and contribute to the failure mode or location. Wrought single piece samples were also CT
scanned but no detectable porosity was found in either of the samples.
4.3.3.4 Fatigue test – single piece wrought Based on the tensile test results carried out prior to the fatigue tests, it was established that
wrought titanium alloys were ductile and would potentially have favourable fatigue
characteristics, hence the loading considered for wrought single piece test specimens started
at a high loading of 50% of the UTS (ultimate tensile strength) as established from the tensile
test results, which translates to 500 MPa. This scenario was repeated twice (Figure 4-49) and
on the third occasion the R factor was increased from -0.1 to -1.0, resulting in a full tension
to compression ratio of the axial loading. The results are as per Table 4-9.
Frequency
Loading
Cycles
Result
Loading R value
W1
500
0.1
30
58,046
Test to failure
Wrought single piece
W2
500
0.1
30
72,014
Test to failure
Wrought single piece
W3
400
1
30
51,955
Test to failure
Wrought single piece
Table 4-9 ASTM E466 fatigue test results for single piece wrought samples Identification Material
4.3.3.5 Post-test SEM – single piece wrought At the completion of the fatigue tests, the fracture surfaces were imaged using scanning
electron microscopy. The objective being to understand the root cause and mode of failure.
In the case of wrought single pieces all failures were ductile and on the plane of maximum
sheer stress at 45 degrees to the axis of axial loading. The summit of the cup and cone
damages were located at points where material defects were present. This is represented in
4 Results and Discussion
Page | 118
the dark patches evident in Figure 4-50.
4 Results and Discussion
Page | 119
Figure 4-49. Single piece wrought samples subjected to 500MPa and R= -0.1 axial fatigue loading. Top image is the top view, and the bottom image is the side view. As evident from these images the failure mode is ductile, with a fracture plane at 45 degrees to the plane of loading.
Figure 4-50. Fatigue test results for single piece wrought samples as per ASTM E466 (Sample ID W1), which failed at 58,046 cycles, shows a complete ductile failure initiated at a material defect as evident in the highlighted images above.
4 Results and Discussion
Page | 120
4.3.3.6 Single piece LPBF fatigue tests Single piece LPBF fatigue samples, provide another perspective to establish the baseline on
the inherent fatigue behaviour characteristics of LPBF parts, and similar to the logic behind
testing single piece wrought samples, enable us to distinguish the parameters affecting the
fatigue performance of welded assemblies from the parameters pertaining to their
constituent parts. A total of two parts were tested, one at 500MPa and R value of -0.1 and a
loading frequency of 30Hz and another part at a lower stress level of 400MPa but and a much
more demanding R value of 1.0 and a frequency of 30Hz.
4.3.3.7 Pre-test porosity analysis – single piece LPBF Prior to the fatigue tests the samples were CT scanned and any major pores that would
contribute to the failure mode were identified.
A total of 85 pores were identified in sample ID LL1, with a mean diameter of 316 microns and
Histogram of Dia (microns) Normal
35
Mean 316.4 StDev 190.5 85 N
30
25
20
15
y c n e u q e r F
10
5
0
0
200
400
600
800
1000
1200
Dia (microns)
a standard deviation of 190.5 microns (Figure 4-51).
4 Results and Discussion
Page | 121
Figure 4-51. Histogram of sample ID LL1 porosity analysis.
4.3.3.8 Fatigue test – single piece LPBF Both parts failed with the abovementioned loadings at much lower cycles than the wrought
samples as summarised in Table 4-10.
Frequency
Loading
Result
Loading R value
0.1
30
6,529
Test to failure
L1
500
LPBF single piece
1.0
30
2,587
Test to failure
L2
400
LPBF single piece
Table 4-10 ASTM E466 fatigue test results for single piece LPBF samples Cycles Identification Material
4.3.3.9 Post-test CT scan – single piece LPBF Upon conclusion of the single piece LPBF fatigue test, the sample was CT scanned. The
distribution of porosities within the sample (Figure 4-52) was such that suggested the more
Figure 4-52. The largest 5 pores in sample ID LL1 were distributed along the entire length of the sample.
4 Results and Discussion
Page | 122
spherical the porosities were, the smaller in diameter they were too (Figure 4-53).
Figure 4-53. Distribution of pores in sample ID LL1.
