- Email: [email protected]
Fatigue behaviour of notched additive manufactured Ti6Al4V with as-built surfaces
Accepted Manuscript Fatigue behaviour of notched additive manufactured Ti6Al4V with as-built surfaces M. Kahlin, H. Ansell, J.J. Moverare PII: DOI: Reference:
S0142-1123(17)30180-9 http://dx.doi.org/10.1016/j.ijfatigue.2017.04.009 JIJF 4319
To appear in:
International Journal of Fatigue
Received Date: Revised Date: Accepted Date:
30 January 2017 18 April 2017 22 April 2017
Please cite this article as: Kahlin, M., Ansell, H., Moverare, J.J., Fatigue behaviour of notched additive manufactured Ti6Al4V with as-built surfaces, International Journal of Fatigue (2017), doi: http://dx.doi.org/10.1016/j.ijfatigue. 2017.04.009
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Fatigue behaviour of notched additive manufactured Ti6Al4V with asbuilt surfaces Authors: M. Kahlina,b, H. Ansella,c, J. J. Moverareb* a
Saab AB, Aeronautics, SE-58188 Linköping, Sweden
b
c
Division of Engineering Materials, Linköping University, SE-581 83 Linköping, Sweden
Division of Solid Mechanics, Linköping University, SE-581-83 Linköping, Sweden
* Corresponding author. [email protected], +46 13 28 11 41 Keywords: Additive manufacturing, Fatigue, Ti6Al4V, Stress concentration, Fatigue notch factor
Abstract Additive manufacturing (AM) allows the manufacturer to produce parts with complex geometries that are difficult to produce with conventional production methods. Generally, AM is considered to have great potential for the aerospace industry by contributing to reduced weight and lower costs. There are a number of challenges to be solved before AM can be fully utilized in the aerospace industry, and the understanding of fatigue behaviour is one of the major challenges. Although the fatigue properties of flat additive manufactured specimens with rough as-built surfaces already have been widely studied, in practice, few aerospace components have a simple flat geometry with no corners or radii that would act as stress concentrations. Therefore, the combined effect on fatigue life of a rough as-built surface and a geometrical notch needs to be established. In this study, the fatigue properties of both laser sintered and electron beam melted Ti6Al4V have been investigated and a combined effect of
1
a rough as-built surface and a geometrical notch has been determined. In addition, hot isostatic pressing was found to have no impact on fatigue life for rough as-built surfaces. These findings can be directly applied to predict fatigue behaviour of an AM industrial component.
1. Introduction Additive manufacturing (AM), 3D-printing, in metal, is a relatively new production method that has evolved from rapid prototyping. AM allows the manufacturer to automatically create physical components directly from a computer model and can be used to produce industrial parts that are both lighter in weight, cheaper to produce, and have complex geometries that are difficult or impossible to produce with conventional methods. The material behaviour of AMproduced components, however, can differ substantially from the behaviour of the same component produced by conventional manufacturing. This difference is due to the complexity of the AM-methods which include layer manufacturing and repeated heating of the material [1]. There are many different methods for metal AM and the two main AM techniques are powder bed fusion (PBF) and directed energy deposition (DED) [1]. This study is limited to the PBF processes of electron beam melting (EBM) and laser sintering (LS). All PBF processes have the same basic principles, in which a computer model is sliced into thin layers which will be manufactured one at a time. A thin layer of metal powder is first spread out on a base plate. This is followed by a melting procedure in which the bottom slice of geometry is melted to a solid layer. All excess powder that is not included in the bottom layer geometry is left un-melted. A new layer of powder is thereafter spread out and the second slice of geometry is melted. This process is repeated until the full geometry of the part has been manufactured. Even though they are similar, the EBM process and the LS processes have different procedures, which results in differences in the quality of the material. The major differences between the EBM and the LS processes are that the EBM is performed in 2
vacuum at elevated temperatures, which effectively reduces the residual stresses in the part produced, while LS processes generally are performed in an inert gas environment with a nonheated building chamber [2]. This results in more residual stresses and post stress relieving heat treatment is therefore necessary for LS material. The EBM process uses larger powder particles and thicker layers compared to the LS processes which results in a coarser surface roughness of the EBM parts [3]. In general, AM is considered to have great potential for use in the aerospace industry, as it would contribute to reduced weight, fuel consumption and production and development costs. The potential for AM to be used for the structural parts of an aircraft is closely related to the possibilities to manufacture complex geometries. A part with a complex geometry will have some areas with rough as-built surfaces, since all surfaces will neither be accessible for machining nor polishing, nor will it be cost efficient to machine all surfaces. In spite the potential of AM, there are still a number of challenges yet to be solved before AM can be fully utilized in the aerospace industry. One of the major challenges is the poor fatigue behaviour of the rough as-built surfaces of AM parts since the fatigue behaviour of AM parts is dominated by the rough surface rather than by internal defects [4]. The roughness of the as-built AM surface has been attributed to several mechanisms, for example, balling stair-stepping, and partially melted powder grains attached to the as-built surface [5]. In general, the fatigue strength for additive manufactured samples can be reduced due to several reasons, for example due to a rough as-built surface, an internal defect or a geometrical notch. All these features acts as stress concentrations thus raising the local stress and causing a fatigue crack initiation.
3
Hot isostatic pressing (HIP) has generally been considered to improve the fatigue properties of additive manufactured materials. However, recent studies indicates that HIP has no enhancing effect on fatigue strength for EBM and LS materials with rough as-built surfaces since the fatigue behaviour will be dominated by the roughness rather than by internal defects [6]. The fatigue properties of un-notched specimens with rough as-built surfaces have already been widely studied [3,6–11], but in practice, few aircraft components have simple geometries without any corners or radii that would act as stress concentrations. Therefore, the combined effect of a rough as-built surface and a geometrical notch needs to be established to enable relevant fatigue predictions for structural aircraft parts. There has been no study to date, to our knowledge, that has considered how the fatigue properties of AM material are affected if a geometrical notch, that is built directly by AM, is combined with the rough as-built AM surface. In the present study, the fatigue behaviour of Ti6Al4V produced both with LS and EBM has been investigated. Furthermore, the effect of a combination of a rough asbuilt surface and a geometrical notch has been established as well as the effect of HIP on specimens with a rough as-built surface. These findings can be used to produce design data for prediction of fatigue behaviour of an industrial part with stress concentrations caused by corners or radii.
2. Materials and Experimental Methods 2.1. Material and Test Specimens
Test specimens of the titanium alloy Ti6Al4V were produced both by an EOS M 290 LS equipment using 30µm layer thickness with an average powder size of 50µm and by an 4
Arcam A2 EBM equipment using layer thickness 50µm and a powder range of 45-100µm. The specimens manufactured by LS were stress relieved (SR) at 650°C for 3h in argon gas followed by sandblasting while the specimens produced by EBM were only blasted with titanium powder to remove loosely bound powder. Generally, these different surface post process procedures are the standard ones for each process. HIP was performed at 920°C and 1000 bar for 2h in argon gas environment. Light optical microscopy was performed on both EBM and LS samples with and without HIP to determine the presence of internal defects. All specimens, presented in this paper, manufactured with LS and EBM, respectively, were built in the same build cycle to be able to compare different test series without the scatter from batch to batch variations. The fatigue specimens produced with both EBM and LS had the loading direction in the Z-direction, the build direction, since this orientation gives the roughest as-built surfaces [1]. Furthermore, the AM specimens were produced with either rough as-built surfaces or machined-and-polished surfaces. Reference specimens were manufactured from a wrought Ti6Al4V Ø30mm diameter bar, in mill annealed condition. A brief summary of the conditions for the fatigue test series can be found in Table 1. Two different specimen geometries, with different theoretical stress concentration factors (Kt), were used for the fatigue specimens: Type 1 had a smooth surface (Kt=1) and type 2 had a geometrical notch (Kt=2.5) with a 0.85mm radius, as illustrated by Figure 1 and Figure 2. Specimens of type 1 and type 2 were produced with either a rough as-build surface or a machined surface. The diameter of specimen type 1 and 2 were Ø6.5-7.0 mm. In this paper the rough surface will be called rough as-built (AB) surface even though blasting and heat treatment were performed as post processes.
5
Figure 1. Fatigue test specimens of type 1 (left) and 2 (right) with both rough as-build surfaces and machinedand-polished surfaces.
Figure 2. Geometrical notch for type 2 specimens. Machined notch (left), LS as-built notch (middle) and EBM as-built notch (right).
