International Journal of Impact Engineering 64 (2014) 20e29

Contents lists available at ScienceDirect

International Journal of Impact Engineering

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

Blast-wave impact mitigation using negative effective mass density

concept of elastic metamaterials
K.T. Tan a, *, H.H. Huang b, C.T. Sun a
a
b

School of Aeronautics and Astronautics, Purdue University, West Lafayette, IN 47907, USA
Department of Engineering Science and Ocean Engineering, National Taiwan University, Taipei 10617, Taiwan

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 13 April 2013
Received in revised form
4 September 2013
Accepted 23 September 2013
Available online 3 October 2013

This paper presents the use of elastic metamaterials for impact attenuation and blast-wave mitigation.
Metamaterials represent a novel and emerging research area where materials exhibit exceptional
properties not commonly found in natural materials. These unique properties are enacted by specically
designed microstructures. In this study, a single-resonator model and a dual-resonator microstructural
design are proposed to exhibit negative effective mass density. The effect of negative effective mass
density is explicitly conrmed by analysis of wave propagation using numerical simulations. Results
evidently show that impact stress wave attenuation occurs over a wider frequency spectrum for the dualresonator model as compared to the narrow band gap of a single-resonator design. Parametric studies of
blast-wave simulation reveal that the mass and number of internal resonators have signicant inuence
over the frequency range of blast-wave attenuation. The effectiveness and performance of the singleresonator and dual-resonator models on blast-wave mitigation are examined and discussed. Finally,
practical ways to design and manufacture elastic metamaterials with negative effective mass density are
presented and explored.
2013 Elsevier Ltd. All rights reserved.

Keywords:
Blast-wave
Negative effective mass density
Elastic metamaterials
Wave propagation
Wave attenuation

1. Introduction
As advanced body armor and head protection gears greatly
reduced soldier fatalities from explosive and ballistic attacks, the
problem of blast-related traumatic brain injury (TBI) in the attack
survivors has become a more severe issue [1]. The impact of blastwave propagation is so lethal to human health that although no
external physical injuries are detected in the affected soldiers,
serious damages are inicted internally, particularly to the brain
where neurological effects can be slow to appear [2]. It has been
postulated that blast waves ripple through the victims torso up
into the brain through major blood vessels [3]. The increasing use of
improvised explosive devices (IEDs) poses a very serious threat in
military conicts. Moreover, in recent times, the exposure to blast
threat extends beyond war zones. Civilians now encounter the
increased risk to terrorist attack by the use of explosive bombs,
dangerous detonation, as well as IEDs. As such, it is important to
develop materials that would absorb or reect the full range of
blast-wave frequencies generated by an explosion. It is critical to
nd ways for a material, not only to stop projectiles like shrapnel or

* Corresponding author. Tel.: 1 765 494 6237; fax: 1 765 494 0307.
E-mail addresses: kttan@purdue.edu, kwektze@hotmail.com (K.T. Tan).
0734-743X/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijimpeng.2013.09.003

bullets, but also to effectively attenuate the injurious effects of

incoming blast and shock wave.
Researchers have been investigating ways to mitigate shock and
blast waves by using different forms of materials and structures.
Nesterenko [4] considered the applications of soft condensed
matter for blast mitigation using simplied approach, based on the
materials ability to absorb signicant energy. Christou et al. [5]
developed a uid-structure interaction computational continuum
model to investigate the attenuation properties of foam-specimens
containing ller materials under shock loading. Dawson [6]
designed a blast mitigation and protection infrastructure consisting of a steel plate backed by a layer of low-density, reticulated,
exible, uid-lled foam. Wadley et al. [7] explored the feasibility
of cellular materials concepts for active mitigation of blast overpressures, in which a deployable, precompressed, cellular medium is released just prior to the arrival of the blast-created impulse. This accelerates an attached buffer toward the blast and
creates momentum opposing that acquired from the blast. Su et al.
[8] proposed a blast wave mitigation device, which makes use of
the repeated reection of the shock wave within the blast mitigation device to signicantly increase the duration of the force on the
base of the cylinder over that of the blast wave. Grujicic et al. [9]
investigated the blast-wave impact-mitigation ability of polyurea
when used as a helmet suspension-pad material and compared

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

with the case of a conventional foam suspension-pad material.

