Biomech Model Mechanobiol (2011) 10:651–661
DOI 10.1007/s10237-010-0263-1
ORIGINAL PAPER
A novel experimental procedure based on pure shear testing
of dermatome-cut samples applied to porcine skin
Marc Hollenstein · Alexander E. Ehret ·
Mikhail Itskov · Edoardo Mazza
Received: 31 May 2010 / Accepted: 15 October 2010 / Published online: 11 November 2010
© Springer-Verlag 2010
Abstract This paper communicates a novel and robust
method for the mechanical testing of thin layers of soft biological tissues with particular application to porcine skin.
The key features include the use of a surgical dermatome
and the highly defined deformation kinematics achieved by
pure shear testing. Thin specimens of accurate thickness were
prepared using a dermatome and were subjected to different
quasi-static and dynamic loading protocols. Although simple
in its experimental realisation, pure shear testing provides
a number of advantages over other classic uni- and biaxial testing procedures. The preparation of thin specimens of
porcine dermis, the mechanical tests as well as first representative results are described and discussed in detail. The
results indicate a pronounced anisotropy between the directions along and across the cleavage lines and a strain ratedependent response.
Keywords Pure shear · Soft tissue testing · Dermatome ·
Skin · Dermis · Anisotropy · Preconditioning
M. Hollenstein (B) · E. Mazza
Institute for Mechanical Systems, ETH Zurich,
8092 Zurich, Switzerland
e-mail: [email protected]
A. E. Ehret · M. Itskov
Department of Continuum Mechanics, RWTH Aachen University,
Aachen, Germany
e-mail: [email protected]
M. Itskov
e-mail: [email protected]
E. Mazza
EMPA, 8600 Duebendorf, Switzerland
e-mail: [email protected]
1 Introduction
Mechanical characterisation of biological tissues plays an
essential role for the understanding of their behaviour and
functioning. Likewise, testing represents the initial step for
the development of constitutive models for these tissues,
which in turn form the basis for computer simulations in
many fields of research and engineering.
Ideally, biological materials are characterised in their
physiological environment, which provides the biological
and mechanical interactions with the surrounding tissues and
fluids. However, the possibilities for in vivo testing are limited by technical feasibility on the one hand and by ethical aspects on the other hand. Typical soft tissue testing
methods for in vivo application include indentation, pulling,
suction or compression tests, e.g. on skin (Dobrev 1999;
Delalleau et al. 2006) and porcine liver (Ottensmeyer 2002),
intraoperative indentation and aspiration experiments on
human liver (Carter et al. 2001; Nava et al. 2008), surface
torsion (Kalanovic et al. 2003; Valtorta and Mazza 2006),
pinching (Brouwer et al. 2001) and elastography methods
(Sandrin et al. 2003; Rouvière et al. 2006; Klatt et al. 2007).
Although similar to physiological loading conditions, the
kinematic states achieved in these tests are generally not welldefined and highly inhomogeneous. Accordingly, when these
experiments are used to determine the parameters of constitutive models, sophisticated and computationally intensive
inverse optimisation routines with numerous degrees of freedom are necessary and the results are often unsatisfactory.
In vitro tests provide the possibility to create experimental
data for kinematically rather simple load cases such as unior biaxial tension, compression or simple shear. Typically,
the respective mechanical boundary value problems can be
solved analytically and thus, based on these data, constitutive
models can be calibrated in a straightforward manner. In vitro
123
652 M. Hollenstein et al.
characterisation of biological tissues is based on methods
and protocols which are well-proven for testing of engineering materials. However, while the latter allow for tailoring
of specimens of any shape and size, the sample geometry for
biological tissues is restricted to its naturally grown dimensions and consequently often of a more primitive form.
A critical issue related to in vitro characterisation is
the artificial behaviour induced by sample preparation and
mounting. Waldman and Lee (2005) recently studied the
effect of sample geometry on results in biaxial tension experiments with cruciform specimens. In their study, the arm
length of the specimens had significant influence on the
apparent extensibility of the material as well as on shearing and rotation of the sample. Butler et al. (1984) and
Zernicke et al. (1984) discussed possible reasons for the large
discrepancies and variations in the reported behaviour of tendons and fascia in uniaxial tension. Apart from many other
factors, they identified the strain measurement technique to
have a major effect. According to their results, at a certain
maximum stress, the strains measured based on grip-to-grip
distance were approximately three times larger than those
measured optically in the central region of the specimen.
These differences may be explained by tissue inhomogeneities on the one hand and clamping effects such as slippage
or local tissue damage on the other hand. Clearly, also separation of the specimens from the tissue itself is prone to
induce artefacts. In fact, by excision, physiological interactions are broken, the tissue structure is locally changed (see
also Butler et al. 1984), artificial boundaries are created and
physiological residual stresses are released. Finally, storage
time and conditions between sample preparation and testing
may have significant effect on the tissue properties.
Uniaxial loading represents the simplest test method from
a technical point of view and has been used extensively for
a long time to characterise soft biological tissues of various
kinds. The load is usually applied such that the extension
occurs under a uniaxial stress state with free lateral contraction. Tensile specimens are realised in different aspect
ratios either as segments of the natural structure, e.g. tendons and muscle fascicles, or as strips cut from planar or
bulky organs such as skin and arterial walls (e.g. Holzapfel et
al. 2005). Compressive properties are studied in either constrained (e.g. Korhonen et al. 2002) or unconstrained (e.g.
