Materials and Design 98 (2016) 344–357

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Selective laser melting (SLM) of AlSi12Mg lattice structures

Martin Leary a,⁎, Maciej Mazur a, Joe Elambasseril a, Matthew McMillan a, Thomas Chirent b, Yingying Sun a,
Ma Qian a, Mark Easton a, Milan Brandt a
a
RMIT Centre for Additive Manufacture, RMIT University, Melbourne, Australia
b
Institut Français de Mécanique Avancée, Clermont-Ferrand, France

a r t i c l e i n f o a b s t r a c t

Article history: Additive manufacture (AM) enables innovative structural design, including the fabrication of complex lattice
Received 26 October 2015 structures with unique engineering characteristics. In particular, selective laser melting (SLM) is an AM process
Received in revised form 26 February 2016 that enables the manufacture of space filling lattice structures with exceptional load bearing efficiency and
Accepted 29 February 2016
customisable stiffness. However, to commercialise SLM lattice structures it is necessary to formally define the
Available online 6 March 2016
manufacturability of candidate lattice geometries, and characterise the associated mechanical response including
Keywords:
compressive strength and stiffness. This work provides an experimental investigation of the mechanical proper-
Additive manufacture ties of SLM AlSi12Mg lattice structures for optimised process parameters as well as the manufacturability of lat-
Lattice structures tice strut elements for a series of build inclinations and strut diameters. Based on identified manufacturability
Selective laser melting limits, lattice topologies of engineering relevance were fabricated, including both stretch-dominated and
Aluminium bending-dominated structures. Manufactured lattice morphology and surface roughness was quantified, as
AlSi12Mg were the associated mechanical properties and deformation and failure behaviour.
Optimisation © 2016 Elsevier Ltd. All rights reserved.
Design for additive manufacture

1. Additive manufacture and selective laser melting
Additive manufacturing (AM) refers to the process of fabricating
near-net components from unit materials [1] in a layerwise manner.
AM differs fundamentally from “subtractive manufacturing methodolo-
gies, such as traditional machining” [2], and has been labelled as a
disruptive technology [3]. The disruptive nature of AM enables a series
of technical and economic advantages, including: enhanced cost-
competitiveness for low volume production [4]; reduced environmental
impact of manufacture [5]; and, increased design complexity [3,6].
AM is based on a digital representation of product geometry. This
representation typically consists of tessellated triangular facets, known
as a Stereolithography (STL) file [7,8]. However, enhanced data struc-
tures are also proposed that are more compatible with specific AM re-
quirements, such as STL 2.0 [9], Additive Manufacturing Format (AMF)
[10] and Bezier STL [11].
Selective laser melting (SLM) is a powder-bed AM technology that
uses a scanning laser to sequentially melt layers of powdered metal
under an inert atmosphere [12]. The significant advantage of SLM tech-
nology includes high flexibility and achievable component complexity,
enabling the fabrication of highly complex lattice structures that are
not otherwise manufacturable [13]. Such lattice structures enable struc-
tural optimisation [14], while reducing associated manufacture time
and materials costs by minimising component volume [15].
⁎ Corresponding author.
E-mail address: martin.leary@rmit.edu.au (M. Leary).
SLM material research is typically focused on titanium alloys due to
their high corrosion resistance, high specific strength [16–22] and bio-
compatibility [23]. These favourable properties have enabled a range
of technical SLM titanium applications, including: customised load bear-
ing medical implants [24,25], lattice structures with mechanical proper-
ties optimised for compatibility with bone [26], high-value aerospace
[27]; space [28], and automotive applications [29].
Although aluminium alloys are not biocompatible, they exhibit a
number of properties that make them eminently suitable for commer-
cial AM applications, including:
• Aluminium (2.7 g/cm3) has significantly lower density than titanium
(4.5 g/cm3), thereby enabling commercial opportunities in scenarios
that benefit from mass reduction. The mass reduction advantage of
AM aluminium lattice structures is especially relevant for non-
stationary scenarios such as automotive and aerospace where the
fuel consumption associated with component mass is significant [30].
• The high strength-to-weight ratio of aluminium enables optimisation
for structural applications, especially for structures subject to bending
loads such as beam and plate applications [31]. This outcome is partic-
ularly relevant to the optimisation of AM lattice structures, which are
subject to significant bending, either during elastic loading (in the
case of under-stiff structures) or plastic collapse (in the case of just-
stiff and over stiff structures) (Section 3).
• The fatigue-strength of aluminium, in combination with low material
density, results in optimal material selection for weight limited design
where the number of loading cycles is below, or close to the

http://dx.doi.org/10.1016/j.matdes.2016.02.127
0264-1275/© 2016 Elsevier Ltd. All rights reserved.
 M. Leary et al. / Materials and Design 98 (2016) 344–357 345

• Importantly, aluminium alloys often provide a lower cost to achieve a

Nomenclature
specified load bearing function in comparison to titanium [30], and
therefore provide an opportunity for the economical commercialisation
Term Definition
of AM technologies, especially in non-stationary applications.
AM Additive manufacture
CT Computed tomography
BCC Body centred cubic unit cell Despite the commercial opportunities associated with aluminium
BCCZ Body centred cubic unit cell with (vertical) Z-struts lattice structural design in SLM, there appears to be sparse available re-
DOE Design of experiments search in comparison with titanium. This lack of research may be in part
FBCCZ Face and body centred cubic unit cell with (vertical) Z- due technical challenges in SLM of aluminium powders, including poor
struts powder flowability, high laser reflectivity and oxidation of aluminium
FCC Face centred cubic unit cell during melting [40].
FCCZ Face centred cubic unit cell with (vertical) Z-struts
FEA Finite element analysis 2. SLM processing and properties
SLM Selective laser melting
SLM is a highly complex multiphysics process, due to: large thermal
Symbols gradients [41]; complex three-dimensional geometries [42]; local
Term Definition overheating [43,44]; local heat transfer paths that are temporally and spa-
Dv Volume weighted percentile particle diameter tially transient [45,46]; and, thermal powder bed resistance that is poorly
D [3,2] Surface area moment mean (Sauter Mean Diameter) understood and subject to significant experimental uncertainty [47,48].
D [4,3] Volume moment mean (De Brouckere Mean Diameter) SLM is highly dimensional, with nearly 130 variables of influence on
E Lattice specimen modulus [MPa] final component quality [49,50].
Es Solid material modulus [MPa] Processing parameters that determine the laser energy density
h Hatch spacing [μm] (Eq. (1)) are: laser power, P, laser scanning speed, v, hatch spacing, h,
j Roughness sample index and the layer thickness, t.
kR Relative stiffness
M Maxwell number Where :

