Materials and Design 155 (2018) 86–98

Contents lists available at ScienceDirect

Materials and Design

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

Fabrication of polymeric lattice structures for optimum energy

absorption using Multi Jet Fusion technology
F.N. Habib ⁎, P. Iovenitti, S.H. Masood, M. Nikzad
Department of Mechanical and Product Design Engineering, Faculty of Science, Engineering and Technology, Swinburne University of Technology, Melbourne 3122, Australia

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Designing polymeric periodic lattice

structures of identical relative density
from six different unit cells
• Successful fabrication of the lattices via
Multi Jet Fusion (MJF)
• Characterization of their quasi-static en-
ergy absorption behaviour using FEA
and experimental validation
• The lattice of octagonal unit cell pro-
vides the optimum energy absorption
characteristics.
• MJF can produce polymeric lattices of
much better quality with dramatically
reduced time than traditional FDM.

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

Article history: Unlike stochastic foams, 3D printed lattice structures can be created with a tailored microstructure to achieve the
Received 24 April 2018 desired global mechanical properties. In this study, quasi-static energy absorption of six polymeric lattice struc-
Received in revised form 25 May 2018 tures of different unit cell topologies and deformation behaviours but with identical relative density are investi-
Accepted 27 May 2018
gated. For the first time, a recently developed powerful 3D printing technology, namely Multi Jet Fusion (MJF), is
Available online 29 May 2018
utilized to produce high quality and performance lattices. Nonlinear finite element analysis (FEA) is used to an-
Keywords:
alyse their compressive response and energy-absorbing characteristics. To validate the FEA results, experimental
Lattice structure compression tests are performed on two types of lattices. The results of this investigation showed that the global
Cell topology energy absorption of lattices can be improved through controlling and manipulating their micro-topology. It is
3D printing found that the lattice structures with bending-dominated deformation have low stiffness and strength but pro-
Multi Jet Fusion vide good energy absorption capability. The stretch and buckling dominated structures are shown to be stiffer
Finite element analysis and stronger but have low energy absorption performance. The findings of this research identify high performing
Energy absorption unit cell geometries that can be used for energy absorption application of lattice structures in the development of
advanced 3D printed structures that are superior to stochastic foams.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction undergo a large compressive strain at nearly a constant stress level,

thus absorbing a large amount of energy without producing a high
One of the major uses of cellular materials is in energy absorption stress on the protected object. This makes them excellent for such appli-
applications, such as packaging and protective devices. They can cations because the aim of using energy-absorbing materials is not just
to absorb a certain amount of energy, but also to keep the transmitted
⁎ Corresponding author. force to the protected object below the threshold that can cause damage
E-mail address: [email protected] (F.N. Habib). or injury [1]. Foams are the most well-known type of cellular materials

https://doi.org/10.1016/j.matdes.2018.05.059
0264-1275/© 2018 Elsevier Ltd. All rights reserved.
 F.N. Habib et al. / Materials and Design 155 (2018) 86–98 87

Table 1
Unit cell and overall specimen design and dimensions for the examined 3D structures of 15% relative density.

Structure type Lattice designation Unit cell design Cell design A 2 × 2 × 2 cells structure showing Overall cubic spacemen
parameters the way the cells joined together

D or l d CAD model Side length

(mm) (mm) (mm)

1 Circular D = 10 1.488 51.488

2 Octagonal l = 4.14 1.626 51.626

3 Strengthened Octagonal l = 4.14 1.214 51.214

4 Kelvin l = 3.54 1.496 51.496

5 Rhombicuboctahedron (RO) l = 4.14 1.134 51.134

6 Cubic l = 10 2.246 52.246

that are widely used. Polymeric and metallic foams are tailored for a showed that the mechanical properties, compressive behaviour and
specific energy absorption application by choosing the right material failure mechanism of 3D printed porous titanium alloy biomaterials
and density. However, the drawback of foams is they are stochastic were highly dependent upon the type and dimensions of the building
and have a random microstructure [2]. As the structure of these mate- unit cells. Craddock et al. [13] indicated that despite the current limita-
rials at the micro level play a big role in their global behaviour and prop- tion of additive manufacturing techniques they can be used for
erties [3,4], researchers have tried to find better alternatives for foams. manufacturing tailored lattice structures for impact absorption.
Lattice materials (structures) fabricated by 3D printing can overcome Vesenjak et al. [14] experimentally studied the influence of cell shape
this limitation to a great extent. By exploiting this advantage offered on the compressive behaviour and energy absorption of 2D polymeric
by 3D printing technology, lattice materials can be made with uniform cellular structures with quadratic and circular cells under quasi-static
and ordered microstructure, their unit cells can be manipulated and op- and dynamic uniaxial compressive loading. Hammetter et al. [15] inves-
timized to attain the desired mechanical properties for a particular ap- tigated the effect of strut slenderness, strut inclination angle, and num-
plication [2]. ber of repeat lattice layers on the mechanics of compressive
Lattice structures are commonly constructed by duplicating three- deformation including the specific strength and energy absorption of
dimensional mesoscale unit cells. They are multi-functional materials pyramidal lattice structures using analytical and numerical methods.
that can offer a range of desirable properties. They offer the potential Ozdemir et al. [5,16] experimentally and numerically studied the energy
to easily engineer specific meso-scale properties (at cell level) to pro- absorption of additively manufactured metallic lattice structures of
duce desirable macro-scale material properties for a wide variety of en- cubic, diamond and re-intrant lattice structures under quasi-static and
gineering applications [5]. dynamic loading. Ullah et al. [17] used single unit cell approach to pre-
Several researchers have investigated the influence of unit cell shape dict the energy absorption capacity of Kagome and atomic truss config-
on compressive strength of lattice structures especially for bio-based urations built through selective laser melting via experimental and
tissue engineering (TE) scaffolds [6–10]. Others have studied the effect numerical studies. Gautam et al. [18] studied the influence of build ori-
of the unit cell on the global deformation and crushing behaviour of cel- entation, truss radius and surface roughness on stiffness, strength and
lular structures. For instance, Li et al. [11] showed that the unit cell energy absorption of ABS Kagome truss single unit cells built via fused
shape design is an effective way to control the mechanical properties deposition modeling (FDM). Ha et al. [19] designed and studied an en-
of the reticulate meshes such as the elastic modulus, compressive ergy absorbing lattice constructed from multiple tetra-beam-plate unit
strength and deformation behaviour. Ashby [3] considered the distinc- cells with negative stiffness behaviour. The lattice was fabricated by se-
tion between stretch-dominated and bending-dominated structures lective laser sintering (SLS) approach and analysed both numerically
the most important concept in analysing the mechanical behaviour of and experimentally. Kaur et al. [20] investigated the deformation of
lattice materials. For a given density, the first is extremely stiff and two types of 3D printed stretch-dominated micro-lattice structures
strong, while the second is compliant and not strong, but absorbs en- made of different polymeric materials using compression testing and
ergy well during compression. Ahmadi et al. [12] experimentally FEA simulation methods. Choy et al. [21] investigated the mechanical
88 F.N. Habib et al. / Materials and Design 155 (2018) 86–98

