pss b
solidi
Phys. Status Solidi B, 1–9 (2012) / DOI 10.1002/pssb.201084223
A three-dimensional rotating rigid
units network exhibiting negative Poisson’s ratios
,1 1,2
Daphne Attard* and Joseph N. Grima
1
Faculty of Science, Department of Chemistry, University of Malta, Msida Malta, MSD 2080, Malta
2
Faculty of Science, Metamaterials Unit, University of Malta, Msida Malta, MSD 2080, Malta
Received 29 June 2011, revised 25 April 2012, accepted 26 April 2012
Published online 25 May 2012
Keywords deformation, 3D models, negative Poisson’s ratio, rotating rigid units
* Corresponding author: e-mail auxetic@um.edu.mt, Phone: þ356 2340 2274, Fax: þ356 2540 1091,
Web: www.auxetic.info
Materials exhibiting auxetic behaviour get fatter when stretched
(i.e. possess a negative Poisson’s ratio). This property has been
closely related to particular geometrical features of a system
and how it deforms. One of the mechanisms which is known
to have a potential to generate such behaviour is that of
rotating rigid units. Several models based on this concept have
been developed, including two-dimensional as well as three-
1 Introduction When a material is uniaxially
stretched, the deformation along the stretching direction is
normally accompanied by changes in its lateral dimensions.
These changes are conveniently measured through the
Poisson’s ratio, a property which quantifies the changes in
the aspect ratios of particular planes of a material when it
is subjected to a uniaxial strain, ei, in that plane. More
specifically, the Poisson’s ratio nij in the Oxi–Oxj plane for
loading in the Oxi direction is defined as:
ej
nij ¼  ði; j ¼ 1; 2; 3Þ;
ei
where ej is the strain in the direction perpendicular to that of
loading. Everyday experience makes us intuitively assume
that materials must become thinner when uniaxially
stretched. In these cases the Poisson’s ratio assumes positive
values. But this is not always the case and in fact, it has long
been known that it is possible for a material to have a negative
Poisson’s ratio. Such materials, more commonly referred to
as auxetic [1], become fatter when stretched.
Apart from being an anomalous property, a negative
Poisson’s ratio is known to impart on materials several
enhanced characteristics such as increased resilience and
indentation resistance [2–6], pore size tunability [7–9] and
better energy absorption and acoustic properties [10–15].
physica
status
www.pss-b.com
basic solid state physics
dimensional (3D) models. In this work, we propose a new 3D
structure constructed from rigid cuboids which also deform
through relative rotation of the units. In particular, analytical
models for the mechanical properties, namely the Poisson’s
ratio and the Young’s moduli, are derived and it is shown that
for loading on-axis, these systems have the potential to exhibit
negative values for all the six on-axis Poisson’s ratios.
ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
These enhanced characteristics make a negative Poisson’s
ratio a highly desirable property in materials as they can be
used in several practical applications ranging from smart
filters to protective equipment.
In recent years various new auxetic materials or
structures have been designed, discovered and/or manufac-
tured, which materials range from macroscale honeycomb
structures [16–18] to nanoscale auxetic crystalline materials
such as silicates [19–27] and zeolites [28–31].
Research carried out in the past decades clearly suggests
that the sign and magnitude of the Poisson’s ratio may be
explained by looking at how particular geometric features
present in the nano- or microstructure of materials deform
when the material is subjected to a uniaxial stress (the
deformation mechanism). These observations triggered
various studies that have attempted to propose, develop or
improve mechanistic models based on particular geometries
which can result in negative Poisson’s ratios. These include
models based on flexure and hinging of re-entrant honey-
combs [2, 16, 17, 32–35], rotation of rigid and semi-rigid units
[23, 25, 36–40], models based on hard spheres [41–43], hard
disks [44, 45], polydisperse systems [46], chaotic systems [47]
dilatory/sliding models [23, 30, 48–54] and models based on
chiral geometries [18, 55, 56]. Of particular interest in this
respect is the fact that the Poisson’s ratio is a scale independent
ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 pss b

solidi
physica

status
2 D. Attard and J. N. Grima: 3D rotating rigid units network exhibiting negative Poisson’s ratios