2.2. Hardness Tests
Vickers hardness testing was performed with a Struers Durascan hardness test equipment using 5 kgf loading (HV5), fulfilling the requirements in ASTM E 384 for minimum distance between indents and to the outer edge. The load was applied in the Z-direction (building direction) of the specimen on polished surfaces and a hardness profile from specimen surface to centre was produced, for each material condition. Three hardness indents were used for each position. Table 1. A summary of conditions for the different fatigue test series. HIP=Hot Isostatic Pressing, SR=Stress relieving
Material EBM
Heat treatment HIP
EBM EBM
HIP none
Surface Machined-andpolished Rough as-built Rough as-built 6
Specimen type 1&2 1&2 1&2
LS
SR+HIP
LS SR+HIP LS SR Wrought bar Mill annealed
Machined-andpolished Rough as-built Rough as-built Machined-andpolished
1&2 1&2 1 1&2
2.3. Surface Roughness Investigations
The surface roughness was investigated, in the Z-direction, using a Nikon Optishot optical microscope with x10 magnification. The surface roughness of cross-sectioned specimens was evaluated with an image analysing technique over an evaluation length of approximate 15mm. The image analyses were performed using an in-house Matlab script [12] that acquires a surface roughness profile from greyscale micrographs and calculates the surface roughness parameters, Rv (maximum profile valley depth within a sampling length) and Ra (arithmetical mean deviation within a sampling length). 2.4. Fatigue Tests
Constant amplitude fatigue testing were performed in a servo hydraulic fatigue test rig with an Instron±50kN load cell and an Instron 8800 controller using stress control with a stress ratio of R=0.1 and a load frequency of 20Hz. All tests were carried out at room temperature. After the tests, the fracture surface of each fatigue specimen was investigated by stereomicroscopy. Representative specimens for each test series were then further investigated by a HITACHI SU-70 field emission gun scanning electron microscope (SEM), operating at 15kV for more detailed imaging of the crack initiation locations.
3. Results 3.1. Hardness
All test series in this study have a uniform hardness throughout the thickness which is illustrated by Figure 3. The EBM and LS specimens subjected to HIP treatment show similar 7
hardness to the wrought reference material, which indicates that the static strengths for these series are approximate the same. The hardness decreases considerably for both the LS and the EBM series after the HIP treatment. The mean values from the Vickers hardness measurements are given in Figure 3 as hardness profiles ranging from specimen edge to centre.
Figure 3. Average hardness for all the fatigue test series. The error bar indicates the maximum and minimum hardness measurement for each position. Three tests per position.
3.2. Surface Roughness
The obtained surface roughness is illustrated as mean values in Figure 4 and examples of the surface profiles used for image analysis are presented in Figure 5. In this study, it was found that the as-built EBM surface in the Z-direction was very rough with an average Rv of 92µm which was more than twice that of the as-built LS surface.
Figure 4. Mean values of surface roughness, in Z-direction, obtained by image analyses. Rv=maximum profile valley depth, Ra =arithmetical mean deviation.
8
Figure 5. Surface profiles of as-built EBM, as-built LS and machined-and-polished wrought material.
3.3. Internal Defects
Microscopy investigations showed a large amount of gas porosity and Lack-of-Fusion (LOF) in the EBM material without HIP while LS material without HIP had considerably smaller and fewer number of pores and LOF. In contrast, the HIP:ed EBM and LS material showed similar quality with only occasional LOF located just below the surface. 3.4. Fatigue 3.4.1. HIP vs no-HIP Material
The fatigue behaviour for AM material with rough as-built surface has been investigated for material with and without HIP treatment and the results are presented in Figure 6 and Figure 7. The fatigue test results indicate that there is no distinct enhancing effect of the HIP on the fatigue life for neither EBM nor LS material.
9
Figure 6. Fatigue life for HIP:ed vs non-HIP:ed AM material with Kt=1(type 1) and a rough as-built (AB) surface.
Figure 7. Fatigue life for HIP:ed vs non-HIP:ed AM material with Kt=2.5(type 2) geometrical notch and a rough as-built (AB) surface.
3.4.2. Machined vs As-built Surfaces
The fatigue properties for both the wrought bar and the AM specimens with machined-andpolished surface have been tested and compared to the fatigue properties of AM material with rough as-built surface, see Figure 8. The rough as-built surface, as illustrated by Figure 12, has a huge impact on the fatigue strength and reduces the fatigue limit with approximate 45% for LS and 75% for EBM material.
Figure 8. Fatigue life for specimens with machined-and-polished (m&p) surfaces vs rough as-built (AB) surfaces. Kt=1(type 1) specimens.
10
3.4.3. Notched vs Un-notched Specimens
The fatigue notch factor, Kf, can deviate from the stress concentration factor, Kt, depending on the notch-sensitivity, q, of the material. The fatigue notch factor, Kf, is defined as the actual reduction of the fatigue limit, see Eq. 1 [13]. In this study the fatigue notch factor Kf has been determined for all fatigue test series using the fatigue limit for machined-andpolished specimens from wrought bar with Kt=1 (type 1) as a base line and calculating the resulting Kf in comparison to this base line for all other test series in accordance to Eq. 2. The fatigue limit was defined as the maximum stress at 5 x 106 cycles to failure. The HIP and noHIP series with a rough as-built surface have been evaluated as one single series since there were none or minor differences between the series. The fatigue notch factors are illustrated in Figure 9 and Figure 10 and summarized in Table 2 and Table 3. The un-notched (Kt=1, type 1) EBM samples that were subjected to HIP, machining-andpolishing showed comparable fatigue behaviour to the un-notched (Kt=1, type 1) wrought bar while the corresponding LS samples had considerably lower fatigue strength. In contrast to this, for material with rough as-built surface, the un-notched (Kt=1, type 1) the LS samples had superior fatigue strength compared to the EBM samples. However, the fatigue strength for Kt=2.5 (type 2) specimens are similar for machined-andpolished EBM and LS material. Moreover, the Kt=2.5 (type 2) EBM and LS specimens with rough as-built surface also show equally good fatigue properties. Kf = (un-notched fatigue limit) / (notched fatigue limit)
(Eq. 1)
Kf = (fatigue limit for wrought bar Kt = 1 (type 1)) / (fatigue limit for each test series)
(Eq. 2)
Table 2. Fatigue notch factor Kf for test series without a geometrical notch (specimen type 1). Kf calculated on maximum stress at a fatigue life of 5x106cycles. HIP= hot isostatic pressing, AB= as-built.
Series without a
Wrought bar,
EBM+HIP, machined-
LS+HIP, machined11
LS + HIP/no HIP,
EBM + HIP/no HIP,
geometrical notch Kf
machinedand-polished and-polished rough AB and-polished surface surface surface surface 1 1.04 1.51 2.77
rough AB surface 4.26
Table 3. Fatigue notch factor Kf for test series with a Kt=2.5 geometrical notch (specimen type 2). Kf calculated on maximum stress at a fatigue life of 5x106cycles. HIP= hot isostatic pressing, AB= as-built.
Series with a Kt=2.5 geometrical notch Kf
Wrought bar, machined surface
EBM+HIP, machined surface
LS+HIP, machined surface
LS + HIP/no HIP, rough AB surface
EBM + HIP/no HIP, rough AB surface
1.61
1.84
1.80
6.15
6.64
12
Figure 9. Fatigue notch factor for LS material compared to wrought material. Kf= Notch fatigue factor, Kt= Stress concentration factor. AB= as-built surface.
Figure 10. Fatigue notch factor for EBM material compared to wrought material. Kf= Notch fatigue factor, Kt= Stress concentration factor. AB= as-built surface.
13
Figure 11. Crack initiation locations for Kt=1 (type 1) specimens; a-b: machined-and-polished wrought bar, c-d: EBM+HIP with rough as-built surface, e-f: LS+HIP with rough as-built surface, g: EBM+HIP with machinedand-polished surface, h: LS+HIP with machined-and-polished surface. Primary crack initiations site is marked with a red circle.
14
Table 4. No. of specimens with failure from surface defects.
Test series LS, rough as-built surface, no HIP LS, rough as-built surface, HIP EBM, rough as-built surface, no HIP EBM, rough as-built surface, HIP EBM, HIP, machined-and-polished LS, HIP, machined-and-polished Wrought bar, machined-and-polished
No. of specimens with surface failure 6 of 6 (100%) 7 of 7 (100%) 7 of 7 (100%) 5 of 5 (100%) 2 of 3 (67%) 1 of 4 (25%) 6 of 7 (86%)
3.4.4. Crack Initiation
All AM fatigue tests specimens with rough as-built surface, both Kt=1 (type 1) and Kt=2.5 (type 2), have one or several crack initiations at the surface, while a fraction of the cracks initiates
at
internal
locations
for
the
machined-and-polished
specimens.
Fracture surfaces and crack initiation locations are presented in Figure 11 and Table 4. A more detailed description of the crack initiation sites for the different conditions is given below. Kt=1 (type 1) specimens The general trend for Kt=1 specimens with rough as-built surfaces is that low fatigue loads gives 1-2 initiation sites while higher loads gives multiple surfaces initiation. The machined AM test specimens have cracks starting either from the polished surface or at internal features. The crack initiations for machined-and-polished LS+HIP material were generally at internal features in contrast to the machined-and-polished EBM+HIP specimens in which the cracks mostly starts from the surface, see Table 4. Generally, the fracture surfaces for the EBM and wrought bar test series are considerably smoother compared to the fracture surfaces of the LS test series. Kt=2.5 (type 2) specimens
15
All specimens both machined and with rough as-built surface have, one or several, crack initiations at the surface at the root of the notch.