Fig. 1. Schematic of building process in Multi Jet Fusion (MJF) system.

properties of density graded metallic lattice structures of cubic and hex- capable of absorbing maximum energy at the lowest transmitted stress
agonal unit cells fabricated by selective laser melting (SLM). Al-Saedi level.
et al. [22] studied the compressive properties and energy absorption ca- This work examines the effect of a variety of 3D unit cell topologies
pability of functionally graded aluminum alloy lattice fabricated by SLM. on the energy absorption behaviour and compressive response of the
Harris et al. [23] experimentally investigated the impact response of ad- respective 3D printed polymeric lattice structures of uniform relative
ditively manufactured stainless steel hybrid lattice materials. Winter density. The aim is the development of a high energy-absorbing lattice
et al. [24] numerically studied the dynamic energy absorption behav- with the closest behaviour to the ideal energy absorber (the highest
iour of a series of two-dimensional stainless steel cellular structures of specific energy absorption at lowest stress level). Full-scale nonlinear fi-
different architectures. Mohsenizadeh et al. [25] experimentally nite element analysis (FEA) method was used to simulate their com-
showed that it is possible to design and additively manufacture light- pressive behaviour under uniaxial quasi-static loading. Experimental
weight polymeric metamaterials for energy absorption capable of re- compressive tests are performed to validate the FEA results with lattice
covering their shape after high deformation. structures printed on a newly developed powerful 3D system, namely
While most of the published works have investigated the behaviour Multi Jet Fusion (MJF), in a new brand of polyamide 12 material namely
of specific types of individual lattice structures with specific unit cell ge- HP PA12A available on the system. A systematic approach is presented
ometry, very few studies have been made to compare the mechanical to characterise and investigate the energy absorption properties of the
response and energy absorption characteristic of 3D printed lattices of 3D lattice structures and a clear criterion is set to identify the best struc-
different types of unit cells made of identical relative density. Re- tures. The global stress-strain behaviour and the energy absorption per-
searchers have also been able to design and 3D print energy absorbing formance of the structures are related to the deformation behaviour of
lattice structures, but still more detailed studies are required to identify their building unit cells. The optimum lattice for energy absorption ap-
the idealised geometry of the unit cell, which will generate a lattice plications is identified.
 F.N. Habib et al. / Materials and Design 155 (2018) 86–98 89

Fig. 2. 3D printed prototypes of the lattice structures on HP Jet Fusion 4200 3D printing machine in polyamide 12 (HP PA12).

2. Design and material Six types of cellular structures were designed from different unit
cells. The unit cell shapes were chosen in a way that provides different
2.1. Lattices design deformation mechanisms of the cell edges, such as bending, stretching
and buckling. Each unit cell had centre line dimensions of 10 × 10
The relative density is the single most important feature of cellular × 10 mm3 and was geometrically symmetric around the three perpen-
solids [1]. It is the ratio of the apparent density of a cellular material, dicular planes intersected at its central point. Each lattice structure
ρ*, to that of the parent solid material from which the cell edges are consisted of 5 × 5 × 5 cells constructed by duplicating unit cells in the
made, ρs. It is also equal to the volume fraction of solid in the cellular three directions. The six structures were designed to have a constant
material. In this study, to investigate the compressive response and en- relative density of 15%. Table 1 shows the computer-aided design
ergy absorption performance of ordered lattice structures, this parame- (CAD) models of the six building unit cells, the way they were joined to-
ter was kept constant. gether, the 3D lattice specimens constructed from each unit cell, and
their dimensions. It can be observed from Table 1 that the unit cells
that have more edge members have their struts' diameter thinner in
order to keep the relative density constant.
The unit cell of type 1 (Table 1) is a bending dominated structure,
built from three perpendicular crossed rings (designated as Circular lat-
tice). In a previous study [26] on the cell geometry effect on in-plane en-
ergy absorption behaviour of 2D honeycomb structures, it was shown
that the octagonal honeycomb could provide the optimum energy ab-
sorption performance after the hexagonal honeycomb. As hexagonal
cell geometry was not easy to pack in three dimensions, and so an octag-
onal cell was chosen for constructing 3D lattice in this work. Thus type 2
was constructed from three crossed octagons, designated as Octagonal
lattice. Its deformation mechanism under uniaxial compressive loading
is a combination of bending and buckling of the cell edges. These two
types of lattices (type 1 and 2) have been also investigated for tissue en-
gineering scaffolds [6]. Some internal struts were added to strengthen
the Octagonal lattice and form type 3 which was designated as
Strengthen Octagonal lattice. It was shown that the vertical edges in
Fig. 3. Stress-strain curves obtained from standard tensile testing of HP PA12 coupons the loading direction, which deform by buckling during the loading,
(specimen Type I, ASTM D638-14). cause instability and undulating behaviour of stress in the plateau
90 F.N. Habib et al. / Materials and Design 155 (2018) 86–98

Table 2
Material properties of HP PA12 determined from the standard tensile test (ASTM D638-14).

