O R I G I N A L C O N T R I B U T I O N
Fatigue response dependence of thickness measurement
methods for additively manufactured E-PBF Ti-6Al-4 V
Mikael Segersäll
1|
Annie Kerwin
2|
Alex Hardaker
2|
Magnus Kahlin
1,3|
Johan Moverare
11Division of Engineering Materials,
Linköping University, Linköping, SE-58183, Sweden
2Manufacturing Technology Centre,
Coventry, CV7 9JU, UK
3Aeronautics, Saab AB, Linköping,
SE-58188, Sweden
Correspondence
Mikael Segersäll, Division of Engineering Materials, Linköping University, Linköping SE-58183, Sweden. Email: mikael.segersall@liu.se
Funding information
Horizon 2020 Framework Programme, Grant/Award Number: 738002
Abstract
Light weight metal parts produced with additive manufacturing have gained increasing interest from the aerospace industry in recent years. However, light weight parts often require thin walls which can have different material proper-ties compared to thick bulk material. In this work, the fatigue properproper-ties of Ti-6Al-4 V produced by electron beam powder bed fusion have been investi-gated for samples with three different wall thicknesses ranging from 1.3 to 2.7 mm and in three different directions; 0, 45, and 90 relative to the build plate. Generally, the 90 specimens show worse fatigue life compared to both 0 and 45. It was found that the fatigue strength is lower for thin samples compared to thicker samples when the stress is calculated from nominal thick-ness or calliper measurements. However, since materials produced by electron beam powder bed fusion often have a rough as-built surface, the load bearing area is not easy to determine. In this paper, four different methods for deter-mining the load bearing area are presented. It is shown that if the surface roughness is considered when calculating the stress levels, the influence from specimen thickness decreases or even disappears.
K E Y W O R D S
additive materials, fatigue, surface roughness, Ti-6al-4v
1
|
I N T R O D U C T I O N
The interest for additive manufacturing (AM), sometimes referred to as 3D printing, of metals has grown since AM gives the manufacturer a higher design freedom compared to conventional manufacturing methods such as casting and machining. AM also gives an engineer the possibility to topologically optimize components to minimize the weight which is of great interest in for
example the aerospace industry. Titanium alloys are an important material group for light weight design in the aerospace industry since they show high strength, excellent corrosion resistance, and low density. Different AM techniques exist where electron beam powder bed fusion (E-PBF) and laser powder bed fusion (L-PBF) are two common techniques.1–3In this study, only E-PBF is considered. The mechanical properties of Ti-6Al-4V produced by E-PBF have been widely studied during the
DOI: 10.1111/ffe.13461
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2021 The Authors. Fatigue & Fracture of Engineering Materials & Structures published by John Wiley & Sons Ltd.
last years (see, e.g., previous works4–9), and it has been shown that a part manufactured by E-PBF has different material properties compared to material from conven-tional methods such as casting or machining. Moreover, the material properties of E-PBF material are dependent on the part geometry and orientation in the build cham-ber. Toh et al.10and Kok et al.11showed that thin-walled E-PBF material had higher microhardness compared to thicker sections. This was attributed to differences in microstructures withα + β structure mixed with acicular α0 for samples up to 1 mm in thickness in contrast to
thicker samples that had onlyα + β structure. In addition to this, the prior β grain interspacing increased with increasing sample thickness which also attributed to the decrease in microhardness.11
The influence from build direction on general mechanical properties, such as tensile properties, of Ti-6Al-4V manufactured by E-PBF has been highlighted by, for example, de Formanoir et al.12and Bruno et al.13 Concerning the anisotropic fatigue response for E-PBF Ti-6Al-4V, several studies have shown that the build direction influences the fatigue life.14–16 For example, Persenot et al.14 investigated the fatigue properties of as-built E-PBF Ti-6Al-4V thin parts manufactured in three different orientations. They concluded that the detrimental effect of the as-built surface state on the fatigue resistance of individual struts is significant and designing lattice structures against fatigue using data obtained on machined samples is unsafe and can lead to erroneous conclusions. The build orientation affects the fatigue properties for samples with as-built surfaces because of its influence on the shape, size, and number of notch-like defects from which fatal cracks initiate. Horizontal and vertical samples were shown to be, respectively, the best and worst samples in terms of fatigue lives. In that study, it was also argued that using roughness measurements of as built AM samples to infer information regarding the fatigue properties is unsafe since these measurements do not reveal the thin and deep notch-like defects present at the as-built surfaces.