property, such that the same geometry-deformation mechan- planes of symmetry and can be described in terms of two
ism model may operate at the macroscale, microscale as well hexagonal honeycomb cells connected perpendicular to each
as at the nanoscale. This is clearly illustrated by rotating rigid other with a face of a cuboidal unit sharing one of the faces of
units systems which have been applied to a wide range of the honeycombs, as shown in Fig. 1a, so that the structure
materials including microstructured polymeric auxetic foams contains a total of ten cuboids. Since the hexagonal cells can
where the rigid units represent mm–mm features in the foams be allowed to have different aspect ratios, cuboids having
[57, 58], and crystalline materials such as zeolites where different sizes can be used, although the connectivity
the rigid units correspond to features at the atomic (nano) reduces the number of independent geometric parameters. In
scale [28, 31, 40, 59]. particular, in the case considered here, three different types of
Using these rotating rigid units models, it has been cuboids are used: two a1  b1  c1,1 four a2  b2  c2 where
shown that auxetic behaviour may be achieved through two- c2 ¼ c1 and four a3  b3  c3 where a3 ¼ a1 arranged as shown
dimensional (2D) models involving rotation of rigid units in Fig. 1b, so that a total of eight geometric parameters
which may also be accompanied by additional modes of are required to fully describe the structure: a1, b1, c1, a2, b2,
deformation. Such 2D models are very useful as they may b3, c3 and u, which is defined as the angle between two
explain the Poisson’s ratios of real three-dimensional (3D) a2  b2  c2 cuboids. It is important to note that this structure
systems by looking at their 2D projection in some particular is tessellatable and in fact can be used as the basic unit cell to
plane. Nevertheless, it is obvious that such models also construct larger structures such as the one shown in Fig. 1c.
have their limitations, such as their inability to exhibit or Note also that since in the presented model, the cuboids
explain concurrent in-plane and out-of-plane auxeticity. are assumed to remain rigid through the deformation, the
Such properties require the use of more complex models geometric parameters ai, bi and ci remain fixed, which means
which must necessarily be of a 3D nature. that the shape and size of any given system will be dependent
The concept of using 3D structures to predict or explain on a single variable, u. In fact, although in Fig. 1b, the
auxetic behaviour is not new. In fact, some of the pioneering geometry is described in terms of u and w, these two
work in the field of auxetics was based on 3D structures. These parameters are dependent on each other, where
include the re-entrant structures proposed by Evans et al. [32],   
the tetrakaidecahedron with re-entrant features proposed by a2 u
’ ¼ 2 sin1 sin : (1)
Choi and Lakes [34], and the 3D rotating and/or dilating c3 2
tetrahedral model proposed and derived by Alderson et al.
[23–25, 37] The latter works are of particular significance not Referring to Fig. 1b, the structure is oriented in space such
only because they have shown that such models can explain that the two cuboids of dimensions a1  b1  c1 have their
extremely well the mechanical properties of real crystalline edges of length b1 aligned parallel to the Ox2 direction while
auxetic materials such as a-cristobalite but also because they the edges of length a1 and c1 are aligned along the Ox1 and Ox3
represent the first attempt to achieve auxetic behaviour in directions, respectively. In this way, one of the remaining two
terms of rotating rigid or semi-rigid units. sets of cuboids are then free to rotate in the Ox1–Ox2 plane
The work presented here aims to further expand on this while the other set can rotate in the Ox2–Ox3 plane.
earlier work by considering 3D models which are constructed Given this orientation, the projections, X1, X2 and X3
from cuboidal shaped units which are connected to each other along the Ox1, Ox2 and Ox3 directions for this unit cell can be
in such a way so as to permit relative rotation of the rigid expressed in terms of the geometric parameters and the
cuboids. In particular, an attempt is made to derive the on-axis variable u (or w) by:
Poisson’s ratios and Young’s moduli of these systems, which     
u u
may be considered to be the 3D equivalents of the 2D systems X1 ¼ a1 þ 2 a2 cos þ b2 sin ; (2)
studied before. Furthermore, an attempt will be made to show 2 2
how 2D models of projections can be used to predict the   
Poisson’s ratio of such systems in particular planes, thus u  ’
confirming the importance of both the 3D and much simpler X2 ¼ 2 b1 þ a2 sin  2 b1 þ c3 sin ; (3)
2 2
2D models to explain auxetic behaviour.
 ’ ’
2 Analytical models In this section, a system con- X3 ¼ c1 þ 2 c3 cos þ b3 sin
2 2
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
structed from cuboids connected at their edges which rotate    ! (4)
with respect to each other is proposed. Expressions for its 2 2 2 u b3 a2 u
 c1 þ 2 c3  a2 sin þ sin :
mechanical properties as a 3D structure are derived and it is 2 c3 2
shown that such a structure may exhibit auxetic behaviour
in any of the three directions. A simplified version of this model, Since the units are assumed to be rigid, the changes dX1,
which considers a 2D projection of the system is also presented. dX2 and dX3 along the Ox1, Ox2 and Ox3 directions can be
2.1 The three-dimensional model As illustrated in 1
It should be noted that cuboids of this type from adjacent cells are
Fig. 1, the unit cell of the modelled 3D structure has three connected together to form a single cuboid of length 2b1.

ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-b.com

 Original
Paper

Phys. Status Solidi B (2012) 3

Figure 1 (online colour at: www.pss-b.com) (a) Two hexagonal honeycombs connected perpendicularly to each other, to which cuboids
can be attached to generate the rotating cubes structure in (b). This structure is tessellatable and can be used to construct larger systems as
shown in (c).

e3
found by differentiating Eqs. (2)–(4) with respect to the n13 ¼ ðn31 Þ1 ¼ 
single variable u so that the resulting strains are given by: e1
   ’ ’
  u
dX1 1 dX11 a2 cos b3 cos  c3 sin
e1 ¼ ¼ du X1 2 2 2
X1 X1 du ¼
X3 ’  
u
  ;
u
(9)
     c3 cos b2 cos  a2 sin
1 u u 2 2 2
¼ b2 cos  a2 sin du; (5)
X1 2 2

      ’ ’
dX2 1 dX22 1 dX2
e2 ¼ ¼ du  d’ e3 X2 b 3 cos  c 3 sin
X2 X2 du X2 d’ n23 ¼ ðn32 Þ1 ¼ ¼ 2   2 :
   ’ (6) e2 X3 ’
a2 u c3 c3 cos
¼ cos du  cos d’; 2
X2 2 X2 2 (10)
 
dX3 1 dX3 1  ’ ’
e3 ¼ ¼ d’ ¼ b3 cos  c3 sin d’ 2.2 The two-dimensional models An alternative
X3 X3 d’ X3 2 2
  ’ ’   approach to model the properties of 3D systems such as the
u one presented in the previous section is to consider only
b3 cos  c3 sin a2 cos du
1 2 2 2 the 2D projection of a particular plane. Such an approach
  ’ :
X3 c3 cos has been used, for example to explain the auxeticity of
2 a-cristobalite in the (100) and (010) plane, a system
(7) constructed from 3D tetrahedra [23–25, 37], in terms of a
2D rotating rectangles model, where the rectangles are the
The corresponding Poisson’s ratios are then given by:
2D projection of the framework in these planes [26, 27].
  In this case, it is interesting to note that it is possible for
u
a2 cos the 3D structure described in the previous section to have
1 e2 X1 2
n12 ¼ ðn21 Þ ¼  ¼       ; simpler 2D counterparts. For example, for some particular
e1 X2 u u
b2 cos  a2 sin planes, the 2D model of the structure would be constructed
2 2 from rectangles connected at their corners instead of
(8) cuboids connected at their edges. As shown in Fig. 2, a 2D

www.pss-b.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

 pss b

solidi
physica

status
4 D. Attard and J. N. Grima: 3D rotating rigid units network exhibiting negative Poisson’s ratios

so that the Poisson’s ratio in the Ox1–Ox2 plane can be given

by:

’
 1 e22 X11 p2 cos
n12 ¼ n21 ¼ ¼  ’ 2 ’ :
e11 X22 q2 cos  p2 sin
2 2
(15)