(a) Physical and elastic properties

Material Elastic modulus Poisson's ratioa Densityb 0.2% offset yield strength

HP PA12 1340 (MPa) 0.33 919 kg/m3 22.8 (MPa)

(b) Post-yield true stress versus true plastic strain

Yield stress (MPa) 22.8 27.6 31.3 35.6 40.9 47.6 51.0 51.7

Plastic strain 0.000 0.007 0.014 0.025 0.045 0.085 0.129 0.150
a
From FDM Nylon12 [37].
b
Measured.

region, which is not preferred for energy absorbing applications [26]. powder material on the building platform. Then, and in one continuous
That's why the added struts in type 3 were inclined and not vertical. Lat- pass, the fusing agent is applied to the material layer where the powder
tice type 4 was built from tetrakaidekahedron cells, known as Kelvin particles are meant to fuse together, and the detailing agent is applied to
cells, which is used for representing, idealizing and modeling of stochas- modify fusing and create fine detail and smooth surfaces. This pass also
tic foams [27–29] due to its bending dominated deformation. Type 5 is combines the printing with fusing energy, and reactions between
constructed from rhombicuboctahedron (RO) cells and it has been in- agents and the material cause the material to selectively fuse together
vestigated for TE scaffold application [6,30,31]. Its deformation mecha- to form the part.
nism is stretching-dominated. Type 6 is constructed from simple cube Fig. 2 shows the printed prototypes of the investigated lattice struc-
cells which deform by buckling of the vertical struts under uniaxial tures built by HP Jet Fusion 4200 3D printing machine in HP PA12 mate-
compressive loading. rial. It was observed that the dimensions, appearance, quality and
All unit cell struts have a circular cross-section of diameter “d”. For details of the parts were very good.
unit cell of the Circular lattice type 1, “D” represents the diameter of
the three crossed rings (refer to Table 1) and for all other unit cell 3. Numerical simulation
types “l” represents the length of the equilateral members of polygons.
Choosing this wide range of different unit cells allows the energy ab- 3.1. Material characterization
sorption behaviour of lattices of different deformation mechanisms to
be compared with each other. To characterise the material properties of HP PA12, alongside with
the lattices, four tensile standard test specimens of Type I with dimen-
2.2. Fabrication of lattices using Multi Jet Fusion process sions described in the standard test method for tensile properties of
plastics, ASTM D638-14 [33], were built on HP Jet Fusion 4200 machine.
In order to investigate the printability of the studied structures, a The MTS Criterion Electromechanical Universal Test System (Model 43)
prototype of each lattice type was built using a new 3D printing technol- with 50 kN load cell and an optical extensometer was used for the ten-
ogy called Multi Jet Fusion (MJF) in a brand of polyamide 12 plastic, sile test procedure at room temperature in accordance with ASTM
namely HP PA12 available for this process. This recently developed tech- D638-14. The resulting nominal stress-strain curves of the four samples
nique can build plastic parts with high quality, functionality and dimen- under uniaxial tensile loading rate of 5 mm/min are shown in Fig. 3. The
sional accuracy much faster than the currently available 3D printing material behaviour of HP PA12 can be described as elastic, power hard-
technologies [32]. Another important advantage of this new process is ening model [34]. It has relatively low yield stress but increases remark-
the parts made by this technique are almost isotropic, and their me- ably in the plastic region. The Abaqus plasticity with isotropic hardening
chanical properties do not depend much on the building orientation. model was used to define the material's behaviour. The elastic proper-
A schematic of the building process in MJF system is shown in Fig. 1. ties determined from the tensile tests are listed in Table 2a. The plastic
In this technique, the building process begins by applying a thin layer of nominal stress-strain data up to the necking obtained from the tensile

Fig. 4. (a) FEA setup and boundary conditions of a lattice structure (b) compressing the lattice by moving the top rigid plate with a constant speed towards the fixed bottom rigid plate.
 F.N. Habib et al. / Materials and Design 155 (2018) 86–98 91

Fig. 5. Schematic compressive stress-strain behaviour of a cellular solid showing the three main regimes and the important parameters (reproduced from [37]).

tests was converted to true stress-strain data using the Eqs. (1) and (2) the parent material and it undergoes plastic deformation (yields), sec-
given in [35] and is listed in Table 2b. These values are used in the FEA to ondly, the geometrical nonlinearities due the large deformation of the
define the elastic and post-yield behaviour of HP PA12 as required by structures, and the third type is the contact nonlinearity due to the in-
the Abaqus material model [36]. teraction of struts during the crushing.
Abaqus/Explicit package [38] was employed to simulate the lattices'
σ true ¼ σ nominal ð1 þ εnominal Þ ð1Þ compressive behaviour under quasi-static uniaxial loading. The material
properties determined from the tensile test and described in Section 3.1
εtrue ¼ ln ð1 þ εnominal Þ ð2Þ were used as the material parameters input. Beam elements are the nat-
ural choice for discretisation of such lattice structures. They are compu-
tationally inexpensive and can be used to simulate models with many
3.2. Finite element analysis of the lattice structures cells in a reasonable time [39]. Hence, the beam elements of type B31
[38], a 2-node linear beam in space, were used to model the lattices.
As the crushing mechanism of the lattices is a very complex process Through convergence studies, it was found that the element global
and involves all three types of nonlinearities, it cannot be simulated size smaller than 1.125 mm could not improve the accuracy of results
with implicit solvers. The three types of nonlinearities involved in the much while significantly increasing the computational time. Thus, all
crushing of these structures include, firstly, the material nonlinearity, the lattice structures were discretised with the element size of
as the local stress in the cell struts exceeds the linear elastic limit of 1.125 mm. The compression simulation was performed using two

Fig. 6. Stress versus strain and energy absorption efficiency versus strain for (a) Circular and (c) Octagonal lattices, and their corresponding energy absorption diagrams for (b) Circular and
(d) Octagonal lattices with plateau end indicated on the diagrams, all based on FEA.
92 F.N. Habib et al. / Materials and Design 155 (2018) 86–98