The influence of the surface roughness on fatigue life of additively manufactured Ti-6Al-4V has been investi-gated by several different authors; see, e.g., previous studies.6,7,17–22Greitemeier et al.18,19and Kahlin et al.6,7 investigated the effect of surface roughness on fatigue life for both the E-PBF and L-PBF process and found that it is the most dominant factor for the fatigue performance for additively manufactured materials with rough as-built surfaces. Furthermore, Greitemeier et al.18 concluded that both the E-PBF and L-PBF processes show two types of roughness: (i) roughness induced due to solidification of the melt pool (primary roughness) and (ii) roughness
induced by partly melted powder particles (secondary roughness). E-PBF samples typically have a higher roughness compared to L-PBF samples due to both larger particle size and higher layer thickness.
For conventional materials which are, for example, machined, the nominal thickness corresponds well to the actual thickness. However, the rough as-built surface of additively manufactured materials can take up a consid-erably amount of the total cross-section. Therefore, there will be a difference between the real stress in a part and the stress based on the nominal cross-section from the 3D model.23,24This difference will increase with decreas-ing build thickness which will result in a higher real stress on thin samples compared to thick samples even if both are loaded with the same stress based on the nomi-nal cross-section.23The effect of sample wall thickness of the fatigue performance of AM materials is therefore an important topic, but further studies are needed to cover all aspects of it.
Razavi et al.23 proposed that an accurate measure-ment of the fracture surfaces would give a more correct size of the load bearing cross-section. In the present study, this method suggested by Razavi et al.23 is compared to other methods that use different ways of measuring the surface roughness, in order to investigate how the thickness dependence of the fatigue properties for E-PBF Ti-6Al-4V could be evaluated for different material directions.
2
|
M A T E R I A L S A N D
E X P E R I M E N T A L M E T H O D S
2.1
|
Materials
Flat Ti-6Al-4V plates were prepared by E-PBF using an ARCAM A2XX equipment. A layer thickness of 70μm and powder size ranging from 45–105 μm was used. The default build parameters recommended by ARCAM for this alloy and equipment were utilized; see Table 1. After printing, the plates were subjected to HIP for 2 h at 920C/103 MPa according to the ASTM standard F2924/ F3001 2012. The plates were printed in the build
T A B L E 1 E-PBF process parameters E-PBF software: EBM ver 3.2.108 Speed function: 36
Line offset: 0.2 mm Beam max current: 17 mA Beam radius: 0.25–0.30
F I G U R E 1 (A) Specimen orientation with respect to the build direction; (B) specimen geometry [Colour figure can be viewed at wileyonlinelibrary. com]
F I G U R E 2 Average surface roughness obtained from the four different methods. The error bars indicate the standard deviation [Colour figure can be viewed at wileyonlinelibrary.com]
direction; see Figure 1a, with three different nominal thicknesses 1.5, 2.1, and 2.7 mm resulting in nine main groups. From the plates, fatigue specimens were extracted by electro discharge machining (EDM) in three directions 0, 45, and 90 relative to the build direction; see Figure 1a. The specimen geometry was adopted from ASTM E466; see Figure 1b. To avoid crack initiation at the corners of the rectangular cross sections, all edges (hence no surfaces) were rounded by a manual grinding process.