n Number of nodes ϒ : energy density J=mm3
N Number of roughness measures per sample, number of P Ρ : laser power ½W 
Υ¼ : ð1Þ
test specimens h  v  t ν : scan speed ½mm=s
ρ Lattice specimen density [kg/m3] t : layer thickness ½mm
ρc Lattice unit cell density [kg/m3] h : hatch spacing ½mm
ρs Lattice solid material density [kg/m3]
P Laser power [W] 2.1. SLM processing of aluminium alloys
Ra The average surface roughness [µm]
Rt The distance between the maximum peak and mini- Aluminium alloys provide significant technical and economic oppor-
mum valley [µm] tunities due to their favourable strength-to-weight and cost-to-weight
Rz The peak to peak roughness [µm] attributes. Aluminium-Silicon alloys are strong candidates for SLM due
s Number of struts to their robust welding characteristics and small difference between
σ Lattice specimen compressive strength [MPa] liquidus and solidus temperatures [51]. However, in-comparison with
σs Solid material compressive strength [MPa] titanium alloys, very little research exists on the processing of alumini-
t Layer thickness [mm] um by SLM [52]. A review of the available literature has identified the
v Scan speed [mm/s] following research of interest:
Υ Laser energy density [J/mm3]
• Kempen et al. [51] found that for AlSi10Mg (mean particle size: 16 and
W Volumetric energy absorption [MJ/m3]
48 μm), the manufactured part quality varies with powder morpholo-
gy, size and chemical composition.
endurance limit [30]. The robust design of fatigue-optimised AM lat- • Olakanmi et al. [53] assessed the effects of SLM process parameters on
tice structures is an active research domain [32,33] that will enable the degree and orientation of porosity in Al‐12Si (Particle size distri-
commercially relevant innovation in the design of fatigue-limited alu- bution: 45‐75 μm).
minium lattice structures. • Dadbakhsh and Hao [54] investigated the manufacturability of nu-
• The high thermal diffusivity of aluminium alloys combined with the merous aluminium-based powders, including commercially pure alu-
capability of AM for complex part geometry enables the fabrication minium, AlSi10Mg, Al Mg1SiCu, mixed with 15 wt% Fe2O3, and found
of high-efficiency thermal devices, including heat sinks and heat ex- AlSi10Mg to have the highest density due to low thermal conductivity
changers [34,35]. and oxide layer breakdown; a nano-scale dendritic structure was ob-
• Aluminium has excellent resistance to corrosion [36], enabling served leading to high hardness.
engineered aluminium lattice structures to be applied in corrosive en- • Dadbakhsh and Hao [55] investigated SLM of numerous aluminium-
vironments, for example in [37]: architectural and marine applica- based powders including: Al‐5FeO3, Al‐10FeO3, Al‐15FeO3, and found
tions; faying surfaces in automotive and aerospace applications, and that Hot Isostatic Pressing (HIP) increased density but did not eliminate
in the design of heat-exchangers in direct contact with corrosive all porosity due to the formation of oxide bands.
fluids. • Buchbinder et al. [56] found that for AlSi10Mg, high laser power and
• Aluminium has relatively high electrical conductivity [38], enabling its scan velocity enabled a build rate of 21mm3/s with 99.5% density
application in electrically conductive devices. The combined electrical, (420 MPa and 145 HV).
structural and lightweight properties of aluminium enable the fabri- • Brandl et al. [57] found that for AlSi10Mg with a particle size distri-
cation of high-value components that combine both electrical and bution of 25–40 μm, post manufacture heat treatment has the
structural function, for example in self-supporting lightweight electri- greatest effect on fatigue limit whilst orientation has a relatively
cal conductors [39]. smaller effect.
346 M. Leary et al. / Materials and Design 98 (2016) 344–357
Table 1
Chemical composition of AlSi12Mg powder and solid material using process parameters of Table 2.
Sample (wt.%) Al Si Cu Fe Mg
Powder Bal 11.7 b0.01 0.15 0.39
Solid Bal 11.6 b0.01 0.15 0.37
• Kempen et al. [58] identified SLM process parameters for AlSi10Mg
that obtain mechanical properties comparable to cast materials:
E = 68 GPa, UTS = 396 MPa, HV = 127.
• Ghosh and Saha [59] assessed crack density and wear performance
for SLM processed aluminium Metal Matrix Composites (MMCs) to
fabricate wear resistant components.
• Wong et al. [60] showed that aluminium alloy 6061 and stainless
steel grade 316 L are viable materials for the fabrication of SLM
processed heat sinks.
• Ameli et al. [61] demonstrated the application of alloy 6061 for fab-
rication of porous wick structures.
• Li et al. [62] investigated the mechanical properties of SLM proc-
essed Al\\12Si, and report as-manufactured yield and tensile
strengths of approximately 225 MPa and 360 MPa, with ductility
of around 5%. This work also reports on the effect of solution heat
treatment on the trade-off between strength and ductility.
The limited research data available for SLM processing of aluminium
may be in part due to technical challenges of processing aluminium in
comparison to titanium or steel powders [40,52,63]. Technical challenges
associated with SLM processing include robust powder deposition and
achieving high fused density.
SLM process repeatability requires that the powder layer be consis-
tently deposited; this is technically challenging with aluminium due to
low density and poor flowability, especially in the presence of humidity
or moisture [52]. This challenge is an active research area and can be ad-
dressed by appropriate machine design and powder specification.
There exist several distinct technical challenges to the dense fusion of
aluminium in the SLM process, including challenges associated with
power absorption, heat diffusion and oxidisation effects. An aluminium
powder layer has relatively high absorptivity due to internal reflections
within the powder layer [64]; however, fused aluminium has relatively
high reflectivity, resulting in a relatively low effective heating for a
given laser power. Relative to titanium-based alloys, aluminium alloys
also have high thermal diffusivity [65], thereby resulting in more rapid
heat transfer from the melt pool [66]. The combined effects of high reflec-
tivity and high thermal diffusivity thereby require commensurately
higher power to achieve powder fusion.
A particular challenge to the fusion of aluminium powder is the pro-
pensity for aluminium to oxidise [67]. Particle surface oxidisation can
occur during powder production, handling, or during SLM processing;
and presents difficulties in robust SLM processing due to: increased resis-
tance to heat diffusion from the particle surface to the core [68], reduced
wettability due to the presence of oxide films, and the potential for poros-
ity and entrapments as the oxide mixes with molten aluminium alloy
[69].
Louvis et al. [52] proposes that oxide films are the technical basis for
the relative difficulty of SLM processing aluminium and studied the ef-
fect of oxide films on the processing of Al‐12Si and 6061 aluminium
powders. With a laser power of 100 W, Louvis et al. were able to achieve
a solid-section Al‐12Si density approaching 90%. By using laser power in
Table 2
Optimal AlSi12Mg SLM processing parameters.
Laser energy Laser power, Hatch spacing,
density, Υ (J/mm3) P (W) h (mm)
40 350 0.19
Zn Cr Ni Mn Ti Sr Zr O
b0.01 b0.001 b0.005 0.053
b0.01 b0.001 b0.005 0.051
excess of 200 W, part densities approaching 100% have been achieved
with aluminium alloys including AlSi10Mg [40].
2.2. AlSi12Mg process parameter optimisation
To characterise the effect of SLM process parameters on the quality
of the SLM manufactured AlSi12Mg, a design of experiments (DOE)
was developed to study the effect of processing parameters on the po-
rosity of 4 mm cubes built with a 50 μm layer thickness.
The powder was provided by SLM Solutions GMBH and was
characterised by laser diffraction [70] to obtain a volume weighted par-
ticle size distribution and associated distribution statistics according to
ISO 9276–2:2014 [71] (Fig. 2). Powder chemical composition of powder
and solid specimens was determined using inductively coupled plasma-
atomic emission spectroscopy (ICP-AES) (Table 1).
Test cubes were manufactured with varying laser power and scan-
ning speed and cube porosity was measured with computed tomogra-
phy (CT) scanning. In summary:
• The process parameters of Table 2 achieve a material density of 99.86%
with no cracks present.
• Etched microstructures (Fig. 1) indicate the semi-elliptic structure of
solidified melt pool typical of SLM processing.
• The track boundary has been partially remelted and heat affected and