4. Discussion 4.1. Comparison to Previous Fatigue Investigations
Fatigue properties of AM Ti6Al4V have previously been investigated in a number of studies. Fatigue limits from relevant studies, that have performed fatigue tests with the stress ratio R=0.1, are compared to the results from the present study in Table 5. The results from these investigations show a large spread in fatigue limit which further enlightens the fact that AM material can have a large variation in properties and that the material quality can differ even though the same alloy and type of AM equipment were used. Table 5. Fatigue limit at 5 x 10 6 cycles for AM Ti6Al4V from previous studies with stress ratio R=0.1. SR= stress relieved, HIP= Hot isostatic pressing, m&p=machined-and-polished, AB=as-built surface, Z=loading in vertical direction (building direction), X=loading in horizontal direction, 45°=loading in 45° direction, n/a= loading direction not available.
AM process and condition EBM, HIP, m&p LS, SR, HIP, m&p EBM, HIP, AB LS, SR, HIP, AB EBM, AB
LS, SR, AB
This study
Literature
800 MPa Z 500-600 MPa Z 200 MPa Z 300 MPa Z 180 MPa Z
650 MPa [6] X 650 MPa [6] X 140 MPa [6] X 185 MPa [6] X 150 MPa [6] SR, X
300 MPa Z
200 MPa [6] X
16
550-600 MPa [7] Z
570 MPa [14] X/Z
200-250 MPa [7] Z 210 MPa [8] 45° 450 MPa [10] Z
340 MPa [3] n/a 550 MPa [3] n/a 200 MPa [9] Z/45
4.2. Effect of Surface Roughness
The surface is generally very rough for additive manufactured materials and is dependent on which AM-process that is used as well as how the part is orientated in the build chamber. A flat surface parallel to the build direction, Z, is 2-3 times as rough as the surface perpendicular to the build direction according to Frazier [1]. Moreover, the magnitude of the as-built surface roughness is also dependent on melting energy, powder size distribution, the thickness of each built layer [15]. Compared to EBM material, the as-built surface of LS material is relatively smooth due to the thinner layers, slower scan speeds and finer powder [3].The surface topography consists of multiple sharp radii of curvature that act as micro notches and lead to local stress concentrations [2,5]. An example of a crack initiation at a micro notch can be seen in Figure 11d. One of the most common ways of presenting the magnitude of the surface roughness is the arithmetic mean value Ra measured with a stylus profilometer. However, this could give misleading results for AM materials since AM surfaces can contain very high peaks, deep valleys and overhangs. Karlsson [16] states that Ra measured with a profilometer could underestimate the surface roughness since the stylus cannot correctly follow the peaks and valleys of the additive manufactured surface. The depth of a surface micro notch is, according to Chan [5], in the range of the maximum surface valley R v. To achieve an adequate surface roughness measurement that could be related to the surface effect on fatigue properties, R v was therefore determined through image analysis in this study. If the surface roughness is compared to the fatigue limit, a distinct trend can be seen, as illustrated in Figure 12, in which the fatigue limit is increased with improved surface quality. This indicates that surface roughness has a major impact on the fatigue life of AM components. However, for the machined-and-polished samples, the fatigue limit for LS material is considerably lower compared to EBM material even though both materials have 17
similar surface roughness and hardness. The explanation could be that the LS material has differences in microstructure or internal features compared to the EBM material. However, Greitemeier et al. [6] compared the fatigue properties of LS and EBM material that were HIP:ed and machined-and-polished and found them to have equally good fatigue properties. This shows that the material quality depends on additional factors and may vary even though the same AM technique and material are used.
Figure 12. Fatigue limit (at 5 x 106 cycles) compared to surfaces roughness. Rv= maximum profile valley depth, AB= as-built.
In a previous study by Chan et al. [2], the micro notches in rough as-built Ti6Al4V surfaces have been evaluated and a local stress concentration factor, Kt, was determined, see Table 6. Chan et al. calculated the stress concentration at a crack initiation site using the surface roughness as notch depth and the radii of the micro notch at the crack initiation position. Table 6. Stress concentration factor, Kt, for surface micro notches, fatigue notch factor, Kf, and the notch sensitivity, q, for of AM material with rough as-built surfaces compared to wrought bar.
Material Kt Kf q LS 5.4 [2] 2.77* 0.40 EBM 9.1[2] 4.26* 0.40 Wrought bar 2.5 1.61** 0.41 (reference) * Kt=1 (type 1) specimen, ** Kt=2.5 (type 2) specimen The fatigue notch factor, Kf, for AM material with rough as-built surfaces from the present study could therefore be assumed to correspond to the local stress concentration from Chan et
18
al. The fatigue behaviour of a notched material can be defined with the notch sensitivity, q, which is given in Eq. 3 [13]. q = (Kf – 1) / (Kt – 1)
(Eq. 3)
The notch sensitivity, q, for Kt=1 (type 1) specimens with rough as-built surface was therefore determined using the local stress concentration from Chan et al. and the fatigue notch factors from the present study and the results are presented in Table 6. The notch sensitivity, q, for wrought bar was also calculated for Kt=2.5 (type 2) specimens as a reference. It appears that the notch sensitivity, q, for EBM and LS Ti6Al4V with rough as-built surface is equal, even though the fatigue properties and the local stress concentration, Kt, differ a great deal. 4.3. Effect of HIP 4.3.1. Hardness
It has previously been reported that the microstructures of EBM and LS Ti6Al4V are coarsening during HIP which leads to lowered static strength of the materials [6]. The coarsening seems to have the same reducing effect on the Vickers hardness, see Figure 3, presented in this study. Previous studies of the effect of HIP on the tensile behaviour, presented in Table 7, show a reduced tensile strength and an increased elongation to fracture after HIP treatment. The effect on tensile behaviour is larger for LS material compared to EBM material which corresponds well with the hardness behaviour presented in the present study. Table 7. Average tensile strength from previous studies of Ti6Al4V with and without HIP. Specimen orientation is in Z direction (build direction). HIP performed at 915-920°C for 2h at 1000 bar. HIP = Hot Isostatic Pressing
AM process / equipment
EBM / not specified EBM / not specified LS / Concept Laser M2
Ultimate tensile strength, MPa Without HIP 953 915 1155 1115*
Ultimate tensile strength, MPa With HIP 942 870 973 19
Elongation to fracture, %
Elongation to fracture, %
Without HIP 13.8 13.1 10.9 11.3*
With HIP 12.9 13.7 19.0
Source
[14] [17] [11]
LS / SLM 250 LS / SLM 250
1080 1040** 1315 1228**
1005
1.6 5** 4 8**
1089
8.3
[18]
13.8
[19]
* Stress relieved at 700°C for 1h, ** heat treated at 800°C for 2h
4.3.2. Fatigue
There are very few studies on the effect of HIP treatment on the fatigue strength for additive manufactured Ti6Al4V with rough as-built surface. In a recent study, however, Greitenmeier et al. [6] compared the fatigue behaviour of both LS and EBM material with rough as-built surface and concluded that HIP did not improve the fatigue properties. The present study confirms these findings for Kt=1 (type 1) samples and showed that the same behaviour also applies to samples with a built notch (Kt=2.5). Consequently, for fatigue analysis of LS or EBM Ti6Al4V with rough as-built surfaces, the same fatigue design data can be used for both HIP:ed and non-HIP:ed parts. Several investigations on EBM [14] and LS [11,18] Ti6Al4V material show that machined EBM/LS Ti6Al4V with HIP treatment can have similar fatigue life as conventional wrought material. The test results for machined EBM samples subjected to HIP, with Kt=1, in the present study show similar fatigue strength to conventional wrought material as illustrated by Figure 10 which correspond well with the previous findings. The fatigue strength for machined LS Kt=1 (type 1) samples subjected to HIP is, however, similar to previous test results [6], see Table 5, but considerably lower than the wrought bar investigated in this study. Microscopy investigations of HIP:ed EBM and LS samples showed neither gas pores nor LOF defects in the bulk material. Moreover, occasional LOF defects were found just below the as-built surface, for both the HIP:ed EBM and LS material, which indicates that these LOF have a direct connection to the surface since HIP would need fully enclosed voids to be effective. For the machined-and-polished samples, however, these surface connected LOF 20
would subsequently be removed by machining and could therefore not have affected the fatigue life. Günther et al. [20] suggests that for HIP:ed AM material, in which large defects like LOF and porosity have been removed, the fatigue crack initiation occurs instead at “single α-phase grains or at clusters of α-phases”. The image of α-phase crack initiation presented by Günther et al. corresponds well with the failure locations of the machined samples in the present study, in which the EBM+HIP and LS+ HIP samples had similar initiation appearances, see Figure 13. This would indicate that these initiations also would originate from α-phases making the fatigue properties of the HIP:ed and machined AM samples, in the present study, dependent on microstructure features rather than internal defects. This further enlights the complexity of AM material properties, since the microstructure features will depend on the thermal history from the production process which would dependent on, for example AM equipment, part geometry and support structure.