3 Results and discussion Typical plots for the six

on-axis Poisson’s ratios of the 3D model are shown in Figs. 3
and 4. In these plots, we consider a total of nine different
cases which cover some of the possible configurations of the
system ranging from the simplest one in (a) which is made up
from identical cubes to a more generalised one in (i) made
Figure 2 (online colour at: www.pss-b.com) A two-dimensional from cuboids of different size. More specifically cases (b) to
version of the 3-D model proposed in this paper. (e) explore the effect of parameter bi on the Poisson’s ratio.
These cases represent structures where the units correspond
to cuboids having an identical square base (the side which
forms part of the honeycomb cell shown in Fig. 1a). In (b) the
representation of the system projected in the Ox1–Ox2 plane cuboids are all identical, i.e. their height bi is the same while
or the Ox2–Ox3 plane is one where the unit cell contains two in (c) and (d) one set of cuboids has a different height from
p1  q1 rectangles and four p2  q2 rectangles. In this case, the other two which are identical and in (e) each set of
the system has two lines of symmetry which are parallel to cuboids have a different height. Cases (f) to (h), on the other
the Ox1 and Ox2 directions. In such a symmetric system, five hand explore the effect of changing the dimensions of the
variables are required to define the geometry: p1, q1, p2, q2 faces of the honeycomb cell while the height of the cuboids bi
and f. If it is assumed that the rectangles remain perfectly are equal and kept fixed. In case (f) the horizontal face, i.e.
rigid during deformation then f will be the only variable the base of the vertical cuboids is a square while that of
parameter. Note that this structure in fact corresponds to the the other two sets of inclined cuboids are rectangular and
projection of one of the faces of the 3D structure oriented identical. In (g) all the cuboids are identical and with a
parallel to Ox1–Ox2 plane or the Ox2–Ox3 plane. rectangular base while in (h) each set of cuboids has different
If this planar structure is oriented in a similar way as its base dimensions.
3D version, i.e. the rectangles of dimensions p1  q1 are From the graphs one can immediately note that in cases
oriented such that q1 is aligned parallel to the Ox2 direction where the cuboids are not identical the range of values that u
while p1 is aligned parallel to the Ox1 direction whilst the can have may be limited. In fact, for geometries having b2
remaining rectangles are allowed to rotate in the Ox1–Ox2 or b3 greater than b1, the minimum value that u can have is
plane, then the dimensions X11 and X22 along the Ox1 and Ox2 limited by the conditions b2 cosðu=2Þ  b1 for b2 > b1 or
directions can be expressed as follows: b3 cos ð’=2Þ  b1 for b3 > b1. Similarly, there may also be
  ’  ’  an upper limit to the u-value which occurs for instances
X11 ¼ p1 þ 2 p2 cos þ q2 sin ; (11) where a2 > c3. For such geometries the maximum value
2 2 that u can have has to satisfy the condition that
sin ðu=2Þ  c3 =a2 . These conditions establish a range of u
 ’ values which will ensure that the structure is physically
X22 ¼ 2 q1 þ p2 sin : (12) realisable.
2 As illustrated in these plots, the equations derived here
suggest that the Poisson’s ratios for the proposed rotating
The strains which result from the deformation of the rigid cuboids structure can be such that it may simul-
structure are then given by: taneously exhibit negative values for all the six on-axis
Poisson’s ratios. In fact, it is evident from these plots, as well
dX11 1   ’ ’
as from the equations, that there is always a plane for which
e1 ¼ ¼ q2 cos  p2 sin d’; (13)
X11 X11 2 2 the Poisson’s ratio is negative and that it is possible to have
systems which exhibit auxeticity in all three directions
 ’ simultaneously. This is particularly true for the lower values
dX22 p2 of u. This is very significant as it clearly shows the potential
e2 ¼ ¼ cos d’; (14)
X22 X22 2 of such 3D models for achieving auxetic behaviour.

ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-b.com

 Original
Paper

Phys. Status Solidi B (2012) 5

Figure 3 (online colour at: www.pss-b.com) Variation of the Poisson’s ratios n12, n13 and n32 with the degree of openness (u) for various
rotating rigid cuboids systems showing that at low values of u, all the systems considered exhibit auxeticity in all three directions
simultaneously.

Before continuing any further, it is worth examining the
manner in which such 3D systems may be constructed. In
particular, it should be noted that on a 2D level, for the
rotating rigid unit mechanism to be able to operate, the units
must necessarily be connected at their corners. Contrary to
this, in systems consisting of 3D units, there is more than one
way how they can be connected to each other – either at their
corners, at their edges or even a combination of both. When
the units are connected at their corners, their movement is
much less constrained and can move in a similar way as if

Figure 4 (online colour at: www.pss-b.com) Variation of the Poisson’s ratios n21, n31 and n23 with the degree of openness (u) for various
rotating rigid cuboids systems showing that at low values of u, all the systems considered exhibit auxeticity in all three directions
simultaneously.

www.pss-b.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

 pss b

solidi
physica

status
6 D. Attard and J. N. Grima: 3D rotating rigid units network exhibiting negative Poisson’s ratios