tends to be coarser than the structure produced from the melt pool.
Tensile specimens (N = 6, Specimen 3, E8M [72]) were subsequent-
ly manufactured with the following properties:
• Yield strength is 236.1 ± 5.4 MPa which is consistent with literature
reported data [56], and exceeds the requirements for die-cast material
[73], and is slightly higher than the reported strength for SLM proc-
essed Al‐12Si [62].
• Tensile strength is 434.1 ± 14.0 MPa, which exceeds data reported in
the literature for AlSi10Mg, i.e. 391 ± 6 (XY direction) to 396 ± 8 MPa
(Z direction) [74], as well as for Al‐12Si [62].
• Strain to failure is 4.6 ± 0.5%, which is comparable with data reported
for Al‐12Si [62], and significantly exceeds the minimum standard for
die-cast AlSi12Mg [73].
The identified parameters were subsequently used in the manufac-
ture of lattice structures used in this work.
3. Structural lattice
The mechanical properties of lattice structures, in particular com-
pressive strength and modulus is dependent on factors such as: cell to-
pology; structural boundary and loading conditions; as well as the
relative density (ρ/ρs) of the lattice density (ρ) to the solid material
density (ρs) [75]. Mechanical properties are particularly influenced by
whether the cell topology results in a loading response of lattice struts
Focal Scan speed, Layer thickness, Relative
offset (mm) v (mm/s) t (μm) density (%)
2 921.05 50 98.86
 M. Leary et al. / Materials and Design 98 (2016) 344–357
(b)
(a)
(b)
(a)
Fig. 1. Optical micrographs of etched AlSi12Mg specimens at two different magnifications.
The arrow (a) indicates the direction normal to the platen. Sample of course
microstructure at track boundary identified by arrows (b).
that is either stretch-dominated or bending-dominated [76]. It is possi-
ble to gain insight into the expected loading response by considering the
degree of strut connectivity and associated degrees of freedom present
in a lattice cell [77]. These can be quantified by the Maxwell number (M)
(Eq. (2)) which quantifies the expected mechanical behaviour in a lat-
tice consisting of a number of interacting struts, s, pin-jointed at a num-
ber of nodes, n, (Fig. 3) [78]:
M ¼ s−2n þ 3 For a two−dimensional truss
ð2Þ
M ¼ s−3n þ 6 For a three−dimensional truss
The combination of struts and nodes (intersections) results in one of
the following mechanical behaviours [79]:
Fig. 2. Volume weighted particle size distribution and associated distribution statistics.
347
Fig. 3. Examples of (i) under-stiff, (ii) just-stiff and (iii) over-stiff truss structures.
• An under-stiff lattice structure (M b 0) occurs when the lattice has too
few struts to be statically determinant and will act as a mechanism
unless moments are transferred at nodes, for example, Fig. 2 (i). An
under-stiff lattice shows bending-dominated behaviour and has high
compliance and relatively low strength.
• A just-stiff lattice structure (M = 0) occurs when the lattice has the
minimum number of struts required for the lattice to be statically
determinant, for example, Fig. 1 (ii). A just-stiff lattice does not re-
quire the transmission of moments at nodes to maintain structural
integrity, and strut loads can be determined analytically by refer-
ence to the equations of equilibrium. A just-stiff lattice has rela-
tively low compliance (high stiffness), and relatively high
strength. Such a structure exhibits stretch-dominated behaviour
and has high structural efficiency.
• An over-stiff lattice structure (M N 0) occurs when the lattice has
an excess of struts required to achieve static equilibrium, for exam-
ple, Fig. 1 (iii). An over-stiff structure does not require the trans-
mission of moments at nodes to maintain structural integrity, and
strut loads can only be calculated by assessing the effect of strut
stiffness. An over-stiff lattice has very low compliance (high stiff-
ness) and high strength due to the reinforcing effect of the redun-
dant strut elements [80]. Such structures exhibit stretch-
dominated behaviour [81].
In regards to the relative density of lattice cells, mechanical proper-
ties such as the relative strength σ/σs (where σ denotes the compres-
sive strength of the lattice structure and σs of the solid material) and
relative modulus E/Es typically exhibit a positive power relationship
with the relative density (ρc/ρs) [75]. Bending-dominated structures
also typically exhibit proportionally lower values of specific compres-
sive strength and modulus compared to stretch-dominated structures;
specific relationships are described in Section 6.
Fig. 4. For a cubic unit cell with body and face nodes, there are 3 possible strut inclination
angles. Blue: 0°, Green: 35.3°, Yellow: 45° Red: 90°.
348 M. Leary et al. / Materials and Design 98 (2016) 344–357
4. Lattice manufacturability
SLM is a thermo-mechanical process with associated manufacturabil-
ity limits. In particular, structures with acutely overhanging faces and geo-
metrical features with excessive heat transfer resistance may not be
manufacturable [12,49]. SLM structures that violate manufacturability
limits require a support structure in order to achieve the required build
geometry. Support structures offer significant design freedom [82]; how-
ever the use of support material can increase surface roughness, compro-
mise part functionality and is difficult to remove especially when
constrained by limited access [83], as in lattice structures. To ensure ro-
bust lattice manufacture, it is necessary to experimentally determine
the minimum achievable strut diameter and strut inclination angles that
can be manufactured without the need for support structures.
Data for Ti64 SLM lattice manufacturability is available, including:
• Successful production of 0.2 mm diameter struts is commonly report-
ed, VanBael et al. [84] were able to produce a strut diameter of 0.1 mm
and Hao and Raymond [85] state that 0.05 diameter mm struts are
possible
• A reported difference between virtual and as built geometry in terms
of geometric accuracy and surface quality. This variation is related to
the limitations of SLM including meltpool characteristics, powder
size [84,86], material type [52] and process parameters [87]. Micro-
CT and optical inspection have shown this variation and identified it
as a function of build angle [84,88].
Despite the technical opportunities associated with aluminium lat-
tice structures, the authors are aware of no formal experimental data
on aluminium lattice manufacturability.
Fig. 5. As manufactured specimens. (i) (BCC) body centred cubic (ii) (BCCZ) body centred cubic with Z struts (iii) (FCC) face centred cubic (iv) (FCCZ) face centred cubic with Z struts
(v) (FBCCZ) face and body centred cubic with Z struts (vi) Inset describes typical cell dimensions.
4.1. Experimental investigation of strut manufacturability
This work responds to this identified deficiency by experimentally
quantifying manufacturability of aluminium lattice strut elements. To
establish the feasible build angle and feature size relevant for AlSi12Mg,
SLM cantilever strut test specimens were manufactured with a broad
design of experiments (DOE) analysis of manufacturability versus
strut diameter and associated inclination. For a unit cell with equal
side length, there exist four possible strut inclination angles: 0, 35.26,
45, and 90° (Fig. 4).
Sample specimens were assessed for manufacturability at these
angles with a range of strut diameters (Table 3). This work experi-
mentally identified that horizontal struts (0° inclination) are not
manufacturable at any assessed diameter; for this case the strut
was not able to remain in contact with part and fused material was
wiped away by the recoater. The assessed strut diameters are manu-
facturable for angles trialled other than 0°, although for these scenar-
ios there is variation in geometric fidelity of the manufactured struts.
In this work the strut length was defined to be significantly larger
than the strut diameter, in order to assess worst-case manufactur-
ability; however research suggests that manufacturability can be im-
proved for low inclination struts by reducing the spanned distance
[50]. Further work is required to extend this broad manufacturability
DOE in order to fully characterise AlSi12Mg SLM manufacturability
for all geometric configurations.
5. Lattice specimen manufacture
Based on the experimental study of aluminium lattice manufactur-
ability, a series of manufacturable lattice structures with varying param-
eters (such as topology, Maxwell Number and strut alignment to the
 M. Leary et al. / Materials and Design 98 (2016) 344–357 349