Figure 13. -Crack initiation for a machined EBM sample subjected to HIP.
4.4. Effect of Geometrical Notch
The fatigue strength for machined-and-polished samples with a geometrical notch (Kt=2.5, type 2), was similar for both EBM and LS material even though the corresponding smooth (Kt=1, type 1) specimens showed considerably difference in fatigue strength. The geometrical notch will guide the crack initiation to the surface, or close to the surface, of the machined
21
notch. Therefore, the similarity for the Kt=2.5 results can most likely be attributed to a volume effect, in which the probability is low that the most severe internal defect in the sample would be located in the volume close to the notch. Moreover, the fatigue behaviour for EBM and LS material with a rough as-built surface in combination with a notch geometry, built directly by the AM process, was also similar even though the LS material had much less rough surface, see Figure 4, and superior fatigue properties for Kt=1 (type 1) samples with rough as-built surface. In this case, that can also most likely be attributed to a volume effect, or an area effect, in which the crack initiation for a notched sample is guided to the highest stress concentration in the rough as-built surface at the notch rather than to the overall most severe surface stress concentration of the sample. Chan [5] used microcomputed tomography to investigate the average local stress concentration for EBM and LS materials with as-built surfaces and found Kt=3.28 for EBM and Kt=2.31 for LS material. These average stress concentrations, for EBM and LS material, are more similar than the most severe stress concentrations, Kt=9.1 (EBM) and Kt=5.4 (LS) that Chan et al. [2] found at the crack initiation locations. There is a larger probability that the local stress concentration inhered from the rough surface, in the notch, is close to the average value rather than equal to the maximum value which could explain the more similar behaviour of the notched EBM and LS material with rough as-built surfaces. The combination effect of a rough as-built surface and a geometrical notch is very important for fatigue life prediction of additive manufactured materials since few aerospace components have a simple flat geometry without corners or radii. The fatigue data presented in this report can therefore be used to produce fatigue design data that will allow more realistic fatigue life predictions for additive manufactured components.
22
5. Conclusion The major goal of this study was to investigate how the fatigue properties of additive manufactured Ti6Al4V, produced with laser sintering (LS) and electron beam melting (EBM), were affected by geometrical notches and rough as-built surfaces as well as a combination of these. Moreover, the effect of hot isostatic pressing (HIP) post heat treatment has been investigated. The findings in this study leads to the following conclusions:
The surface roughness is the single most severe factor for fatigue for additive manufactured materials with a rough as-built surface which results in a fatigue notch factor, Kf, of 2.77 for LS material and 4.26 for EBM material, which has a rougher surface.
The combined effect of a rough as-built surface and a geometrical notch (Kt=2.5) gives a fatigue notch factor, Kf, of 6.15 for LS material and 6.64 for EBM material.
The notch sensitivity, q, for EBM and LS Ti6Al4V with rough as-built surface is comparable, 0.40, even though the fatigue properties and the local stress concentrations, Kt, are very different.
LS and EBM material show similar fatigue behaviour for specimens with geometrical notches, stress concentration factor Kt=2.5, for both machined surfaces and rough asbuilt surfaces even though the corresponding un-notched samples, Kt=1, showed very different fatigue behaviour. This different in behaviour is most likely due to a volume effect in which the crack initiation site for Kt=2.5 specimens is guided to the built notch area and not to the overall most severe surface stress concentration.
23
HIP post heat treatment gives no improvement in fatigue life for additive manufactured materials with rough as-built surface since the rough surface is more severe for crack initiation than the internal defects.
Material produced with EBM was comparable in fatigue to conventional wrought bar due to the reduction of internal defects through HIP and improved surface finish through machining and polishing.
In order to improve the fatigue strength of EBM and LS materials, the severity of the micro notches in the rough surfaces need to be reduced. This could be achieved by further work either within process optimization or by surfaces post processes. Machining is not an option for complex geometries due to the lack accessibility and high cost. Therefore, other cost efficient tool-less post processes are needed to improve the surface quality before AM parts can be fully utilized in structural aircraft parts for serial production.
Acknowledgements The authors are grateful to Saab AB, the Swedish Foundation for Strategic Research and the European commission, through the Clean Sky 2 programme, for financial support for the presented study. Mid Sweden University is acknowledged for contribution of test samples.
References [1]
Frazier WE. Metal additive manufacturing: A review. J Mater Eng Perform 2014;23:1917–28.
[2]
Chan KS, Koike M, Mason RL, Okabe T. Fatigue life of titanium alloys fabricated by additive layer manufacturing techniques for dental implants. Metall Mater Trans A Phys Metall Mater Sci 2013;44:1010–22.
[3]
Rafi HK, Karthik N V., Gong H, Starr TL, Stucker BE. Microstructures and mechanical properties of Ti6Al4V parts fabricated by selective laser melting and electron beam melting. J Mater Eng Perform 2013;22:3872–83.
[4]
Uriondo A, Esperon-Miguez M, Perinpanayagam S. The present and future of additive manufacturing in the aerospace sector: A review of important aspects. Proc Inst Mech Eng Part G J Aerosp Eng 2015;229:2132–47.
24
[5]
Chan KS. Characterization and analysis of surface notches on Ti-alloy plates fabricated by additive manufacturing techniques. Surf Topogr Metrol Prop 2015;3:44006.
[6]
Greitemeier D, Palm F, Syassen F, Melz T. Fatigue performance of additive manufactured TiAl6V4 using electron and laser beam melting. Int J Fatigue 2017;94:211–7.
[7]
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– 10.
[8]
Wycisk E, Emmelmann C, Siddique S, Walther F. High Cycle Fatigue (HCF) Performance of Ti-6Al-4V Alloy Processed by Selective Laser Melting. Adv Mater Res 2013;816–817:134–9.
[9]
Wycisk E, Solbach A, Siddique S, Herzog D, Walther F, Emmelmann C. Effects of Defects in Laser Additive Manufactured Ti-6Al-4V on Fatigue Properties. Phys Procedia 2014;56:371–8.
[10]
Gong H, Rafi K, Starr T, Stucker BE. Effect of Defects on Fatigue Tests of As-Built Ti-6Al-4V Parts Fabricated by Selective Laser Melting. Proc. Solid Free. Fabr. Symp., 2012, p. 499–506.
[11]
Kasperovich G, Hausmann J. Improvement of fatigue resistance and ductility of TiAl6V4 processed by selective laser melting. J Mater Process Technol 2015;220:202–14.
[12]
Eriksson R. Thermal Barrier Coatings: Durability Assessment and Life Prediction. Dissertation No. 1527, Linköping University; 2013.
[13]
Theodore N. High Cycle Fatigue: A Mechanics of Materials Perspective. 1st ed. Elsevier Ltd; 2006.
[14]
Svensson M, Ackelid U, Ab A. Titanium Alloys Manufactured with Electron Beam Melting Mechanical and Chemical Properties. Med. Device Mater. V Proc. from Mater. Process. Med. Devices Conf. 2009, 2009, p. p 189-194.
[15]
Kranz J, Herzog D, Emmelmann C. Design guidelines for laser additive manufacturing of lightweight structures in TiAl6V4. J Laser Appl 2015;27:S14001.
[16]
Karlsson J. Optimization of Electron Beam Melting for Production of Small Components in Biocompatible Titanium Grades. Disseratation No. 1206, Uppsala University; 2015.
[17]
Facchini L, Magalini E, Robotti P, Molinari A. Microstructure and mechanical properties of Ti-6Al-4V produced by electron beam melting of pre-alloyed powders. Rapid Prototyp J 2009;15:171–8.
[18]
Leuders S, Thöne M, Riemer A, Niendorf T, Tröster T, Richard H a., et al. On the mechanical behaviour of titanium alloy TiAl6V4 manufactured by selective laser melting: Fatigue resistance and crack growth performance. Int J Fatigue 2013;48:300–7.
[19]
Leuders S, Lieneke T, Lammers S, Tröster T, Niendorf T. On the fatigue properties of metals manufactured by selective laser melting – The role of ductility. J Mater Res 2014:1–9.
[20]
Günther J, Krewerth D, Lippmann T, Leuders S, Tröster T, Weidner A, et al. Fatigue life of additively manufactured Ti–6Al–4V in the very high cycle fatigue regime. Int J Fatigue 2017;94:236–45.
