DOI: 10.1002/adem.

201180057

RESEARCH ARTICLE
Structure and Mechanical Performance of a ‘‘Modern’’
Fish Scale**
By Deju Zhu, Cesar Fuentes Ortega, Ramak Motamedi,
Lawrence Szewciw, Franck Vernerey and Francois Barthelat*

Protective materials and structures found in natural organisms may inspire new armors with improved
resistance to penetration, flexibility, light weight, and other interesting properties such as transparency
and breathability. All these attributes can be found in teleost fish scales, which are the most common
types of scales in modern fish species. In this work, we have studied the structure and mechanics of fish
scales from striped bass (Morone saxatilis). This scale is about 200–300 mm thick and consists of a hard
outer bony layer supported by a softer cross-ply of collagen fibrils. Perforation tests with a sharp needle
indicated that a single fish scale provides a high resistance to penetration which is superior to
polystyrene and polycarbonate, two engineering polymers that are typically used for light transparent
packaging or protective equipment. Under puncture, the scale undergoes a sequence of two distinct
failure events: First, the outer bony layer cracks following a well defined cross-like pattern which
generates four ‘‘flaps’’ of bony material. The deflection of the flaps by the needle is resisted by the collagen
layer, which in biaxial tension acts as a retaining membrane. Remarkably this second stage of the
penetration process is highly stable, so that an additional 50% penetration force is required to eventually
puncture the collagen layer. The combination of a hard layer that can fail in a controlled fashion with a
soft and extensible backing layer is the key to the resistance to penetration of individual scales.

Nature increasingly serves as a model and inspiration to scales, which have received relatively little attention in the
scientists and engineers, and biomimetics has the potential to past.[1–5] This type of scale displays interesting combinations
lead to novel engineering materials and systems with new of flexibility, strength, resistance to penetration, light weight,
combinations of properties, multi-functionalities, adaptabil- and transparency. Fish scales exhibit large variations in shape,
ity, and environmental sustainability. In this work, we have size, and arrangement. The general classification includes
studied the structure and mechanics of modern teleost fish cosmoid, ganoid, placoid, and elasmoid (cycloid and ctenoid)
found in the modern teleost class of fishes.[6] The ‘‘primitive’’
[*] Dr. D. Zhu, C. F. Ortega, R. Motamedi, L. Szewciw, cosmoid and ganoid scales are bulky, bony scales which offer
Dr. F. Barthelat very effective protective properties, through a multilayered
Department of Mechanical Engineering, structure capable of a variety of dissipative mechanisms.[3]
McGill University, Montreal, QC, (Canada) However, over the course of evolution the reduction of the
E-mail: francois.barthelat@mcgill.ca integumental skeleton has improved swimming perfor-
Dr. F. Vernerey mance,[3,7] and the ‘‘ancient’’ cosmoid and ganoid scales have
Department of Civil Engineering, been replaced by the thinner, more flexible teleost scales.[8]
University of Colorado, Boulder, CO, (USA) Teleost scales have excellent hydrodynamic properties[9,10]
[**] The authors wish to acknowledge the support of the National and provide a protective layer resisting penetration.[3,7,8]
Science Foundation under award CMMI 0927585 and of Currey, in a review article on mineralized tissues, noted that
Faculty of Engineering at McGill University. Atomic emission some fish scales are so tough that they cannot be easily
spectroscopy tests were performed by Monique Riendeau, Dept. of fractured even after immersion in liquid nitrogen.[11] At larger
Mining & Materials Engineering, and Ranjan Roy and Andrew lengths, the arrangement of the scales provides a flexible skin
Golsztajn, Dept. of Chemical Engineering, McGill University. that allows for changes in shape. In fact, the scaled skin has
This paper was amended in issue 4 of Advanced Engineering been shown to play a critical structural role in fish locomotion
Materials because there was a mistake in the Early View by regulating wave propagation[12–14] and by storing mechan-
publication. ical energy in order to make swimming more efficient.[15]

ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4 ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com B185
 D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .
RESEARCH ARTICLE

Fig. 1. The hierarchical structure of a teleost fish scale from striped bass, M. saxatilis. (a) Whole fish, (b) staggered multiple scales, (c) an individual scale, (d) cross-section of a scale,
(e) cross-ply collagen structure, (f) collagen fibrils.