Fig. 6. μCT scan data of a 2 × 2 × 2 array subset of the manufactured lattice specimen array.

loading direction) were selected for experimental testing in order to as-
sess mechanical performance under compressive loading (Table 4). Lat-
tice nomenclature is based on conventions in crystallography with
additional designation for vertical (Z) struts [89]. Specific lattice struc-
tures were selected due to the following range of expected properties:
• Body centred cubic (BCC): High compliance and low relative density
(Table 5).
• Face centred cubic (FCC): High compliance.
• Body centred cubic with vertical struts (BCCZ) and Face centred cubic
with vertical struts (FCCZ): Theoretically under-stiff but high compres-
sive strength and stiffness along z-direction struts.
• Face and body centred cubic with vertical struts (FBCCZ): Theoretically
under-stiff but high strength and stiffness along z-direction struts, with
body centred struts resisting loads in x and y-directions. Comparatively
high relative density.

Fig. 7. Representative optical microscope image of the lower and upper faces of 35.26°, 45°and 90° struts.
350 M. Leary et al. / Materials and Design 98 (2016) 344–357
Fig. 8. Summary surface roughness data (error bars indicate ±1 SD).
The selected lattice structures were successfully manufactured (Fig. 4)
with the previously identified processing parameters (Table 2). A strut di-
ameter of 1.0 mm was selected due to manufacturability constraints; 10
cells were specified with a cell size of 7.5 mm such that the overall lattice
structure size (75 mm cubed) allowed all specimens to be manufactured
concurrently on one 250 mm square build plate.
The complex transient thermal effects, layerwise geometry and sto-
chastic effects characteristic of SLM manufacture of aluminium, result in
variation between the intended CAD geometry (Table 5) and the as-
manufactured geometry. To provide insight into variation of lattice
geometry, destructive and non-destructive methods were used to qual-
itatively assess the as-manufactured lattice morphology, and the associ-
ated strut roughness.
5.1. Geometric fidelity of as-manufactured lattice geometry
Manufactured specimens were inspected using micro Computed To-
mography (μCT) to provide quantitative insight into cell morphology
Fig. 9. Uniaxial compression stress-strain responses of tested lattice specimens.
Fig. 10. Volumetric energy absorption (W) of tested lattice specimens.
based on a 2 × 2 × 2 cell array. In particular μCT analysis indicates that
(Fig. 6):
• Downward facing surfaces consistently display higher than average
surface roughness; specific analysis of strut surface roughness was
completed to quantify the effect of geometry and inclination on
strut roughness (Section 5.2)
• Visible geometric defects occur when struts intersect at nodes. These
defects appear to be asymmetric even for lattice structures with nom-
inally symmetric geometry. The effect of node geometry on structural
integrity has been assessed by computational analysis of the μCT data
(Section 6.2).
5.2. Surface roughness evaluation
The struts of cell topologies assessed in this work exhibit build an-
gles of 35.26°, 45° and 90° to the platen (Table 5). To assess the associ-
ated surface roughness, three struts were mechanically removed from
the manufactured lattice array for each build angle. An optical micro-
scope (5× magnification, 200 ms exposure time, transmission settings)
was used to capture a shadow profile as an image with a nominal sur-
face length of 1.41 mm (Fig. 7). Four angular orientations of interest
were assessed: upper and lower surfaces have surface normal oriented
 M. Leary et al. / Materials and Design 98 (2016) 344–357 351

Inspection of the surface roughness data (Fig. 8) indicates:

• Lower strut surfaces have an order of magnitude higher surface

roughness than the other sampled surfaces. The increased roughness
is due to preferential particle adhesion on downward facing surfaces
on account of limited heat transfer between the powder bed and
solid material [12,84,88].
• The average, Ra, and peak-to-peak, Rz, roughness of lower surfaces is
highest for 35.26° struts. For example (Table 6), for 35.26° struts,
Ra = 120.2 ± 39.7 μm and Rz = 456.5 ± 113.6 μm; for 45° struts,
Ra = 43.52 ± 20.56 μm and Rz = 235.5 ± 118.2 μm.
• Side and upper surfaces have average and peak-to-peak surface rough-
ness that is statistically indistinguishable.

Fig. 11. Loading response for a BCCZ specimen identifying Young's modulus as well as
average of loading and unloading moduli for 1% and 2% strain.
away from, and toward the platen, respectively; side surfaces have sur-
face normal parallel to the platen. Note the 90° struts do not exhibit
upper or lower faces and the reported result is the average for all four
measured perpendicular faces. Average roughness, Ra, and peak-to-
peak roughness, Rz, was assessed for each orientation. Two positions
were measured for 3 struts, resulting in 6 samples for every combina-
tion of angle and orientation (Table 6).
6. Experimental results
The manufactured lattice specimens (Fig. 5) were tested in quasi-
static uniaxial compression in a calibrated universal testing machine
to quantify compressive strength and modulus as well as failure
modes. A strain rate of 1 × 10−3 s−1 was used according to applicable
testing standards [91]. The strain was determined from machine cross-
head displacement and the compressive stress was calculated based on
applied compression force and the initial XY plane cross-sectional base
area of each lattice specimen. Tests were conducted on two repetitions
of each lattice specimen:

1X N • Test #1: Compression to 15% strain. The data was used to characterise
Ra ¼ jy j ð3Þ the compressive strength, and volumetric energy absorption.
N i¼1 i
• Test #2: Compression to the onset of collapse (approximately 1–2%
strain beyond peak stress). Test results were used to characterise the
where yi is the distance between the local position of the surface and Young's, 1% strain and 2% strain moduli and to obtain images of failure
the nominal surface. mode associated with collapse (Fig. 9).

Rt ¼ maxðyi Þ− minðyi Þ ð4Þ

6.1. Test results

1X
j
Rz ¼ R ð5Þ
j i¼1 ti The uniaxial compression stress-strain responses for the tested 103
cell lattice specimens (corresponding to Test #1, Section 6) are shown
where Rti is each unique value of Rt samples. in Fig. 9. Fig. 10 shows the corresponding volumetric energy absorption

Fig. 12. Mechanical properties of tested lattice specimens. (i) Absolute compressive strength and modulus (ii) specific compressive strength and specific modulus.
352 M. Leary et al. / Materials and Design 98 (2016) 344–357

Fig. 13. Failure summary of tested lattice specimens, including: initial state, 2% strain, 4% strain, prior to collapse, collapse and progressed failure.

Fig. 14. Detail of observed deformation and failure modes of tested lattice specimens at collapse.
 M. Leary et al. / Materials and Design 98 (2016) 344–357 353

Fig. 15. Relative modulus and relative compressive strength versus relative density for tested lattice cell specimens, including typical trends for bending and stretch dominated cellular
structures. (i) Relative Young's modulus (ii) relative modulus (1% strain) (ii) relative modulus (2% strain) (iv) Relative compressive strength.