Van Loocke et al. 2008) compression tests. A general problem with compression tests is friction between sample and
supports, which complicates the interpretation of the results
(cf. Miller 2005). Planar tissues are frequently tested in biaxial tests implemented either as biaxial loading of cruciform
(Waldman and Lee 2005) or rectangular specimens (Lanir
and Fung 1974; Sacks 2000), or in inflation tests (Wineman
et al. 1979). A major problem concerns the fixation of the planar specimens as discussed later in this paper. Finally, simple
shear provides a method to characterise soft tissues and has
for example been used to investigate myocardium (Dokos et
al. 2002) and ligaments (Bonifasi-Lista et al. 2005).
Facing the importance of experimental data under different loading conditions for the formulation of constitutive
models, it is surprising that the classic pure shear test, one of
the simplest available tests, is rarely used in biomechanics.
This test is usually realised by extending a thin rectangular
specimen with a high width-to-length ratio on a uniaxial testing device. In fact, to the best of our knowledge, there are
only two systematic applications of this test to soft tissues.
Chu et al. (1972) performed in vivo and in vitro measurements in pure shear on cat mesentery in order to analyse the
effects of perfusion, blood pressure, respiratory movements,
and excision of the samples from the underlying organ on
the obtained results. Recently, Gao et al. (2009) studied the
behaviour of bulky isotropic liver parenchyma and emphasised the simplicity and reliability of the pure shear test. We
remark that the deformation kinematics of pure shear has
been realised in more complicated testing set-ups, where the
lateral sample edges were fixed by sutures (Lanir and Fung
1974) or a biaxial testing device has been used (Sacks 2000)
to restrict the lateral contraction.
In our work, we apply pure shear to anisotropic planar
tissue samples of accurately defined thickness. Porcine dermis was chosen as a representative planar connective tissue
with properties that are characteristic of the majority of soft
collagenous tissues. These are strong non-linearity, very
slight compressibility, viscoelasticity, softening and pronounced anisotropy. The preferred directions of the anisotropy coincide approximately with the optically observable
cleavage lines of the skin (Cox 1942), which in turn result
from the underlying diamond shape arrangement of collagen
fibre bundles (Ridge and Wright 1966). Since there is a substantial amount of histological and biochemical similarities
between porcine and human skin, the pig represents a useful animal model for skin research such as wound-healing
studies (cf. Vardaxis et al. 1997).
2 Materials and methods
2.1 Sample preparation
The porcine dermis has a thickness of about 1–2 mm and is
located just underneath the 30–140 µm thick epidermis (cf.
Vardaxis et al. 1997). Dermal tissue is mainly composed of a
dense three-dimensional network of collagen fibres and fibre
bundles crossing each other in two main directions (Meyer
et al. 1982). Depending on the anatomical site, also a substantial amount of elastic fibres are present (Meyer et al.
1981). Specifically, collagen and elastin are reported to constitute about 70–80% and about 4% of the dry weight, respectively (Mathews 1975; Fung 1993). Beneath the dermis, there
123
Pure shear testing of porcine skin 653
follows the subcutaneous layer, which primarily consists of
fatty tissue and can take a thickness of over 12 mm depending on sex, anatomical site and nutritional state of the pig (cf.
Vardaxis et al. 1997).
Skin pieces from the snout and head region of female
domestic pigs (sus scrofa domestica) were obtained from
the slaughterhouse immediately after animal slaughter. The
pieces were approximately 300×300 mm large with a thickness of about 10–15 mm, leaving the tissue down to the subcutaneous layer intact. The skin pieces were transported in a
cold box and then stored at 4◦C in a refrigerator wrapped in
physiological saline-soaked cloths. The samples were tested
within 60 h of animal death. We remark that this time span
may influence the tissue properties and should be reduced in
a comparative study.
2.1.1 Preparation of the skin pieces
Skin pieces with a relatively uniform thickness were selected,
carefully rinsed with water until clean and then dried with
tissue paper. Subsequently, the pieces were shaved, first with
an electrical hair trimmer to remove larger bristles and finally
with a regular razor. After shaving, the cleavage lines (Langer
1861) became clearly visible and allowed for the estimation
of the local preferred directions of fibre orientation. Pieces
not directly further processed with the dermatome were again
stored in saline-soaked cloths at 4◦C.
2.1.2 Sample extraction by means of a dermatome
We used a wooden board with a rectangular foam damping pad of 80×120 mm attached to it as a mounting rack
for dermatome cutting. The foam pad provided a compliant
base for the skin pieces during cutting, similar to the soft fat
and muscle tissue compound supporting the skin in situ. The
cleaned and shaved skin pieces were mounted under slight
tension over the damping pad and nailed on the circumference to the wooden board.
For sectioning the skin into precisely defined layers, we
used an air-driven dermatome with a 80 mm width plate
installed (Fig. 1a). The dermatome-cut thickness was set to
500 µm. According to standard surgical procedures, paraffine oil was applied to the skin in order to reduce friction
between skin surface, dermatome and blade. The first layer
contained the epidermis and was disposed. Sections two and
three consisted completely of dermal tissue and were used
for testing. We note that the first layer contracted considerably after excision revealing the presence of residual strains
in situ. The effect was much less pronounced for the subsequent dermal layers.