2.2
|
Surface roughness investigations
When evaluating results from fatigue testing, there is a question about what cross-section area should be consid-ered in order to calculate the appropriate stress value. This is not straight forward when taking a rough surface (as from E-PBF in this case) into account. To do as correct an estimation of the cross-section as possible, several different methods to determine the surface roughness were used. In the first method, hereafter called method 1, light-optical microscope (LOM) images of cross-sectioned specimens were taken with a Leica DM6–Compound microscope. After cross-sectioning, the samples were mechanically grinded and polished before etching. The surface roughness was evaluated, for one specimen from each group, over a length of about 8 mm along the loading direction with an image analyzing technique using an in-house Matlab script.25 The Rv
(maximum valley depth of the roughness profile) and Rz
(maximum height of the roughness profile) values were evaluated. For the second method, method 2, 3D-topography images of the specimen surfaces were
created using the Leica DM6–Compound with an analyz-ing software. From the z-stack images, the surface roughness was evaluated over a length of 15 mm in the load direction of the specimens. For the third method, method 3, a conventional surface roughness measure-ment profilometer (MarSurf PS 10) was used to measure the surface roughness. The profilometer needle is moved over the surface along the loading direction, and the surface roughness is calculated by the instrument. Finally, for method 4, the thickness of the net cross-section area was determined by evaluating the fracture surfaces of the fatigue tested samples after failure. For almost every sample, the thickness of the fracture surface was measured at 10 random positions, and the average value was calculated. The roughness of each sample was then evaluated by comparing the measured thickness from the fracture surfaces to the calliper measurements of the un-tested specimens. Since the fatigue testing in the present study was performed in the high cycle regime, the amount of plastic deformation during the tests is neglectable, and the fracture surface will thus be a good representation of the load bearing area of the samples during fatigue.
2.3
|
Fatigue testing
Room temperature fatigue testing with an R ratio of 0.1 and load frequency of 10 Hz was performed in an MTS hydraulic fatigue test rig with a ±10 kN load cell and an EDC-580 V controller. For each main group, 4–7 specimens were fatigue tested. Different stress amplitudes were used in order to cover the range from 50.000 to 5.000.000 cycles to failure. Any test that survived 5.000.000 cycles without failure was interrupted and considered as a run-out sample. Those samples were later retested at a higher stress amplitude to generate the maximum amount of data with the samples available. In total, 49 fatigue tests were performed.
In order to take the surface roughness into account when using methods 1–4, the net cross-sectional thick-ness for each individual specimen was defined as the calliper measurement for the specimen minus two times the Rzmeasurement (factor 2 accounts for both opposite
surfaces of the specimen).
2.4
|
Microscopy
Cross-sections and fracture surfaces were studied by LOM in the Leica DM6–Compound microscope as well as with scanning electron microscopy (SEM). Here, both a Hitachi SU70 FEG-SEM and JEOL SEM were used. The
F I G U R E 3 Cross-section of specimen (0, 2.1 mm) showing shadowing problems at surface roughness determination
cross-sections were polished and etched with Kroll's reagent prior to the investigations.
3
|
R E S U L T S A N D D I S C U S S I O N
3.1
|
Surface roughness investigation
Four methods were used to measure the surface rough-ness of the specimens, and Figure 2 illustrates the average maximum peak to valley distance (often referred to as Rzor Sz) for the different methods, directions, and
thicknesses, respectively. The highest values were obtained by method 4, and a reasonable explanation for this is that the fracture surface most likely represents the worst-case scenario, i.e., the section of the specimen with the smallest cross-sectional area. Further, it is seen that method 1 results in higher values compared to method 2. One explanation for this is that there might be shadowing problems in method 2 where the surface roughness is evaluated by creating 3D images with a z-stack technique. Figure 3 shows the typical cross-section of a specimen where the rough surface is charac-terized by pits which will not be detected by method 2. However, when using method 1, all pits in the image will be detected and will contribute to a higher surface roughness value. The average surface roughness value obtained from the conventional technique, method 3, is lower compared to all other methods, indicating that con-tact techniques are not suitable for determining the roughness of AM samples.
In the recent study by Razavi et al.,23 it was found that a thin build thickness resulted in higher surface roughness. However, no influence from specimen thick-ness (or build orientation) on surface roughthick-ness was
found in this study. This seems to be valid for all four measurement methods.