25
Highlights
The rough as-built surface of AM materials is the most severe factor for fatigue The fatigue reduction of AM materials was quantified by fatigue notch factor (Kf) A combined effect with geometrical AM notches and rough surface was determined The notch sensitivity (q) is for AM material is equal to wrought material HIP does not improve fatigue properties for AM material with rough surfaces
26
S0142-1123(17)30180-9 http://dx.doi.org/10.1016/j.ijfatigue.2017.04.009 JIJF 4319
To appear in:
International Journal of Fatigue
Received Date: Revised Date: Accepted Date:
30 January 2017 18 April 2017 22 April 2017
Please cite this article as: Kahlin, M., Ansell, H., Moverare, J.J., Fatigue behaviour of notched additive manufactured Ti6Al4V with as-built surfaces, International Journal of Fatigue (2017), doi: http://dx.doi.org/10.1016/j.ijfatigue. 2017.04.009
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Fatigue behaviour of notched additive manufactured Ti6Al4V with asbuilt surfaces Authors: M. Kahlina,b, H. Ansella,c, J. J. Moverareb* a
Saab AB, Aeronautics, SE-58188 Linköping, Sweden
b
c
Division of Engineering Materials, Linköping University, SE-581 83 Linköping, Sweden
Division of Solid Mechanics, Linköping University, SE-581-83 Linköping, Sweden
* Corresponding author. [email protected], +46 13 28 11 41 Keywords: Additive manufacturing, Fatigue, Ti6Al4V, Stress concentration, Fatigue notch factor
Abstract Additive manufacturing (AM) allows the manufacturer to produce parts with complex geometries that are difficult to produce with conventional production methods. Generally, AM is considered to have great potential for the aerospace industry by contributing to reduced weight and lower costs. There are a number of challenges to be solved before AM can be fully utilized in the aerospace industry, and the understanding of fatigue behaviour is one of the major challenges. Although the fatigue properties of flat additive manufactured specimens with rough as-built surfaces already have been widely studied, in practice, few aerospace components have a simple flat geometry with no corners or radii that would act as stress concentrations. Therefore, the combined effect on fatigue life of a rough as-built surface and a geometrical notch needs to be established. In this study, the fatigue properties of both laser sintered and electron beam melted Ti6Al4V have been investigated and a combined effect of
1
a rough as-built surface and a geometrical notch has been determined. In addition, hot isostatic pressing was found to have no impact on fatigue life for rough as-built surfaces. These findings can be directly applied to predict fatigue behaviour of an AM industrial component.
1. Introduction Additive manufacturing (AM), 3D-printing, in metal, is a relatively new production method that has evolved from rapid prototyping. AM allows the manufacturer to automatically create physical components directly from a computer model and can be used to produce industrial parts that are both lighter in weight, cheaper to produce, and have complex geometries that are difficult or impossible to produce with conventional methods. The material behaviour of AMproduced components, however, can differ substantially from the behaviour of the same component produced by conventional manufacturing. This difference is due to the complexity of the AM-methods which include layer manufacturing and repeated heating of the material [1]. There are many different methods for metal AM and the two main AM techniques are powder bed fusion (PBF) and directed energy deposition (DED) [1]. This study is limited to the PBF processes of electron beam melting (EBM) and laser sintering (LS). All PBF processes have the same basic principles, in which a computer model is sliced into thin layers which will be manufactured one at a time. A thin layer of metal powder is first spread out on a base plate. This is followed by a melting procedure in which the bottom slice of geometry is melted to a solid layer. All excess powder that is not included in the bottom layer geometry is left un-melted. A new layer of powder is thereafter spread out and the second slice of geometry is melted. This process is repeated until the full geometry of the part has been manufactured. Even though they are similar, the EBM process and the LS processes have different procedures, which results in differences in the quality of the material. The major differences between the EBM and the LS processes are that the EBM is performed in 2
vacuum at elevated temperatures, which effectively reduces the residual stresses in the part produced, while LS processes generally are performed in an inert gas environment with a nonheated building chamber [2]. This results in more residual stresses and post stress relieving heat treatment is therefore necessary for LS material. The EBM process uses larger powder particles and thicker layers compared to the LS processes which results in a coarser surface roughness of the EBM parts [3]. In general, AM is considered to have great potential for use in the aerospace industry, as it would contribute to reduced weight, fuel consumption and production and development costs. The potential for AM to be used for the structural parts of an aircraft is closely related to the possibilities to manufacture complex geometries. A part with a complex geometry will have some areas with rough as-built surfaces, since all surfaces will neither be accessible for machining nor polishing, nor will it be cost efficient to machine all surfaces. In spite the potential of AM, there are still a number of challenges yet to be solved before AM can be fully utilized in the aerospace industry. One of the major challenges is the poor fatigue behaviour of the rough as-built surfaces of AM parts since the fatigue behaviour of AM parts is dominated by the rough surface rather than by internal defects [4]. The roughness of the as-built AM surface has been attributed to several mechanisms, for example, balling stair-stepping, and partially melted powder grains attached to the as-built surface [5]. In general, the fatigue strength for additive manufactured samples can be reduced due to several reasons, for example due to a rough as-built surface, an internal defect or a geometrical notch. All these features acts as stress concentrations thus raising the local stress and causing a fatigue crack initiation.
3
Hot isostatic pressing (HIP) has generally been considered to improve the fatigue properties of additive manufactured materials. However, recent studies indicates that HIP has no enhancing effect on fatigue strength for EBM and LS materials with rough as-built surfaces since the fatigue behaviour will be dominated by the roughness rather than by internal defects [6]. The fatigue properties of un-notched specimens with rough as-built surfaces have already been widely studied [3,6–11], but in practice, few aircraft components have simple geometries without any corners or radii that would act as stress concentrations. Therefore, the combined effect of a rough as-built surface and a geometrical notch needs to be established to enable relevant fatigue predictions for structural aircraft parts. There has been no study to date, to our knowledge, that has considered how the fatigue properties of AM material are affected if a geometrical notch, that is built directly by AM, is combined with the rough as-built AM surface. In the present study, the fatigue behaviour of Ti6Al4V produced both with LS and EBM has been investigated. Furthermore, the effect of a combination of a rough asbuilt surface and a geometrical notch has been established as well as the effect of HIP on specimens with a rough as-built surface. These findings can be used to produce design data for prediction of fatigue behaviour of an industrial part with stress concentrations caused by corners or radii.
2. Materials and Experimental Methods 2.1. Material and Test Specimens
Test specimens of the titanium alloy Ti6Al4V were produced both by an EOS M 290 LS equipment using 30µm layer thickness with an average powder size of 50µm and by an 4
Arcam A2 EBM equipment using layer thickness 50µm and a powder range of 45-100µm. The specimens manufactured by LS were stress relieved (SR) at 650°C for 3h in argon gas followed by sandblasting while the specimens produced by EBM were only blasted with titanium powder to remove loosely bound powder. Generally, these different surface post process procedures are the standard ones for each process. HIP was performed at 920°C and 1000 bar for 2h in argon gas environment. Light optical microscopy was performed on both EBM and LS samples with and without HIP to determine the presence of internal defects. All specimens, presented in this paper, manufactured with LS and EBM, respectively, were built in the same build cycle to be able to compare different test series without the scatter from batch to batch variations. The fatigue specimens produced with both EBM and LS had the loading direction in the Z-direction, the build direction, since this orientation gives the roughest as-built surfaces [1]. Furthermore, the AM specimens were produced with either rough as-built surfaces or machined-and-polished surfaces. Reference specimens were manufactured from a wrought Ti6Al4V Ø30mm diameter bar, in mill annealed condition. A brief summary of the conditions for the fatigue test series can be found in Table 1. Two different specimen geometries, with different theoretical stress concentration factors (Kt), were used for the fatigue specimens: Type 1 had a smooth surface (Kt=1) and type 2 had a geometrical notch (Kt=2.5) with a 0.85mm radius, as illustrated by Figure 1 and Figure 2. Specimens of type 1 and type 2 were produced with either a rough as-build surface or a machined surface. The diameter of specimen type 1 and 2 were Ø6.5-7.0 mm. In this paper the rough surface will be called rough as-built (AB) surface even though blasting and heat treatment were performed as post processes.
5
Figure 1. Fatigue test specimens of type 1 (left) and 2 (right) with both rough as-build surfaces and machinedand-polished surfaces.
Figure 2. Geometrical notch for type 2 specimens. Machined notch (left), LS as-built notch (middle) and EBM as-built notch (right).