Rectangular specimens were punched out of the dermis
sections by the use of a die cutter with dimensions 30×60mm
(Fig. 1b). For punching, the dermis sections were carefully
Fig. 1 a Separation of 500 µm thick dermis sections from skin pieces
by means of a surgical dermatome. b Cutting die for the pure shear
specimens
spread on saline-wetted tissue paper on an acrylic plate. The
die was aligned with its long side either along or perpendicular to the optically observable cleavage lines in order
to obtain transversal and longitudinal specimens. Pressing
the punching tool onto the plate, the specimens were excised
including the tissue paper as support.
For recovery from the preparation process, the ready-made
samples were stored, wrapped in saline-soaked towels at 4◦C
for at least 2 h. 15 min before testing, the samples were taken
to room temperature.
2.2 Experimental set-up
We performed the pure shear tests on a custom-made testing
device. Two hydraulic actuators, each with 2.7 kN capacity
and an available piston rod stroke of 100 mm, were mounted
opposing, horizontally on a massive steel plate. A hydraulic
power unit provided the two actuators with oil at 200 bar
continuous pressure. Load cells with a capacity of 100 N
were installed at the end of each actuator piston rod. Custom-made tissue clamps were directly attached to the load
cells. The clamps made of titanium offered a clamping surface of 60×10 mm equipped with sandpaper of grade P320
to improve the grip. For contactless measurement of the
in-plane strain field in the central region of the sample, a
video extensometer system for triggered image acquisition
was installed. The extensometer CCD camera equipped with
a telecentric lens was mounted on a custom-made positioning frame for easy and accurate alignment. With a field of
view of 32×32 mm, a resolution of 30 µm per pixel was
obtained. A 4” white light LED-ring for optimal illumination of the measuring field was mounted at the end of the
telecentric lens. The images were stored as 8-bit greyscale
pictures. To isolate the test rig from ambient vibrations, the
123
654 M. Hollenstein et al.
Table 1 Technical specifications as for the dermatome and the experimental set-up including the video extensometer system
Unit/software Type Manufacturer Specifications
Dermatome Air dermatome Zimmer
Hydraulic actuators 242 Actuator MTS Systems 2.7 kN capacity, 100 mm stroke
Hydraulic power unit 505 SilentFlo 200 bar pressure, 40 l/min capacity
External controller 793 FlexTest GT
Control software 793 with MPT
Load cells SMT S-Type Interface 100 N capacity
Video extensometer system uniDAC FAST Chemnitzer Werkstoffmechanik
Image acquisition software VEDDAC cam 3.2
Image correlation software VEDDAC 4.0 Digital image correlation, sub-pixel algorithm
CCD camera Pike F-100B Allied Vision Technologies 2/3” monochrome CCD, 1,000×1,000 pixels,
60 Hz frame rate, 8 bit greyscale pictures
Telecentric lens NT55-349 Edmund Optics 32×32 mm field of view
Vibration control table RS 2000, I-2000 Newport
Manufacturer details: Zimmer Inc., Warsaw, IN, USA; MTS Systems Corp., Eden Praire, MN, USA; Interface Inc., Scottsdale, AZ, USA; Chemnitzer Werkstoffmechanik GmbH, Chemnitz, Germany; Allied Vision Technologies GmbH, Stadtroda, Germany; Edmund Optics GmbH, Karlsruhe,
Germany; Newport Corp., Irvine, CA, USA
actuator positioning plate was fixed on a vibration control
table. The technical details are summarised in Table 1.
2.3 Test realisation
The load cells were reset just before clamping of the tissue
samples. A small supporting table was installed in between
the clamps once the actuator piston rods were brought to their
initial testing position (Fig. 2a), leaving a free gauge length
of 10 mm. This table allowed precise and gentle spreading
of the specimens on the support and the lower grip faces and
carfully removing the tissue paper support with forceps.

clamping support (Fig. 2b), the sample was hanging slightly
slack in-between the two clamps. To straighten the sample,
the active actuator was retracted at 0.05 mm/s until a preload
of 0.05 N was reached. This state demonstrated to be well
reproducible, and we defined it as the reference state. Nominal strains and stresses were computed with respect to this
state. The time between beginning of sample installation and
test onset did not exceed 5 min.
The force acting on the passive actuator and the displacement of the active actuator relative to its initial position were
where the initial cross-sectional area A0 (0.5×60 mm2) was
given, and the clamp displacement of the active actuator d(t),
the free gauge length at preload l0 and the tensile force on
the passive actuator F(t) were measured.
2.4 Pure shear test protocol
We performed the tests in displacement-controlled mode,
where the protocol was defined based on nominal strain and
123
Pure shear testing of porcine skin 655
Fig. 3 Test protocol for the quasi-static preconditioning (Q) and cyclic
loading (V). Nominal strain control is plotted over test time. While the
upper load reversal points were predefined, the lower reversal points
were enforced when the measured force became smaller than 0.05 N in
order to prevent the tissue from becoming slack
Table 2 Strain level and strain rate specifications for the test protocol
(Fig. 3)
q1 q2 q3 q4 q5 v1 v2
[-] [-] [-] [-] [-] [-] [-]
0.057 0.086 0.114 0.143 0.171 q3 q4
r0 r1 r2 r3 r4 r5
[1/s] [1/s] [1/s] [1/s] [1/s] [1/s]
0.001 0.002 0.005 0.01 0.02 0.05
strain rate. The lower cycle reversal points were enforced
when the measured force became smaller than 0.05 N in
order to prevent the tissue from becoming slack during the
cyclic loading.