connected by a ball joint, which means that this connection and obviously, since n21 is the reciprocal of n12 , the same
may impart on the model two degrees of freedom which may conditions apply for a negative n21 .
be represented through the two Euler angles. On the other If this information is transposed to the 3D model, then
hand, when connected at their edges, the units are this requirement becomes:
constrained to rotate in one plane only and thus only one  
degree of freedom may be associated with this mode of b2 u
n12 : > tan ; (17)
connection as is the case in the system presented here. In fact, a2 2
such ‘edge hinges’ can be considered as being analogous to
the hinges in 2D systems so that if one looks at the structure and
from a direction which is parallel to the hinging axis, then b3 ’
one would effectively observe a ‘2D system’ made from n23 : > tan ; (18)
c3 2
units connected at their corners. Although, prima facie this
may appear as offering no additional benefits to the 2D or in terms of u :
systems, it should be remembered that 3D polyhedral units   2  1=2
have edges lying in different planes, thus facilitating the b3 u c3 2 u
n23 : > sin  sin : (19)
construction of 3D systems. c3 2 a22 2
If one had to look at the equations for the mechanical
properties of the 2D and 3D models derived, one will note the Note that once again, the conditions for negative n21
remarkable similarity in form between the expressions for and n32 are the same as those for negative n12 and n23,
the Poisson’s ratios derived in the two approaches. In respectively, since these are the reciprocal of each other.
particular, the 2D expression for the Poisson’s ratio n12 has Thus, if both of these conditions are satisfied, then it is easy to
the same form as the equations for n12 and n32 in the 3D show that for loading in the Ox2 direction, the Poisson’s ratio
model and can be directly used to express these Poisson’s can be simultaneously negative in the Ox1–Ox2 plane and
ratios by replacing the 2D geometric parameters pi, qi and f Ox2–Ox3 plane. Also, in the case of the Poisson’s ratio in the
with the respective parameters ai, bi, ci, u and/or w. Similar Ox1–Ox3 plane, in the absence of the simple 2D model
arguments can be made for the Poisson’s ratios in these equivalent, one must look directly at the 3D model. Here,
planes for loading in the Ox2 direction. This is significant as it one notes that for negative n13, the requirements are similar
clearly shows the important role of the very simple 2D model to those above, and for it to be negative, there are two
to explain the properties of more complex 3D structures conditions which need to be satisfied simultaneously:
and gives added confidence to similar work based on this 8
approach [26, 27, 38, 40, 60]. However, it should also     2  !1=2
>
> b u b u c u
> > tan
> 2
and
3
> sin 3
 sin 2
be noted that it is not always possible to use this simple >
> a 2 c 2 2 2
>
> 2 3 a 3
approach. For example, it should be noted that the <
Poisson’s ratios in the Ox1–Ox3 plane, cannot be derived as n13 : or
>
>
easily in terms of a simple 2D model involving rotations. >
>
>     2  !1=2
>
> b2 u b3 u c3 2 u
In this respect, if one had to use the 2D approach, an >
: < tan and < sin  sin :
analysis of the deformations projected in the Ox1–Ox3 a2 2 c3 2 a22 2
plane would have indicated a dilation type mechanism, (20)
something which would be analogous to the pioneering
work by Wojciechowski and Branka [49] and Rothenburg Thus, it can be shown from the equations that given
et al. [48]. Obviously, this dilatory-like deformation is in that both of the conditions tan ðu=2Þ < b2 =a2 and
reality a consequence of the rotational deformations which tan ð’=2Þ < b3 =c3 are satisfied, then all the on-axis Poisson’s
occur in the other two orthogonal planes: the negative ratios are negative.
Poisson’s ratio in the Ox1–Ox3 plane is resulting due to the At this point, it is important to understand what happens
fact that the 3D topology of the system is such that the extent when the structure deforms in order to understand when it
of dilation of the elements parallel to the Ox1 direction can exhibit auxetic behaviour. When a tensile force is
(a 2D projection of units rotating in an orthogonal plane) applied along the Ox2 direction, the structure responds
are dependent on that of the elements parallel to the Ox3 by increasing u and w and the inclined cuboids rotate
direction (also a 2D projection of units rotating in an accordingly away from the vertical cuboids to occupy more
orthogonal plane). space. The largest possible dimensions of the unit cell in
If one analyses the expressions presented above in more the Ox1 direction is determined by the diagonal of the
detail, one may note that the 2D models suggests that for n12, inclined cuboids lying in the Ox1–Ox2 plane and correspond
since cosð’Þ is always positive, then a negative n12 requires to a configuration where the diagonal of the inclined cuboids
that the denominator in Eq. (15) is positive, i.e. lies horizontally, i.e. when tan ðu=2Þ ¼ b2 =a2 . Referring to
 ’ Eq. (8), these conditions in fact correspond to a continuous
q2 transition of the Poisson’s ratio n21 in the Ox1–Ox2
n12 : > tan ; (16)
p2 2 plane from a negative to a positive value. Thus, until the

ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-b.com

 Original
Paper

Phys. Status Solidi B (2012) 7

structure reaches this point, it is auxetic in the relevant plane.