(W) of the lattice specimens, defined as the energy absorption per lat- 6.2. Numerical analysis
tice specimen solid material volume (Eq. (6)). The energy absorption
is plotted up to a strain integration limit, ϵs, of 10% as well the strain cor- Numerical analysis of μCT data was completed to provide insight
responding to the compressive strength achieved by the test specimen into: the fundamental deformation and failure mechanisms; associated
(Fig. 8). mechanical property trends; and, to allow comparison with observed
experimental data.
Where :

Cell boundary conditions vary significantly depending on boundary
W : Volumetric energy absorption MJ=m3 conditions imposed on the external lattice boundary as well as cell po-
ϵs   ϵ : Strain ½m=m sition within the bulk cellular array. For example, a unit cell within a
W¼
ρ
ρs ∫ 0
σdϵ
MJ
m3
ϵs : Strain integration limit ½m=m
σ : Stress ½MPa
ρ
: Lattice specimen relative density ½%
ρs
ð6Þ

Fig. 11 shows a representative stress-strain curve obtained from a

Test #2 (Section 6) BCCZ specimen which identifies Young's Modulus,
and the average of loading and unloading moduli at 1% and 2% strain.
The response is typical of the other tested lattice specimens. Fig. 12 sum-
marises compressive strength and modulus results (including specific
values) for all tested lattice specimens; associated failure modes are
photographically documented in Figs. 13 and 14.
Fig. 15 (i–iii) compares the relative Young's, 1% and 2% strain moduli

ðE Es Þ against the theoretical relative density ðρρ Þ for the tested lattice
s

specimens. The relative compressive strength ðσ =σ s Þ is compared

against the measured relative density ðρρ Þ in Fig. 15 (iv). Expected theo-
s

retical trends are presented for bending and stretch dominated cellular Fig. 16. Relative compressive stress versus relative density for experimental (EXP) and
structures in which mechanical properties show a positive power rela- numerical (FEA) results. Numerical results include constrained (CON.) and unconstrained
tionship with the relative density [75]. (UNCON.) boundary conditions.
354 M. Leary et al. / Materials and Design 98 (2016) 344–357
bulk cellular array will be constrained by interaction with neighbouring
cells, while a unit cell on an lattice boundary may be unconstrained de-
pending on its specific location [89]. To gain insight into the feasible cell
response, the following extreme in boundary conditions were assessed
individually:
• Constrained boundary condition, whereby nodes on the lateral lat-
tice boundary are constrained to disallow lateral translation. The
constrained scenario represents lattice cells fully within bulk lat-
tice structures as well as external lattice structures that are subject
to rigid external constraints.
• Unconstrained boundary condition, whereby nodes on the lateral
lattice boundary are free to translate. The unconstrained scenario
represents lattice cells on the external lattice surface that are not
subject to external constraints.
For both cases, linear-elastic analysis was completed with solid ele-
ments loaded to a strain of 5% of total cell size, based on μCT data
(Section 5.1); VonMises stress and relative stiffness, kR, are shown
(Table 7) normalised to results for BCC, which displays the lowest stiff-
ness of all assessed topologies. Numerical and experimental data are
compared in terms of relative compressive strength and relative density
in Fig. 16.
The numerical analysis results indicate that:
• Higher stiffness is associated with lattices including vertical struts
aligned with the loading direction confirming the associated high re-
sistance to compressive loads observed in Section 6.
• Stress concentrations are present at the intersection of diagonal struts
due to the presence of small fillet radii. Increased filleting may im-
prove strength by reducing stress concentrations, particularly in lat-
tice structures without vertical struts where stress concentrations
are particularly noticeable due to pronounced bending at intersecting
strut nodes.
• Specimens with constrained boundary constraints exhibit expectedly
higher stiffness values than unconstrained specimens. The results for
constrained boundary conditions are in general closer to the experi-
mental results, suggesting the constrained boundary conditions are
more representative of experimental conditions.
• The loading response trends in Fig. 16 generally align with theoretically
expected trends for stretch and bending dominated cellular structures
in which compressive strength increases with relative density [75].
6.3. Discussion of experimental results
The following characteristics are observed for the experimental and
numerical lattice specimen test results:
• BCC specimens exhibit the lowest absolute compressive strength and
Moduli of all candidate lattice cell structures. Similarly low values are
Table 3
Manufacturability of cantilever strut inclination specimens.
Ranking indices: red: failure, green: manufacturable.
Strut diameter (mm)
Strut angle
(degrees)
0.5 1.0 2.0 3.0
0
35.3
45
90
exhibited by FCC specimens (Fig. 12). These results align with the ex-
pected high structural compliance of the BCC and FCC lattice topolo-
gies due to their: under-stiff behaviour, lack of struts aligned in
loading direction and low relative density (Table 4).
• The low compliance of BCC and FCC specimens corresponds to a stress
concentration at the strut node intersection in the associated numer-
ical results (Table 7). These specimens are not capable of absorbing
useful amounts of energy during deformation (Fig. 8). Of the highly
compliant lattice topologies, FCC specimens show notably higher spe-
cific compressive strength and modulus values compared to BCC spec-
imens, indicating superior structural efficiency, however with lower
overall ductility (Fig. 9).
• The highest overall compressive strength and observed modulus was
achieved by FBCCZ specimens (Fig. 12), attributable to their increased
strut connectivity, presence of Z-axis struts aligned with loading di-
rection and correspondingly high density. Numerical analysis indi-
cates that peak stresses occur in the Z-struts in the FBCCZ lattice;
however, it is apparent that the body, and face centred elements with-
in the FBCCZ lattice assist in supporting the Z-struts as the associated
stress concentration is lower than for either FCZ or BCZ specimens
(Table 7). Of all experimental results, FBCCZ is the only lattice struc-
ture that displays a crushing failure mode – all other specimens fail
by shearing of lattice cells (Fig. 11).
• When observing specific strength and moduli values (Fig. 11 (ii)),
FCCZ specimens show the highest specific strength and modulus rela-
tive to other specimens, despite having lower absolute values com-
pared to FBCCZ specimens (Fig. 12 (i)). This outcome indicates that
FCCZ provides superior performance for compressive load scenarios
aimed at maximizing strength- and stiffness-to-weight ratio. Similar
performance has been noted for titanium lattice structures [89].
• The ductility of the tested specimens was approximately 6–8% and all
specimens showed a large ratio of collapse stress to the plateau stress
(average stress post initial peak) where high initial strength is follow-
ed by drastic collapse after reaching maximum compressive strength
(Fig. 7).
• Young's modulus was found to be always lower than the Moduli at 1%
and 2% strain (Fig. 12) indicating that localised plasticity is occurring
in the lattice specimens at stresses below the compressive strength.
Similar observations have been made for metallic foams, for example
[75], and recently in titanium alloy SLM lattice structures [89]. This out-
come is particularly important as a number of publications relevant to
additively manufactured lattices associate the Young's modulus with
the expected in-situ lattice stiffness, for example [92–94] despite the
observation demonstrated in this work that the Young's modulus may
underestimate stiffness for repeated elastic loading of lattice structures.
• Of the investigated lattice, the volumetric energy absorption is highest
for FBCCZ specimens (Fig. 10). However, FBCCZ is also associated with
high modulus and compressive strength (Fig. 12) and if applied to ab-
sorb crushing stress, the force transferred to the protected structure
by the lattice will also be the highest of the tested specimens.
• The response of tested lattice specimens exhibits trends in: relative
Table 4
Cell topology, Maxwell number, and load-aligned struts for candidate cubic unit cell lattice
structures.
BCC BCCZ FCC FCCZ FBCCZ
Cell type
Struts, s 8 12 16 20 28
Struts, s 9 9 12 12 13
Maxwell
number, M -13 -9 -14 -10 -5
Strut aligned to
NO YES NO YES YES
load direction
 M. Leary et al. / Materials and Design 98 (2016) 344–357 355