The protocol consisted of a quasi-static preconditioning
of 5 cycles at strain rate r0 on the strain levels q1 to q5 (Q)
and 5 ramp cycles at increasing strain rates r1 to r5 on the
strain levels v1 and v2, respectively (V, Fig. 3 and Table 2).
The highest strain level q5 was chosen clearly below the
rupture strain in preceding tension to failure experiments,
and q1 to q4 were obtained by appropriate scaling. In order
to determine the strain rate r0 for quasi-static testing, we ran
successive load cycles with gradually decreasing velocity.
Below a certain limit, a further decrease in the strain rate
had negligible effect on the force response. A strain rate of
r0 = 0.001s–1 was identified to be sufficiently small.
2.5 Deformation kinematics
We consider the uniform extension of a body along three
orthogonal axes specified by the orthonormal vectors ei, i =
1, 2, 3. In this case, the deformation gradient takes the form

where λi are the principal stretches. Assuming that the deformation is isochoric, they are related by the incompressibility
constraint
J = det F = λ1λ2λ3 = 1. (3)
In the experiments, a thin specimen in the plane spanned
by e1 and e2 is stretched by λ in the direction of e1. Considering samples with a high aspect ratio, i.e. the ratio between free
length and width, uniform extension with free lateral contraction is obtained. Assuming incompressible behaviour (3), the
principal stretches are given by λ1=λ, λ2 and λ3=1/(λ1λ2).
On the other hand, with a decreasing aspect ratio, lateral contraction in the e2-direction becomes increasingly restricted,
so that in the limit λ2=1. This state is referred to as pure
shear or strip biaxial extension. The respective stretches are
given by λ1=λ, λ2=1, λ3=1/λ.
In the present study, pure shear specimens with an aspect
ratio of 1:6 were employed, which is smaller than the one
recently applied by Gao et al. (2009), who worked with an
aspect ratio of 1:4.
2.6 In-plane strain evaluation
A commercial image correlation software was used to trace
a set of user-defined measuring points over the video extensometer images and determine their displacements. Based
on this discrete displacement field, the in-plane strains were
approximated by use of the finite element method.
The in-plane position x (X, t) of a material particle of the
specimen in the current configuration at time t is given by
x (X, t) = X + u (X, t) , (4)
where X is the in-plane position of the particle in the reference configuration, and u (X, t) is the corresponding in-plane
displacement. The in-plane part F2D of the deformation gradient (2) is thus obtained by

where I2D is the second-order identity tensor in two dimensions. The finite element interpolation of the displacement
field based on a linear quadrilateral element yields
u (X) =
4a=1
N (a)u(a), (6)
123
656 M. Hollenstein et al.
Fig. 4 Sketch of the quadrilateral element used to calculate in-plane
strains. The nodal positions and displacements were obtained by averaging over nine adjacent measuring points (pixels)
with u(a) being the in-plane displacements of the element
nodes with respect to an image coordinate system and N (a)
representing the shape functions. The deformation gradient
(5) is thus approximated by

a=1
The explicit dependence on time t has been omitted for
the sake of brevity.
In this work, we used one single element adequately
spanned over the free area of the specimens. To reduce the
influence of stochastic errors, the positions and displacements of the four element nodes were obtained by averaging
over nine adjacent pixels arranged in a 3×3 pattern, respectively (Fig. 4).
3 Results
A representative set of data is presented for a longitudinal
and a transversal sample, which were both obtained from the
same skin piece. The time span between these two tests was
less than 1.5 h.
3.1 Quasi-static preconditioning
A distinct anisotropic response of the tissue was observed
(Fig. 5). The tissue was considerably stiffer along the cleavage lines than perpendicular to them. The specimens showed
substantial preconditioning effects such as softening,residual
deformations and a tendency to stabilise after some cycles.
Even for loading at low strain rates, the stabilised loading–unloading cycles demonstrate substantial hysteresis, a
phenomenon usually referred to as pseudo-elasticity in bio-
(b)
(a)
Fig. 5 Nominal stress response of a longitudinal (a) and transversal
(b) porcine dermis sample to the quasi-static preconditioning protocol
Q in a pure shear experiment
mechanics (Fung 1993). All response curves are characterised by the J-shape typical for the majority of soft biological
tissues with a clearly distinguished linear region at higher
stretch ratios.
3.2 Viscoelastic behaviour
The response of the dermal specimens generally depended
on the strain rate (Fig. 6). It was observed that during cyclic
loading, the response tends to become stiffer with increasing
strainrate(Fig. 6 top). Similar behaviour was observedfor the
peak stresses reached at the peak strains v1 and v2 after stabilisation. For further investigation of this issue, we considered
the ratio between these peak stresses Pi and P1 in the respective last cycle for each strain rate ri, i = 1, 2, 3, 4. Plotting
this ratio against the strain rate suggests that the influence
of the strain rate on the peak stress decreases with the strain
level (Fig. 7). The first five cycles at v1 and v2 with a strain
rate of 0.002 s–1 deserve closer attention. While at the level
123
Pure shear testing of porcine skin 657
(b)
(a)
Fig. 6 Nominal stress response of a longitudinal and transversal porcine dermis sample to the protocol V at strain levels v1 (a) and v2 (b).