However, if u allows, stretching any further will cause the
diagonal to go past this point so the respective unit cell
dimension decreases and the structure starts to behave
conventionally in this plane. Beyond this point the Poisson’s
ratio becomes increasingly positive, towards a value of:
" #
X2 b 2 c3
n21 ¼   pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ;
X1 a 2 a2 2  c 3 2
when a2 > c3 and to very high values when c3  a2 as u tends Figure 5 (online colour at: www.pss-b.com) The 3D system pro-
to its maximum value. The Ox2–Ox3 plane behaves in a posed and modelled by Grima [52] and Gaspar et al. [61].
similar way where the Poisson’s ratios tend to very high
values when a2  c3 and to:
" #
the cuboidal units, for example through the ‘chiral
X2 b 3 a2
n23 ¼   pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; mechanism’.
X1 c 3 c 3 2  a2 2 In view of all this, it is obvious that the 3D system
when c3 > a2 as u tends to its maximum value. proposed and modelled here is only one of the many
Similarly, on stretching in the Ox1 direction, u also structures which may be constructed from connected
increases and the Ox1–Ox2 plane is also auxetic. However, cuboids. It is not within the scope of this work to fully
when the structure reaches the conformation where the investigate and characterise such systems but Fig. 6
diagonal of the cuboids in this plane becomes aligned with illustrates one other system which can be constructed from
the horizontal, then the structure becomes locked in that cuboids, where the cuboids are connected at their corners
configuration and can be stretched no further. However, this rather than from their edges. Preliminary investigations
barrier can be overcome by applying a moment to the cuboids using finite element methods on such a system where the
to force them past this barrier after which the structure shows cuboids are ‘fused’ together suggest that it can also exhibit
conventional behaviour in this plane. In fact, this change in negative Poisson’s ratios.
sign of the Poisson’s ratio is characterised by an asymptotic Also, in this derivation, a highly idealised scenario was
transition at the point where b2 =a2 ¼ tan ðu=2Þ. Following considered since it was assumed that the cuboidal units
this transition, the Poisson’s ratio becomes highly positive (or their 2D rectangular projection) remain perfectly rigid,
and starts to decrease to a lower value of: something which in real scenarios is difficult to achieve. In
" #1 this respect it should be highlighted that the expressions for
X1 b 2 c3 the Poisson’s ratio derived here are scale independent, i.e.
n12 ¼   pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ;
X2 a 2 a2 2  c 3 2
if a2 > c3 or 0 if c3 > a2, as u reaches its maximum value. A
similar analysis also applies for loading in the Ox3 direction.
In the case of the Ox1–Ox3 plane, the Poisson’s ratio n13
can be expressed in terms of the Poisson’s ratio for the other
two planes as follows:
n12
n13 ¼ ðn31 Þ1 ¼  :
n32
This relation suggests that when n12 tends to 0 or n32
approaches very high magnitudes, n13 tends to 0. Similarly,
n13 tends to high magnitudes of Poisson’s ratio when n32
approaches 0 or when n12 tends towards large magnitudes.
Before concluding, it should be noted that 3D systems
may not only be constructed from polyhedra, but other
shapes can be used including 2D plates (example polygons),
and 1D fibres or rod-like units. For example, the system in
Fig. 5 modelled by Grima [52] and Gaspar et al. [61]
represents an example of one such 3D system made from
cuboids and fibrils, which system achieves auxeticity Figure 6 (online colour at: www.pss-b.com) A 3D system that can
through the re-entrant mechanism. Nevertheless, as be constructed from cuboids that are connected from their corners
explained by Grima [52], this system could easily be and which has a potential to exhibit auxetic behaviour. Different
converted to one which would have involved rotation of colours are used to distinguish adjacent unit cells.

www.pss-b.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

 pss b

solidi
physica

status
8 D. Attard and J. N. Grima: 3D rotating rigid units network exhibiting negative Poisson’s ratios