Table 5 Table 7
Test specimen geometric parameters. Numerical analysis of candidate lattice structures. Image shows Von Mises stress and rel-
ative stiffness, kR, both normalised to data for unconstrained BCC.
Cell Unit cell Specimen Cell Strut Number
Topology theoretical relative theoretical relative Size diameter of cells Unconstrained Constrained
density, ρc [%] density, ρ [%] [mm] [mm]

BCC 8.5% 8.5% 7.5 1.0 103

BCCZ 10.8% 9.7% BCC
FCC 11.6% 7.5%
FCCZ 14.7% 8.7%
FBCCZ 20.6% 16.1%

density, relative modulus and compressive strength which are associat-

ed with the presence of vertical Z-struts aligned with the compressive kR 1.0 3.5
loading direction (Fig. 15). Lattice topologies which include vertical
struts (BCCZ, FCCZ, and FBCCZ) correspond with the expected behav- BCZ
iour of stretch dominated structures (suggesting that the topologies
are more subject to stretch dominated deformation mechanisms),
whereas lattice topologies without vertical axis struts (BCC, FCC) do
not adhere to this trend.

kR 8.1 11.1

7. Concluding remarks FBCZ

Although some researchers have reported difficulty in processing

SLM aluminium, the processing variables optimised for AlSi12Mg
allowed robust manufacture of structural lattice. Limitations of manu-
facturability were documented, including minimum strut diameter
and allowable inclination angle for the proposed processing parameters. kR 9.6 15.3
Based on these identified processing limits, a number of relevant lattice
structures were defined, including structures with low Maxwell num- FCZ
ber (intending to provide high compliance) as well as structures with
higher Maxwell number and struts aligned in the loading direction
(these lattice elements provide high energy absorption and stiffness).
Sample specimens were successfully manufactured for all identified
lattice structures - the specimens displayed ductility of approximately
6–8%, and all specimens display classic lattice structural behaviour, in- kR 8.0 10.9
cluding a large ratio of collapse stress to plateau stress (Fig. 7). Based
on this work, the manufacturability of SLM AlSi12Mg lattice structures
FC
was theoretically and experimentally assessed for varying cell topology,
including surface morphology and roughness; failure mechanisms and
energy absorption characteristics. These novel outcomes assist to enable
commercial application of SLM AlSi12Mg lattice structures.
In particular, the energy absorption capacity of the lattice structures
of interest displayed considerable variation. The associated data allows kR 1.5 6.2
the prediction of energy absorption forces of structural AlSi12Mg lattice
structures – in particular for the FBCCZ structure, which was
demonstrated to provide stable crushing behaviour and excellent ener-
gy absorption characteristics.
Table 6 Experimental observation indicates that Young's Modulus is always
Average, Ra, and peak-to-peak, Rz, roughness of strut elements (Fig. 8). SD denotes stan- lower than the Moduli at 1% and 2% strain. This effect is due to local plas-
dard deviation. Measurements in μm.
ticity not observed in bulk structures, and although typically reported
Build angle for metallic foams, does not appear to be commonly reported for AM
(degrees) structures. This observation is therefore particularly important for AM
35.26 45 90 lattice designers and researchers as the reported Young's modulus
Ra lower 120.2 43.5 –
may underestimate stiffness for actual in-service loading.
SD 39.7 20.6 – Although all tested lattices are classified as bending dominated
Ra sides 8.1 9. 2 8.2 (Table 4) according to Maxwell's stability criteria (Section 3), the results
SD 3.7 1.9 3.0 highlight that the criteria is a necessary but not sufficient condition for
Ra upper 8.8 8.9 –
predicting mechanical response as the significant effect of loading direc-
SD 2.6 1.5 –
Rz lower 456.5 235.5 – tion on cell rigidity is not directly accommodated by the Maxwell criteria.
SD 113.6 118.2 –
Rz sides 56.2 57.2 48.9
SD 22.1 15.3 11.8 References
Rz upper 53.0 60.5 –
[1] Y. Zhai, D.A. Lados, J.L. LaGoy, Additive manufacturing: making imagination the
SD 10.3 16.0 –
major limitation, JOM 66 (5) (2014) 808–816.
356 M. Leary et al. / Materials and Design 98 (2016) 344–357
[2] ASTM standard F2792-12a, F2792-12a, Standard Terminology for Additive
Manufacturing Technologies, ASTM International, West Conshohocken, Pa, 2012.
[3] I. Gibson, D.W. Rosen, B. Stucker, Additive Manufacturing Technologies, Springer,
New York, 2010.
[4] D.E. Cooper, M. Stanford, K.A. Kibble, G.J. Gibbons, Additive manufacturing for prod-
uct improvement at Red Bull Technology, Mater. Des. (41) (2012) 226–230.
[5] A. Drizo, J. Pegna, Environmental impacts of rapid prototyping: an overview of re-
search to date, Rapid Prototyp. J. 12 (2) (2006) 64–71.
[6] G. Guan, et al., Evaluation of selective laser sintering processes by optical coherence
tomography, Mater. Des. 88 (2015) 837–846.
[7] R. Gibson, Stucker, Additive Manufacturing Technologies — Rapid Prototyping to Di-
rect Digital Manufacturing2010.
[8] M. McMillan, et al., Programmatic lattice generation for additive manufacture, Proc.
Technol. 20 (2015) 178–184.
[9] J.D. Hiller, H. Lipson, STL 2.0: a Proposal for a Universal Multi-Material Additive
Manufacturing File Format, In Proceedings of the Solid Freeform Fabrication Sympo-
sium, 2009.
[10] W.E. Frazier, Metal additive manufacturing: a review, J. Mater. Eng. Perform. 23 (6)
(2014) 1917–1928.
[11] R. Paul, Modeling and Optimization of Powder Based Additive Manufacturing (AM)
Processes, University of Cincinnati, 2013.
[12] J.P. Kruth, L. Froyen, J. Van Vaerenbergh, P. Mercelis, M. Rombouts, B. Lauwers, Selec-
tive laser melting of iron-based powder, J. Mater. Process. Technol. 149 (1) (2004)
616–622.
[13] C. Chu, G. Graf, D.W. Rosen, Design for additive manufacturing of cellular structures,
Comput.-Aided Des. Applic. 5 (5) (2008) 686–696.
[14] M. Leary, L. Merli, F. Torti, M. Mazur, M. Brandt, Optimal topology for additive man-
ufacture: a method for enabling additive manufacture of support-free optimal struc-
tures, Mater. Des. 63 (2014) 678–690.
[15] S. Mellor, L. Hao, D. Zhang, Additive manufacturing: a framework for implementa-
tion, Int. J. Prod. Econ. 149 (2014) 194–201.