Fig. 7. Stress versus strain and energy absorption efficiency versus strain for (a) Strengthened Octagonal and (c) Kelvin lattices, and their corresponding energy absorption diagrams for
(b) Strengthened Octagonal and (d) Kelvin lattices with plateau end indicated on the diagrams, all based on FEA.

rigid plates at the top and bottom of the lattice structures. All degrees of contacting surfaces during the crushing, a general contact algorithm
freedom of the bottom plate were set to zero. The top plate was allowed (Explicit) was defined for the interaction of the lattice struts with them-
to move just in the vertical direction and compressed the lattices when selves and with the top and bottom plates.
moving downwards at a constant speed towards the bottom plate, as il- The compression simulation of each structure was performed up to
lustrated in Fig. 4a and b. In order to prevent the penetration of the densification. The force and displacement data were obtained for

Fig. 8. Stress versus strain and energy absorption efficiency versus strain for (a) RO and (c) Cubic lattices, and their corresponding energy absorption diagrams for (b) RO and (d) Cubic
lattices with plateau end indicated on the diagrams, all based on FEA.
 F.N. Habib et al. / Materials and Design 155 (2018) 86–98 93

Table 3 4. Energy absorption characterization of cellular structures

Energy absorption and compressive properties of the six HP PA12 lattices derived from the
FEA stress-strain diagrams.
4.1. Densification strain
Type Lattice Plateau At the plateau end
designation stress, Plateau stress and densification strain are the most relevant proper-
Strain, Stress, Energy Efficiency,
σpl
εcd σ absorbed, E (%) ties of a cellular structure to its energy absorption characteristic [34].
(MPa)
(MPa) W The plateau of the stress-strain curve of a cellular structure ideally ter-
(J/cm3)
minates at the densification strain, εd. Theoretically, εd is equal to the
1 Circular 0.79 0.68 0.99 0.51 51 porosity (1 − ρ*/ρs) as it is the limiting strain at which all the pore
2 Octagonal 1.05 0.74 1.40 0.75 54 space has been squeezed out. However, in reality, the cell edges jam to-
3 Strengthened 0.85 0.67 0.96 0.55 57
Octagonal
gether at a smaller strain than this [1].
4 Kelvin 0.89 0.71 1.10 0.59 53 The optimum use of such structures is up to the plateau end (also re-
5 RO 0.67 0.70 0.96 0.46 47 ferred to as onset of densification), after which the stress increases sig-
6 Cubic 0.48 0.85 1.46 0.41 27 nificantly and become almost vertical at the densification strain [1],
without much increase in energy absorption.
The plateau end (onset of densification) strain and the densification
1000 evenly spaced time intervals during the loading cycle as suggested strain relate to different stages of interaction between cell edges, and
by ISO 13314:2011 [40]. hence to different points on the compressive stress-strain curve (refer
to Fig. 5). However many publications do not distinguish between
3.3. Numerical stress-strain diagrams these two critical strains [42]. It was shown that the method based on
the energy absorption efficiency diagram (as discussed in Section 4.5)
The compressive behaviour of cellular structures is commonly pre- gives unique, precise and consistent results to determine the plateau
sented as the compressive stress-strain diagram, in which values of end strain, εcd [42]. According to this method, the plateau end strain of
compressive stress, σ, are plotted as ordinates against corresponding each stress-strain curve was calculated as the strain that provides the
values of compressive strain, ε, as abscissas [41]. Here, the compressive highest energy absorption efficiency. Figs. 6–8 illustrate the plateau
stress, σ, is the global stress and defined as the compressive load carried end of each stress-strain curve determined by this method and the pla-
by the cellular solid specimen at any given moment divided by the orig- teau end strain, εcd, for all lattices is listed in Table 3. Some other param-
inal cross-sectional area of the specimen perpendicular to the loading eters are listed in the table, which will be discussed in detail in the next
direction expressed in force per unit area. The compressive strain, ε, is sections.
defined as the dimensionless ratio of reduction in the gage length at
any given moment to the gage length of the cellular solid specimen 4.2. Plateau stress
along the loading direction.
Fig. 5 shows the schematic stress-strain behaviour of a typical cellu- From Fig. 5 it can be observed that the amount of energy absorbed in
lar material under compressive loading. When an open cellular solid or the initial elastic regime is very small and most of the energy is absorbed
foam is compressed, initially the cell edges bend giving a linear elastic in the plateau regime nearly at a constant stress level. Hence, the energy
deformation. At a certain stress level, the cell edges begin collapsing absorption capability and characteristic of any cellular structure mainly
by elastic buckling, plastic yielding or brittle fracture, depending upon depend on the plateau stress magnitude and the plateau length. For a
the cell material properties. Subsequent collapsing of the cells continues given relative density, the higher, the more stable (nearly constant)
at an approximately constant stress, termed plateau stress. When all cell and the longer the stress plateau is, the better is the energy absorption
edges collapse and contact each other, the densification regime begins, performance.
and the stress increases rapidly. At the end of this final stage, the densi- For a typical cellular solid, as shown in Fig. 5, the plateau regime
fied cellular solid behaves more like the dense parent material [3]. starts from the crush (yield) strain, εy, representing the initiation of a
The compressive stress-strain diagrams of the investigated six lattice new deformation mechanism of the cell edges, and ends at a critical
structures in this work under quasi-static displacement-controlled strain, εcd, representing the end of stress plateau or onset of densifica-
loading, as obtained from the FEA method described in Section 3.2 are tion. The plateau stress is given by the equation [42],
presented in Figs. 6–8. The aforementioned three main regimes of the R εcd
compressive behaviour of such structures, namely elastic, plateau and εy σ ðεÞdε
σ pl ¼ ð3Þ
densification regimes, are apparent for each structure. Figs. 6–8 also εcd −εy
show the efficiency versus strain and absorbed energy versus peak
stress (energy absorption diagram) for all six structures that will be or simply it is the average of stress in the plateau region. For the studied
discussed in detail in later sections. lattice structures, the crush strain εy was determined as the first point