they can be used to downscale the design to the micro- or
nanoscale so as to achieve a system which can be considered
as an auxetic material rather than an auxetic structure. [13]
Furthermore, the models derived here may find use to explain
the auxeticity in already existent (man-made or naturally [14]
occurring) auxetics. Obviously, in such systems, it is even
[15]
more unrealistic to assume that behaviour is as idealistic as
the one described here, and it is clear that the model [16]
presented here would need to be extended to include
additional modes of deformation. [17]
4 Conclusions In this work, it has been shown that the [18]
concept of using rotating rigid units for achieving auxetic [19]
behaviour in 2D may easily be transposed to 3D hence
obtaining auxetic systems which may exhibit auxeticity in [20]
more than one plane. In fact, a simple system constructed
[21]
from rigid cuboidal units connected together at their edges
was proposed and it was shown through detailed analytical [22]
modelling that it may exhibit simultaneously negative
Poisson’s ratios in its three major planes. [23]
It was also shown that in such systems, there may still be
an important role for the simpler 2D modelling approach [24]
which considers only particular projections of the more
complex 3D systems, although in some cases (or projections) [25]
this may be more difficult to accomplish.
In view of the high extent of auxeticity of this system, [26]
particularly its ability to exhibit simultaneously negative
[27]
Poisson’s ratio nij (i,j ¼ 1, 2, 3), it is hoped that this work will
stimulate additional work which will attempt to design and [28]
construct real auxetics at any scale (the Poisson’s ratio is a
scale independent property) which are based on the models [29]
presented here.
[30]
References
[31]
[1] K. E. Evans, M. A. Nkansah, I. J. Hutchinson, and S. C.
Rogers, Nature 253, 124 (1991). [32]
[2] R. S. Lakes, Science 235, 1038–1040 (1987).
[3] N. Chan and K. E. Evans, J. Cellular Plast. 34, 231–260 [33]
(1998).
[4] K. L. Alderson, A. P. Pickles, P. J. Neale, and K. E. Evans, [34]
Acta Metall. Mater. 42, 2261–2266 (1994).
[5] K. L. Alderson, A. Fitzgerald, and K. E. Evans, J. Mater. Sci. [35]
35, 4039–4047 (2000).
[6] C. W. Smith, F. Lehman, R. J. Wootton, and K. E. Evans, [36]
Cell. Polym. 18, 79–101 (1999). [37]
[7] A. Alderson, J. Rasburn, S. Ameer-Beg, P. G. Mullarkey,
W. Perrie, and K. E. Evans, Ind. Eng. Chem. Res. 39, 654–
665 (2000). [38]
[8] A. Alderson, J. Rasburn, K. E. Evans, and J. N. Grima,
Membr. Technol. 137, 6–8 (2001). [39]
[9] J. Rasburn, P. G. Mullarkey, K. E. Evans, A. Alderson,
S. Ameer-Beg, and W. Perrie, AiChE J. 47, 2623–2626 [40]
(2001).
[10] C. P. Chen and R. S. Lakes, Cell. Polym. 8, 343–369 (1989). [41]
[11] C. P. Chen and R. S. Lakes, J. Eng. Mater. Technol. (ASME) [42]
118, 285–288 (1996).
[12] B. Howell, P. Prendergast, and L. Hansen, Acoustic Beha- [43]
viour of Negative Poisson’s Ratio Materials (David Taylor
Research Center, Bethesda Md, Ship Materials Engineering
Dept., 1991).
F. Scarpa and G. Tomlinson, J. Sound Vib. 230, 45–67
(2000).
F. Scarpa, L. G. Ciffo, and Y. R. Yates, Smart Mater. Struct.
13, 49–56 (2004).
F. Scarpa, W. A. Bullough, and P. Lumley, Proc. Inst. Mech.
Eng. C: J. Mech. Eng. Sci. 218, 241–244 (2004).
L. J. Gibson, M. F. Ashby, G. S. Schajer, and C. I. Robertson,
Proc. R. Soc. Lond. A 382, 25–42 (1982).
I. G. Masters and K. E. Evans, Compos. Struct. 35, 403–422
(1996).
F. Scarpa, Compos. Sci. Technol. 70, 1033 (2010).
A. Yeganeh-Haeri, D. J. Weidner, and J. B. Parise, Science
257, 650–652 (1992).
N. R. Keskar and J. R. Chelikowsky, Nature 358, 222–224
(1992).
N. R. Keskar and J. R. Chelikowsky, Phys. Rev. B 48, 227–
233 (1993).
H. Kimizuka, H. Kaburaki, and Y. Kogure, Phys. Rev. Lett.
84, 5548–5551 (2000).
A. Alderson and K. E. Evans, Phys. Chem. Miner. 28, 711–
718 (2001).
A. Alderson and K. E. Evans, Phys. Rev. Lett. 89, 225503
(2002).
A. Alderson and K. E. Evans, J. Phys.: Condens. Matter 21,
025401 (2009).
J. N. Grima, R. Gatt, A. Alderson, and K. E. Evans, J. Mater.
Chem. 15, 4003–4005 (2005).
J. N. Grima, A. Alderson, and K. E. Evans, Mater. Sci. Eng. A
423, 219–224 (2006).
J. N. Grima, R. Jackson, A. Alderson, and K. E. Evans, Adv.
Mater. 12, 1912–1928 (2000).
J. N. Grima, R. Gatt, V. Zammit, J. J. Williams, A. Alderson,
and R. I. Walton, J. Appl. Phys. 101, 86102 (2007).
J. N. Grima, P. S. Farrugia, C. Caruana, R. Gatt, and
D. Attard, J. Mater. Sci. 43, 5962–5971 (2008).
J. J. Williams, C. W. Smith, K. E. Evans, Z. A. D. Lethbridge,
and R. I. Walton, Acta Mater. 55, 5697–5707 (2007).
K. E. Evans, M. A. Nkansah, and I. J. Hutchinson, Acta
Metall. Mater. 42, 1289–1294 (1994).
K. E. Evans, A. Alderson, and F. R. Christian, J. Chem. Soc.
Faraday Trans. 91, 2671–2680 (1995).
J. B. Choi and R. S. Lakes, Int. J. Mech. Sci. 37, 51–59
(1995).
A. Bezazi, F. Scarpa, and C. Remillat, Compos. Struct. 71,
356–364 (2005).
O. Sigmund, Mech. Mater. 20, 351–368 (1995).
A. Alderson, K. L. Alderson, K. E. Evans, J. N. Grima, M. R.
Williams, and P. J. Davies, Comput. Methods Sci. Technol.
10, 117–126 (2004).
J. N. Grima and K. E. Evans, J. Mater. Sci. Lett. 19, 1563–
1565 (2000).
J. N. Grima, A. Alderson, and K. E. Evans, Phys. Status Solidi
B 242, 561–575 (2005).
J. N. Grima, V. Zammit, R. Gatt, A. Alderson, and K. E.
Evans, Phys. Status Solidi B 224, 866–882 (2007).
K. W. Wojciechowski, Phys. Lett. A 137, 60–64 (1989).
M. Kowalik and K. W. Wojciechowski, Phys. Status Solidi B
242, 626–631 (2005).
K. V. Tretiakov and K. W. Wojciechowski, Phys. Status
Solidi B 244, 1038–1046 (2007).

ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-b.com

 Original
Paper

Phys. Status Solidi B (2012) 9

[44] K. W. Wojciechowski, J. Phys. A 36, 11765–11778 (2003). [53] A. Alderson, Chem. Ind. 384–391 (1999).
[45] K. W. Wojciechowski, K. V. Tretiakov, and M. Kowalik, [54] P. V. Pikhitsa, Phys. Rev. Lett. 93, 15505 (2004).
Phys. Rev. E 67, 036121 (2003). [55] D. Prall and R. S. Lakes, Int. J. Mech. Sci. 39, 305–314 (1997).
[46] J. W. Narojczyk and K. W. Wojciechowski, Phys. Status [56] J. N. Grima, R. Gatt, and P. S. Farrugia, Phys. Status Solidi B
Solidi B 245, 2463–2468 (2008). 245, 511–520 (2008).
[47] V. V. Novikov and K. W. Wojciechowski, Phys. Solid State [57] J. N. Grima, A. Alderson, and K. E. Evans, J. Phys. Soc. Jpn.
41, 1970–1975 (1999). 74, 1341–1342 (2005).
[48] L. Rothenburg, A. A. Berlin, and R. J. Bathurst, Nature 354, [58] S. A. McDonald, N. Ravirala, P. J. Withers, and A. Alderson,
470–472 (1991). Scr. Mater. 60, 232–235 (2009).
[49] K. W. Wojciechowski and A. C. Branka, Mol. Phys. Rep. 6, [59] J. N. Grima, V. Zammit, R. Gatt, D. Attard, C. Caruana, and
71 (1994). T. G. Chircop Bray, J. Non-Cryst. Solids 354, 4214–4220
[50] R. Lakes and K. W. Wojciechowski, Phys. Status Solidi B (2008).
245, 545–551 (2008). [60] J. J. Williams, C. W. Smith, K. E. Evans, Z. A. D. Lethbridge,
[51] K. E. Evans and A. Alderson, Adv. Mater. 12, 617–628 (2000). and R. I. Walton, Chem. Mater. 19, 2423–2434 (2007).
[52] J. N. Grima, New Auxetic Materials, Ph.D. dissertation, [61] N. Gaspar, C. W. Smith, A. Alderson, J. N. Grima, and K. E.
University of Exeter (2000). Evans, J. Mater. Sci. 46, 372–384 (2011).

www.pss-b.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim