Int J Adv Manuf Technol (2013) 69:1299–1309

DOI 10.1007/s00170-013-5106-7

ORIGINAL ARTICLE

A comparison of the tensile, fatigue, and fracture behavior

of Ti–6Al–4V and 15-5 PH stainless steel parts made
by selective laser melting
H. Khalid Rafi & Thomas L. Starr & Brent E. Stucker

Received: 14 February 2013 / Accepted: 28 May 2013 / Published online: 16 June 2013
# Springer-Verlag London 2013

Abstract In this work, Ti–6Al–4V and 15-5 PH steel samples 1 Introduction

were fabricated using selective laser melting (SLM) and their
tensile, fatigue, and fracture properties were analyzed and Selective laser melting (SLM) is an additive manufacturing
compared. The tensile properties were compared with respect process for building metallic parts [1]. The process is
to the build orientation. The horizontally built samples showed attracting considerable interest because of its capability to
relatively better tensile properties as compared with the verti- quickly fabricate complex metallic parts directly from a
cally built samples. Fatigue performance was studied for the CAD model. In this process, the parts are built in a layer by
vertical build orientation and compared with standard wrought layer fashion when a high-power laser beam scans over a
material data. The tensile and fatigue performance of SLM- metallic powder bed and selectively melts the powder. The
built materials were comparable to their respective standard melted powder solidifies and cools quickly forming a layer
wrought materials. Fractography was carried out for all tensile of the part to be produced. On completion of each layer,
and fatigue-tested samples. The fatigue fracture behavior of Ti– powder is added and laser scanning is repeated to build the
6Al–4V was different from 15-5 PH steel. For Ti–6Al–4V, the subsequent layers until the product is completed. The un-
fatigue crack initiation occurred deep in the subsurface whereas melted powder in the bed acts as a support as well as a path
for PH steel the fatigue crack was initiated from the surface. for heat conduction. Residual stresses built up in the part
necessitate that supports be used to attach the part to a rigid
Keywords Fatigue . Fracture . Fractography . Additive base-plate so that the part does not move in the powder bed
manufacturing . Selective laser melting . Ti–6Al–4V . 15-5 or distort significantly. Like other additive manufacturing
PH steel . Direct metal laser sintering techniques, layer-wise production enables SLM to produce
complex geometrical features that cannot be obtained using
Abbreviations conventional methods.
SLM Selective laser melting The influence of process parameters and scanning
OM Optical microscopy strategy on the development of the microstructure dur-
SEM-SE Scanning electron microscopy–secondary ing SLM of Ti–6Al–4V was studied by Thijs et al. [2].
electron The microstructure was significantly affected by the
YS Yield stress high localized heat inputs and the very short interaction
UTS Ultimate tensile strength times. The local heat transfer conditions determine the
orientation of the grains. The process conditions like
scanning velocity, hatch spacing (the distance between
H. K. Rafi : T. L. Starr : B. E. Stucker
two adjacent scan vectors), and scanning strategy also
J.B. Speed School of Engineering, University of Louisville,
Louisville, KY 40292, USA have a significant role in determining the grain orienta-
tion and grain size. Because of the very high cooling
B. E. Stucker (*) rate, the resulting microstructure was composed of an acicular
Computer Aided Engineering, Department of Industrial
martensitic phase. Previous work performed by Yadroitsev
Engineering, JB Speed School of Engineering,
University of Louisville, Louisville, KY 40292, USA et al. [3], Morgan et al. [4], and Yasa et al. [5] have provided
e-mail: [email protected] details on the influence of substrate, energy input, laser
1300 Int J Adv Manuf Technol (2013) 69:1299–1309