Fig. 9. Comparison of FEA efficiency parameter at the plateau end between the MJF HP PA12 (a) Octagonal and (b) Kelvin lattices.
94 F.N. Habib et al. / Materials and Design 155 (2018) 86–98

Fig. 10. Schematic (a) stress-strain, (b) energy absorption (energy absorbed vs peak stress) and (c) efficiency vs strain diagrams of an ideal energy absorber.

on the stress-strain diagram at which an increase in strain occurs with- The energy absorption diagram for the studied lattice structures is
out an increase in stress if that occurred before the strain of 0.1. Other- obtained by plotting the energy absorbed, W, against the peak produced
wise, as in Circular and Kelvin lattices, the εy was taken as the strain of stress, σp, up to each of 1000 strain intervals. The energy absorption di-
0.1 [41]. And, the efficiency diagram (Section 4.5) was used to deter- agrams of the studied lattice structures are shown in Figs. 6–8. It can be
mine the plateau end strain, εcd. The calculated plateau stress for all observed that the plateau end, which was calculated as the point that
studied lattice structures is listed in Table 3. Among all structures, the provides the highest energy absorption efficiency, represents the shoul-
Octagonal lattice provides the highest plateau stress of 1.05 MPa. der point on the energy absorption diagram precisely, which is consis-
tent with the Maiti et al. [45] and Gibson and Ashby's [1] explanation.
4.3. Energy absorption
4.5. Energy absorption efficiency
The energy absorbed per unit volume, W, by a cellular structure
under compressive loading up to a strain ε is represented by the area The efficiency parameter is the ratio of the energy absorbed by a ma-
under the compressive stress-strain curve up to the strain ε and given terial or structure up to a strain ε to the energy absorbed by an ideal en-
by [43]. ergy absorber when both produce the same peak stress [46].
The efficiency parameter at the plateau end for the Octagonal and
Z ε Kelvin lattices is illustrated in Fig. 9. It is equal to the area under each
W¼ σ ðε Þdε ð4Þ curve, W, up to the plateau end divided by the area under their corre-
0
sponding ideal energy absorber (the green rectangular) which has a
height of magnitude equal to the stress at the plateau end and a length
The cumulative area under each curve up to each 1000 strain inter- of 1 (100% strain).
vals was calculated in Microsoft Excel using a numerical method, The efficiency parameter, E, at a strain, ε is given by [46]:
namely trapezoidal rule [44]. The energy absorption of each lattice at Rε
the end of plateau region, which is the optimum use of each structure σ ðε Þdε
E¼ 0
ð5Þ
[1] is listed in Table 3. From the table, it can be seen that the Octagonal σp  1
lattice is capable of absorbing the highest energy of 0.75 J/cm3.
where σp is the peak produced stress up to a given strain, ε.
4.4. Energy absorption diagram In other words, the efficiency parameter, E, at any strain, ε, is simply
equal to the ratio of the energy absorbed to the peak produced stress σp
The energy absorption diagram [45] is a widely used engineering de- up to the strain ε.
sign tool for characterizing the energy absorption behaviour of mate- Herein the ideal energy absorber needs to be defined. Such a struc-
rials and structures. It presents the relation between the two most ture or material is the one which provides a rectangular stress-strain
important parameters for protective devices and packaging materials. curve under compressive loading [34] as shown in Fig. 10a. An ideal
It describes the amount of energy absorbed per unit volume, W, by a cel- structure does not deform initially (shows a rigid behaviour with no ini-
lular material up to a strain ε as a function of the corresponding peak tial elastic region) until the stress reaches the compressive strength of
produced stress, σp, on the protected object. From the diagram, it can the energy absorber where the cells begin collapsing at a constant stress
be determined whether an energy absorber is capable of absorbing a (plateau stress) up to 100% strain. It has the longest possible stress pla-
particular amount of energy within the limited stress threshold to teau, from the strain interval 0 to 1 without densification. In reality,
keep the protected object safe. there are foams made with a relative density as small as 0.001 [1],
which theoretically means it can undergo up to 99.9% strain. The energy
absorption diagram of an ideal energy absorber is shown in Fig. 10b.
The schematic energy absorption diagram of an ideal energy ab-
sorber is shown in Fig. 10b. It can be seen that there is no energy absorp-
tion up to a certain stress (plateau stress), then the energy absorption
goes up vertically at that constant stress. The efficiency of an ideal en-
ergy absorber (Fig. 10c) has a linear relationship with strain. It starts
from 0% at 0 strain and reaches 100% at 100% strain.
The efficiency-strain diagrams for the lattice structures are shown
on the stress-strain diagrams in Figs. 6–8 and the plateau end strain
was determined as the strain which provides the highest energy ab-
sorption efficiency as suggested by Li [42]. Among the six lattices, the
Strengthened Octagonal and Octagonal lattices provide the highest en-
ergy absorption efficiency of 57% and 54% as listed in Table 3.
Some researchers have investigated the energy absorption capability
Fig. 11. Mechanical compression test setup. of such materials using the so-called ideality parameter [26,43,46],
 F.N. Habib et al. / Materials and Design 155 (2018) 86–98 95

Fig. 12. Comparing the simulation (FEA) with the experimental (Exp) compressive stress-strain curves of HP PA12 lattices of (a) Octagonal and (b) Kelvin unit cells.

which is the ratio between the energy absorbed by a cellular solid to the It can be noticed from Fig. 12 that the FEA approach used is in good
energy absorbed by an ideal energy absorber when both are com- agreement with the experimental results. However, as reported in the
pressed to the same strain, and produce the same peak stress [46]. How- previous studies [37,48] the FEA approach predicts a higher densifica-
ever, it was found that the ideality parameter cannot provide tion strain, much closer to the theoretical prediction [37] than the ex-
satisfactory results regarding the capability of cellular structures for ab- perimental results, which causes some variance in the FEA and
sorbing energy [43]. experimental results. The experimental energy absorption efficiency
for the Octagonal lattice is found to be 48% versus 54% predicted by
the FEA method, and for the Kelvin lattice, they were 49% vs 53%
5. Experimental validation respectively.