The black arrows indicates the direction of stabilisation during the five
cycles at a strain rate of 0.002 s–1
v1 the curves stabilise towards a stiffer behaviour, only slight
stabilisation to a softer behaviour occurs at the higher strain
level v2. If noticeable, the stabilisation direction is indicated
by arrows in Fig. 6.
We remark that the data for the strain rate r5 were not
considered in the results due to a too small number of data
points per cycle.
3.3 Comparison of nominal and in-plane strains
Nominal (εn) and optically measured in-plane strains (εpa)
were compared based on the difference ρa = εpa – εn. Likewise, the deviation ρl of the optically determined lateral strain
εl
p from the ideal case of zero contraction was considered, so
that ρl = εpl . Thereby, the in-plane strains were determined
from the deformation gradient (7) evaluated at the element
centroid according to Sect. 2.6.
Fig. 7 Dependence of the peak stress in the fifth loading cycle on
strain rate for the two strain amplitudes v1 and v2. The graph displays
the ratio between the peak stresses Pi obtained in the last load cycle for
each strain rate ri , i=1, 2, 3, 4, respectively, and the peak stress P1 in
the slowest cycle with rate r1
Figure 8 shows the progression of ρa and ρl at the reversal points of the quasi-static preconditioning protocol (Q)
for one representative sample. The difference between the
nominal and the in-plane strain is more or less symmetrically distributed around zero and does not exceed 0.0068
(5% of the applied nominal strain). This indicates that the
observed deviations are supposedly of a stochastic nature,
in addition influenced by the digital image correlation on
wet tissue surfaces. Thus, no systematic slippage of the sample could be detected with the current evaluation method,
which implies that the clamping worked satisfactorily. Furthermore, ρl does not exceed 0.009 (5% of the applied nominal strain).
Fig. 8 Difference ρa between nominal and measured axial in-plane
strains and lateral in-plane strains ρl evaluated at the reversal points of
the quasi-static preconditioning protocol (Q)
123
658 M. Hollenstein et al.
4 Discussion
4.1 Mechanical behaviour of porcine dermis
Like the majority of soft biological tissues, porcine dermis
shows transient preconditioning behaviour in cyclic loading
experiments. The softening characteristics resemble observations in rubber-like materials (see e.g. Itskov et al. 2006). This
concerns in particular the dependence on the strain history
known as Mullins effect (Mullins 1947) and the substantial
amount of residual strains accumulating during the course
of the experiment. Similar behaviour has been reported for
murine skin samples excised longitudinally from the abdominal wall (Muñoz et al. 2008). The observed softening may
be explained by damage on the one hand and alterations of
the microstructure on the other hand. Since softening occurs
even at small strains within the range of physiological loading, damage is unlikely in this domain, well possible, however, for higher strains. Microstructural alterations comprise
particularly the alignment, reorientation and uncrimping of
the collagen network (see e.g. Sellaro et al. 2007). The high
content of collagen in the dermal layer might thus as well
explain the remarkably large residual strains observed for
this type of tissue.
Cyclic loading at varying strain rates reveals viscoelastic properties of the dermal tissue. Soft biological tissues are
often reported to be insensitive to strain rate over a wide range
(e.g. Fung 1993). Our results suggest that the sensitivity to
strain rate depends on the strain amplitude applied during the
cyclic loading. Generally, we observed an increase of both,
the overall stiffness (Fig. 6) and the peak stress (Fig. 7), with
respect to strain rate. The rate dependence was, however, less
pronounced for cyclic loading at the higher strain level v2.
Moreover, the slope of the peak stress decreases with strain
rate (Fig. 7) and, indeed, the curves may approach a limit at
higher rates. Also, the different stabilisation behaviour during the initial cycles at the strain rate of 0.002 s–1 (Fig. 6)
suggests a dependence of the viscoelastic behaviour on the
strain amplitude.
The observed strain and strain rate-dependent viscoelastic
properties suggest different mechanisms taking place on the
microstructural level. For moderate strains, the interaction
of the proteoglycan-rich ground substance with the orienting
and uncrimping collagen fibres might dominate the response.
For large strains or at high loading rates, where the fibres are
not given the time for uncrimping in the viscous gel-like
ground substance, the mechanical properties of the collagen
fibres itself may become crucial. These assumptions serve to
explain the observed behaviour and are supported by recent
experimental findings in tendons (Puxkandl et al. 2002).
Comparing the longitudinal and transverse response
(Fig. 7), the fibre orientation appears to have a minor effect on
the increase in the peak stress ratio indicating that anisotropy
is substantially less pronounced in the viscoelastic than in the
purely elastic properties.
4.2 Pure shear testing in comparison with uniaxial and
biaxial tension
Pure shear testing provides a number of advantages over classic uni- and biaxial testing procedures. Since most soft tissues are subject to biaxial loading in vivo, the pure shear test
is much closer to physiological conditions than the uniaxial tension test, while maintaining experimental simplicity.