3.2
|
Fatigue testing
From the introduction, it is clear that the two most important factors which cause additively manufactured Ti-6Al-4V components to in general have an inferior fatigue performance are as follows:
1. High surface roughness originating from the melt pool geometry or partly melted powder particles on the surface.
2. Internal defects aligned along or with a higher projec-tion area on the planes perpendicular to the building orientation.
For thin-walled structures, the effect of both surface roughness and internal defects will be more critical since any features will have a more significant impact on the load bearing capacity by reducing cross-section area more compared to bulk samples. However, this depends very much on how the cross-section is defined and how the surface roughness is taken into account.
Fatigue testing was performed for three different thicknesses and three different orientations. All samples were tested with the rough as built surfaces obtained from E-PBF; see Figure 4. From the 3D-topography images of the surfaces, one can see that there is a contri-bution to the surface roughness from the melt pool layers which are oriented differently compared to the fatigue load direction for each orientation, respectively. For the 0 specimens, the melt pool layers are oriented along the load direction. The 45 specimens have their melt pool
F I G U R E 4 Three-dimensional-topography images of the surface roughness of as-built E-PBF samples. Note that the colors in the scales are not identical [Colour figure can be viewed at wileyonlinelibrary.com]
(Q) (R)
F I G U R E 5 Fatigue results for 1.5, 2.1, and 2.7 mm thick specimens with 0, 45, and 90build orientation. Cross-section area is determined by (A–C) nominal, (D–E) caliper, (G–I) method 1—LOM, (J–L) method 2—3D scan, (M–O) profilometer, and (P–R) net cross-sectional measurement [Colour figure can be viewed at wileyonlinelibrary.com]
layers oriented 45 from the load direction while 90 specimens have their melt pool layers oriented perpen-dicular to the load direction.
In Figure 5, the fatigue data are presented as S-N cur-ves for all tested specimens, where the stress is calculated from nominal thickness (i.e., the thickness defined in the
F I G U R E 6 Summary of fatigue results for evaluation method 4 based on the net cross section area measurements evaluated from the fracture surfaces [Colour figure can be viewed at wileyonlinelibrary.com]
F I G U R E 7 LOM images of cross sections: (A) 0, (B) 45, and (C) 90(all with thicknesses 2.1 mm). *The build orientation is oriented 45from the image plane [Colour figure can be viewed at wileyonlinelibrary.com]
CAD drawing sent to the E-PBF machine for printing), thickness from calliper measurements, and stress calcu-lated from methods 1–4, respectively. For the calliper measurements, the average thickness was determined using a calliper with cylindrical measuring rods with flat ends and a diameter of 6 mm. When nominal stress is plotted versus cycles to failure, a clear build direction dependence is visible where 90 specimens show worse fatigue life compared to both the 0 and 45 directions, independent of specimen thickness. When comparing the 0 and 45 directions, no clear difference in fatigue life can be observed. The 3D-topography images of the rough as built surface of the E-PBF samples (see Figure 4) reveal that there is a contribution to the roughness from the melt pool layers. When looking at the surface topog-raphy of the 90 specimens, it is obvious that here the fatigue crack can initiate and propagate more easily perpendicular to the loading direction which seems to lead to a low fatigue life. For the 0 and 45 specimens, the orientation of the melt pool layers is not as
detrimental as for the 90 direction. This influence from build orientation is consistent with previous studies where the influence was attributed to surface defects.14 In that study, no difference between build directions was found when machined specimens were fatigue tested. This build direction dependence where the 90specimens show worse fatigue properties compared to 0 and 45 specimens seems independent from how the stress is calculated.
The thickness dependence on fatigue life indicates that generally the fatigue strength is lower for the thinnest samples compare to the two other thicknesses when nominal thickness is considered. For both the 45 and 90 oriented specimens, a thickness of 1.5 mm leads to shortest fatigue life compared to both 2.1 and 2.7 mm when the nominal stress is considered. From this figure, it is also visible that for the 0 specimens, the influence from thickness on fatigue life is small. For 45specimens, a thickness of 2.7 mm results in slightly better fatigue lives compared to a thickness of 2.1 mm, while there is a
negligible difference between 2.1 and 2.7 mm for the 90 specimens. For 0specimens, no clear difference between the thicknesses can be observed.
All trends described above remain the same regard-less of whether the stress is calculated from nominal thickness or by calliper measurements. However, all the differences seem to be slightly amplified when stress is calculated from calliper measurements compared to stress from nominal thicknesses. In general, the calliper measurements yield a higher thickness compared to the nominal value and the difference increases for decreasing nominal thickness. The average difference in thickness is 0.138, 0.173, and 0.196 mm for a nominal thickness of 2.7, 2.1, and 1.5 mm, respectively.
Previous studies have shown that the surface rough-ness influences the fatigue life,18,19 and if one compen-sates for the surface roughness in the stress calculations, this seems to influence the fatigue results where the thickness dependence decreases or sometimes even disappears; see plots for methods 1–4. For example, if determining the surface roughness with method 4 (where the highest value of Rzwas obtained), the influence from
specimen thickness is negligible for all orientations. These results differ from the study by Razavi et al.23 where a lower fatigue strength was found for thin specimens compared to thicker specimens. In that study, thinner samples showed a higher surface roughness, with randomly distributed deep micro-notches, which was one explanation for the lower fatigue strength. Another
reason was the higher surface to volume ratio for thin samples. In our study, the orientation difference remains after the roughness compensation, showing a lower fatigue life for the 90 orientation where the samples are along the building direction and have the least favorable orientation of the melt pool layers; see Figure 6. It should, however, be noted that the difference in fatigue properties for different build thicknesses to some extent also can depend on differences in microstructure and microhardness which previously have been reported by Kok et al.11and Toh et al.10
3.3
|
Microstructure and crack initiation
locations
The build orientation highly influences the surface topog-raphy of the specimens; the direction of the melt pool layers is clearly visible in Figure 4, and the difference in surface roughness for different orientations is displayed in Figure 7. Optical microscopy reveals an α + β Widmanstätten microstructure that is similar for all spec-imens; see Figures 7 and 8. Columnar grains along the building direction can be seen in the bulk material for 90 specimens, Figure 7c. Similar grain structure is expected also for the other orientations, but is not visible in Figure 7a,b due to the difference of the cross-section plane relative to the building direction. The elongated grain structure is further revealed in Figure 9. No attempt
F I G U R E 9 Elongated grain structure in a 90specimen with thickness 2.1 mm [Colour figure can be viewed at
has been made to quantitatively measure the size of the grains or the α + β Widmanstätten, but from the from the micrographs, no significant difference in
microstructure was observed for specimens with different thickness or for different positions within the samples. Thus, the only influence from microstructure on the
F I G U R E 1 0 (A) Surface connected LOFs of 90specimen with thickness 1.5 mm; (B) magnification of (A)
F I G U R E 1 1 Typical fracture surface appearance for samples with different orientation and thickness [Colour figure can be viewed at wileyonlinelibrary.com]
fatigue results would come from the elongated grain structure and then depend on the orientation on the sample. Internal defects, such as pores or LOF's, were not detected in the bulk material. This was expected since the HIP treatment should close any internal porosity. Only near-surface LOFs (see Figure 11) which have an open channel to the surface are likely to remain after HIP. These surface-connected LOFs are typically
observed in the build plane and often serve as crack initi-ation points during fatigue loading. One typical fracture surface from each group of samples can be found in Figure 10. For all specimens in this study, the fatigue cracks initiate from the surface and propagate into the material before final failure occurs; see Figures 12 and 13. In Figure 12a, which is a cross-section image, a secondary fatigue crack has initiated at a surface pit just
F I G U R E 1 3 Crack initiation for a 90 specimen with thickness 1.5 mm [Colour figure can be viewed at wileyonlinelibrary.com]
below an un-melted particle. Figure 12b instead shows an image of the fracture surface where a secondary crack starts.