5.1. Compression tests of lattices 5.2. MJF process vs. FDM

To further investigate the compression response of lattice structures A prototype of each lattice also made by fused deposition modeling
experimentally and validate the FEA compression results, two types of (FDM) process. Fortus 450mc FDM by Stratasys, which is one of the
lattices (Octagonal and Kelvin) with the highest energy absorption pre- most powerful FDM systems currently available on the market, was
dicted by FEA method (Table 3) were selected for compression tests. used for printing the lattices. Careful attention was paid to choose ap-
Three specimens of each of these two lattices were tested under propriate process parameters to make sure the lattices would be printed
quasi-static uniaxial displacement-controlled loading. The same afore- properly. The lattices are constructed from cylindrical struts of small di-
mentioned universal testing machine (MTS Criterion Model 43) was ameters ranging from 1.134 to 2.246 mm (refer to Table 1). To be able to
used for the compression tests. The tests were performed at room tem- print the cross-section of these cylindrical struts appropriately the
perature by placing a specimen on the lower fixed platform. The cross- smallest tip size (T12) available for Nylon12 (with the layer thickness
head was moving downwards at a constant speed of 5 mm/min in of 0.1778 mm), and the smallest raster width of 0.254 mm were used.
accordance with the standard test method for compressive properties An Octagonal lattice sample made by this process (FDM) is compared
of rigid cellular plastics ASTM 1621-16 [47]. Fig. 11 shows the compres- to the same lattice made by MJF process in Fig. 13 and optical micro-
sion test setup of an Octagonal lattice sample. The force-displacement graphs of a sample of each 3D printing process are shown in Fig. 14. It
data is recorded during the test on a connected computer. This data can be observed that the MJF part's appearance and quality are much
was converted to stress-strain data in the same way as explained in better compared to the FDM sample. The time needed to print 4 lattices
Section 3.3. The experimental stress-strain curves of three samples of on FDM was 93 h and 16 min while for 12 lattices it took 3 h and 39 min.
both Octagonal and Kelvin lattices are compared to their FEA results in Which means the MJF required 0.013 fraction of the time required for
Fig. 12. the FDM process. The quasi-static compressive stress-strain curves of

Fig. 13. An Octagonal lattice sample built by HP Jet Fusion 4200 (MJF) compared to a sample built by Fortus 450mc (FDM).
96 F.N. Habib et al. / Materials and Design 155 (2018) 86–98

Fig. 14. Optical micrographs of (a) MJF sample and (b) FDM sample.

MJF vs. FDM Octagonal lattice samples are compared in Fig. 15. It can be minimum energy absorption of 0.46 J/cm3 and 0.41 J/cm3 and minimum
seen that the MJF sample provides much higher modulus, crush efficiency of 47% and 27% respectively.
strength and energy absorption than the FDM sample. It was also ob- It was observed that when keeping the relative density constant,
served that the FDM samples tear apart after the compression tests adding extra inclined struts to strengthen the Octagonal lattice of type
whereas the MJF samples showed no signs of rupture in the lattices 2, caused an increase in efficiency from 54% to 57%. This is due to the
struts after being fully compressed, and they even had good recovery more stable stress plateau of the Strengthened Octagonal compared to
behaviour, as shown in Fig. 16. the Octagonal lattice. However, the energy absorption dropped signifi-
cantly from 0.75 to 0.55 J/cm3, caused by a shorter plateau.
6. Discussion of results The nonlinear FEA method used in this study provided predictions
which were in good agreement with the experimental test results ob-
From the comparison of MJF with the FDM lattices and their com- tained from two types of lattices built by MJF. However, the FEA results
pression behaviour it was observed that the newly developed MJF pro- predicted a slightly higher densification strain, and thus longer stress
cess, could produce the lattice structures with much higher quality and plateau, which resulted in overestimating energy absorption and effi-
functionality in 0.013 fraction of time needed for FDM process. ciency. The methodology adopted in this study to determine the plateau
The analysis of the FEA data, compression response and energy ab- end (onset of densification) by using the efficiency method, was pre-
sorption behaviour of the lattices showed that the lattices which deform cisely consistent with the shoulder point on the energy absorption dia-
by bending under the compression loading such as Circular, Octagonal, gram as reported by Maiti et al. [45] and Gibson and Ashby [1]. The
Strengthened Octagonal and Kelvin lattices generally provide good en- Octagonal and Kelvin lattices investigated experimentally in this study
ergy absorption characteristic (high energy absorption and efficiency). were found to provide higher energy absorption efficiency, 48% and
Among the six studied structures, the Octagonal lattice is identified as 49% respectively, when compared to the Octet-truss and all other cellu-
the optimum structure with the closest performance to the ideal energy lar materials reported in [25] including the Aluminum foam and
absorber. It provides the highest plateau stress of 1.05 MPa, the highest honeycomb.
energy absorption per unit volume of 0.75 J/cm3, and the second highest
efficiency of 54% as listed in Table 3. The high energy absorption perfor- 7. Conclusions
mance of the 3D Octagonal lattice was found to be consistent with the
results of 2D Octagonal honeycomb reported in [26]. The compressive This investigation aimed to develop a high energy-absorbing lattice
response of the RO type lattice is stretching dominated and the Cubic with the closest behaviour to the ideal energy absorber. The compres-
lattice is buckling dominated, which makes them stiff and strong but sive stress-strain response and energy absorption characteristics of lat-
unsuitable for energy absorption due to the significant softening of tice structures from six different types of unit cells with identical
these structures after the yield point. These two lattices have the relative density were evaluated and compared. A recently developed
3D printing process, namely Multi Jet Fusion (MJF), was successfully
used to print high-quality prototypes of the lattices in a brand of poly-
amide (nylon) 12. Full-scale nonlinear FEA was used to investigate the
compressive behaviour of lattices. The mechanical tests performed on
two types of the lattices, Octagonal and Kelvin lattices, were in good
agreement with the FEA predictions. Clear trends were observed on
the influence of unit cell configuration on plateau stress, energy absorp-
tion and efficiency. In terms of plateau stress, the lattices with bending-
dominated deformation were found to be compliant and display a rela-
tively stable plateau region resulting in higher energy absorption and ef-
ficiency. On the other hand, those with stretching and buckling
dominated deformation were found to be much stiffer and stronger,
but soften after yielding, thus causing a lower energy absorption and
efficiency.
The results have identified that under quasi-static loading the Octag-
onal lattice offers the optimum energy absorption characteristic among
Fig. 15. Quasi-static experimental (Exp) compressive stress-strain curves of MJF vs FDM the six types of studied unit cell structures. It provides a high plateau
Octagonal lattice samples. stress, energy absorption and efficiency of 1.05 MPa, 0.75 J/cm3 and
 F.N. Habib et al. / Materials and Design 155 (2018) 86–98 97