The sample size of 30×60 mm with a final testing aspect
ratio of 1:6 turned out to be a good compromise between
approximation of the ideal case of pure shear and optimal
use of the available sample material. This is confirmed by the
optical measurements of the in-plane strains (cf. Sect. 3.3).
Although we did not measure stresses in the lateral direction,
their presence was indicated by a specimen that failed in a
fracture mode, in which the tissue ruptured perpendicular to
the axis of loading (Fig. 9).
The test can be performed on a standard uniaxial testing machine by adapting the clamping interfaces to mount
pure shear samples. Fixation of the samples itself is easily accomplished by metal grips equipped with sandpaper
strips. The wide shape of the pure shear specimens, however, allows for an enlarged contact area between sandpaper
and tissue. Accordingly, better fixation is obtained as a result
of the increased frictional area, see Sect. 3.3. Moreover, the
shorter free lateral boundaries in comparison with uniaxial
tension specimens imply a reduced influence of the sample
edges, which are particularly prone to be flawed by sample
preparation as discussed in Sect. 1.
Fig. 9 Failure of a transversal specimen: the tissue ruptured perpendicular to the axis of loading, which indicates the presence of significant
stresses in the lateral direction during the pure shear tests
123
Pure shear testing of porcine skin 659
Uniaxial tension specimens are not trivial to handle. For
example, their lateral edges tend to roll in due to slight
residual stresses in the tissue, and when mounting the samples to the testing device, already gravity deforms the soft
specimens. This is not only cumbersome from a technical point of view but also complicates the definition of the
initial dimensions of the specimen such as length, width
and cross-sectional area, which turns out to be a challenging issue. After mounting the specimen, a slight preload is
applied, which is negligible in comparison with the target
loads reached in the course of the experiments. The dimensions measured at this preload are taken as reference values for the calculation of the strains. As a result of the
low initial stiffness of soft biological tissues and the small
cross-sectional area of uniaxial tensile specimens, already
these small forces can cause substantial deformations. Moreover, these deformations are to a large extent affected by the
above-mentioned cutting artefacts. As a result, the definition of the reference state is somewhat imprecise and not
well repeatable, which causes variations in the stress–strain
curves of different specimens. Since pure shear specimens
of the same thickness provide much more resistance, both,
to the unavoidable initial loads and the applied preload, the
problems mentioned above are considerably less significant.
From a kinematical point of view, stress or strain controlled biaxial tests provides a useful testing modality which
allows for different ratios between tension or extension in
two orthogonal in-plane directions. In fact, these tests are
standard in many fields of material characterisation such as
rubber or metal testing. For soft biological tissues, however,
sample preparation, and in particular mounting, poses a fundamental challenge. Either cruciform specimens are rigidly
clamped in a testing device with orthogonal loading axes
(Waldman and Lee 2005) or square cut samples are flexibly
fixed by sutures (Lanir and Fung 1974; Demer and Yin 1983),
surgical staples (Sacks 1999), tiny fishing hooks (Boriek et al.
2000) or magnetic tapes (Humphrey et al. 1986). The preparation of cruciform specimens requires a relatively large
section of the tissue with preferably homogeneous properties. Moreover, the arms of the cruciform specimens affect
the strain field in the central sample area of interest to a large
extent so that cases with big strain differences in the two
extension directions are difficult to realise. Flexible mounting by sutures or hooks, on the other hand, induces strong
local deformations, an inhomogeneous strain field as well as
damage at the points where the fixation devices enter the tissue. At higher loads, stress concentration around these points
may cause premature failure. In both gripping techniques,
inhomogeneity and local anisotropy of the tissue may lead to
shearing and rigid rotations of the tissue (see the discussion
in Waldman and Lee 2005).
A major advantage of pure shear tests is the highly defined
kinematic state obtained. A homogenous state of deformation
provided, the only degree of freedom is given by the controllable axial stretch, while the lateral direction is constrained
and the thickness change follows from the incompressibility
constraint. This clearly simplifies the parameter identification process.
As a critical remark, we would like to emphasise that pure
shear tests alone are not sufficient to fully characterise the
anisotropic behaviour of soft biological tissues, just as little
as other standard tests (see Holzapfel and Ogden 2009, for a
discussion of this issue). However, well-defined experimental procedures and accurate data form an essential basis.
In the current study, the tests were not conducted submerged in a saline bath, which would have provided optimal
humidification and a controllable temperature. The oscillatory movement of the clamps would have caused wave
formation and rendered optical displacement measurement
impossible. Nevertheless, the saline solution sprayed onto
the specimen accumulated in the narrow gap between the
metal grips and formed a closed fluid film on top of the sample so that satisfactory wetting of the tissue was ensured.
4.3 Testing of dermatome-cut tissue samples
Mechanical characterisation of biological tissues in vitro is a
controversial issue. On the one hand, excision of the sample
from its physiological surroundings changes the mechanical properties, which leads to an immediate relief of residual stresses and creates artificial tissue boundaries. On the
other hand, in vitro testing allows for standardised and
well-defined experimental conditions. As a necessary
requirement, the specimens have to be prepared carefully
with accurate geometry. The dermatome allows to create tissue sections of small uniform thickness in the range between
100–1,000 µm, depending on the used dermatome. Moreover, the punching tools guarantee a precise geometry of
the specimens in the loading plane. The dermatomes are
designed to perform extremely smooth sectioning in plastic and reconstructive surgery and for this reason are predestined for the preparation of evenly shaped specimens without
significant damage. Dermatomes and blades exist in different sizes so that the method may be applied to a variety of
tissues. The cutting technique is simple, fast and does not
require surgical skills.