Grujicic et al. [10] also studied the effect of the prior processing and
the resulting microstructure on the blast performance of all-metal
sandwich-structures with an auxetic-hexagonal core. By far, the
principal of blast-wave mitigation is either using deformation of
materials to absorb energy, which results in vast material damage,
or using anti-momentum principle to oppose the incoming blastwave, which requires somewhat complex structures. Moreover,
these concepts cannot be designed to tailor attenuation of specic
frequency ranges, or cover the full range of blast-wave frequencies.
In this paper, we investigate the use of elastic metamaterials to
mitigate blast-wave and attenuate stress wave propagation. Metamaterials represent a novel and emerging eld in materials engineering, and are characterized by the fact that they possess
exceptional material properties not commonly found in natural
materials. These material properties are enacted, not due to
chemical composition, but by specially designed man-made microstructures. Metamaterials originally started in the eld of electromagnetic waves where researchers investigated negative
electrical permittivity, negative magnetic permeability and negative refractive index [11,12]. Motivated by the mathematical analogy between acoustic and electromagnetic waves, the counterpart
acoustic metamaterials are recently explored [13e16]. This new
branch of acoustic metamaterials, also known as mechanical or
elastic metamaterials, consists of tailored microstructures that
exhibit unusual mechanical properties such as negative effective
modulus and/or negative effective mass density [13e16]. This paper
aims to employ the effective negative mass density concept of
acoustic/elastic metamaterials for blast-wave impact mitigation.
The original single mass-in-mass resonator model, which offers
negative mass property over a specic frequency range, is rst
introduced. A newer dual-resonator model is then described. This
model maximizes the negative mass density over a wider frequency
range, and aptly corresponds to the wide frequency nature of blastwave problem. The transient and dynamic response of the metamaterial is then analyzed by examining a lattice system model that
consists of mass-in-mass units to exhibit negative effective mass
density. The effectiveness of the negative mass models on impact
wave attenuation is demonstrated by computational simulation
using numerical analysis. The use of elastic metamaterials for blastwave mitigation is further presented and discussed by understanding how different internal resonator parameters affect stress
wave propagation.

21

2. Analytical negative mass models

2.1. Single-resonator model
The concept of negative effective mass density can be achieved
by using a mass-in-mass unit microstructure [15,17]. Fig. 1 shows
the microstructure of a single-resonator spring-mass system. The
outer unit cell has a mass m1 and displacement u1, while the internal resonator has a mass m2 and displacement u2. The internal
resonator is coupled to the outer mass by a linear spring of stiffness
k2. By considering the free body diagrams of the masses m1 and m2,
we can obtain the equations of forces as

1 F k2 u2  u1
m1 u

(1)

2 k2 u1  u2
m2 u

(2)

Assuming that the displacement of masses follow harmonic

wave behavior, similar to that of the applied force F, we have

Ft FO eiut

(3)

b g eiut
ug x; t u

(4)

whereby g 1, 2 in this case.

By solving the above equations, we can simplify the relation to

u2 m2
b1
0 FO m1 2 2 2 u2 u
u2  u

p
where u2
k2 =m2 is the local resonance frequency of the internal resonator mass m2.
To obtain an effective mass behavior of the microstructure, the
following equation must be satised.

b1
FO meff u2 u

(6)

Physically, this means that the motion of outer mass, m1 is

similar to that of an equivalent effective mass, meff, depicted schematically in Fig. 1. Solving Equations (5) and (6), we attain the
effective mass of the microstructure as

u1

u1

meff

m1

k2

F (t )

(5)

u2
m2

FO e

i t

F (t )

FO e

i t

Fig. 1. Microstructure of single-resonator spring-mass system and its effective mass; the motion of outer mass m1 equals to that of the effective mass meff.

22

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

Fig. 2. Plot of normalized effective mass meff/mst against normalized operating frequency u/u2.

meff m1

u22 m2
u22  u2

(7)

It is clear that the effective mass, meff, is frequency u dependent.

Further manipulation of the equation gives the normalized effective mass, meff/mst as follows:

"
#
meff
q
u=u2 2
1
mst
1 q 1  u=u2 2

(8)

where mst m1 m2 is the static mass of the microstructure and

q m2/m1 is the ratio of internal mass m2 to that of the outer mass
m1.
From Equation (8), it is obvious that the function of meff/mst
against u/u2 is dependent only on a single parameter q. u/u2 simply