In Figure 13a, at least two crack initiation points can be observed at the fracture surface at low magnification. At higher magnification, Figure 13b, it is visible that that the crack has initiated at a surface defect, which is also confirmed by the final image Figure 13c showing fatigue striations in the fracture surface. These defects are sur-face connected LOFs that, most likely, will reduce the fatigue properties of 90 specimens in comparison to 0 and 45specimens due to the unfavorable orientation rel-ative to the load, leading to a higher stress concentration. This is consistent with the fatigue results in Figure 6. In addition, for the 90 specimens, no non-destructive method for surface roughness measurement method is capable of capturing these surface connected LOFs (or valleys) in order to compensate for this in the stress calculations.
4
|
C O N C L U S I O N S
The alloy Ti-6Al-4V produced by E-PBF has been studied, and the fatigue properties have been investigated for samples with three different wall thicknesses ranging from 1.5 to 2.7 mm and in three different directions: 0, 45, and 90relative to the build plate. The main findings are as follows:
• Results indicate that generally the fatigue strength is lower for the thinner samples compared to thicker samples when nominal thickness or calliper measure-ments are used to calculate the corresponding stress. However, if the surface roughness is considered when calculating the stress levels, the influence from speci-men thickness decreases or even disappears.
• All non-destructive testing methods for measuring the surface roughness seem to underestimate the impact of surface roughness on fatigue life for thin-walled structures.
• Samples loaded in the 90 orientation (along the build
direction) show a reduced fatigue life compared to the 0and 45directions, and the reason can be attributed to the detrimental orientation of the melt pool layers and associated near-surface LOFs of the 90-oriented specimens.
A C K N O W L E D G M E N T S
The work has been performed within the project AddMan funded by the Clean Sky 2 joint undertaking under the European Union's Horizon 2020 Framework Programme (research and innovation program) under
grant agreement No. 738002. Ester Lundin at Linköping University is acknowledged for supplementary thickness determinations.
D A T A A V A I L A B I L I T Y S T A T E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.
A U T H O R C O N T R I B U T I O N S
MS and JM carried out the experiments. AK and AH were responsible for printing the test specimens. MS wrote the manuscript with support from JM, MK, and AK. All authors were active in the discussion of the results.
O R C I D
Mikael Segersäll https://orcid.org/0000-0002-7606-5244
Johan Moverare https://orcid.org/0000-0002-8304-0221
R E F E R E N C E S
1. Frazier WE. Metal additive manufacturing: a review. J Mater Eng Perform. 2014;23(6):1917-1928.
2. Herzog D, Seyda V, Wycisk E, Emmelmann C. Additive manufacturing of metals. Acta Mater. 2016;117:371-392. 3. Körner C. Additive manufacturing of metallic components by
selective electron beam melting—a review. Int Mater Rev. 2016; 61(5):361-377.
4. Tan X, Kok Y, Tan YJ, et al. Graded microstructure and mechanical properties of additive manufactured Ti-6Al-4V via electron beam melting. Acta Mater. 2015;97:1-16.
5. Hrabe N, Gnäupel-Herold T, Quinn T. Fatigue properties of a titanium alloy (Ti–6Al–4V) fabricated via electron beam melting (EBM): effects of internal defects and residual stress. Int J Fatigue.2017;94:202-210.
6. Kahlin M, Ansell H, Moverare JJ. Fatigue behaviour of additive manufactured Ti6Al4V, with as-built surfaces, exposed to variable amplitude loading. Int J Fatigue. 2017;103: 353-362.
7. Kahlin M, Ansell H, Moverare JJ. Fatigue behaviour of notched additive manufactured Ti6Al4V with as-built surfaces. Int J Fatigue.2017;101:51-60.
8. Edwards P, O'Conner A, Ramulu M. Electron beam additive manufacturing of titanium components: properties and perfor-mance. J Manuf Sci Eng. 2013;135(6):061016-1-061016-7. 9. Rafi HK, Karthik NV, Gong H, Starr TL, Stucker BE.
Micro-structures and mechanical properties of Ti6Al4V parts fabri-cated by selective laser melting and electron beam melting. J Mater Eng Perform. 2013;22(12):3872-3883.
10. Toh WQ, Wang P, Tan X, Nai MLS, Liu E, Tor SB. Microstruc-ture and wear properties of electron beam melted Ti-6Al-4V parts: a comparison study against as-cast form. Metals (Basel). 2016;6(11):9-13.