In particular, in the example of the skin, the proposed
method allows for a layer-by-layer characterisation in order
to create a detailed three-dimensional mechanical map of
the tissue. Beyond that, the thin, fairly translucent specimens are suitable for further histological investigations and
microscopical studies. This includes light scattering methods
(Ferdman and Yannas 1993) as well as polarisation and confocal laser scanning microscopy (Meyer et al. 1982; Vardaxis
et al. 1997).
123
660 M. Hollenstein et al.
5 Conclusions
We presented a robust and highly defined testing method for
soft biological tissues. The key features include the use of a
surgical dermatome and pure shear testing. The results suggest that the ideal pure shear state was adequately achieved
by simple and inexpensive means with a standard materials
testing machine available in most laboratories. As a novel
method for sample preparation, a surgical dermatome was
used in combination with punching dies. This allowed to
excise specimens of very accurate thickness and geometry.
In the present work, this preparation procedure was applied to
porcine skin, but it is applicable to a variety of other soft tissues. The precise geometry of the samples and the accurate
control over initial and boundary conditions (cf. Sect. 4.2)
in pure shear indicate a general reduction in the systematic
errors in comparison with classic uniaxial and biaxial testing.
This will have to be proven statistically on a broader database
in the future. This concerns both, the defined slicing procedure as well as the pure shear testing. A first set of pure shear
data on porcine dermis is presented herein and is available
for interpretation. The observed behaviour is to a large extent
consistent with typical characteristics of soft tissues. A more
detailed analysis of the results incorporating modelling of
the dermis by means of an adequate constitutive formulation
will follow and be provided elsewhere.
Acknowledgments We are grateful to Prof. Dr. med. A. Prescher,
University Hospital Aachen, for advising us of the use of a dermatome
in order to cut thin skin sections. We thank Dr. med. M. Guggenheim,
University Hospital Zurich, for providing the dermatome, explaining its
use and for his kind support concerning questions on anatomy and histology of the skin. Partial funding was provided by the Swiss National
Science Foundation (NCCR CO-ME). The authors would also like to
gratefully acknowledge both anonymous reviewers for many helpful
critical comments on the first version of this paper.
References
Bonifasi-Lista C, Lake SP, Small MS, Weiss JA (2005) Viscoelastic
properties of the human medial collateral ligament under longitudinal, transverse and shear loading. J Orthop Res 23(1):67–76
Boriek AM, Kelly NG, Rodarte JR, Wilson TA (2000) Biaxial constitutive relations for the passive canine diaphragm. J Appl Physiol
89(6):2187–2190
Brouwer I, Ustin J, Bentley LS, A Dhruv N, Tendick F (2001) Measuring in vivo animal soft tissue properties for haptic modeling in
surgical simulation. Stud Health Technol Inform 81:69–74
Butler DL, Grood ES, Noyes FR, Zernicke RF, Brackett K
(1984) Effects of structure and strain measurement technique on
the material properties of young human tendons and fascia. J Biomech 17(8):579–596
Carter FJ, Frank TG, Davies PJ, McLean D, Cuschieri A (2001) Measurements and modelling of the compliance of human and porcine
organs. Med Image Anal 5(4):231–236
Chu BM, Fraher WG, Wayland H (1972) Hysteretic behavior of soft
living animal tissue. Ann Biomed Eng 1(2):182–203
Cox HT (1942) The clevage lines of the skin. Br J Surg 29(144):234–
240
Delalleau A, Josse G, Lagarde J, Zahouani H, Bergheau J (2006) Characterization of the mechanical properties of skin by inverse analysis
combined with the indentation test. J Biomech 39(9):1603–1610
Demer LL, Yin FCP (1983) Passive biaxial mechanical properties of
isolated canine myocardium. J Physiol 339:615–630
Dobrev HP (1999) In vivo study of skin mechanical properties in
patients with systemic sclerosis. J Am Acad Dermatol 40(3):436–
442
Dokos S, Smaill B, Young A, LeGrice I (2002) Shear properties of passive ventricular myocardium. Am J Physiol Heart Circ Physiol
283(6):H2650–H2659
Ferdman A, Yannas I (1993) Scattering of light from histologic sections: a new method for the analysis of connective tissue. J Invest
Dermatol 100(5):710–716
Fung YC (1993) Biomechanics: mechanical properties of living tissues:
2nd edn. Springer, New York
Gao Z, Lister K, Desai J (2009) Constitutive modeling of liver tissue:
experiment and theory. Ann Biomed Eng 83(2):502–516
Holzapfel GA, Ogden RW (2009) On planar biaxial tests for anisotropic
nonlinearly elastic solids. A continuum mechanical framework.