2.2. Hardness Tests
Vickers hardness testing was performed with a Struers Durascan hardness test equipment using 5 kgf loading (HV5), fulfilling the requirements in ASTM E 384 for minimum distance between indents and to the outer edge. The load was applied in the Z-direction (building direction) of the specimen on polished surfaces and a hardness profile from specimen surface to centre was produced, for each material condition. Three hardness indents were used for each position. Table 1. A summary of conditions for the different fatigue test series. HIP=Hot Isostatic Pressing, SR=Stress relieving
Material EBM
Heat treatment HIP
EBM EBM
HIP none
Surface Machined-andpolished Rough as-built Rough as-built 6
Specimen type 1&2 1&2 1&2
LS
SR+HIP
LS SR+HIP LS SR Wrought bar Mill annealed
Machined-andpolished Rough as-built Rough as-built Machined-andpolished
1&2 1&2 1 1&2
2.3. Surface Roughness Investigations
The surface roughness was investigated, in the Z-direction, using a Nikon Optishot optical microscope with x10 magnification. The surface roughness of cross-sectioned specimens was evaluated with an image analysing technique over an evaluation length of approximate 15mm. The image analyses were performed using an in-house Matlab script [12] that acquires a surface roughness profile from greyscale micrographs and calculates the surface roughness parameters, Rv (maximum profile valley depth within a sampling length) and Ra (arithmetical mean deviation within a sampling length). 2.4. Fatigue Tests
Constant amplitude fatigue testing were performed in a servo hydraulic fatigue test rig with an Instron±50kN load cell and an Instron 8800 controller using stress control with a stress ratio of R=0.1 and a load frequency of 20Hz. All tests were carried out at room temperature. After the tests, the fracture surface of each fatigue specimen was investigated by stereomicroscopy. Representative specimens for each test series were then further investigated by a HITACHI SU-70 field emission gun scanning electron microscope (SEM), operating at 15kV for more detailed imaging of the crack initiation locations.
3. Results 3.1. Hardness
All test series in this study have a uniform hardness throughout the thickness which is illustrated by Figure 3. The EBM and LS specimens subjected to HIP treatment show similar 7
hardness to the wrought reference material, which indicates that the static strengths for these series are approximate the same. The hardness decreases considerably for both the LS and the EBM series after the HIP treatment. The mean values from the Vickers hardness measurements are given in Figure 3 as hardness profiles ranging from specimen edge to centre.
Figure 3. Average hardness for all the fatigue test series. The error bar indicates the maximum and minimum hardness measurement for each position. Three tests per position.
3.2. Surface Roughness
The obtained surface roughness is illustrated as mean values in Figure 4 and examples of the surface profiles used for image analysis are presented in Figure 5. In this study, it was found that the as-built EBM surface in the Z-direction was very rough with an average Rv of 92µm which was more than twice that of the as-built LS surface.
Figure 4. Mean values of surface roughness, in Z-direction, obtained by image analyses. Rv=maximum profile valley depth, Ra =arithmetical mean deviation.
8
Figure 5. Surface profiles of as-built EBM, as-built LS and machined-and-polished wrought material.
3.3. Internal Defects
Microscopy investigations showed a large amount of gas porosity and Lack-of-Fusion (LOF) in the EBM material without HIP while LS material without HIP had considerably smaller and fewer number of pores and LOF. In contrast, the HIP:ed EBM and LS material showed similar quality with only occasional LOF located just below the surface. 3.4. Fatigue 3.4.1. HIP vs no-HIP Material
The fatigue behaviour for AM material with rough as-built surface has been investigated for material with and without HIP treatment and the results are presented in Figure 6 and Figure 7. The fatigue test results indicate that there is no distinct enhancing effect of the HIP on the fatigue life for neither EBM nor LS material.
9
Figure 6. Fatigue life for HIP:ed vs non-HIP:ed AM material with Kt=1(type 1) and a rough as-built (AB) surface.
Figure 7. Fatigue life for HIP:ed vs non-HIP:ed AM material with Kt=2.5(type 2) geometrical notch and a rough as-built (AB) surface.
3.4.2. Machined vs As-built Surfaces
The fatigue properties for both the wrought bar and the AM specimens with machined-andpolished surface have been tested and compared to the fatigue properties of AM material with rough as-built surface, see Figure 8. The rough as-built surface, as illustrated by Figure 12, has a huge impact on the fatigue strength and reduces the fatigue limit with approximate 45% for LS and 75% for EBM material.
Figure 8. Fatigue life for specimens with machined-and-polished (m&p) surfaces vs rough as-built (AB) surfaces. Kt=1(type 1) specimens.
10
3.4.3. Notched vs Un-notched Specimens
The fatigue notch factor, Kf, can deviate from the stress concentration factor, Kt, depending on the notch-sensitivity, q, of the material. The fatigue notch factor, Kf, is defined as the actual reduction of the fatigue limit, see Eq. 1 [13]. In this study the fatigue notch factor Kf has been determined for all fatigue test series using the fatigue limit for machined-andpolished specimens from wrought bar with Kt=1 (type 1) as a base line and calculating the resulting Kf in comparison to this base line for all other test series in accordance to Eq. 2. The fatigue limit was defined as the maximum stress at 5 x 106 cycles to failure. The HIP and noHIP series with a rough as-built surface have been evaluated as one single series since there were none or minor differences between the series. The fatigue notch factors are illustrated in Figure 9 and Figure 10 and summarized in Table 2 and Table 3. The un-notched (Kt=1, type 1) EBM samples that were subjected to HIP, machining-andpolishing showed comparable fatigue behaviour to the un-notched (Kt=1, type 1) wrought bar while the corresponding LS samples had considerably lower fatigue strength. In contrast to this, for material with rough as-built surface, the un-notched (Kt=1, type 1) the LS samples had superior fatigue strength compared to the EBM samples. However, the fatigue strength for Kt=2.5 (type 2) specimens are similar for machined-andpolished EBM and LS material. Moreover, the Kt=2.5 (type 2) EBM and LS specimens with rough as-built surface also show equally good fatigue properties. Kf = (un-notched fatigue limit) / (notched fatigue limit)
(Eq. 1)
Kf = (fatigue limit for wrought bar Kt = 1 (type 1)) / (fatigue limit for each test series)
(Eq. 2)
Table 2. Fatigue notch factor Kf for test series without a geometrical notch (specimen type 1). Kf calculated on maximum stress at a fatigue life of 5x106cycles. HIP= hot isostatic pressing, AB= as-built.
Series without a
Wrought bar,
EBM+HIP, machined-
LS+HIP, machined11
LS + HIP/no HIP,
EBM + HIP/no HIP,
geometrical notch Kf
machinedand-polished and-polished rough AB and-polished surface surface surface surface 1 1.04 1.51 2.77
rough AB surface 4.26
Table 3. Fatigue notch factor Kf for test series with a Kt=2.5 geometrical notch (specimen type 2). Kf calculated on maximum stress at a fatigue life of 5x106cycles. HIP= hot isostatic pressing, AB= as-built.
Series with a Kt=2.5 geometrical notch Kf
Wrought bar, machined surface
EBM+HIP, machined surface
LS+HIP, machined surface
LS + HIP/no HIP, rough AB surface
EBM + HIP/no HIP, rough AB surface
1.61
1.84
1.80
6.15
6.64
12
Figure 9. Fatigue notch factor for LS material compared to wrought material. Kf= Notch fatigue factor, Kt= Stress concentration factor. AB= as-built surface.
Figure 10. Fatigue notch factor for EBM material compared to wrought material. Kf= Notch fatigue factor, Kt= Stress concentration factor. AB= as-built surface.
13
Figure 11. Crack initiation locations for Kt=1 (type 1) specimens; a-b: machined-and-polished wrought bar, c-d: EBM+HIP with rough as-built surface, e-f: LS+HIP with rough as-built surface, g: EBM+HIP with machinedand-polished surface, h: LS+HIP with machined-and-polished surface. Primary crack initiations site is marked with a red circle.
14
Table 4. No. of specimens with failure from surface defects.
Test series LS, rough as-built surface, no HIP LS, rough as-built surface, HIP EBM, rough as-built surface, no HIP EBM, rough as-built surface, HIP EBM, HIP, machined-and-polished LS, HIP, machined-and-polished Wrought bar, machined-and-polished
No. of specimens with surface failure 6 of 6 (100%) 7 of 7 (100%) 7 of 7 (100%) 5 of 5 (100%) 2 of 3 (67%) 1 of 4 (25%) 6 of 7 (86%)
3.4.4. Crack Initiation
All AM fatigue tests specimens with rough as-built surface, both Kt=1 (type 1) and Kt=2.5 (type 2), have one or several crack initiations at the surface, while a fraction of the cracks initiates
at
internal
locations
for
the
machined-and-polished
specimens.
Fracture surfaces and crack initiation locations are presented in Figure 11 and Table 4. A more detailed description of the crack initiation sites for the different conditions is given below. Kt=1 (type 1) specimens The general trend for Kt=1 specimens with rough as-built surfaces is that low fatigue loads gives 1-2 initiation sites while higher loads gives multiple surfaces initiation. The machined AM test specimens have cracks starting either from the polished surface or at internal features. The crack initiations for machined-and-polished LS+HIP material were generally at internal features in contrast to the machined-and-polished EBM+HIP specimens in which the cracks mostly starts from the surface, see Table 4. Generally, the fracture surfaces for the EBM and wrought bar test series are considerably smoother compared to the fracture surfaces of the LS test series. Kt=2.5 (type 2) specimens
15
All specimens both machined and with rough as-built surface have, one or several, crack initiations at the surface at the root of the notch.