pulsing, and laser irradiation parameters on process stability a maximum power of 200 W was used to fabricate the parts.
and dimensional accuracy of the final product. Process parameters used for building the parts are given in
The mechanical properties of SLM fabricated parts are of Table 1. The build chamber was maintained at a pressure of
great importance when used for performance-critical appli- 35 mbar above atmosphere with a constant supply of argon
cations. For example, the high cycle fatigue experienced by gas. When the pressure exceeds 35 mbar, any over pressure
aircraft engine components can lead to premature initiation is exhausted. A parallel scan strategy with alternating scan
of fatigue cracks at small microstructural discontinuities and direction was adopted and for the successive layers the
their rapid propagation results in catastrophic failure of the scanning direction was rotated to 180°. For tensile studies,
component [6]. Though SLM gives high degrees of geomet- specimens were fabricated in both the vertical and horizontal
rical freedom, its high thermal stresses and localized rapid orientations. For fatigue studies, the specimens were fabri-
cooling during processing may cause the formation of non- cated in the vertical orientation. Ti64 specimens were heat
equilibrium phases and porosity. In addition, the “stair-case treated at 650 °C for 4 hours in an argon filled furnace for
effect” [2] and the “balling effect” [7] can also have an stress relief and cooled down in the furnace to room temper-
impact on mechanical properties. Mechanical testing of as- ature. PH1 steel samples were heat treated to 482 °C for
built Ti–6Al–7Nb parts made by SLM showed the influence precipitation hardening prior to tensile testing. Tensile test-
of build direction on tensile properties [8]. The tensile strength ing was carried out on an Instron 50 kN test machine (model:
of SLM fabricated Ti–6Al–4V was shown to be higher than 5569A) and the tensile specimens were prepared as per
the strength of hot worked parts due to the martensitic micro- ASTM E8M. The fatigue specimens were prepared as per
structure. However, the ductility was lower [9]. ASTM E466. Fatigue testing was carried out within the high
In this work, the tensile and fatigue deformation behavior cycle fatigue range using an Instron 10 kN fatigue machine
of Ti–6Al–4V (herein after referred to as Ti64) and 15-5 PH (model: Electropulse E10000). The test was performed at R=
(herein after referred to as PH1) stainless steel solid cylin- 0.1 using a 50 Hz sinusoidal wave form under load control to
drical parts made by SLM are investigated and the results are a maximum of 107 cycles. R=0.1 corresponds to a tension–
correlated with corresponding fracture behavior. tension cycle in which the minimum stress is equal to 0.1
times the maximum stress. By keeping the stress ratio R to a
lower value, the mean stress remains low and therefore high
2 Materials and methods stress amplitudes can be sustained by the material without
fracture. The fatigue test was conducted at an ambient tem-
The materials used for this study are Ti64 and PH1 stainless perature of 22 °C and a relative humidity of 45 %.
steel. Ti64 is the workhorse material for biomedical implants The tensile and fatigue fracture surfaces were examined
and aerospace components due to its high specific strength, using a FEI Nova Nano-scanning electron microscope
biocompatibility, stiffness, and good corrosion resistance. equipped with an EDS analysis system. Metallographic sam-
The alloy PH1 stainless steel is commonly used in applica- ples were prepared from the specimens for optical
tions such as aircraft components and fabricated parts in high micrography and SEM using standard metallographic sam-
pressure corrosive environments including valves, shafts, ple preparation methods. Ti64 samples were etched with
fasteners and gears [10]. Kroll’s reagent and PH1 samples were etched with modified
Specimens were fabricated from Ti64 (chemical composi- Fry’s reagent.
tion (in weight percent): 5.5 Al, 3.5 V, 0.04 N, 0.08 O, and
balance Ti) and PH1 (chemical composition (in weight per-
cent): 16.4 Cr, 4 Ni, 3.7 Cu, 1 Mn, 1.3 Si, 0.045 C, 0.17 N, 3 Results and discussion
and balance Fe) steel powders supplied by EOS. The average
particle size was measured using a Microtrac S3000 particle 3.1 Microstructures
size analyzer. For Ti64, the average particle size was 38 μm
and for PH1 powders the particle size was 43 μm. Ti64 The optical microstructure and the scanning electron micro-
powders exhibited spherical morphology whereas PH1 pow- structure (secondary electron mode) of the Ti64 sample fabri-
ders were irregular. An EOS M270 direct metal laser cated using SLM is shown in Fig. 1a, b, respectively. The
sintering machine equipped with a ytterbium fiber laser with figures show columnar grains with martensitic laths originating

Table 1 Process parameters

used to build the samples Material thickness Scan speed (mm/s) Power (W) Hatch spacing (mm) Layer (μm)

Ti64 1,250 170 0.10 30

PH1 800 195 0.10 40
Int J Adv Manuf Technol (2013) 69:1299–1309 1301

Fig. 1 a Optical micrograph of

Ti64 part built in a vertical
orientation (longitudinal cross
section; arrow indicates build
direction). b High-magnification
SEM-SE image of Ti64 part built
in a vertical orientation

from the columnar grain boundaries. The columnar grains are the melt pool boundaries undisturbed. Fine columnar grains
oriented along the build direction which is the direction of heat originate from these melt pool boundaries and bundles of
flow. The columnar grain boundaries represent the prior β columnar grains grow along the build direction, cutting
grains created upon solidification. Typically upon cooling, β across several layers due to variations in localized thermal
transforms to α at a temperature of 875 °C. However, because conditions. However, Thijis et al. [2] observed that dark
of the high cooling rate experienced during the SLM process, bands representing melt pools are visible in Ti64 as well
β is completely transformed to ά martensite rather than when more energy is applied.
transforming to α.
The optical microstructure of a PH1 sample aged at 3.2 Tensile characteristics
482 °C is shown in Fig. 2a. The microstructure consists of
martensite with some amount of retained austenite formed The tensile results for Ti64 and PH1 steels in both vertical and
between the martensitic laths. A high-magnification scan- horizontal orientations are shown in Table 2 and the corre-
ning electron microscopy–secondary electron (SEM-SE) im- sponding stress–strain curves are shown in Fig. 3. Three tests
age of the PH1 steel matrix is shown in Fig. 2b. As observed were carried out in each case and the mean value is reported.
in Ti64, the martensitic microstructure in PH1 steel is an The tensile properties for both the materials fabricated by
outcome of the very high cooling rate involved in the SLM SLM are found to be higher than that of the standard mate-
process. Presence of retained austenite in the PH1 steel rials from the ASM handbook [11, 12]. Vilaro et al. also
microstructure is the result of the lowering of the Ms temper- reported higher yield and ultimate strengths for as-fabricated
ature. Because of the localized variations in processing con- SLM Ti64 specimens [13]. The improved tensile properties
ditions and the nature of solidification the Ms temperature were attributed to the martensitic microstructure as com-
could have shifted below room temperature. The actual mech- pared to the α structure in standard Ti64 materials.
anism which lowers the Ms temperature during the SLM The ultimate tensile stress (UTS) for horizontally built Ti64
process is not fully clear and requires further investigation. samples was found to be marginally higher compared to their
Another striking difference in the microstructures be- vertically built counterpart. This variation in tensile properties
tween Ti64 and PH1 is that in Ti64 the prior β columnar grain with respect to the build orientation can be attributed to the
boundaries are more prominent and the melt pool boundaries fabrication defects and the orientation of the defects as a
are almost invisible whereas the PH1 microstructure is com- function of the loading axis [13]. While loading a vertically
prised of well-defined melt pool boundaries rather than co- built sample for tensile testing, the layers within the specimen
lumnar boundaries. This difference is due to the intermediate will be perpendicular to the tensile loading axis. Therefore any
phase transition from β→α→α′ in SLM Ti64 at 875 °C. fabrication defect present in the layer will also be perpendic-
Such a phase transition does not occur in steels, which keeps ular to the loading axis. In such condition, the defect opens up

Fig. 2 a Optical micrograph of

PH1 part built in a vertical
orientation (longitudinal cross
section; arrow indicates build
direction; short arrowheads
show melt pool boundaries). b
High-magnification SEM-SE
image of PH1 part built in a
vertical orientation
1302 Int J Adv Manuf Technol (2013) 69:1299–1309

Table 2 Tensile properties of Ti64 and PH1 stainless steels at two different orientations

Material Built orientation Yield stress (MPa; 0.2 % offset) UTS (MPa) Strain at break (%)

Ti64 Horizontal 1,195 (19.89) 1,269 (9.57) 5 (0.52)

Vertical 1,143 (38.34) 1,219 (20.15) 4.89 (0.65)
ASM Handbook (1993) 885 (cast and annealed condition) 930 (cast and annealed condition)
PH1 Horizontal 1,297 (1.01) 1,450 (2.06) 12.53 (1.07)
Vertical 1,100 (38.79) 1,467 (17) 14.92 (0.32)
ASM Handbook (1993) 1,170 (peak aged condition) 1,310 (peak aged condition)