Mathe Mech Solids 14:474–489
Holzapfel GA, Sommer G, Gasser CT, Regitnig P (2005) Determination of layer-specific mechanical properties of human coronary arteries with nonatherosclerotic intimal thickening and
related constitutive modeling. Am J Physiol Heart Circ Physiol
289(5):H2048–H2058
Humphrey JD, Vawter DL, Vito RP (1986) Mechanical-behavior of
excised canine visceral pleura. Ann Biomed Eng 14(5):451–466
Itskov M, Haberstroh E, Ehret A, Voehringer M (2006) Experimental
observation of the deformation induced anisotropy of the Mullins
effectin rubber. KGK Kautschuk Gummi Kunststoffe 3(3):93–96
Kalanovic D, Ottensmeyer M, Gross J, Buess G, Dawson S (2003) Independent testing of soft tissue visco-elasticity using indentation and
rotary shear deformations. Stud Health Technol Inform 94:137–
143
Klatt D, Hamhaber U, Asbach P, Braun J, Sack I (2007) Noninvasive
assessment of the rheological behavior of human organs using
multifrequency mr elastography: a study of brain and liver viscoelasticity. Phys Med Biol 52(24):7281–7294
Korhonen RK, Laasanen MS, Töyräs J, Rieppo J, Hirvonen J, Helminen
HJ, Jurvelin JS (2002) Comparison of the equilibrium response of
articular cartilage in unconfined compression, confined compression and indentation. J Biomech 35(7):903–909
Langer K (1861) Zur Anatomie und Physiologie der Haut. I. Über
die Spaltbarkeit der Cutis In: British Journal of Plastic Surgery
31, 3–8 (1978). Sitzungsbericht der Mathematisch-naturwissenschaftlichen Classe der Kaiserlichen Academie der Wissenschaften
44(1):19–46
Lanir Y, Fung YC (1974) Two-dimensional mechanical properties of
rabbit skin-II. Experimental results. J Biomech 7(2):171–182
Mathews MB (1975) Connective tissue, molecular biology, biochemistry and biophysics. Springer, Berlin
Meyer W, Neurand K, Radke B (1981) Elastic fibre arrangement in the
skin of the pig. Arch Dermatol Res 270(4):391–401
Meyer W, Neurand K, Radke B (1982) Collagen fibre arrangement in
the skin of the pig. J Anat 139(148):134
Miller K (2005) Method of testing very soft biological tissues in compression. J Biomech 38(1):153–158
Mullins L (1947) Effect of stretching on the properties of rubber. J Rubber Res 16:275–289
Muñoz M, Bea J, Rodríguez J, Ochoa I, Grasa J, Pérez del Palomar A,
Zaragoza P, Osta R, Doblaré M (2008) An experimental study of
the mouse skin behaviour: damage and inelastic aspects. J Biomech
41(1):93–99
123
Pure shear testing of porcine skin 661
Nava A, Mazza E, Furrer M, Villiger P, Reinhart W (2008) In vivo
mechanical characterization of human liver. Med Image Anal
12(2):203–216
Ottensmeyer M (2002) Tempest I-D: an instrument for measuring solid
organ soft tissue properties. Exp Tech 26(3):48–50
Puxkandl R, Zizak I, Paris O, Keckes J, Tesch W, Bernstorff S, Purslow
P, Fratzl P (2002) Viscoelastic properties of collagen: synchrotron
radiation investigations and structural model. Philos Trans R Soc
B 357(1418):191–197
Ridge MD, Wright V (1966) The directional effects of skin. A bio-engineering study of skin with particular reference to Langer’s lines. J
Invest Dermatol 46(4):341–346
Rouvière O, Yin M, Dresner M, Rossman P, Burgart L, Fidler J,
Ehman R (2006) MR elastography of the liver: preliminary results.
Radiology 240(2):440–448
Sacks MS (1999) A method for planar biaxial mechanical testing that
includes in-plane shear. J Biomech Eng 121(5):551–555
Sacks MS (2000) Biaxial mechanical evaluation of planar biological
materials. J Elast 61(1–3):199–246
Sandrin L, Fourquet B, Hasquenoph J, Yon S, Fournier C, Mal F,
Christidis C, Ziol M, Poulet B, Kazemi F, Beaugrand M, Palau
R (2003) Transient elastography: a new noninvasive method for
assessment of hepatic fibrosis. Ultrasound Med Biol 29(12):
1705–1713
Sellaro TL, Hildebrand D, Lu Q, Vyavahare N, Scott M, Sacks
MS (2007) Effects of collagen fiber orientation on the response
of biologically derived soft tissue biomaterials to cyclic loading.
J Biomed Mater Res 80(1):197–205
Valtorta D, Mazza E (2006) Measurement of rheological properties of
soft biological tissue with a novel torsional resonator device. Rheol
Acta 45(5):677–692
Van Loocke M, Lyons C, Simms C (2008) Viscoelastic properties of
passive skeletal muscle in compression: stress-relaxation behaviour and constitutive modelling. J Biomech 41(7):1555–1566
Vardaxis N, Brans T, Boon M, Kreis R, Marres L (1997) Confocal
laser scanning microscopy of porcine skin: implications for human
wound healing studies. J Anat 190:601–611
Waldman SD, Lee JM (2005) Effect of sample geometry on the apparent
biaxial mechanical behaviour of planar connective tissues. Biomaterials 26(35):7504–7513
Wineman A, Wilson D, Melvin JW (1979) Material identification of
soft-tissue using membrane inflation. J Biomech 12(11):841–850
Zernicke RF, Butler DL, Grood ES, Hefzy MS (1984) Strain topography of human tendon and fascia. J Biomech Eng 106(2):177–180
123