represents the normalized operating frequency of the system. Plots

of meff/mst against u/u2 with varying q values are presented in Fig. 2.
It is evident that a narrow band gap region exists where the
effective mass becomes negative near to the local resonance frequency of the internal mass m2, when u/u2 1. It has been shown
in earlier work that this negative effective mass region corresponds
to the band gap region of the dispersion curve when wave propagation is considered [15]. It is clear that when the forcing frequency
u approaches from above, the effective mass becomes negative. It is
of interest to note that, if the mass becomes negative, the acceleration would be in the opposite direction to the applied force according to the Newtons second law of motion, and the response
amplitude would be reduced. This effect is greatly magnied as the
input frequency u approaches the local frequency u2, It is clear that
the negativity of the effective mass is the result of our attempt to
use a single mass to represent a two-mass system. A similar
negative mass density phenomenon exists when the lattice system
is modeled either as a classical elastic solid or a microstructure
continuum model [15]. A corresponding band gap is observed in the
dispersion curve between the acoustical and optical modes.
p
Moreover, since the wave speed is given as n
E=r, this means
that wave cannot propagate in elastic solids when effective mass
density is negative. This also implies that stress wave cannot be
transmitted through the metamaterial structure when the frequency u is near to the local resonance frequency of the internal
resonator u2. When the operating frequency approaches zero
(u / 0), the dynamic effective mass meff expectedly approaches
static mass mst. From Fig. 2, it is also clear that by increasing q
values, meaning to increase m2 and/or decrease m1, the band gap
region can be broadened. Thus, increasing q value is one way of
maximizing the band gap region of the acoustic/elastic metamaterial. However, the negative mass region still lies signicantly
close to u/u2 1 and the broadened region of negative mass shows
minimal effect due to small amplitude in negativity, thus causing
little reduction in response amplitude during wave propagation.
The physical meaning and mechanism of negative effective mass
density can be well understood when we observe wave transmission through a one-dimensional lattice system consisting of
mass-in-mass units [18]. The spring-mass model represents a
simple analytical model. It is not the intention of the authors to
discretize materials into atomic or molecular level. The simple
analytical spring-mass model can be feasibly realized in a practical
material system design described in Section 5 of this paper. The

u3

u3

k1 k2

F (t )

mEFF

m3

u1

u 2 m1
m2

FO e

i t

F (t )

FO e

i t

Fig. 3. Microstructure of dual-resonator spring-mass system and its effective mass; the model resembles that of mass-in-(mass-in-mass) system.

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

23

negative effective mass density arises from the negative total momentum of the unit cell with positive velocity elds due to local
resonance, which has been conrmed by experimental study [19]. It
has been shown that such system is capable of attenuating wave
propagation. However, one major limitation of the single-resonator
model is that the negative effective mass exhibits only within a very
narrow band gap region, specically near the resonance frequency
of the internal resonator. This limits the use of such system in applications where broad band gap is required, like wide frequency
range in blast/shock wave.
2.2. Dual-resonator model
A dual-resonator microstructure model is recently proposed
[20]. The dual-resonator model is basically a mass-in-(mass-inmass) unit, having two coupled internal resonators, as depicted in
Fig. 3. The internal resonator is mass m2, the middle mass is m1 and
the outer mass is m3. The analytical model of this microstructure
can be similarly derived by considering the equations of motions
for the three masses and equating the behavior of the entire
microstructure as an effective mass mEFF (Fig. 3). The equations of
forces considering the free body diagrams of the masses give the
following equations

3 F k1 u1  u3
m3 u

(9)

1 k2 u2  u1 k1 u3  u1
m1 u

(10)

2 k2 u1  u2
m2 u

(11)

Using the same assumptions as the single-resonator model, we

can solve and manipulate the equations to give the normalized
effective mass, mEFF/mST of the dual-resonator model as below:

q1
mEFF

mST q1 q3 q1 q3
(

q3

q1 q3 q1 q3

1 q1 u=u2 2
h
ih
i
1u=u2 2 1 d1 d1 =q1 u=u2 2  d1

(12)
where mST m1 m2 m3 is the static mass of the microstructure,
p
u2 k2 =m2 is the local resonance frequency of internal mass
m2, q1 m2/m1 is the ratio of internal mass m2 to that of the middle
mass m1, q3 m2/m3 is the ratio of internal mass m2 to that of the
outer mass m3, and d1 k2/k1 is the ratio of spring stiffness connecting the internal mass k2 to that of the middle mass k1. It is clear
from Equation (12) that the function of mEFF/mST against u/u2 is
dependent on three parameters, namely: q1, q3 and d1, and their
effects are shown in Fig. 4(a)e(c). The graphs clearly show that a
dual-resonator microstructure generates negative effective mass
over a larger frequency spectrum, particularly at two distinctive
asymptotic regions. From Fig. 4(a), we can see that the second band
gap increases drastically with decreasing d1. Too large d1 value
results in a single band gap system when m1 and m2 are coupled
with a very high k2 stiffness. Increasing q1 values shifts the second
band gap to higher frequency range (Fig. 4(b)). Depending on
application, larger q1 values provide broader and further band
gaps. Small q1 values result in larger band gap region and closer
band gap differences. Too small q1 values again result in a single
band gap system. Fig. 4(c) demonstrates that the inuence of q3
resembles that of q in the single-resonator model. Basically, the
asymptotic regions are not affected by q3 and large q3 values increases both band gap regions. By combining the effects of all 3

Fig. 4. Plot of mEFF/mST against u/u2, with varying (a) d1, (b) q1, (c) q3 values.