Post-test SEM – single piece LPBF 4.3.3.10 Scanning electron microscopy (SEM) images of the LPBF single piece part also showed similar
results to the CT scan data, that the porosities within this part were distributed over the entire
body of the part and also contributed significantly to the brittle failure of the part. Multiple
pores can be spotted on the face of the fracture (Figure 4-54). The edges of the fracture plane
are also jagged and suggest that surface roughness had a contribution to the shape of the
4 Results and Discussion
Page | 123
breakage area.
4 Results and Discussion
Page | 124
Figure 4-54. LPBF single piece fatigue sample fractured surface. Note the multiple pores on the surface of the failure plane and edges of the fracture (bottom right)
Welded wrought to LPBF assembly fatigue tests 4.3.3.11 Probably one of the most significant areas of focus of this research were the study of the
behaviour of autogenous laser welded wrought to LPBF assemblies. A total of 12 samples
were tested in this configuration, with fatigue loadings ranging from 600MPa all the way
down to 200MPa, to get a comprehensive understanding of the fatigue behaviour of such
assemblies. 600MPa axial fatigue loading translates to an approximate 65% of the mean
ultimate tensile strength (UTS) of 923MPa as per the tensile test results earlier. Other loading
stresses used were 500MPa, 400MPa, 300MPa and 200MPa (which translates to
approximately 22% value of the mean UTS of such welded assemblies).
Pre-test porosity analysis – wrought to LPBF 4.3.3.12 As per usual of this research, prior to fatigue tests, samples were CT scanned to identify
significant pores and material inconsistency. The CT scan results from a batch of 3 samples is
summarised as follows:
Figure 4-55: Annotation of five largest pores along the assembly. Note that the LPBF
side is taking most of the image area with the heat affected zones and fusion zone at
the lower end of the images.
Figure 4-56: Pore diameter to sphericity distribution of the samples.
Figures 4-57 and 4-58: Pore diameter, volume and Z-axis distribution of the samples.
Note again that the Z-axis originates at the wrought side HAZ and proceeds to the FZ
and LPBF HAZ and LPBF base metal.
4 Results and Discussion
Page | 125
Figure 4-59: Pore size histogram, including count of pores and standard deviations.
Figure 4-55. CT scan images of samples (from top to bottom and left to right) LW10, LW11 and LW12 (alongside an image to show the orientation of the axis, note Z-axis is zero at the bottom and increases upwards)
4 Results and Discussion
Page | 126
LW10
LW11
Figure 4-56. Pore diameter to sphercity distribution graph (from top to bottom) LW10 and LW11.
4 Results and Discussion
Page | 127
LW12
LW 10 Porosity Diameter
2.50E+08
2.00E+08
) 3
1.50E+08
l
1.00E+08
m µ ( e m u o V
5.00E+07
0.00E+00
0
50
100
150
200
250
300
Z-axis (vox)
Porosity Diameter
Figure 4-56. (continued) Pore diameter to sphercity distribution graph LW12.
4 Results and Discussion
Page | 128
Figure 4-57. Pore diameter, volume and Z-axis distribution of samples (from top to bottom) LW10, LW11 and LW12
LW11 Porosity Diameter
2.50E+08
2.00E+08
) 3
1.50E+08
l
1.00E+08
m µ ( e m u o V
5.00E+07
0.00E+00
0
50
100
150
200
250
300
Z-axis (vox; 1 vox = 25.967microns)
LW11 Porosity Diameter
LW12 Porosity Diameter
1.60E+08
1.40E+08
1.20E+08
) 3
1.00E+08
8.00E+07
l
m µ ( e m u o V
6.00E+07
4.00E+07
2.00E+07
0.00E+00
0
50
100
150
200
250
300
350
Z-axis (vox; 1 vox = 25.967microns)
LW12 Porosity Diameter
4 Results and Discussion
Page | 129
Figure 4-58. Pore diameter, volume and Z-axis distribution of samples (from top to bottom) LW11 and LW12
Histogram of LW10 Normal
Histogram of LW11 Normal
70
60
Mean 267.5 StDev 219.9 76 N
Mean 313.7 StDev 240.8 82 N
60
50
50
40
40
30
30
y c n e u q e r F
y c n e u q e r F
20
20
10
10
0
0
0
400
800
1200
1600
2000
0
400
1200
1600
LW10
800 LW11
H
21WL ,11WL ,01WL fo margotsi lamroN
Histogram of LW12 Normal
5300.0
40
Mean 304.0 StDev 232.8 91 N
0300.0
elbairaV 01W L 11WL 21WL
5200.0
30
NveDtSnaeM 679.9125.76 2 288.0427.313 198.2320.403
0200.0
20
y t i s n e D
5100.0
y c n e u q e r F
0100.0
10
5000.0
0
0000.0
0
300
600
900
1200
1500
1800
0
003
006
0021
0051
008 1
LW12
009 ata
D
4 Results and Discussion
Page | 130
Figure 4-59. Pore size distribution histogram (from top to bottom and left to right) LW10, LW11 and LW12. As can be seen in the summary graph the quantity of pores, their size and distribution are relatively consistent, with the porosities concentrated at the fusion zone. Yet we see that the fatigue damage mainly occurs at the heat affected zone, due to the more important defining factor of microstructure.