Values in bracket shows standard deviation

at relatively lower stress levels. When a horizontally built layers become perpendicular to the tensile loading axis, in
sample is subjected to tensile testing, the layers become par- the case of vertically built samples, nucleation, and coales-
allel to the tensile loading axis. Hence any defect present in the cence of voids becomes faster at the interface causing a
layers will also be parallel to the tensile loading axis, making reduction in tensile strength.
opening of these defects more difficult when subjected to The tensile behavior of PH1samples fabricated in hori-
tensile testing. In addition, some microstructural discontinuity zontal and vertical orientations can be expected to be similar
can be expected at the interface between consecutive layers as to that observed in the case of Ti64. However, it is interesting
there will be a change in grain orientation after each layer. This to note that the yield stress (YS) for horizontally built sam-
makes the interface weaker. When the interface between ples are higher than for vertically built counterparts but the

Fig. 3 Stress–strain curves for SLM built parts obtained from tensile testing. a Ti64 parts built in a horizontal orientation. b Ti64 parts built in a
vertical orientation. c PH1 parts built in a horizontal orientation. d PH1 parts built in a vertical orientation
Int J Adv Manuf Technol (2013) 69:1299–1309 1303

UTS remains almost the same for both orientations. This is

because of the strain hardening effect occurring after the
yield point stress. Strain hardening is associated with the
martensitic microstructure and the precipitation of fine scale
copper precipitates upon ageing [14]. It was reported that
ageing 15-5 PH steels at 482 °C result in precipitation of fine
copper precipitates (about 4–6 nm) along the martensitic lath
boundaries. Once the yielding starts, the fine copper pre-
cipitates introduce maximum strain hardening in the matrix
[14–16]. The UTS of alloy Ti64 is marginally higher than the
YS (about 7 %), indicating the work hardening rate beyond
the yield point is low. In contrast, for PH1 steels the UTS is
about 15 % higher than the YS, which is an indication of its Fig. 4 Maximum stress vs. number of cycles for SLM built Ti64
work hardening behavior. samples
The tensile properties of metals are quite structure sensi-
tive. The grain orientation, grain size and texture that result can be attributed to the lamellar and acicular microstructure
from solidification of Ti64 are controlled primarily by the resulting from the SLM process, which also contributed to
thermal conditions that exist locally at the start of solidifica- the improved tensile strength. Ti64 exhibits superior high
tion while the fine scale microstructure is controlled primar- cycle fatigue life when the slip length is small [19]. The fine
ily by the post solidification cooling rate [17]. Rapid cooling lamellar structure acts as effective obstacles to dislocation
rates are typical with SLM processes and earlier studies also motion and hence reduces the dislocation slip length.
reported that Ti64 processed by SLM exhibit an acicular A comparison of fatigue performance of PH1 steels with
martensitic structure instead of the two-phase α hcp and β the MMPDS data for a standard material showed a sharp
bcc structure that would be present at equilibrium conditions drop in stress with increased life time. The endurance limit
[2, 18]. The presence of fine lamellar and acicular micro- was reduced by about 20 % compared with the MMPDS
structure can improve the tensile properties due to the fact endurance limit. This can be attributed to the differences in
that the crack path deviations and bifurcations are more surface roughness conditions. The surface roughness of the
numerous in these structures. Therefore the SLM fabricated samples corresponding to the MMPDS data was 0.2 μm Ra;
Ti64 parts show better strength as compared to the parts whereas, the samples used for the current work were CNC
fabricated by conventional methods. The properties can be turned to a surface roughness of 3 μm Ra. Otherwise peak
further improved by adopting optimized build methods that aged PH steel is expected to show better fatigue perfor-
can avoid the formation of pores and brittle non-equilibrium mance. The peak aged condition has higher yield strength
phases. The same holds true in the case of PH1 stainless steel
as well. The microstructure of PH1 steel is martensitic and
the ageing treatment results in the precipitation of fine cop-
per precipitates along the lath boundaries [14–16]. The pre-
cipitates cause the matrix to become severely strained and
effectively hinder dislocation motion. Here, the primary
strengthening mechanism is by precipitation hardening un-
like the strengthening mechanism in Ti64 which is due to the
strain field induced by the martensite and the dislocation
density inside the laths.