parameters (q1, q3 and d1), we can specically tailor an acoustic/

elastic metamaterial to achieve wave attenuation over designed
band gap regions.
Elastic metamaterials created by incorporating locally resonant
microstructures in a relatively stiff matrix material forbids dynamic
disturbances with certain frequencies from propagating into them.
This unusual dynamic behavior can be quantied by the negative
effective mass density. The identity of the internal resonator mass
m2 would be ignored and its effect would be absorbed by the
introduction of an effective mass meff as depicted in Fig. 1. The stop
band for mechanical waves can be optimized to the incoming
disturbance by tailoring the local resonance frequency of the resonators. The attenuation is signicant if the forcing frequency is
close to the local resonance frequency due to the sequestration of
kinetic energy by the resonator mass. The energy sequestration
and rejection achieved in the elastic metamaterials creates an
apparent damping mechanism without the use of any dissipative
component.

24

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

3. Numerical simulation

o
tt
t

F Fmax e
F 0

3.1. Impact wave attenuation

A one-dimensional spring-mass lattice system is constructed as
shown schematically in Fig. 5(a). The masses are mL 0.03 kg, and
the springs have stiffness of kL 1  107 N/m. The rst 20 unit cells
are considered Medium 1 and the last 20 unit cells are denoted as
Medium 2. The middle 10 unit cells represent the Metamaterial
region, where the microstructural design can consist of either the
single-resonator (Fig. 1) or the dual-resonator (Fig. 3) design, to
exhibit negative effective mass density. Each metamaterial unit cell
is explicitly modeled with springs and undeformable rigid masses
as shown in Figs. 1 and 3. An impact pulse is given to unit cell #1
and the transient dynamic response of wave propagation, due to
the effect of metamaterial region, is analyzed before reection occurs at the right end of the model (unit cell #50). In this simulation,
the impact pulse is a prescribed displacement-controlled superposition of two simple harmonic waves (u1 100 rad/s and
u2 300 rad/s), signifying a more complex wave form than a single
sinusoidal pulse. The impact pulse has a short duration of around
31.4 ms, as illustrated in Fig. 5(b). The model is built using Abaqus
CAE and run by Abaqus dynamic explicit.
3.2. Blast-wave mitigation
To demonstrate blast-wave mitigation, an extended 1D springmass lattice system consisting of 400 unit cells is used. To model
the elastic metamaterial region, internal resonators (using singleresonator or dual-resonator model) are placed inside the 51st to
60th unit cells to exhibit the effect of negative effective mass
density and to mitigate blast-wave propagation. Each metamaterial
unit cell is explicitly modeled with springs and undeformable rigid
masses as shown in Figs. 1 and 3. As such, Medium 1 will consist of
the rst 50 unit cells, and Medium 2 is made up of the remaining
unit cells (from unit cell #60 to #400), beyond the metamaterial
region. It is well known that blast loading produced by detonation
of an explosive is characterized by a very short time duration and
extremely sudden and high load peak. In this simulation, the forcecontrolled blast pulse (Fig. 6(a)) is given to unit cell #1 as the input
forcing history, based on the following equation:

t  to
t < to

(13)

where Fmax 1000 N, to 0.5 ms, td 0.1 ms.

The blast pulse used in this study (Fig. 6(a)) reasonably resembles a realistic air blast. In terms of duration, the pulse duration
of 0.5 ms used corresponds well with references [5,9] and is within
the range of practical problems used in references [7,8]. The
negative phase of air blast is assumed negligible in this study due to
weak rarefaction [7]. Su et al. [8] similarly ignored the negative
phase in their blast simulation. Although the negative phase might
affect the material behavior, but in this study of demonstrating
blast wave frequency attenuation, the results are expected to show
similar distribution of a wide range of frequency content, with
observable dips at specic frequencies due to local resonance frequency of the internal resonator. The force parameters used in this
study is hypothetically determined with the mere intention to
demonstrate the effect on the frequency dips due to resonator
design.
Due to the blast pulse, the stress wave generated in the model
has a wide frequency domain ranging from 0 to 5000 Hz, as displayed in Fig. 6(b). The frequency spectrum is calculated using Fast
Fourier Transform (FFT) by analyzing the velocity prole of the
100th unit cell (#100), so as to study the effect of the negative
effective mass density of metamaterials on frequency attenuation
in blast-wave mitigation. The model is similarly constructed using
Abaqus CAE and run on Abaqus dynamic explicit.
4. Results and discussion
4.1. Impact wave attenuation
Figs. 7e9 show how the microstructural design of metamaterials can affect impact stress wave propagation. The lattice
system is subjected to an impact pulse mentioned earlier in
Fig. 5(b). The transient and dynamic response of the model is
analyzed using numerical simulation. Fig. 7 portrays time snapshot of the wave propagation from left to right. When there is no
internal resonator in the material (unit cells #21 to #30 containing
no internal mass), meaning that the material is not a metamaterial,

Fig. 5. (a) One-dimensional spring-mass lattice system, (b) impact pulse used in simulation.

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

25

Fig. 6. (a) Blast pulse used in simulation, (b) Frequency domain of blast simulation calculated by Fast Fourier Transform (FFT).

it is clear that the impact wave passes through the material without
any inuence and attenuation, as depicted in Fig. 7. In this study,
the material properties (unit cell mass and spring stiffness)
throughout various sections of the model, like Medium 1, Metamaterial and Medium 2 (Fig. 5(a)), are kept the same, so as to
neglect the effect of impedance mismatch of different materials,
focusing the study only on the effect of negative effective mass
density due to internal resonator design.
Fig. 8 shows the wave propagation results when the metamaterial is made up of the single resonator design, with a target
frequency of 100 rad/s. This means that each of the unit cell #21 to
#30 has
an internal resonator mass (m2 0.03 kg and
p
u2 k=m 100 rad=s). It is observed that the wave propagation snap-shots at time t 0.0290 s and t 0.0374 s (Fig. 8(a) and
(b)) are the same as that of Fig. 7(a) and (b) of the non-metamaterial
design. However, when the impact stress wave passes through the
metamaterial, there is a slight decrease in the displacement
amplitude of the wave form, thus implying a slight attenuation of
stress wave (Fig. 8(c)). Reection away from the metamaterial can
be seen particularly at t 0.0674 s (Fig. 8(e)). As the impact pulse is
composed of two different harmonic frequencies (Fig. 5(b)), a
single-resonator design is insufcient to have signicant effect on
stress wave attenuation. In reality, an impact pulse will consist of
more than one basic harmonic frequency, thus the narrow negative
effective mass density band gap of a single-resonator design is seen
to be ineffective.

In Fig. 9, the performance of a dual-resonator metamaterial is

evaluated. Once again, it is clear that the wave propagation snapshots at time t 0.0290 s and t 0.0374 s (Fig. 9(a) and (b)) are
the same as that of the non-metamaterial (Fig. 7(a) and (b)), since
the wave has not reached the metamaterial. However, at
t 0.0474 s, it is evident that there is a signicant decrease in the
amplitude of the wave form, thus inferring a signicant attenuation
of stress wave in the metamaterial (Fig. 9(c)). From Fig. 9(d) and (e),
it is also evident that very little amount of stress wave passes
through beyond the metamaterial. In fact, there is a signicant
amount of reection from the metamaterial. In this case, the dualresonator metamaterial works well and is able to effectively
attenuate the input excitation wave that spans over 2 broad frequencies. In transient impact loading, internal resonators have
demonstrated to be effective in reducing the displacement of the
protected structure (Medium 2 in this study), and attenuate a
specically-designed range of frequency where the negative
effective mass density is exhibited.
4.2. Blast-wave mitigation
The frequency domain (obtained by FFT) of the lattice system
subjected to a blast pulse is analyzed to understand how specially
designed microstructural internal resonator can affect blast-wave
mitigation of specic frequency ranges. In Fig. 10, we can see the
effect of frequency range and the inuence of the number of

Fig. 7. Time snap shot of impact wave propagation through a non-metamaterial.

26

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

Fig. 8. Time snap shot of impact wave propagation through a single-resonator metamaterial.

internal resonators. Notations for various cases are explained as

follow: 10-500 Hz-q 1 represents the case with 10 unit cells
(#51 to #60) whereby each unit cell has a single internal resonator
of m2 0.03 kg and k2 29,6088 N/m, hence resulting in a
designed resonator local frequency of u2 500 Hz, and since
m2 m1 0.03 kg, the mass ratio q m2/m1 1 (m1 is the mass of
each outer unit cell, as shown in Fig. 1). Fig. 10 also compares cases
1-500 Hz-q 1 and 1-1500 Hz-q 1 whereby only 1 unit cell (#51)
has an internal resonator. In the cases where the targeted frequency
is at 1500 Hz, only the spring stiffness of the internal resonator is
changed (k2 2,664,793 N/m), keeping m2 and the mass ratio q the
same. A summary of the parameters used in various cases are given
in Table 1.
From Fig. 10, it is evident that there are dips in the amplitude of
the frequency spectrum at specic frequency ranges. These dips, in
fact, correspond to the local frequency of the internal resonators
accordingly, at 500 Hz and 1500 Hz in this study. It is clear that the
results of the frequency domain plot (Fig. 10) reect the analytical
results of the negative effective mass density of the metamaterial
(Fig. 2). As mentioned previously, the effective mass density of the
metamaterial becomes negative at the asymptotic region near the
local resonance frequency of the internal mass m2 when u/u2 1.

This means that waves at these frequency ranges (1 < u/u2  1.3,
when q 1) are unable to propagate through the material. As such,
from the analytical model (Fig. 2), the negative mass region when
q 1 and u2 500 Hz is 500 Hz < u  650 Hz. This corresponds
very well with the case 10-500 Hz-q 1. However, for the case of 1500 Hz-q 1, we can only observe a very sharp dip near 500 Hz,
where the mass negativity is most paramount. This reveals that
having more internal resonators can effectively broaden the frequency range of the wave attenuation, due to more kinetic energy
absorbed by more internal resonators. Study made earlier has
pointed to the same conclusion [18]. For the cases at higher
designed local frequency of the internal resonators (1500 Hz), there
seems to be wave attenuation over a wider frequency range. This is
because when q 1, the condition that 1 < u/u2  1.3 remains, thus
when u2 1500 Hz, the negative mass region naturally extends
over 1500 Hz < u  1950 Hz. As such this leads to a wider band gap
region at high frequency domain when mass ratio q is kept the
same.
The results from Fig. 11 show the effect of increasing the mass of
internal resonators. In this case, m2 is increased to 0.09 kg, resulting
in a higher mass ratio of q 3. From Fig. 2, we can see that negative
mass occurs over 1 < u/u2  2 when q 3. This analytical result

Fig. 9. Time snap shot of impact wave propagation through a dual-resonator metamaterial.

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

Fig. 10. Frequency domain (by FFT) showing the effect of frequency range and number
of internal resonators.

agrees very well with the blast simulation results in Fig. 11. For
q 3, the negative mass region when u2 500 Hz and u2 1500 Hz
is 500 Hz < u  1000 Hz and 1500 Hz < u  3000 Hz respectively.
This implies that a greater mass ratio is more effective in attenuating stress wave propagation. The effect of frequency range and
number of internal resonators are similarly observed in Fig. 11, as in
Fig. 10.
Due to the linearity of this study, we can achieve broader frequency range of wave attenuation by simply adding layers of internal resonators with different designed local frequencies. Fig. 12
demonstrates how the performance of a model with two layers of
internal resonators (10 resonators in each layer: rst layer designed
at 500 Hz and second layer at 1500 Hz) is effectively the same as
summing up the performance of 2 models with single layer of internal resonators. There is also no apparent difference when the
two layers are in reversed order; meaning whether the 500 Hz layer
rst received the impact wave or the 1500 Hz layer rst received
the impact wave.
Fig. 13 evaluates the wave attenuation performance between the
dual-resonator model and a double-layered single-resonator
model. The detailed parameters of the internal resonators are
presented in Table 2. It is clear that wave attenuation frequency
ranges of these two models are comparable. This implies that the
dual-resonator model is more effective, as it requires fewer unit
cells (outer mass) to achieve similar efciency as a double-layered
single-resonator model. This also means that by using a dualresonator model, the metamaterial will be resultantly thinner and
thus lighter. The mass saving is primarily due to the use of lesser
outer unit cells. If we try to further reduce the mass of outer unit
cells in the double-layered single-resonator model by half
(m1 0.015 kg), so that the total mass is the same as that of the
dual-resonator model, it is observed that a broader frequency range

27

Fig. 11. Frequency domain (by FFT) showing the effect of mass of internal resonators.

is being attenuated. This is because by reducing the outer mass, the

mass ratio q is sequentially increased, resulting in wider negative
mass region (Fig. 2). The same broader band gap can be attained by
similarly reducing the outer mass of the dual-resonator model. This
again concludes that the dual-resonator model is always thinner
and lighter than the double-layered single-resonator model of the
same wave attenuation performance.
It is interesting to note that the simple analytical model discussed in this study assumes linearity. However, for the negative
effective mass density property to be exhibited, the model need not
be linear. One can further expand the mass-in-mass analytical
model by using a non-linear spring, or using damping components
(closer to real application). Similar negative effective mass property
will be achieved. In this study, we merely demonstrate the capability of elastic metamaterial for blast wave mitigation by showing
the simplest case of using linear springs in a linear system. To
consider the deformation ranges of the material under certain
design environment, one needs to account for the allowable
displacement amplitude of the resonators and the stretched/
unstretched length of spring attachments in practical material
system design (discussed in the next section).
5. Practical material system designs
This section presents a summary and outlook of practical ways
to design and manufacture elastic metamaterials with negative
effective mass density. Depending on the size of its microstructure, a metamaterial may be viewed as macroscopic structure
rather than a conventional material. The designs are analogous to
the analytical mass-in-mass model described in this paper,

Table 1
Parameters of internal resonators used in blast-simulation.
Simulation case
notation

m2 (kg)

k2 (N/m)

u2 (Hz)

Number of
resonators

10-500 Hz-q 1
10-1500 Hz-q 1
1-500 Hz-q 1
1-1500 Hz-q 1
10-500 Hz-q 3
10-1500 Hz-q 3
1-500 Hz-q 3
1-1500 Hz-q 3
500 Hz 1500 Hz
1500 Hz 500 Hz

0.03
0.03
0.03
0.03
0.09
0.09
0.09
0.09
0.03, 0.03
0.03, 0.03

296,088
2,664,793
296,088
2,664,793
888,264
7,994,380
888,264
7,994,380
296,088, 2,664,793
2,664,793, 296,088

500
1500
500
1500
500
1500
500
1500
500, 1500
1500, 500

10
10
1
1
10
10
1
1
10 10
10 10

Fig. 12. Frequency domain (by FFT) showing the effect of double-layer of internal
resonators.

28

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

be the pioneer paper in using metamaterials for impact blast wave

mitigation. The work on elastic metamaterials is still at its infancy,
with much room for exploration especially in practical designs and
experimental testing.
6. Conclusions

Fig. 13. Frequency domain (by FFT) showing the effect of dual-resonator and reduced
outer-mass.

whereby an internal resonator exists inside an outer mass.

A possible metamaterial with negative effective mass density may
be constructed in the form of a composite material consisting of
periodic mass-in-mass microstructures embedded in a matrix
material [18]. The composites that Liu et al. [21] used have a simple
microstructural unit consisting of a solid core material with a
relatively high density and a coating of elastically soft material. In
their experiments, they used centimeter-sized lead balls as the core
inner mass material (internal mass m2), coated with a 2.5 mm layer
of silicone rubber (spring k2). The coated spheres are then arranged
in simple cubic crystal structure with epoxy as the hard matrix
material (outer mass m1). Lai et al. [22] further proposes a realistic
elastic metamaterial unit that is designed according to the physical
model whereby the internal mass is made of steel (m2), coated with
soft silicone rubber (k2) and embedded within a foam host (m1).
Chen et al. [23] has used physical spring-mass resonators in a
sandwich beam to build specimens with negative effective mass
density so as to attenuate wave propagation within the band gaps.
In their case, internal mass m2 and spring k2 can be designed
accordingly within the unit resonator cell casing and outer mass m1
is the sandwich foam core. In the recent work done by Sun et al.
[24], negative effective mass structures are made up of resonators
which are of the coil spring-mass type and the outer structure is
made of polycarbonate (outer mass m1) with a guide rod to give
lateral support to the resonator mass. It was also found that
cantilever beams can be used in place of coil springs to reduce the
size of the negative mass structure. Another way to manufacture
elastic metamaterial is by using multi-material 3D printing technique to create complex microstructure generating the same
negative effective mass property. Reports regarding on-going work
will soon be published.
This paper provides a conceptual design for blast wave mitigation using elastic metamaterials in which negative effective mass
density is exhibited. This paper shows theoretical proof and computer simulation validation. The authors believe this paper would

Table 2
Parameters of internal resonators used in dual-resonator and double-layered singleresonator model.
Simulation case

m2, m1 (kg)

Dual-resonator
0.03, 0.03
Double-layered
0.03, 0.03
single-resonator

k2, k1 (N/m)

u2 (Hz)

296,088, 1,184,352
391,577, 1,566,306

500e1500 10
575, 1150 10 10

Number of
resonators

In this work, we have explored the use of elastic metamaterials

for blast-wave impact mitigation. Elastic or acoustic metamaterials
exhibit negative effective mass density near the local frequency of
the specially designed microstructural resonators. This unique
property results in the prohibition of stress wave propagation
across the metamaterial. In this study, we proposed a singleresonator and a dual-resonator analytical model to enact the
negative effective mass property. Numerical simulation is used to
demonstrate wave propagation in the metamaterial model. Time
snap shots of impact wave propagation evidently show that elastic
metamaterial offers protection to the structure beyond the metamaterial by attenuating the impeding stress wave. Impact wave
results also demonstrate the effectiveness of the dual-resonator
model compared to the single-resonator model. In blast-wave
simulation, we performed parametric studies to investigate the
effect of internal resonators on blast frequency range attenuation.
Results reveal that having more internal resonators, or increasing
the mass of the internal resonators cause wider frequency range
wave attenuation. The single-resonator model can be employed by
having double or multiple layers of differently designed internal
resonators to achieve broader band gap. However, the dualresonator is always more advantageous by being lighter and
thinner, while offering comparable wide frequency wave attenuation performance. Lastly, practical ways to design and manufacture
elastic metamaterials with negative effective mass density are
presented and explored.where d1 k2/k1, q1 m2/m1, q3 m2/m3.
Acknowledgments
K.T. Tan gratefully appreciates the nancial support provided by
Agency for Science, Technology and Research (A*STAR, Singapore).
H.H. Huang acknowledges the nancial support from the National
Science Council, Taiwan, grant NSC-101-2218-E-002-013-MY3. C.T.
Sun would like to acknowledge the support by an AFOSR grant
#FA9550-10-1-0061.
References
[1] Okie S. Traumatic brain injury in the war zone. N Engl J Med 2005;352:2043e7.
[2] Taber KH, Warden DL, Hurley RA. Blast-related traumatic brain injury: what is
known? J Neuropsychiatry Clin Neurosci 2006;18:141e5.
[3] Bhattacharjee Y. Shell shock revisited: solving the puzzle of blast trauma.
Science 2008;319:406e8.
[4] Nesterenko VF. Shock (blast) mitigation by soft condensed matter. In:
Sen S, Hunt ML, Hurd AJ, editors. Granular Material-Based Technologies,
vol. 759. Pittsburgh, PA: MRS; 2003MM4.3.1e4.3.12 (arXiv:cond-mat/
0303332v2).
[5] Christou GA, Young LR, Goel R, Vechart AP, Jerusalem A. Shock attenuation of
PMMA sandwich panels lled with soda-lime glass beads: a uid-structure
interaction continuum model simulation. Int J Impact Eng 2012;47:48e59.
[6] Dawson MA. Composite plates with a layer of uid-lled, reticulated foam for
blast protection of infrastructure. Int J Impact Eng 2009;36:1288e95.
[7] Wadley HNG, Dharmasena KP, He MY, McMeeking RM, Evans AG, BuiThanh T, et al. An active concept for limiting injuries caused by air blasts. Int J
Impact Eng 2010;37:317e23.
[8] Su Z, Peng W, Zhang Z, Gogos G, Skaggs R, Cheeseman B. Numerical simulation
of a novel blast wave mitigation device. Int J Impact Eng 2008;35:336e46.
[9] Grujicic M, Bell WC, Pandurangan B, He T. Blast-wave impact-mitigation
capability of polyurea when used as helmet suspension-pad material. Mater
Des 2010;31:4050e65.
[10] Grujicic M, Galgalikar R, Snipes JS, Yavari R, Ramaswami S. Multi-physics
modeling of the fabrication and dynamic performance of all-metal auxetichexagonal sandwich-structures. Mater Des 2013;51:113e30.

K.T. Tan et al. / International Journal of Impact Engineering 64 (2014) 20e29

[11] Pendry JB. Negative refraction makes a perfect lens. Phys Rev Lett 2000;85:
3966e9.
[12] Valentine J, Zhang S, Zentgraf T, Ulin-Avila E, Genov DA, Bartal G, et al. Threedimensional optical metamaterials with negative refractive index. Nature
2008;455:376e9.
[13] Li J, Chan CT. Double-negative acoustic metamaterial. Phys Rev E 2004;70:
055602.
[14] Fang N, Xi D, Xu J, Ambati M, Srituravanich W, Sun C, et al. Ultrasonic metamaterials with negative modulus. Nat Mater 2006;5:452.
[15] Huang HH, Sun CT, Huang GL. On the negative effective mass density in
acoustic metamaterials. Int J Eng Sci 2009;47:610e7.
[16] Huang HH, Sun CT. Theoretical investigation of the behaviour of an acoustic
metamaterial with extreme Youngs modulus. J Mech Phys Solids 2011;59:
2070e81.
[17] Milton GW, Willis JR. On modications of Newtons second law and linear
continuum elastodynamics. Proc R Soc A 2007;463:855e80.

29

[18] Huang HH, Sun CT. Wave attenuation mechanism in an acoustic metamaterial
with negative effective mass density. New J Phys 2009;11:013003.
[19] Yao S, Zhou X, Hu G. Experimental study on negative effective mass in a 1D
mass-spring system. New J Phys 2008;10:043020.
[20] Tan KT, Huang HH, Sun CT. Optimizing the band gap of effective mass negativity in acoustic metamaterials. Appl Phys Lett 2012;101:241902.
[21] Liu Z, Zhang X, Mao Y, Zhu YY, Yang Z, Chan CT, et al. Locally resonant sonic
materials. Science 2000;289:1734e6.
[22] Lai Y, Wu Y, Sheng P, Zhang ZQ. Hybrid elastic solids. Nat Mater 2011;10:
620e4.
[23] Chen JS, Sharma B, Sun CT. Dynamic behaviour of sandwich structure containing spring-mass resonators. Compos Struct 2011;93:2120e5.
[24] Sun CT, Manimala JM, Huang HH. Development of negative effective mass
structures. Final technical report submitted as part of the NextGen Aeronautics Inc. prime contract for the DARPA Structural-Logic program; Mar
2013. p. 1e70.