Fatigue test – wrought to LPBF 4.3.3.13 Welded wrought to LPBF samples were subjected to high cycle axial stress loadings, ranging
from 65% to 22% of UTS, i.e., 600MPa to 200MPa (Table 4-11), resulting in the failure of the
parts at cycles correlated to the loadings (Figure 4-61 through 4-64). Plotting the logarithmic
value of failure cycles to the maximum tensile axial loading resulted in the graph in Figure 4-
60.
Cycles at Failure Max Stress (MPa) R 200 144,795 200 55,513 200 149,920 400 20,347 400 16,949 400 23,352 600 60 500 45 500 7,164 500 9,496 300 39,245 300 28,266
Failure Location LPBF HAZ FZ FZ LPBF BM LPBF BM LPBF BM FZ FZ LPBF HAZ FZ LPBF HAZ LPBF BM
-0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1
Table 4-11 ASTM E466 fatigue test results for LPBF to wrought welded assemblies (FZ: fusion zone; HAZ: heat affected zone; BM: base metal); Loading factor R = -0.1; i.e. tension and 10% compression loading cycle. Configuration Sample ID LPBF to W LW1 LPBF to W LW2 LPBF to W LW3 LPBF to W LW4 LPBF to W LW5 LPBF to W LW6 LPBF to W LW7 LPBF to W LW8 LPBF to W LW9 LPBF to W LW10 LPBF to W LW11 LPBF to W LW12
LPBF to wrought welded assemblies
700
600
500
400
) a P M n
300
i S ( s s e r t S
200
100
0 1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
Cycles (N) - Logarithmic scale
4 Results and Discussion
Page | 131
Figure 4-60. Welded LPBF to wrought fatigue behaviour is best represented by using an S-N curve, with the number of cycles to failure in logarithmic scale in the abscissa and the stress levels on the ordinate.
LPBF side ↓
Wrought side ↓
LW3
LW3
LW2
LW2
LW1
LW1
4 Results and Discussion
Page | 132
Figure 4-61. Failed welded LPBF to wrought samples subject to 200MPa (sample ID from top to bottom LW3, LW2 and LW1)
LPBF side ↓
Wrought side ↓
LW11
LW12
LW7
4 Results and Discussion
Page | 133
Figure 4-62. Failed welded LPBF to wrought samples subject to 300MPa (top 2, sample ID from top to bottom LW11 and LW12) and 600MPa (bottom sample LW7)
Wrought side ↓
LPBF side ↓
LW4
LW4
LW5
LW5
LW6
LW6
4 Results and Discussion
Page | 134
Figure 4-63. Failed welded LPBF to wrought samples subject to 400MPa (sample ID from top to bottom LW4, LW5 and LW6)
LPBF side ↓
Wrought side ↓
LW8
LW8
LW9
LW9
LW10
LW10
4 Results and Discussion
Page | 135
Figure 4-64. Failed welded LPBF to wrought samples subject to 500MPa (sample ID from top to bottom LW8, LW9 and LW10)
Post-test CT scan – wrought to LPBF 4.3.3.14 The majority of the fatigue failures occurred either on the LPBF side of the welded assemblies
or at the fusion zone (FZ). tests were CT-scanned, and porosities and other material anomalies
were captured as per Figures 4-65 through 4-70.
Figure 4-65. LW1 (left) and LW2 (right), both failed at the LPBF side of the welded assembly, but the largest pores were identified to be at the FZ as per CT scans.
4 Results and Discussion
Page | 136
Figure 4-66. LW3 (left) and LW4 (right), both failed at the LPBF side of the welded assembly, but the largest pores were identified to be at the FZ as per CT scans.
Figure 4-67. LW5 (left) and LW6 (right), both failed at the fusion zone (FZ) of the welded assembly and the largest pores were also identified to be at the FZ as per CT scans.
Figure 4-69. LW9 (left) failed at the LPBF side of the welded assembly, but the majority of the largest pores were at the FZ. LW10 (right), both failed at the LPBF side of the welded assembly, but the largest pores were identified to be at the FZ as per CT scans.
4 Results and Discussion
Page | 137
Figure 4-68. LW7 (left) and LW8 (right), both failed at the fusion zone (FZ) of the welded assembly and the largest pores were also identified to be at the FZ as per CT scans.
Figure 4-70. LW11 (left) and LW12 (right) both failed at the LPBF side of the welded assembly and as evident in the images, in both assemblies the porosities are concentrated at the FZ and HAZ.
Post-test SEM – wrought to LPBF 4.3.3.15 With the fatigue failures occurring at either the LPBF side of the welded assemblies or at the
fusion zone, it was crucial to try to capture any surface porosities using scanning electron
microscopy (SEM), hence all samples were imaged, and any surface anomalies were
highlighted as per Figures 4-71 through to 4-81. Pores have surface imperfections, which are
clearly visible on the SEM images in following sections, such imperfections are regarded as
stress concentration areas and crack initiation points that can propagate in a cyclic fatigue
loading scenario. The more spherical a pore is, the more even the stress distribution and the
4 Results and Discussion
Page | 138
less spherical the stress loading gets more localised.
Figure 4-71. Fracture surface on LPBF base-metal side of the welded assembly LW1, which failed under 200MPa loading, at 144,795 cycles and R=0.1 shows a brittle failure mode, with multiple points of crack initiations at the perimeter of the failure plane.
4 Results and Discussion
Page | 139
Figure 4-72. Fracture surface at the fusion zone of the welded assembly LW2, which failed under 200MPa axial stress, after 55,513 loading cycles and R=0.1 shows was porosities which acted as multiple points of crack initiations.
4 Results and Discussion
Page | 140
4 Results and Discussion
Page | 141
Figure 4-73. Fracture surface at the fusion zone of the welded assembly LW3, which failed at 200MPa axial stress, after 149,920 loading cycles and R=0.1. Note the porosities and enlarged image of one of the porosities showing crack initiation points at the surface of the pore.
Figure 4-74. Fracture surface on LPBF base-metal side of the welded assembly LW4, which failed under 400MPa loading, at 20,347 cycles and R=0.1. Note the pore at the centre of the fracture surface and multiple fracture initiation points at the perimeter of the failure zone.
4 Results and Discussion
Page | 142
4 Results and Discussion
Page | 143
Figure 4-75. Fracture surface on LPBF base-metal side of the welded assembly LW5, under 400MPa loading, at 16,949 cycles and R=0.1. Two major porosities can be clearly seen on these images; that are on the plane of failure and would have contributed to the failure of the sample based on the shape of the failed surface.
4 Results and Discussion
Page | 144
Figure 4-76. Fracture surface on LPBF base-metal side of the welded assembly LW6, under 400MPa loading, at 23,352 cycles and R=0.1.
Figure 4-77. Fracture surface at fusion zone of welded assembly LW7, under the highest load applied during these tests of 600MPa. The part failed after only 60 cycles. Large porosities can be observed on the failure surface.
4 Results and Discussion
Page | 145
4 Results and Discussion
Page | 146
Figure 4-78. The contribution of surface roughness acting as crack initiation source can be seen by observing the perimeter of the fracture surface of LW9 which failed at 500MPa after 7,164 loading cycles
Figure 4-79. Large pores and their inherent surface cracks that have led to the failure of sample LW10 at 500MPa of stress after 9,496 loading cycles.
4 Results and Discussion
Page | 147
4 Results and Discussion
Page | 148
Figure 4-80. Surface roughness and inconsistencies have played a significant role in crack initiation on sample LW11 at 300MPa axial stress after 39,245 loading cycles.
Figure 4-81. Surface conditions can also be observed to have had a significant effect on the failure mode of LW12 at 300MPa axial stress after 28,266 loading cycles. Crack initiation porosities highlighted in (bottom right) image.
4 Results and Discussion
Page | 149
Welded LPBF to LPBF assembly fatigue tests 4.3.3.16 Another novel aspect of this research is the study of mechanical properties and fatigue
behaviour of welded LPBF to LPBF titanium alloys. This is a topic that has not got much
attention in academic literature, as of the time of writing this thesis.
A total of 9 samples were studied in this configuration, with axial fatigue loadings ranging
from 500MPa all the way down to 100MPa. 500MPa translates to approximately 50% of UTS
(mean value of 1032MPa) as per tensile test results. A fatigue life of 1,000,000 cycles was
considered to be infinite life for this study, which was achieved at 100MPa axial loading of
such welded assemblies.
Pre-test porosity analysis – LPBF to LPBF 4.3.3.17 LPBF to LPBF welded assemblies were CT scanned to identify any inherent porosities during
the build of the base-metal parts as well as any porosities created during the welding process.
B
]xov[ z retneC sv ]³mµ[ emuloV fo tolP elbbu )snorcim( aiD :ezis elbbuB
00000009
00000008
00000007
] ³
00000006
00000005
00000004
l
m µ [ e m u o V
00000003
00000002
00000001
0
0
001
002
003
4
00
C
]xov[ z retne
Histogram of Dia (microns) Normal
35
Mean 328.6 StDev 233.1 87 N
30
25
20
15
y c n e u q e r F
10
5
0
0
300
600
900
1200
1500
Dia (microns)
Sample IDs LL7 and LL8 CT scan images and porosity data are shown in Figures 4-82 and 4-83.
Figure 4-82. Sample ID LL7, note the concentration of porosities at z-axis position around FZ and the HAZ.
Both samples in this example demonstrated a concentration of porosities within the fusion
4 Results and Discussion
Page | 150
zone (FZ) and heat affected zones (HAZ).
B
]xov[ z retneC sv ]³mµ[ emuloV fo tolP elbbu )snorcim( aiD :ezis elbbuB
00000005
00000004
] ³
00000003
l
00000002
m µ [ e m u o V
00000001
0
0
001
002
004
005
006
7
00
003 C
]xov[ z retne
Histogram of Dia (microns) Normal
40
Mean 310.2 StDev 180.6 147 N
30
20
y c n e u q e r F
10
0
0
150
300
750
900
600
450 Dia (microns)
Figure 4-83. Sample ID LL8, shows similar pattern for detectable porosities, with the concentration being at the FZ and HAZ.
Fatigue test – LPBF to LPBF 4.3.3.18 LPBF to LPBF welded assemblies went through the same test regime as the wrought to LPBF
welded assemblies. Results of the fatigue tests are summarised in in Table 4-12 and the S-N
curve (Figure 4-84) and the broken samples are depicted in Figures 4-85 through to 4-88
4 Results and Discussion
Page | 151
(grouped by loading).
Cycles at failure Max Stress (MPa) R 100 1,000,001 200 117,055 200 161,869 200 3,965 300 4,094 300 2,264 400 19,149 400 16,544 500 10,746
Configuration LPBF to LPBF LPBF to LPBF LPBF to LPBF LPBF to LPBF LPBF to LPBF LPBF to LPBF LPBF to LPBF LPBF to LPBF LPBF to LPBF
-0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1
Failure Zone No failure Base Metal Base Metal Fusion Zone Fusion Zone Fusion Zone Base Metal Base Metal Base Metal
Table 4-12 ASTM E466 fatigue test results for LPBF to LPBF welded assemblies Sample ID LL6 LL1 LL2 LL9 LL7 LL8 LL3 LL4 LL5
LPBF to LPBF welded assemblies
600
500
400
) a P M n
300
i S ( s s e r t S
200
100
0 1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
1.E+07
Cycles (N) - Logarithmic scale
4 Results and Discussion
Page | 152
Figure 4-84. Welded LPBF to LPBF fatigue behaviour is best represented by using an S-N curve, with the number of cycles to failure in logarithmic scale in the abscissa and the stress levels on the ordinate.