4. Discussion 4.1. Comparison to Previous Fatigue Investigations
Fatigue properties of AM Ti6Al4V have previously been investigated in a number of studies. Fatigue limits from relevant studies, that have performed fatigue tests with the stress ratio R=0.1, are compared to the results from the present study in Table 5. The results from these investigations show a large spread in fatigue limit which further enlightens the fact that AM material can have a large variation in properties and that the material quality can differ even though the same alloy and type of AM equipment were used. Table 5. Fatigue limit at 5 x 10 6 cycles for AM Ti6Al4V from previous studies with stress ratio R=0.1. SR= stress relieved, HIP= Hot isostatic pressing, m&p=machined-and-polished, AB=as-built surface, Z=loading in vertical direction (building direction), X=loading in horizontal direction, 45°=loading in 45° direction, n/a= loading direction not available.
AM process and condition EBM, HIP, m&p LS, SR, HIP, m&p EBM, HIP, AB LS, SR, HIP, AB EBM, AB
LS, SR, AB
This study
Literature
800 MPa Z 500-600 MPa Z 200 MPa Z 300 MPa Z 180 MPa Z
650 MPa [6] X 650 MPa [6] X 140 MPa [6] X 185 MPa [6] X 150 MPa [6] SR, X
300 MPa Z
200 MPa [6] X
16
550-600 MPa [7] Z
570 MPa [14] X/Z
200-250 MPa [7] Z 210 MPa [8] 45° 450 MPa [10] Z
340 MPa [3] n/a 550 MPa [3] n/a 200 MPa [9] Z/45
4.2. Effect of Surface Roughness
The surface is generally very rough for additive manufactured materials and is dependent on which AM-process that is used as well as how the part is orientated in the build chamber. A flat surface parallel to the build direction, Z, is 2-3 times as rough as the surface perpendicular to the build direction according to Frazier [1]. Moreover, the magnitude of the as-built surface roughness is also dependent on melting energy, powder size distribution, the thickness of each built layer [15]. Compared to EBM material, the as-built surface of LS material is relatively smooth due to the thinner layers, slower scan speeds and finer powder [3].The surface topography consists of multiple sharp radii of curvature that act as micro notches and lead to local stress concentrations [2,5]. An example of a crack initiation at a micro notch can be seen in Figure 11d. One of the most common ways of presenting the magnitude of the surface roughness is the arithmetic mean value Ra measured with a stylus profilometer. However, this could give misleading results for AM materials since AM surfaces can contain very high peaks, deep valleys and overhangs. Karlsson [16] states that Ra measured with a profilometer could underestimate the surface roughness since the stylus cannot correctly follow the peaks and valleys of the additive manufactured surface. The depth of a surface micro notch is, according to Chan [5], in the range of the maximum surface valley R v. To achieve an adequate surface roughness measurement that could be related to the surface effect on fatigue properties, R v was therefore determined through image analysis in this study. If the surface roughness is compared to the fatigue limit, a distinct trend can be seen, as illustrated in Figure 12, in which the fatigue limit is increased with improved surface quality. This indicates that surface roughness has a major impact on the fatigue life of AM components. However, for the machined-and-polished samples, the fatigue limit for LS material is considerably lower compared to EBM material even though both materials have 17
similar surface roughness and hardness. The explanation could be that the LS material has differences in microstructure or internal features compared to the EBM material. However, Greitemeier et al. [6] compared the fatigue properties of LS and EBM material that were HIP:ed and machined-and-polished and found them to have equally good fatigue properties. This shows that the material quality depends on additional factors and may vary even though the same AM technique and material are used.
Figure 12. Fatigue limit (at 5 x 106 cycles) compared to surfaces roughness. Rv= maximum profile valley depth, AB= as-built.
In a previous study by Chan et al. [2], the micro notches in rough as-built Ti6Al4V surfaces have been evaluated and a local stress concentration factor, Kt, was determined, see Table 6. Chan et al. calculated the stress concentration at a crack initiation site using the surface roughness as notch depth and the radii of the micro notch at the crack initiation position. Table 6. Stress concentration factor, Kt, for surface micro notches, fatigue notch factor, Kf, and the notch sensitivity, q, for of AM material with rough as-built surfaces compared to wrought bar.
Material Kt Kf q LS 5.4 [2] 2.77* 0.40 EBM 9.1[2] 4.26* 0.40 Wrought bar 2.5 1.61** 0.41 (reference) * Kt=1 (type 1) specimen, ** Kt=2.5 (type 2) specimen The fatigue notch factor, Kf, for AM material with rough as-built surfaces from the present study could therefore be assumed to correspond to the local stress concentration from Chan et
18
al. The fatigue behaviour of a notched material can be defined with the notch sensitivity, q, which is given in Eq. 3 [13]. q = (Kf – 1) / (Kt – 1)
(Eq. 3)
The notch sensitivity, q, for Kt=1 (type 1) specimens with rough as-built surface was therefore determined using the local stress concentration from Chan et al. and the fatigue notch factors from the present study and the results are presented in Table 6. The notch sensitivity, q, for wrought bar was also calculated for Kt=2.5 (type 2) specimens as a reference. It appears that the notch sensitivity, q, for EBM and LS Ti6Al4V with rough as-built surface is equal, even though the fatigue properties and the local stress concentration, Kt, differ a great deal. 4.3. Effect of HIP 4.3.1. Hardness
It has previously been reported that the microstructures of EBM and LS Ti6Al4V are coarsening during HIP which leads to lowered static strength of the materials [6]. The coarsening seems to have the same reducing effect on the Vickers hardness, see Figure 3, presented in this study. Previous studies of the effect of HIP on the tensile behaviour, presented in Table 7, show a reduced tensile strength and an increased elongation to fracture after HIP treatment. The effect on tensile behaviour is larger for LS material compared to EBM material which corresponds well with the hardness behaviour presented in the present study. Table 7. Average tensile strength from previous studies of Ti6Al4V with and without HIP. Specimen orientation is in Z direction (build direction). HIP performed at 915-920°C for 2h at 1000 bar. HIP = Hot Isostatic Pressing
AM process / equipment
EBM / not specified EBM / not specified LS / Concept Laser M2
Ultimate tensile strength, MPa Without HIP 953 915 1155 1115*
Ultimate tensile strength, MPa With HIP 942 870 973 19
Elongation to fracture, %
Elongation to fracture, %
Without HIP 13.8 13.1 10.9 11.3*
With HIP 12.9 13.7 19.0
Source
[14] [17] [11]
LS / SLM 250 LS / SLM 250
1080 1040** 1315 1228**
1005
1.6 5** 4 8**
1089
8.3
[18]
13.8
[19]
* Stress relieved at 700°C for 1h, ** heat treated at 800°C for 2h
4.3.2. Fatigue
There are very few studies on the effect of HIP treatment on the fatigue strength for additive manufactured Ti6Al4V with rough as-built surface. In a recent study, however, Greitenmeier et al. [6] compared the fatigue behaviour of both LS and EBM material with rough as-built surface and concluded that HIP did not improve the fatigue properties. The present study confirms these findings for Kt=1 (type 1) samples and showed that the same behaviour also applies to samples with a built notch (Kt=2.5). Consequently, for fatigue analysis of LS or EBM Ti6Al4V with rough as-built surfaces, the same fatigue design data can be used for both HIP:ed and non-HIP:ed parts. Several investigations on EBM [14] and LS [11,18] Ti6Al4V material show that machined EBM/LS Ti6Al4V with HIP treatment can have similar fatigue life as conventional wrought material. The test results for machined EBM samples subjected to HIP, with Kt=1, in the present study show similar fatigue strength to conventional wrought material as illustrated by Figure 10 which correspond well with the previous findings. The fatigue strength for machined LS Kt=1 (type 1) samples subjected to HIP is, however, similar to previous test results [6], see Table 5, but considerably lower than the wrought bar investigated in this study. Microscopy investigations of HIP:ed EBM and LS samples showed neither gas pores nor LOF defects in the bulk material. Moreover, occasional LOF defects were found just below the as-built surface, for both the HIP:ed EBM and LS material, which indicates that these LOF have a direct connection to the surface since HIP would need fully enclosed voids to be effective. For the machined-and-polished samples, however, these surface connected LOF 20
would subsequently be removed by machining and could therefore not have affected the fatigue life. Günther et al. [20] suggests that for HIP:ed AM material, in which large defects like LOF and porosity have been removed, the fatigue crack initiation occurs instead at “single α-phase grains or at clusters of α-phases”. The image of α-phase crack initiation presented by Günther et al. corresponds well with the failure locations of the machined samples in the present study, in which the EBM+HIP and LS+ HIP samples had similar initiation appearances, see Figure 13. This would indicate that these initiations also would originate from α-phases making the fatigue properties of the HIP:ed and machined AM samples, in the present study, dependent on microstructure features rather than internal defects. This further enlights the complexity of AM material properties, since the microstructure features will depend on the thermal history from the production process which would dependent on, for example AM equipment, part geometry and support structure.