3.3 Fatigue performance

The fatigue performance with respect to the maximum stress

versus the number of cycles of vertically built Ti64 and PH1
stainless steels are given in Figs. 4 and 5, respectively. A
comparison with Metallic Materials Properties Development
and Standardization (MMPDS Handbook, Battelle Memori-
al Institute, Columbus, Ohio, April 2010) fatigue data show
better fatigue performance for Ti64 specimens built using Fig. 5 Maximum stress vs. number of cycles for SLM built PH1
SLM as compared with Ti64 that is cast and annealed. This stainless steel samples
1304 Int J Adv Manuf Technol (2013) 69:1299–1309

such that dislocations in the matrix are more difficult to move built samples (Fig. 7). A transgranular fracture surface with a
and thus the resistance to crack initiation is enhanced and the network of dimples observed at the central region shows a
fatigue life is increased. mixed mode of brittle and ductile failure mechanisms
(Fig. 7b). Pockets of dimples and fine microcracks can
3.4 Fractography be seen in different regions of the transgranular fracture
surface (Fig. 7c). Shallow dimple networks, an indication of
3.4.1 Tensile fracture—Ti64 ductile failure, observed in the overloading area represent
shear deformation coupled with less intense plastic deforma-
Examination of tensile fracture surfaces of both the vertical tion (Fig. 7d).
and horizontal specimens revealed the role played by micro- SEM observations on fracture surfaces revealed the pres-
structural features on strength and ductility of SLM built parts. ence of some randomly distributed secondary particles.
Representative fractographs of the tensile fracture surface of SEM-EDS (Fig. 8) analysis suggests these particles are
the Ti64 samples for the vertical orientation are shown in Fig. 6 non-equilibrium phases containing C and Ti, exhibiting
and for the horizontal orientation are shown in Fig. 7. The non-stochiometry in elemental composition. These particles
overall view of tensile fracture surfaces of vertical orientation do not seem to be coherent with the matrix and hence will not
samples display a rough and layered topography (Fig. 6a). be beneficial for the mechanical properties. These particles
Figure 6b shows the main cause of fracture, microvoid coa- can act as stress risers and may act as crack initiation sites
lescence, as the case in common structural alloys. There exists during tensile and fatigue testing.
a high degree of deformation around the secondary particles
and the space between the particles has undergone plastic 3.4.2 Tensile fracture—PH1 stainless steel
deformation. However, the entire fracture surface did not
show any indication of severe plastic deformation because The tensile fracture behavior of PH1 stainless steel, in both
of the low work hardening rate of the alloy. The voids present the vertical and horizontal orientation, is shown in Figs. 9
in some locations also acted as crack initiation sites. A crater- and 10, respectively. A wide difference can be observed in
like feature is formed due to the pull out of material surround- fracture surface appearance between the horizontal and ver-
ing the void (Fig. 6c). A population of fine microscopic voids tically built samples. This is consistent with the differences
and shallow dimples cover the overloading area, indicating a observed in the tensile testing results.
ductile rupture failure mechanism (Fig. 6d). For vertically built samples the fracture surface is charac-
The fractographs of the horizontally built samples showed terized by macro cracks and brittle cleavage fracture (Fig. 9a).
similar fracture modes to those observed in the vertically Large micropores form large dimples, indicating localized

Fig. 6 Tensile fracture surface

of Ti64 sample built in the
vertical orientation; SEM-SE
image. a Overall view of the
fracture surface. b Magnified
view of location 1 in (a). c
Magnified view of location 2 in
(a). d Magnified view of
location 3 in (a)
Int J Adv Manuf Technol (2013) 69:1299–1309 1305

Fig. 7 Tensile fracture surface

of Ti64 sample built in a
horizontal orientation; SEM-SE
image. a Overall view of
fracture surface. b Magnified
view of location 1 in (a). c
Magnified view of location 2 in
(a). d Magnified view of
location 3 in (a)

plastic deformation around the pores. The cracks observed are more frequently encountered in titanium alloys [6]. Fa-
intergranular in nature (Fig. 9b). Cleavage and quasi-cleavage tigue crack initiation depends on factors such as micro-
facets are very prominent covering the entire fracture surface structural inhomogenity, surface condition, presence of
(Fig. 9c). flaws and presence of secondary particles in the matrix.
The fracture surface of horizontally built samples is rough Depending upon the processing conditions, alloy Ti64
and exhibits a river flow pattern with ductile ridges forming has different microstructures: equiaxed, acicular, lamel-
around the smaller cleavage facets (Fig. 10b). The fracture lar and bimodal. Irrespective of the nature of the mi-
occurred by the coalescence of microcracks and micro-voids. crostructure, fatigue crack initiation can occur in the
As noted in the case of Ti64, the orientation of defects with subsurface. Ivanova et al. observed that in the high
tensile loading axis could be the reason for the observed cycle fatigue regime the preferred crack initiation sites
differences in fracture surface appearance. are within the alpha grains in an equi-axed grain struc-
ture. The crack nucleates along the slip bands within the
3.4.3 Fatigue fracture alpha grain and results in cleavage fracture [21]. For a
bimodal microstructure the crack initiates by the crack-
The fracture surfaces of the high cycle fatigue specimens ing of suitably oriented α grains and subsequent growth
were studied using SEM to identify the sites of micro-crack along the planar slip bands within interconnected α
initiation, the region of stable crack growth and the overload grains. Crack initiation occurs by cross-colony slip-band
region morphology. In general, high cycle fatigue is depen- fracture for the lamellar microstructure [6]. The microstruc-
dent on the initiation and propagation of fatigue cracks, ture of Ti64 parts made by SLM is fully lamellar because of
where crack initiation takes up much of the overall fatigue the faster cooling rates and directional solidification asso-
life [20]. ciated with the process. Oh et al. suggested that fatigue
crack initiation occurs mainly at the prior β grain boundary
Ti64 An overview of the fracture surface of fatigue tested for an electron beam welded and then annealed Ti64 sample.
Ti64 samples is shown in Fig. 11. Fatigue crack initiation, They also observed that the large micropores present in the
propagation and final rupture are shown in Fig. 11b–d, weld act as fatigue crack initiation sites. However, the
respectively. Fatigue crack initiation was found to occur micropores did not have any effect on crack propagation
at the subsurface of all the specimens tested. The cracks [20]. In titanium alloys, an increase in oxygen content can
initiated at distances ranging from ∼250 μm to 1 mm in also lead to the embrittlement of the alpha phase. Another
the subsurface. Though crack initiation at the subsurface possible reason for crack initiation from the subsurface
is not common in the fatigue of many metallic materials, it is could be the surface condition. While processing the
1306 Int J Adv Manuf Technol (2013) 69:1299–1309

Fig. 8 SEM-EDS analysis of the secondary particle and the Ti64 matrix

sample if compressive stress is introduced on the surface PH1 stainless steel The fatigue fracture surface of a PH1
unintentionally by the processing conditions, then the pro- stainless steel sample is shown in Fig. 12. The entire fracture
pensity of crack initiation can shift from the surface to the surface is free from any visible flaws at low magnification
subsurface. However, in the current study, the primary and showed a crescent shaped fracture surface which clearly
reason for crack initiation at the subsurface is likely due to distinguishes the regions of crack initiation, steady crack
the presence of a stress-raising defect (Fig. 11b). The crack growth and final rupture. The resistance to fatigue crack
propagated radially outward from the defect and was domi- initiation will be high for PH1 stainless steels heat treated
nated by a transgranular mode of cracking with some fine to 482 °C, though the resistance to fatigue crack growth is
secondary cracks across the crack propagation path (Fig. 11c). low [22]. Contrary to the fracture behavior of Ti64, crack
The transgranular failure could be due to separation at alpha initiation occurred at the surface for PH1 steels (Fig. 12b).
prime martensite plate interfaces. The fracture surface is fac- When the matrix is free from any voids or incoherent
eted with localized striations and the facets formed on the secondary particles, then the propensity for crack initia-
fracture surface are an outcome of crack deflection. The over tion will be higher at the surface. Fatigue cracks initi-
loading region examined at higher magnification exhibited ated from slip bands at the surface followed by a region
small shallow dimples, which are indicative of ductile fracture of stable crack growth. The slip bands are formed at the
because of void coalescence (Fig. 11d). surface by a mechanism of intrusion/extrusion as a result of
Int J Adv Manuf Technol (2013) 69:1299–1309 1307