4 Results and Discussion
Page | 153
Figure 4-85. Top and side view of LPBF to LPBF welded assemblies subjected to 200MPa axial loading cycles to failure. From top to bottom, sample IDs: LL1, LL2 and LL9
Figure 4-86. Top and side view of LPBF to LPBF welded assemblies subjected to 300MPa axial loading cycles to failure. Top LL8 and bottom LL7.
4 Results and Discussion
Page | 154
Figure 4-87. Top and side view of LPBF to LPBF welded assemblies subjected to 400MPa axial loading cycles to failure. Top LL4 and bottom LL3.
4 Results and Discussion
Page | 155
Figure 4-88. Top and side view of LPBF to LPBF welded samples, subject to 500MPa (top, LL5) and 100MPa (bottom, LL6). The sample subject to 100MPa didn’t fail after 1,000,001 cycles, which for the purpose of this study is considered the limit for infinite life.
4 Results and Discussion
Page | 156
Post-test CT scan – LPBF to LPBF 4.3.3.19 All parts that failed were CT scanned to find the correlation of the porosities with the failure
Figure 4-89. (left to right and top to bottom): LL1, LL2, LL3, LL4, LL5, LL7. Note that sample ID LL6 did not fail.
4 Results and Discussion
Page | 157
mode and root cause. (Figures 4-89 and 4-90).
Figure 4-90. (left) sample ID LL8 and (right) sample ID LL9
Post-test SEM – LPBF to LPBF 4.3.3.20 Post-test scanning electron microscopy (SEM) images provided a clear view of root cause of
each welded assemblies failure (Figures 4-91 through 4-99). This is evident via the shape of
the fractured area, angle of plane of failure, existence of porosities, their distribution,
4 Results and Discussion
Page | 158
sphericity, inclusions, microcracks and other failure surface characteristics.
Figure 4-91. Sample LL1, subjected to 200MPa of axial load, survived 117,055 cycles. The perpendicular cut surface suggests the brittle nature of the fracture and the pores at the perimeter, are sources of crack initiation.
4 Results and Discussion
Page | 159
Figure 4-92. Sample LL2, subjected to 200MPa of axial loading, underwent 161,869 cycles prior to fracture. As typical in LPBF to LPBF welded assemblies in this research, the failure mode is demonstrated to be brittle, with multiple points of fracture initiation at the perimeter of the welded assembly.
4 Results and Discussion
Page | 160
Figure 4-93. Sample LL3, subjected to 400MPa of axial stress loading, underwent 19,149 cycles prior to failure. Bottom right picture shows the starting point of a microcrack that contributed to the brittle fracture of the part.
4 Results and Discussion
Page | 161
Figure 4-94. Large pores pertaining to incompletely melted particles can be seen as dark patches in the images of sample LL4, which was subjected to 400MPa axial loading and maintained its integrity up to 16,544 cycles.
4 Results and Discussion
Page | 162
Figure 4-95. Sample LL5, which was subjected to the highest axial loading of this research at 500MPa, withstood 10,746 cycles before showing multiple failures, including a secondary fracture on the side of the body, as seen in the bottom two images.
4 Results and Discussion
Page | 163
Figure 4-96. LL6, close up image of the fusion zone (FZ) and heat affected zones (HAZ). Note that albeit surface anomalies, the welded assembly was able to maintain its integrity after 1,000,001 cycles of 100MPa axial loading.
4 Results and Discussion
Page | 164
Figure 4-97. Sample LL7, which failed at the fusion zone at 300MPa axial loading after 4,094 cycles, shows a large cluster of porosities at the centre and perimeter of the fusion zone.
4 Results and Discussion
Page | 165
Figure 4-98. Sample LL8 which also failed at the fusion zone at 300MPa axial loading after only 2,264 cycles, has multiple large pores on the surface and perimeter of the fractured area, contributing to the failure of the part.
4 Results and Discussion
Page | 166
Figure 4-99. Sample LL9, which failed at the fusion zone after 3,965 cycles at 200MPa axial force, also demonstrates large pores scattered all over the face of the fractured surface and perimeter.