Figure 13. -Crack initiation for a machined EBM sample subjected to HIP.
4.4. Effect of Geometrical Notch
The fatigue strength for machined-and-polished samples with a geometrical notch (Kt=2.5, type 2), was similar for both EBM and LS material even though the corresponding smooth (Kt=1, type 1) specimens showed considerably difference in fatigue strength. The geometrical notch will guide the crack initiation to the surface, or close to the surface, of the machined
21
notch. Therefore, the similarity for the Kt=2.5 results can most likely be attributed to a volume effect, in which the probability is low that the most severe internal defect in the sample would be located in the volume close to the notch. Moreover, the fatigue behaviour for EBM and LS material with a rough as-built surface in combination with a notch geometry, built directly by the AM process, was also similar even though the LS material had much less rough surface, see Figure 4, and superior fatigue properties for Kt=1 (type 1) samples with rough as-built surface. In this case, that can also most likely be attributed to a volume effect, or an area effect, in which the crack initiation for a notched sample is guided to the highest stress concentration in the rough as-built surface at the notch rather than to the overall most severe surface stress concentration of the sample. Chan [5] used microcomputed tomography to investigate the average local stress concentration for EBM and LS materials with as-built surfaces and found Kt=3.28 for EBM and Kt=2.31 for LS material. These average stress concentrations, for EBM and LS material, are more similar than the most severe stress concentrations, Kt=9.1 (EBM) and Kt=5.4 (LS) that Chan et al. [2] found at the crack initiation locations. There is a larger probability that the local stress concentration inhered from the rough surface, in the notch, is close to the average value rather than equal to the maximum value which could explain the more similar behaviour of the notched EBM and LS material with rough as-built surfaces. The combination effect of a rough as-built surface and a geometrical notch is very important for fatigue life prediction of additive manufactured materials since few aerospace components have a simple flat geometry without corners or radii. The fatigue data presented in this report can therefore be used to produce fatigue design data that will allow more realistic fatigue life predictions for additive manufactured components.
22
5. Conclusion The major goal of this study was to investigate how the fatigue properties of additive manufactured Ti6Al4V, produced with laser sintering (LS) and electron beam melting (EBM), were affected by geometrical notches and rough as-built surfaces as well as a combination of these. Moreover, the effect of hot isostatic pressing (HIP) post heat treatment has been investigated. The findings in this study leads to the following conclusions:
The surface roughness is the single most severe factor for fatigue for additive manufactured materials with a rough as-built surface which results in a fatigue notch factor, Kf, of 2.77 for LS material and 4.26 for EBM material, which has a rougher surface.
The combined effect of a rough as-built surface and a geometrical notch (Kt=2.5) gives a fatigue notch factor, Kf, of 6.15 for LS material and 6.64 for EBM material.
The notch sensitivity, q, for EBM and LS Ti6Al4V with rough as-built surface is comparable, 0.40, even though the fatigue properties and the local stress concentrations, Kt, are very different.
LS and EBM material show similar fatigue behaviour for specimens with geometrical notches, stress concentration factor Kt=2.5, for both machined surfaces and rough asbuilt surfaces even though the corresponding un-notched samples, Kt=1, showed very different fatigue behaviour. This different in behaviour is most likely due to a volume effect in which the crack initiation site for Kt=2.5 specimens is guided to the built notch area and not to the overall most severe surface stress concentration.
23
HIP post heat treatment gives no improvement in fatigue life for additive manufactured materials with rough as-built surface since the rough surface is more severe for crack initiation than the internal defects.
Material produced with EBM was comparable in fatigue to conventional wrought bar due to the reduction of internal defects through HIP and improved surface finish through machining and polishing.
In order to improve the fatigue strength of EBM and LS materials, the severity of the micro notches in the rough surfaces need to be reduced. This could be achieved by further work either within process optimization or by surfaces post processes. Machining is not an option for complex geometries due to the lack accessibility and high cost. Therefore, other cost efficient tool-less post processes are needed to improve the surface quality before AM parts can be fully utilized in structural aircraft parts for serial production.
Acknowledgements The authors are grateful to Saab AB, the Swedish Foundation for Strategic Research and the European commission, through the Clean Sky 2 programme, for financial support for the presented study. Mid Sweden University is acknowledged for contribution of test samples.
References [1]
Frazier WE. Metal additive manufacturing: A review. J Mater Eng Perform 2014;23:1917–28.
[2]
Chan KS, Koike M, Mason RL, Okabe T. Fatigue life of titanium alloys fabricated by additive layer manufacturing techniques for dental implants. Metall Mater Trans A Phys Metall Mater Sci 2013;44:1010–22.
[3]
Rafi HK, Karthik N V., Gong H, Starr TL, Stucker BE. Microstructures and mechanical properties of Ti6Al4V parts fabricated by selective laser melting and electron beam melting. J Mater Eng Perform 2013;22:3872–83.
[4]
Uriondo A, Esperon-Miguez M, Perinpanayagam S. The present and future of additive manufacturing in the aerospace sector: A review of important aspects. Proc Inst Mech Eng Part G J Aerosp Eng 2015;229:2132–47.
24
[5]
Chan KS. Characterization and analysis of surface notches on Ti-alloy plates fabricated by additive manufacturing techniques. Surf Topogr Metrol Prop 2015;3:44006.
[6]
Greitemeier D, Palm F, Syassen F, Melz T. Fatigue performance of additive manufactured TiAl6V4 using electron and laser beam melting. Int J Fatigue 2017;94:211–7.
[7]
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– 10.
[8]
Wycisk E, Emmelmann C, Siddique S, Walther F. High Cycle Fatigue (HCF) Performance of Ti-6Al-4V Alloy Processed by Selective Laser Melting. Adv Mater Res 2013;816–817:134–9.
[9]
Wycisk E, Solbach A, Siddique S, Herzog D, Walther F, Emmelmann C. Effects of Defects in Laser Additive Manufactured Ti-6Al-4V on Fatigue Properties. Phys Procedia 2014;56:371–8.
[10]
Gong H, Rafi K, Starr T, Stucker BE. Effect of Defects on Fatigue Tests of As-Built Ti-6Al-4V Parts Fabricated by Selective Laser Melting. Proc. Solid Free. Fabr. Symp., 2012, p. 499–506.
[11]
Kasperovich G, Hausmann J. Improvement of fatigue resistance and ductility of TiAl6V4 processed by selective laser melting. J Mater Process Technol 2015;220:202–14.
[12]
Eriksson R. Thermal Barrier Coatings: Durability Assessment and Life Prediction. Dissertation No. 1527, Linköping University; 2013.
[13]
Theodore N. High Cycle Fatigue: A Mechanics of Materials Perspective. 1st ed. Elsevier Ltd; 2006.
[14]
Svensson M, Ackelid U, Ab A. Titanium Alloys Manufactured with Electron Beam Melting Mechanical and Chemical Properties. Med. Device Mater. V Proc. from Mater. Process. Med. Devices Conf. 2009, 2009, p. p 189-194.
[15]
Kranz J, Herzog D, Emmelmann C. Design guidelines for laser additive manufacturing of lightweight structures in TiAl6V4. J Laser Appl 2015;27:S14001.
[16]
Karlsson J. Optimization of Electron Beam Melting for Production of Small Components in Biocompatible Titanium Grades. Disseratation No. 1206, Uppsala University; 2015.
[17]
Facchini L, Magalini E, Robotti P, Molinari A. Microstructure and mechanical properties of Ti-6Al-4V produced by electron beam melting of pre-alloyed powders. Rapid Prototyp J 2009;15:171–8.
[18]
Leuders S, Thöne M, Riemer A, Niendorf T, Tröster T, Richard H a., et al. On the mechanical behaviour of titanium alloy TiAl6V4 manufactured by selective laser melting: Fatigue resistance and crack growth performance. Int J Fatigue 2013;48:300–7.
[19]
Leuders S, Lieneke T, Lammers S, Tröster T, Niendorf T. On the fatigue properties of metals manufactured by selective laser melting – The role of ductility. J Mater Res 2014:1–9.
[20]
Günther J, Krewerth D, Lippmann T, Leuders S, Tröster T, Weidner A, et al. Fatigue life of additively manufactured Ti–6Al–4V in the very high cycle fatigue regime. Int J Fatigue 2017;94:236–45.
25
Highlights
The rough as-built surface of AM materials is the most severe factor for fatigue The fatigue reduction of AM materials was quantified by fatigue notch factor (Kf) A combined effect with geometrical AM notches and rough surface was determined The notch sensitivity (q) is for AM material is equal to wrought material HIP does not improve fatigue properties for AM material with rough surfaces
26