Fig. 9 Tensile fracture surface

of a PH1 stainless steel sample
built in the vertical orientation;
SEM-SE image. a Overall view
of the fracture surface. b
Magnified view of location 1 in
(a). c Magnified view of
location 2 in (a)

continuous cyclic loading. Subsurface fatigue crack initiation properties. The fatigue properties of Ti64 were controlled by
was noticed in 17-4 PH stainless steels, which has similar the subsurface microstructural features whereas for PH1 steel
properties to PH1 stainless steels, when the fatigue test was the fatigue properties were controlled by the surface condi-
carried out at higher temperatures of 300 and 400 °C [23]. The tions. More studies are required in the future to understand the
stable crack growth region showed a featureless surface with fatigue crack initiation behavior of Ti64 parts made by SLM
the crack propagating through the crystallographic planes corresponding to different surface conditions and different
(Fig. 12c). The fracture surface does not contain fatigue stri- processing conditions.
ations and the fracture occurred through a shear mechanism.
The final failure was characterized by ductile rupture and also
contained some features of quasi-cleavage (Fig. 12d). 4 Summary and conclusion
In this study, the tensile and fatigue properties of Ti64 and
PH1 stainless steels fabricated by SLM were analyzed and the The following conclusions can be drawn from the current
properties were correlated with their respective fracture be- study:
havior. The orientation of defects perpendicular to the loading
axis in the case of vertically built SLM samples makes tensile 1. The tensile properties of SLM built Ti64 and PH1 steel
failure easier when compared with the horizontally built ori- samples are comparable to or even better than wrought
entation. Because of rapid cooling during SLM, the samples material. The tensile properties are affected by the build
have a fine microstructure resulting in good tensile and fatigue orientation. The horizontal build orientation is slightly

Fig. 10 Tensile fracture surface

of a PH1 stainless steel sample
built in the horizontal
orientation; SEM-SE image. a
Overall view of the fracture
surface. b Magnified view of the
fracture surface
1308 Int J Adv Manuf Technol (2013) 69:1299–1309

Fig. 11 Fatigue fracture surface

of a Ti64 sample; SEM-SE
image. a Overall view of the
fracture surface. b Magnified
view of location 1 in (a)
showing fatigue crack initiation.
c Magnified view of location 2
in (a) showing fatigue
propagation. d Magnified view
of location 3 in (a) showing
failure at overloading area

better than the vertical build orientation in terms of 4. For both the horizontal and the vertical built Ti64
tensile strength. samples, tensile failure is a mix of ductile and brittle
2. The fatigue performance of SLM fabricated Ti64 samples modes.
are better than standard Ti64 material due to the fine 5. For both horizontal and vertical built PH1 steel samples,
lamellar and acicular microstructure. tensile failure is predominantly by cleavage fracture.
3. The endurance limit of SLM fabricated PH1 steels is 6. For Ti64, fatigue crack initiation occurred at the subsur-
20 % less than standard PH1 steels, which is likely due to face whereas for PH1 steels crack initiation occurred at
the difference in surface finish. the surface.
Fig. 12 Fatigue fracture surface
of a PH1 steel sample; SEM-SE
image. a Overall view of the
fracture surface. b Magnified
view of location 1 in (a)
showing fatigue crack initiation.
The encircled region shows the
shear plane formation due to
intrusion/extrusion. c Magnified
view of location 2 in (a)
showing fatigue crack
propagation. d Magnified view
of location 3 in (a) showing
rupture at overloading area
Int J Adv Manuf Technol (2013) 69:1299–1309 1309