[16] J. Zhang, et al., Fatigue crack propagation behaviour in wire + arc additive
manufactured Ti–6Al–4V: effects of microstructure and residual stress, Mater. Des.
90 (2016) 551–561.
[17] H. Gong, et al., Influence of defects on mechanical properties of Ti–6Al–4V compo-
nents produced by selective laser melting and electron beam melting, Mater. Des.
86 (2015) 545–554.
[18] P. Li, et al., Critical assessment of the fatigue performance of additively
manufactured Ti–6Al–4V and perspective for future research, Int. J. Fatigue 85
(2016) 130–143.
[19] T. Tancogne-Dejean, et al., Probabilistic fracture of Ti–6Al–4V made through additive
layer manufacturing, Int. J. Plast. 78 (2016) 145–172.
[20] A.J. Sterling, et al., Fatigue behavior and failure mechanisms of direct laser deposited
Ti–6Al–4V, Mater. Sci. Eng. A 655 (2016) 100–112.
[21] Q.C. Liu, J. Elambasseril, S.J. Sun, M. Leary, M. Brandt, P.K. Sharp, The effect of
manufacturing defects on the fatigue behaviour of Ti–6Al–4V specimens fabricated
using selective laser melting, Adv. Mater. Res. 891 (2014) 1519–1524.
[22] Y.Y. Sun, et al., The influence of As-built surface conditions on mechanical properties
of Ti–6Al–4V additively manufactured by selective electron beam melting, JOM
(2016) 1–8.
[23] P.A. Kobryn, S.L. Semiatin, The laser additive manufacture of Ti–6Al–4V, JOM 53 (9)
(2001) 40–42.
[24] J. Imanishi, P.F. Choong, Three-dimensional printed calcaneal prosthesis following
total calcanectomy, Int. J. Surg. Case Rep. 10 (83–87) (2015).
[25] X. Wang, et al., Topological design and additive manufacturing of porous metals for
bone scaffolds and orthopaedic implants: a review, Biomaterials 83 (2016)
127–141.
[26] J. Parthasarathy, B. Starly, S. Raman, A design for the additive manufacture of func-
tionally graded porous structures with tailored mechanical properties for biomedi-
cal applications, J. Manuf. Process. 13 (2) (2011) 160–170.
[27] M. Brandt, S.S., M. Leary, S. Feih, J. Elambasseril, Q. Liu, High-Value SLM Aerospace
Components: From Design to Manufacture, Adv Mater Res, Trans Tech Publications,
Switzerland, Vol. 633 2013, pp. 135–147.
[28] J.M. Waller, B.H. Parker, K.L. Hodges, E.R. Burke, J.L. Walker, E.R. Generazio, (), Non-
destructive Evaluation of Additive Manufacturing, NASA, 2014.
[29] D. Cooper, J. Thornby, N. Blundell, R. Henrys, M.A. Williams, G. Gibbons, Design and
manufacture of high performance hollow engine valves by Additive Layer
Manufacturing, Mater. Des. 69 (2015) 44–55.
[30] M. Leary, in: R.E. Lumley (Ed.), Materials Selection and Substitution Using Alumin-
ium Alloys, in Fundamentals of Aluminium Metallurgy: Production, Processing and
Applications, Woodhead publishing, 2011.
[31] M.F. Ashby, Multi-objective optimization in material design and selection, Acta
Mater. 48 (1) (2000) 359–369.
[32] E. Masoumi Khali Abad, S. Arabnejad Khanoki, D. Pasini, Fatigue design of lattice ma-
terials via computational mechanics: application to lattices with smooth transitions
in cell geometry, Int. J. Fatigue 47 (2013) 126–136.
[33] A. Foroozmehr, M. Badrossamay, E. Foroozmehr, Finite element simulation of selec-
tive laser melting process considering optical penetration depth of laser in powder
bed, Mater. Des. 89 (2016) 255–263.
[34] K.J. Maloney, et al., Multifunctional heat exchangers derived from three-dimensional
micro-lattice structures, Int. J. Heat Mass Transf. 55 (9–10) (2012) 2486–2493.
[35] C.S. Roper, Multiobjective optimization for design of multifunctional sandwich panel
heat pipes with micro-architected truss cores, Int. J. Heat Fluid Flow 32 (1) (2011)
239–248.
[36] J.R. Davis, in: J.R. Davis, A.S.M. international (Eds.),Aluminum and Aluminum Alloys,
1993.
[37] C. Vargel, Corrosion of Aluminium, Elsevier, 2004.
[38] E.A. Avallone, T. Baumeister, A. Sadegh, Marks' Standard Handbook For Mechanical
Engineers, Mcgraw-Hill, 2006.
[39] Ishibashi, K., Kazumi Mochizuki, and Yasunobu Kondo, Electric Wire or Cable. 2013:
U.S. Patent Application 13/867,275.
[40] L. Thijs, et al., Fine-structured aluminium products with controllable texture by se-
lective laser melting of pre-alloyed AlSi10Mg powder, Acta Mater. 61 (5) (2013)
1809–1819.
[41] L. Facchini, Microstructure and Mechanical Properties of Biomedical Alloys Produced
by Rapid Manufacturing Techniques, University of Trento, 2010.
[42] J.P. Kruth, P. Mercelis, J. Van Vaerenbergh, T. Craeghs, Feedback Control of Selective
Laser Melting. In 3rd International Conference on Advanced Research in Virtual and
Rapid Prototyping, 2007 (Leiria, Portugal).
[43] E. Foroozmehr, R. Kovacevic, Effect of path planning on the laser powder deposition
process: thermal and structural evaluation, Int. J. Adv. Manuf. Technol. 51 (5–8)
(2010) 659–669.
[44] N. Karapatis, et al., Thermal Behavior of Parts Made by Direct Metal Laser Sintering, 9
Th SFF Symposium, Austin, 1998.
[45] T. Craeghs, et al., Detection of process failures in layerwise laser melting with optical
process monitoring, Phys. Procedia 39 (2012) 753–759.
[46] B.A. Szost, et al., A comparative study of additive manufacturing techniques: Resid-
ual stress and microstructural analysis of CLAD and WAAM printed Ti–6Al–4 V com-
ponents, Mater. Des. 89 (2016) 559–567.
[47] M.R. Alkahari, T. Furumoto, T. Ueda, A. Hosokawa, R. Tanaka, A. Aziz, M. Sanusi, Ther-
mal conductivity of metal powder and consolidated material fabricated via selective
laser melting, Key Eng. Mater. 523 (2012) 244–249.
[48] A.V. Gusarov, et al., Contact thermal conductivity of a powder bed in selective laser
sintering, Int. J. Heat Mass Transf. 46 (6) (2003) 1103–1109.
[49] R.O.a.E. C., Rapid Manufacturing of Lattice Structures with Selective Laser Melting,
Proc. of SPIE 6107, Laser-Based Micropackaging, 2005.
[50] F.G. Bachmann, et al., Rapid Manufacturing of Lattice Structures with Selective Laser
Melting, Lasers and Applications in Science and Engineering, 2006 (6107: p.
61070K-61070K-12).
[51] K. Kempen, L. Thijs, E. Yasa, M. Badrossamay, W. Verheecke, J.P. Kruth, Process Op-
timization and Microstructural Analysis for Selective Laser Melting of AlSi10Mg. In
Solid Freeform Fabrication Symposium, 2011.
[52] E. Louvis, P. Fox, C.J. Sutcliffe, Selective laser melting of aluminium components, J.