4 Results and Discussion
Page | 167
4.4 References
AWS, AWS- 2014, Guide for the Fusion Welding of Titanium and Titanium Alloys, G 2.4M:2014,
Leuders, S, Vollmer, M, Brenne, F, Tröster, T & Niendorf, T 2015, 'Fatigue Strength Prediction for Titanium Alloy TiAl6V4 Manufactured by Selective Laser Melting', Metallurgical and Materials Transactions A, vol. 46, no. 9, pp. 3816-3823.
Liinalampi, S, Remes, H & Romanoff, J 2019, 'Influence of three-dimensional weld undercut geometry on fatigue-effective stress', Welding in the World, vol. 63, no. 2, pp. 277-291.
Ninh Nguyen, T & Wahab, MA 1995, 'A theoretical study of the effect of weld geometry parameters on fatigue crack propagation life', Engineering Fracture Mechanics, vol. 51, no. 1, pp. 1-18.
Yung, WKC, Ralph, B, Lee, WB & Fenn, R 1997, 'An investigation into welding parameters affecting the tensile properties of titanium welds', Journal of Materials Processing Technology, vol. 63, no. 1, pp. 759-764.
4 Results and Discussion
Page | 168
5. Conclusions 5.1 Chapter overview Additive manufacturing provides many benefits to the application of titanium alloys, including
the cost-effective fabrication of parts in complex geometrical shapes and with minimal
material wastage. These benefits are sometimes overshadowed by dimensional limitations,
and relatively slow speeds of mass production, inherent to additive manufacturing, compared
to traditional subtractive manufacturing methods. This research explores the possibility of
overcoming such limitations by welding additively manufactured titanium alloy parts to
wrought and nominally larger components of same material and investigates the governing
parameters that influence the success of such an endeavour. To the date of this writing, no
prior literature exists, to the best knowledge of the author, which addresses the study of laser
5.2 Welding parameters Based on literature available on the fabrication of Ti6Al4V grade 5 titanium alloys, a set of
beam welding of additively manufactured titanium alloys with wrought titanium alloys.
standard parameters were chosen for the fabrication of additively manufactured titanium
alloy samples. Also based on the literature available on welding of wrought titanium alloy
sheets, various welding parameters were elected and included in a design of experiment at
the pilot stage of this research. The variation of such parameters and their effects of on weld
morphology, porosity distribution and hardness profile were documented. The main
parameters effecting the characteristics of an autogenous welded LPBF (laser powder bed
fusion), and wrought assembly were established to be energy density and the build quality
of the LPBF parts. Additively manufactured parts using laser powder bed fusion (LPBF)
technology are prone to process parameter defects that are inherent to the nature of the
build process. Process induced defects such as lack of complete melting of the feedstock
powder, inclusions, porosities and creation of microcracks being the main types of such
defects. The rough surface of additively manufactured parts due to the granular texture
resulting from the layer-by-layer melting of powder feedstock is also another contributor to
the overall quality of the LPBF manufactured parts. In this research, it has been demonstrated
that where the welded joint strength is sufficient and free from defects, failure of the
5 Conclusions
Page | 169
assemblies is most likely to happen at areas within the LPBF base metal that have initial build
quality issues. Energy density can be controlled via the power output of the laser and the
scanning speed of the laser. LPBF build quality depends on many parameters such as
feedstock quality, laser power, and hatch spacing, that have been thoroughly examined by
prior literature available. High quality LPBF are homogeneous and dense and have minimal
discontinuities, inclusions or microcracks.
Too little energy density will prevent the creation of the weld keyhole and consequently there
will not be a full penetration weld across the entire depth of the welded assembly. On the
other hand, too much energy density during welding will result in the increased evaporation
of the molten pool, consequently increasing the risk of porosities and increasing the width of
the heat affected zones (HAZ), leading to a decrease in ductility and increase in
embrittlement, thus heightening the risk of failure at the welded joint. High energy density of
laser beam welds creates narrow, concentrated, parallel walled fusion zones that are best
optimised for full penetration, but minimised to limit the width of the heat affected zones.
Using statistical methods; it was established that there is a direct linear relationship between
5.3 Tensile strength and fatigue performance Laser powder bed fusion (LPBF) parts undergo cyclic thermal heating and cooling during the
energy density and the width of the heat affected zones.
build process, which results in changes in the microstructure and eventual increase in overall
tensile strength of the base metal, due to the creation of martensitic microstructure resulting
from rapid cooling of the molten pool. Martensite microstructures result in reduction in
ductility. Full penetration welding of LPBF and wrought parts results also contributes to the
creation of further martensite microstructure and increase in the tensile strength of the joint.
The increase in tensile strength also is accompanied with the increase in material
microhardness, with the highest microhardness being recorded at the fusion zone, then the
LPBF and then wrought side of the assembly.
During this research it was demonstrated that LPBF to wrought welded assemblies have
superior tensile performance with the failures mostly being recorded at the wrought and
weaker side of the assembly. This can be attributed to the distribution of higher
5 Conclusions
Page | 170
microhardness and tensile strength at the LPBF base metal and fusion zone in the assemblies.
Fatigue performance is highly dependent on the presence of material discontinuities, such as
porosities, inclusions, microcracks and any other points of stress concentration, as well the
presence of brittle alpha-prime martensite. This research demonstrated that both LPBF to
wrought and LPBF to LPBF welded assemblies are most vulnerable to porosities at the LPBF
5.4 Microstructural parameters Welding LPBF and wrought titanium alloys induces further thermal stresses resulting in
base material and LPBF HAZ, which has a high concentration of alpha-prime martensite.
acicular (needle-shaped) alpha-prime crystals in the fusion zone and a mixture of martensitic
alpha-prime, acicular alpha and primary alpha at the heat affected zones, this is due to
quenching below the beta-transus temperature. Brittle alpha prime martensite is most
dominant at the HAZ of LPBF assembly and are most prone to fatigue failure.
Porosity and microstructure of the assembly define the mechanical tensile strength and
fatigue performance of the welded structure, as evident in the failure modes witnessed
during the tests. Fusion zones and LPBF side of a welded assembly showed greater strength
during tensile tests, due to the existence of martensite phases in such areas. Excessive heat
concentration and remelting of LPBF at welded joints with other LPBF created the greatest
opportunity for failure via the creation of large pores and material discontinuity, resulting in
points of high stress concentration and opportunity for crack growth and propagation.
High welding energy densities beyond what is required for the full penetration of the weld,
will result in the increased share of elongated acicular alpha-prime martensite which exhibit
brittle behaviour. Excessive energy densities also result in the increase in the width of the
heat affected zone (HAZ), which hosts a combination of primary alpha, hexagonal close
packed (HCP) alpha-prime and acicular alpha and demonstrate brittle behaviour. Excessive
energy densities also increase the evaporation of material and increase the chance of
porosities and eventual creation of microcracks. Therefore, it is recommended that
autogenous weld parameters be chosen such that optimise for minimal energy densities,
whilst guaranteeing full depth penetration. The welding parameters for this investigation
5 Conclusions
Page | 171
were 2200 Watts at 1200mm/ min for LPBF Ti6Al4V alloys of 4mm thickness.
5.5 Future work The following work is proposed to enhance the understanding of challenges in manufacturing
5.5.1 Dissimilar alloys By investigating the prospects of having large assemblies made by joining titanium alloys in
large hybrid additive and traditional fabricated assemblies:
wrought and additively manufactured format together, new opportunities for more complex
configurations come to mind, such as joining titanium Ti6Al4V alloys to other alloys such as
5.5.2 Post weld processing With the limitation of as-built welded assemblies documented in this research, it would be
Aluminium or copper-rich alloys in wrought or additively manufactured format.
beneficial to investigate the effects of post processing activities, such as heat treatment on
tensile properties and fatigue life of wrought and additively manufactured parts. The
5.5.3 Build parameters effect on weldability Investigate effects of base metal build parameters on weldability, tensile and fatigue
economics and feasibility of such postprocessing should be taken into account.
properties of the final assembly. This study would expand the design of experiments to
include the parameters used to fabricate the base metal (BM) additively manufactured parts
5.5.4 Impact of geometry on weld performance Investigate effects of the geometry of LPBF and wrought parts on the weldability, tensile and
and find their impact on the final mechanical and fatigue performance.
fatigue properties of the final assembly. This study would expand the design of experiments
to include the geometry of the parts being welded together to find the impact of various
5.5.5 Enhance statistical confidence of results Increase the statical confidence of the research by increasing the number of parts fabricated
orientations and shapes of the base metal on the final mechanical and fatigue performance.
5 Conclusions
Page | 172
and tests conducted.
5 Conclusions
Page | 173