Acknowledgments The authors acknowledge Office of Naval Re- the correlation between fracture toughness and precipitation hard-
search (ONR), USA for support through grant N00014-09-1-0147. ening heat treatments in 15-5PH stainless steels. Eng Fail Anal
We thank Chris Scherzer and Gary Graf of the Rapid Prototyping 14:626–631
Center, University of Louisville for providing material testing support. 11. ASM Handbook Vol. 1, ASM International, Materials Park, Ohio,
p. 2071
12. ASM Handbook Vol. 2, ASM International, Materials Park, Ohio,
p. 1894
References 13. Vilaro T, Colin C, Bartout JD (2011) As-fabricated and heat treated
microstructures of the Ti–6Al–4V alloy processed by selective laser
melting. Metall Mater Trans A 42:3190
1. Abe F, Osakada K, Shiomi M, Uematsu K, Matsumoto M (2001) 14. Habibi Bajguirani HR (2002) The effect of ageing upon the micro-
The manufacturing of hard tools from metallic powders by selective structure and mechanical properties of type 15-5PH stainless steel.
laser melting. J Mater Process Tech 111:210–213 Mat Sci Eng A 338:142–159
2. Thijs L, Verhaeghe F, Craeghs T, Humbeeck JV, Kruth JP (2010) A 15. Ozbaysal K, Inal OT (1994) Age-hardening kinetics and micro-
study of the microstructural evolution during selective laser melting structure of PH15-5 stainless steel after laser melting and solution
of Ti–6Al–4V. Acta Mater 58:3303–3312 treating. J Mater Sci 29:1472–1480
3. Yadroitsev I, Bertrand P, Smurov I (2007) Parametric analysis of 16. Anil kumar V, Karthikeyan MK, Gupta RK, Ginoprakash,
the laser melting process. Appl Surf Sci 253(19):8064–8069 Ramkumar P (2012) Aging behavior in 15-5 PH precipitation
4. Morgan R, Sutcliffe CJ, O’Neill W (2004) Density analysis of hardening martensitic stainless steel. Mater Sci forum 710:483–488
direct metal laser remelted 316L stainless steel cubic primitives. J 17. Kobryn PA, Semiatin SL (2003) Microstructure and texture evolu-
Mater Sci 39(4):1195–1205 tion during solidification processing of Ti–6Al–4V. J Mater Process
5. Yasa E, Deckers J, Kruth JP (2011) The investigation of the influ- Tech 135:330–339
ence of laser re-melting on density, surface quality and microstruc- 18. Yasa E, Deckers J, Kruth JP, Rombouts M, Luyten J (2010) Charpy
ture of selective laser melting parts. Rapid prototyping J 17(5):312– impact testing of metallic selective laser melting parts. Virtual and
327 Physical Prototyping 5(2):89–98
6. Nalla RK, Boyce BL, Campbell JP, Peters JO, Ritchie RO (2002) 19. Chen W, Boehlert CJ, Howe JY, Payzant EA (2011) Elevated-
Influence of microstructure on high-cycle fatigue of Ti–6Al–4V: temperature mechanical behavior of as-cast and wrought Ti–6Al–
bimodal vs. lamellar structures. Metall Mater Trans A 33A:899– 4V–1B. Metall Mater Trans A 42A:3046
918 20. Oh J, Kim NJ, Lee S, Lee W (2003) Correlation of fatigue proper-
7. Dongdong G, Shen Y (2009) Balling phenomena in direct laser ties and microstructure in investment cast Ti–6Al–4V. Mat Sci Eng
sintering of stainless steel powder: metallurgical mechanisms and A 340:232–242
control methods. Mater Des 30:2903–2910 21. Ivanova SG, Biederman RR, Sisson RD Jr (2002) Investigation of
8. Chlebus E, Kuznicka B, Kurzynowski T, Dybala B (2011) Micro- fatigue crack initiation in Ti–6Al–4V during tensile-tensile fatigue.
structure and mechanical behavior of Ti–6Al–7Nb alloy produced J Mater Eng Perform 11(2):226–231
by selective laser melting. Mater Charact 62:488–495 22. Ando k, Fuseya A, Tubota M, Chu M (1999) A fracture mechanical
9. Facchini L, Magalini E, Robotti P, Molinari A, Hoges S, behavior of aged 15Cr–5Ni stainless steel. Transaction of the 15th
Wissenbach K (2010) Ductility of a Ti–6Al–4V alloy produced by international conference on structural mechanics in reactor technol-
selective laser melting of pre-alloyed powders. Rapid Prototyping J ogy, Seoul, Korea, 15–20 August
16(6):450–459 23. Wu JH, Lin CK (2002) Tensile and fatigue properties of 17-4 PH
10. Abdelshehid M, Mahmodieh K, Mori K, Chen L, Stoyanov P, stainless steel at high temperatures. Metall Mater Trans A 33A:1715–
Davlantes D, Foyos J, Ogren J, Clark R Jr, Es-Said OS (2007) On 1724