Fig. 16. Post-compression test samples built by MJF compared to an FDM sample.

54% respectively. The results of this study reveal a strong relationship [10] I. Zein, D.W. Hutmacher, K.C. Tan, S.H. Teoh, Fused deposition modeling of novel
scaffold architectures for tissue engineering applications, Biomaterials 23 (4)
between the morphological and energy absorption properties of the lat- (2002) 1169–1185.
tice structures. The findings give a greater understanding on designing [11] E. Linul, D.A. Serban, T. Voiconi, L. Marsavina, T. Sadowski, Energy - absorption and
and 3D printing of lattice structures with the closest behaviour to the efficiency diagrams of rigid PUR foams, in: L. Marsavina (Ed.), Proceedings of the
14th Symposium on Experimental Stress Analysis and Materials Testing 2014,
ideal energy absorber and can help to develop advanced lattice mate- pp. 246–249.
rials that are superior to the currently used stochastic foams. This [12] S.M. Ahmadi, S.A. Yavari, R. Wauthle, B. Pouran, J. Schrooten, H. Weinans, A.A.
study focussed only on using the quasi-static displacement loading to Zadpoor, Additively manufactured open-cell porous biomaterials made from six dif-
ferent space-filling unit cells: the mechanical and morphological properties, Mate-
examine the energy absorption behaviour of the studied structures. rials 8 (4) (2015) 1871–1896.
However, in impact applications, these structures will be exposed to [13] J. Brennan-Craddock, D. Brackett, R. Wildman, R. Hague, The Design of Impact Ab-
various dynamic loading conditions depending on the application, sorbing Structures for Additive Manufacture, Journal of Physics: Conference
SeriesIOP Publishing, 2012.
which may cause a difference in their energy absorption behaviour.
[14] M. Vesenjak, L. Krstulovic-Opara, Z. Ren, Z. Domazet, Cell shape effect evaluation of
Hence, further studies are needed to investigate their performance polyamide cellular structures, Polym. Test. 29 (8) (2010) 991–994.
under specified dynamic loading conditions. [15] C.I. Hammetter, R.G. Rinaldi, F.W. Zok, Pyramidal lattice structures for high strength
and energy absorption, J. Appl. Mech. 80 (4) (2013), 041015. .
[16] Z. Ozdemir, A. Tyas, R. Goodall, H. Askes, Energy absorption in lattice structures in
dynamics: nonlinear FE simulations, Int. J. Impact Eng. 102 (2017) 1–15.
References [17] I. Ullah, M. Brandt, S. Feih, Failure and energy absorption characteristics of advanced
3D truss core structures, Mater. Des. 92 (2016) 937–948.
[1] L.J. Gibson, M.F. Ashby, Cellular Solids: Structure and Properties, second edition [18] R. Gautam, S. Idapalapati, S. Feih, Printing and characterisation of Kagome lattice
Cambridge University Press, Cambridge, 1997 1–510. structures by fused deposition modelling, Mater. Des. 137 (2018) 266–275.
[2] E.B. Duoss, T.H. Weisgraber, K. Hearon, C. Zhu, W. Small, T.R. Metz, J.J. Vericella, H.D. [19] C.S. Ha, R.S. Lakes, M.E. Plesha, Design, fabrication, and analysis of lattice
Barth, J.D. Kuntz, R.S. Maxwell, C.M. Spadaccini, T.S. Wilson, Three-dimensional exhibiting energy absorption via snap-through behavior, Mater. Des. 141
printing of elastomeric, cellular architectures with negative stiffness, Adv. Funct. (2018) 426–437.
Mater. 24 (31) (2014) 4905–4913. [20] M. Kaur, T.G. Yun, S.M. Han, E.L. Thomas, W.S. Kim, 3D printed stretching-dominated
[3] M.F. Ashby, The properties of foams and lattices, Philos. Trans. R. Soc. A-Math. Phys. micro-trusses, Mater. Des. 134 (2017) 272–280.
Eng. Sci. 364 (1838) (2006) 15–30. [21] S.Y. Choy, C.-N. Sun, K.F. Leong, J. Wei, Compressive properties of functionally graded
[4] V. Srivastava, R. Srivastava, On the polymeric foams: modeling and properties, J. lattice structures manufactured by selective laser melting, Mater. Des. 131 (2017)
Mater. Sci. 49 (7) (2014) 2681–2692. 112–120.
[5] Z. Ozdemir, E. Hernandez-Nava, A. Tyas, J.A. Warren, S.D. Fay, R. Goodall, I. Todd, H. [22] D.S. Al-Saedi, S. Masood, M. Faizan-Ur-Rab, A. Alomarah, P. Ponnusamy, Mechanical
Askes, Energy absorption in lattice structures in dynamics: experiments, Int. J. Im- properties and energy absorption capability of functionally graded F2BCC lattice fab-
pact Eng. 89 (2016) 49–61. ricated by SLM, Mater. Des. 144 (2018) 32–44.
[6] F.N. Habib, M. Nikzad, S.H. Masood, A.B.M. Saifullah, Design and development of [23] J. Harris, R. Winter, G.J. Mcshane, Impact response of additively manufactured me-
scaffolds for tissue engineering using three-dimensional printing for bio-based ap- tallic hybrid lattice materials, Int. J. Impact Eng. 104 (2017) 177–191.
plications, 3D Print. Addit. Manuf. 3 (2) (2016) 119–127. [24] R. Winter, M. Cotton, E. Harris, D. Eakins, G. McShane, High resolution simulations of
[7] A. Kantaros, N. Chatzidai, D. Karalekas, 3D printing-assisted design of scaffold struc- energy absorption in dynamically loaded cellular structures, Shock Waves 27 (2)
tures, Int. J. Adv. Manuf. Technol. 82 (1–4) (2016) 559–571. (2017) 221–236.
[8] P. Pooyan, R. Tannenbaum, H. Garmestani, Mechanical behavior of a cellulose- [25] M. Mohsenizadeh, F. Gasbarri, M. Munther, A. Beheshti, K. Davami, Additively-
reinforced scaffold in vascular tissue engineering, J. Mech. Behav. Biomed. Mater. manufactured lightweight Metamaterials for energy absorption, Mater. Des. 139
7 (2012) 50–59. (2018) 521–530.
[9] T. Serra, M. Ortiz-Hernandez, E. Engel, J.A. Planell, M. Navarro, Relevance of PEG in [26] F.N. Habib, P. Iovenitti, S.H. Masood, M. Nikzad, Cell geometry effect on in-plane en-
PLA-based blends for tissue engineering 3D-printed scaffolds, Mater. Sci. Eng. C 38 ergy absorption of periodic honeycomb structures, Int. J. Adv. Manuf. Technol. 94
(2014) 55–62. (5–8) (2018) 2369–2380.
98 F.N. Habib et al. / Materials and Design 155 (2018) 86–98

[27] A. Fereidoon, S.A. Taheri, Using finite element method to analyze the effect of micro- [38] SYSTEMES, D., Simulia, ABAQUS/CAE User's Manual. ABAQUS Documentation V6,
structure on energy absorption properties of open cell polymeric foams, J. Cell. Plast. 2014 7.
48 (3) (2012) 257–270. [39] H. Altenbach, A. Öchsner, Cellular and Porous Materials in Structures and Processes,
[28] R.M. Sullivan, L.J. Ghosn, B.A. Lerch, A general tetrakaidecahedron model for open- vol. 521, Springer Science & Business Media, 2011.
celled foams, Int. J. Solids Struct. 45 (6) (2008) 1754–1765. [40] I. Standard, ISO 13314: 2011 (E) Mechanical Testing of Metals—Ductility Testing—
[29] G. Maîtrejean, P. Terriault, D. Devís Capilla, V. Brailovski, Unit cell analysis of the Compression Test for Porous and Cellular Metals. Ref Number ISO. 13314(13314),
superelastic behavior of open-cell tetrakaidecahedral shape memory alloy foam 2011 1–7.
under quasi-static loading, Smart Mater. Res. 2014 (2014). [41] ASTM, D., 1621, Standard Test Method for Compressive Properties Of Rigid Cellular
[30] N. Chantarapanich, P. Puttawibul, S. Sucharitpwatskul, P. Jeamwatthanachai, S. Plastics, American Society for Testing and Materials, New York, 2010.
Inglam, K. Sitthiseripratip, Scaffold library for tissue engineering: a geometric eval- [42] Q.M. Li, I. Magkiriadis, J.J. Harrigan, Compressive strain at the onset of densification
uation, Comput. Math. Methods Med. 2012 (2012). of cellular solids, J. Cell. Plast. 42 (5) (2006) 371–392.
[31] A. Boccaccio, M. Fiorentino, A.E. Uva, L.N. Laghetti, G. Monno, Rhombicuboctahedron [43] M. Avalle, G. Belingardi, R. Montanini, Characterization of polymeric structural
unit cell based scaffolds for bone regeneration: geometry optimization with a foams under compressive impact loading by means of energy-absorption diagram,
mechanobiology–driven algorithm, Mater. Sci. Eng. C 83 (2018) 51–66. Int. J. Impact Eng. 25 (5) (2001) 455–472.
[32] HP Jet Fusion 3D printing. [13/03/2018], Available from www.hp.com/go/3DPrint. [44] S.C. Chapra, R.P. Canale, Numerical Methods for Engineers, vol. 2, McGraw-Hill, New
[33] A. Standard, D638-14 Standard Test Method for Tensile Properties of Plastics, ASTM York, 2015.
International, West Conshohocken, PA, 2014. [45] S. Maiti, L. Gibson, M. Ashby, Deformation and energy absorption diagrams for cel-
[34] G. Lu, T. Yu, in: T.X. Yu (Ed.), Energy absorption of structures and materials, lular solids, Acta Metall. 32 (11) (1984) 1963–1975.
Woodhead, Cambridge, 2003. [46] J. Miltz, O. Ramon, Energy absorption characteristics of polymeric foams used as
[35] W.D. Callister, D.G. Rethwisch, Materials Science and Engineering, 8th ed.vol. 5, cushioning materials, Polym. Eng. Sci. 30 (2) (1990) 129–133.
John Wiley & Sons NY, USA/Versailles, 2011. [47] A.I. Publisher, Standard Test Method for Compressive Properites of Rigid Celluar
[36] SYSTEMES, D., Simulia, Abaqus 6.12 Analysis User's Manual, Volume III: Materials, Plastics: ASTM D1621-16, SAI Global, Perth, W.A., 2012
2012. [48] S.D. Papka, S. Kyriakides, In-plane crushing of a polycarbonate honeycomb, Int. J.
[37] F.N. Habib, P. Iovenitti, S.H. Masood, M. Nikzad, In-plane energy absorption evalua- Solids Struct. 35 (3) (1998) 239–267.
tion of 3D printed polymeric honeycombs, Virtual Phys. Prototyp. 12 (2) (2017)
117–131.