Mater. Process. Technol. 211 (2) (2011) 275–284.
[53] E.O. Olakanmi, R.F. Cochrane, K.W. Dalgarno, Densification mechanism and micro-
structural evolution in selective laser sintering of Al–12Si powders, J. Mater. Process.
Technol. 211 (1) (2011) 113–121.
[54] S. Dadbakhsh, L. Hao, Effect of Al alloys on selective laser melting behaviour and mi-
crostructure of in situ formed particle reinforced composites, J. Alloys Compd. 541
(2012) 328–334.
[55] S. Dadbakhsh, L. Hao, Effect of hot isostatic pressing (HIP) on Al composite parts
made from laser consolidated Al/Fe2O3 powder mixtures, J. Mater. Process. Technol.
212 (11) (2012) 2474–2483.
[56] D. Buchbinder, et al., High power selective laser melting (HP SLM) of aluminum
parts, Phys. Procedia 12 (2011) 271–278 (Part A(0)).
[57] E. Brandl, et al., Additive manufactured AlSi10Mg samples using Selective Laser
Melting (SLM): Microstructure, high cycle fatigue, and fracture behavior, Mater.
Des. 34 (2012) 159–169.
[58] K. Kempen, et al., Mechanical properties of AlSi10Mg produced by selective laser
melting, Phys. Procedia 39 (2012) 439–446.
[59] S.K. Ghosh, P. Saha, Crack and wear behavior of SiC particulate reinforced aluminium
based metal matrix composite fabricated by direct metal laser sintering process,
Mater. Des. 32 (1) (2011) 139–145.
[60] M. Wong, et al., Selective laser melting of heat transfer devices, Rapid Prototyp. J. 13
(5) (2007) 291–297.
[61] M. Ameli, et al., A novel method for manufacturing sintered aluminium heat pipes
(SAHP), Appl. Therm. Eng. 52 (2) (2013) 498–504.
[62] X.P. Li, et al., A selective laser melting and solution heat treatment refined Al–12Si
alloy with a controllable ultrafine eutectic microstructure and 25% tensile ductility,
Acta Mater. 95 (2015) 74–82.
[63] K. Bartkowiak, et al., New developments of laser processing aluminium alloys via
additive manufacturing technique, Phys. Procedia 12 (2011) 393–401.
[64] A.V. Gusarov, J.P. Kruth, Modelling of radiation transfer in metallic powders at laser
treatment, Int. J. Heat Mass Transf. 48 (16) (2005) 3423–3434.
[65] F. Incropera, D. DeWitt, Introduction to Heat Transfer, Wiley, 1985.
[66] P. Fischer, et al., A model for the interaction of near-infrared laser pulses with metal
powders in selective laser sintering, Appl. Phys. A Mater. Sci. Process. 74 (4) (2002)
467–474.
[67] A. Rai, et al., Understanding the mechanism of aluminium nanoparticle oxidation,
Combust. Theor. Model. 10 (5) (2006) 843–859.
[68] Z.A. Munir, Analytical treatment of the role of surface oxide layers in the sintering of
metals, J. Mater. Sci. 14 (11) (1979) 2733–2740.
[69] J. Campbell, Castings, Butterworth-Heinemann, 2003.
[70] Standardization, T.I.O.F., ISO 13320:2009 Particle Size Analysis – Laser Diffraction
Methods, 2009.
[71] Standardization, T.I.O.F., ISO 9276–2:2014 Representation of Results of Particle Size
Analysis – Part 2: Calculation of Average Particle Sizes/Diameters and Moments
from Particle Size Distributions, 2014.
[72] ASTM, E8/E8M - 11 standard test methods for tension testing of metallic materials.
[73] W. Hesse, Aluminium-Schlüssel, Beuth Verlag, 2012.
[74] K. Kempen, T.L., J. Van Humbeeck, J.P. Kruth, Mechanical properties of AlSi10Mg pro-
duced by selective laser melting, Phys. Procedia 39 (2012) 439–446.
 M. Leary et al. / Materials and Design 98 (2016) 344–357
[75] L.J. Gibson, M.F. Ashby, Cellular Solids: Structure and Properties, Cambridge univer-
sity press, 1997.
[76] M.F. Ashby, D. Cebon, Materials selection in mechanical design, J. Phys. IV 3 (C7)
(1993) (p. C7-1).
[77] S. Xu, et al., Design of lattice structures with controlled anisotropy, Mater. Des. 93
(2016) 443–447.
[78] V.S. Deshpande, N.A. Fleck, M.F. Ashby, Effective properties of the octet-truss lattice
material, J. Mech. Phys. Solids 49 (8) (2001) 1747–1769.
[79] V.S. Deshpande, M.F. Ashby, N.A. Fleck, Foam topology: bending versus stretching
dominated architectures, Acta Mater. 49 (06) (2001) 1035–1040.
[80] D.L. Schodek, Structures, Prentice-Hall, Englewood Cliffs, NJ, 1980.
[81] M.F. Ashby, et al., Metal Foams: a Design Guide: a Design Guide, Elsevier, 2000.
[82] M. Leary, et al., Feasible Build Orientations for Self-Supporting Fused Deposi-
tion Manufacture: a Novel Approach to Space-Filling Tesselated Geometries,
Advanced Materials Research, Trans Tech Publ., 2013.
[83] N. Vanderesse, et al., Image analysis characterization of periodic porous materials
produced by additive manufacturing, Mater. Des. 92 (2016) 767–778.
[84] S. Van Bael, et al., Micro-CT-based improvement of geometrical and mechanical con-
trollability of selective laser melted Ti6Al4V porous structures, Mater. Sci. Eng. A 528
(24) (2011) 7423–7431.
[85] L. Hao, D. Raymond, Design and Additive Manufacturing of Cellular Lattice Struc-
tures. … on Advanced Research in Virtual and …, 2011 249–254.
[86] A.B. Spierings, M. Schneider, R. Eggenberger, Comparison of density measurement
techniques for additive manufactured metallic parts, Rapid Prototyp. J. 17 (5)
(2011) 380–386.
357
[87] G. Strano, et al., Surface roughness analysis, modelling and prediction in selective
laser melting, J. Mater. Process. Technol. 213 (4) (2013) 589–597.
[88] G. Pyka, et al., Surface roughness and morphology customization of additive
manufactured open porous Ti6Al4V structures, Materials 6 (10) (2013) 4737–4757.
[89] M. Mazur, M. Leary, S. Sun, D. Shidid, M. Brandt, Deformation and Failure Behaviour
of Ti–6Al–4V Structural Lattices Processed by Selective Laser Melting (SLM)Journal
of Advanced Manufacturing Technologies 2015.
[91] ASTM Standard E9–09, E9–09 standard test method compression testing of metallic
materials at room temperature. ASTM International: West Conshohocken, Pa.
[92] J. Wieding, A. Wolf, R. Bader, Numerical optimization of open-porous bone scaffold
structures to match the elastic properties of human cortical bone, J. Mech. Behav.
Biomed. Mater. 37 (2014) 56–68.
[93] O. Cansizoglu, et al., Properties of Ti–6Al–4V non-stochastic lattice structures fabri-
cated via electron beam melting, Mater. Sci. Eng. A 492 (1) (2008) 468–474.
[94] C. Yan, et al., Evaluations of cellular lattice structures manufactured using selective
laser melting, Int. J. Mach. Tools Manuf. 62 (2012) 32–38.