11. Kok YH, Tan XP, Loh NH, Tor SB, Chua CK. Geometry depen-dence of microstructure and microhardness for selective elec-tron beam-melted Ti–6Al–4V parts. Virtual Phys Prototyp. 2016; 11(3):183-191.
12. de Formanoir C, Michotte S, Rigo O, Germain L, Godet S. Electron beam melted Ti-6Al-4V: microstructure, texture and mechanical behavior of the as-built and heat-treated material. Mater Sci Eng a. 2016;652:105-119.
13. Bruno J, Rochman A, Cassar G. Effect of build orientation of electron beam melting on microstructure and mechanical prop-erties of Ti-6Al-4V. J Mater Eng Perform. 2017;26(2):692-703. 14. Persenot T, Burr A, Martin G, Buffiere JY, Dendievel R,
Maire E. Effect of build orientation on the fatigue properties of as-built Electron Beam Melted Ti-6Al-4V alloy. Int J Fatigue. 2019;118:65-76.
15. Seifi M, Dahar M, Aman R, Harrysson O, Beuth J, Lewandowski JJ. Evaluation of orientation dependence of fracture toughness and fatigue crack propagation behavior of As-deposited ARCAM EBM Ti-6Al-4V. JOM. 2015;67(3): 597-607.
16. Chern AH, Nandwana P, McDaniels R, et al. Build orientation, surface roughness, and scan path influence on the microstruc-ture, mechanical properties, and flexural fatigue behavior of Ti–6Al–4V fabricated by electron beam melting. Mater Sci Eng a. 2020;772:138740.
17. Kahlin M, Ansell H, Basu D, et al. Improved fatigue strength of additively manufactured Ti6Al4V by surface post processing. Int J Fatigue. 2020;134:105497.
18. Greitemeier D, Dalle Donne C, Syassen F, Eufinger J, Melz T. Effect of surface roughness on fatigue performance of additive manufactured Ti–6Al–4V. Mater Sci Technol. 2016;32(7): 629-634.
19. Greitemeier D, Palm F, Syassen F, Melz T. Fatigue perfor-mance of additive manufactured TiAl6V4 using electron and laser beam melting. Int J Fatigue. 2017;94:211-217.
20. Bagehorn S, Wehr J, Maier HJ. Application of mechanical sur-face finishing processes for roughness reduction and fatigue
improvement of additively manufactured Ti-6Al-4V parts. Int J Fatigue. 2017;102:135-142.
21. Denti L, Bassoli E, Gatto A, Santecchia E, Mengucci P. Fatigue life and microstructure of additive manufactured Ti6Al4V after different finishing processes. Mater Sci Eng a. 2019;755(March): 1-9.
22. Nicoletto G, Konecˇna R, Frkan M, Riva E. Influence of layer-wise fabrication and surface orientation on the notch fatigue behavior of as-built additively manufactured Ti6Al4V. Int J Fatigue. 2020;134:105483.
23. Razavi SMJ, Van Hooreweder B, Berto F. Effect of build thickness and geometry on quasi-static and fatigue behavior of Ti-6Al-4V produced by Electron Beam Melting. Addit Manuf. 2020;36:101426.
24. Pegues J, Roach M, Scott Williamson R, Shamsaei N. Surface roughness effects on the fatigue strength of additively man-ufactured Ti-6Al-4V. Int J Fatigue. 2018;116:543-552.
25. Eriksson R, Sjöström S, Brodin H, Johansson S, Östergren L, Li XH. TBC bond coat-top coat interface roughness: influence on fatigue life and modelling aspects. Surf Coatings Technol. 2013;236:230-238.
How to cite this article: Segersäll M, Kerwin A, Hardaker A, Kahlin M, Moverare J. Fatigue response dependence of thickness measurement methods for additively manufactured E-PBF Ti-6Al-4 V. Fatigue Fract Eng Mater Struct. 2021;44: 1931–1943.https://doi.org/10.1111/ffe.13461