1. The Hierarchical Structure of Fish Skin (atomic emission spectroscopy), we measured an average
hydroxyapatite mass fraction of 46% for the whole scale.
In this study we have investigated the structure and The density of collagen (1.33
103 kg  m3) and hydroxya-
mechanics of a single teleost (ctenoid) fish scale from striped patite (3.17
103 kg  m3)[2] were used to estimate the volume
bass Morone saxatilis. Like many other structural biological fraction of hydroxyapatite as 26%. In another experiment, we
materials,[16–18] the structure of teleost fish scales displays a separated a scale into two samples by dissecting a few plies off
characteristic hierarchical structure, built over several distinct the collagen layer. The upper and lower samples gave
length scales (Figure 1). At the macroscopic level, the scales are hydroxyapatite mass contents of 50% and 14%, respectively
staggered and cover most of the body of the fish (Figure 1a,b). (30% and 6% in volume fractions), confirming that the upper
This arrangement provides a continuous barrier to penetration region of the scale is significantly more mineralized than the
and flexural compliance. When the fish is highly curved (at the lower region. These results are consistent with reports of a
end of a swimming stroke), the scales interact more strongly, general 20–35% percentage points difference in mineralization
which stiffens the skin in flexion.[19] The skin then acts as an between the bony and collagen layers.[25] Bony and collagen
‘‘external tendon,’’ storing mechanical energy, which can be layers are cross-ply layered composites, each ply being made
recovered to facilitate the onset of the next stroke.[15,20] At the of parallel collagen fibrils rotated across layers by angles that
mesoscale level, an individual scale from an adult striped bass can vary from species to species.[25–29] In striped bass, we
is a thin plate with an irregular pentagonal shape, about 10 mm found that the basal layer is formed of 20–25 plies about
in diameter (Figure 1c). 4–5 mm thick each (Figure 1d), where the collagen fibrils are
The posterior area of the scale displays rough patterns rotated by 90 degrees from one ply to the next (Figure 1e).
(ctenii) which offer attractive hydrodynamic properties,[9,10] Cross-ply collagen structures are typically found in natural
while the anterior area consists of grooves in the radial tissues that undergo multiaxial stresses (shell of soft-shelled
direction (radii) and ridges that form circular rings (circuli) turtles,[30] human annulus fibrosus[31]). Several authors have
around a central area called the ‘‘focus.’’[21] Radii and circuli discussed the importance of the plywood nature of the
possibly provide increased flexibility and anchoring of the collagen layer to whole fish scale mechanical proper-
scale, respectively.[10] Teleost scales are composed of collagen ties,[2,5,8,32–34] in general by providing the scale with strength
fibrils type-I, and are partially mineralized with hydroxya- along multiple directions. At smaller length scales, individual
patite (16–59% mineral content in weight[2,5,22–24]). The outer collagen fibrils, about 50–200 nm in diameter, can be observed
layer of the scale is significantly more mineralized and often on a cross-section of the scale (Figure 1f). Interestingly, we
referred to as ‘‘bony layer,’’ whereas the inner layer (‘‘basal’’ found that in striped bass the 90 degree cross-ply is achieved
or ‘‘collagen’’ layer) is mineralized mostly near the bony layer, by alternating layers composed of radial fibrils (‘‘R’’ layers)
but with mineralization pockets proceeding well into the with layers made of circumferential fibrils (‘‘C’’ layers), both
collagen layer.[7,23] In striped bass, bony and collagen layers layers being organized around the focus of the scale (Figure 2).
have approximately the same thickness (100 mm). Using AES This arrangement is consistent with the growth of individual

B186 http://www.aem-journal.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4
D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .

RESEARCH ARTICLE
0.20 mm. Using this technique, samples were cut at 0, 45, and
908 from the longitudinal direction (anteroposterior axis) of
the fish.
The samples were then mounted on a miniature loading
stage (Ernest F. Fullam Inc., Latham, NY), which was placed
under an upright, reflected light microscope (BX-51M,
Olympus, Markham, Canada) equipped with a CCD camera
(RETIGA 2000R, Qimaging, Surrey, Canada) in order to
monitor deformations and failure modes of the specimens. All
specimens were loaded in tension at a rate of 0.005 mm  s1
(corresponding to a strain rate of 1.25
103 s1) up to
complete failure. Images were taken throughout the entire test
every 10 s using the CCD camera. The images were used to
measure the deformation and strain values of the samples
using digital image correlation[37] and to monitor failure
modes such as debonding of the bony layer and pullout of the
collagen fibrils. All samples were kept in hydrated conditions
during preparation and testing.
The resulting stress–strain curves (Figure 3a) display an
initial quasi-linear region, with an initial modulus in the
600–850 MPa range. The material softens slightly before
reaching a maximum stress of 30–50 MPa (depending on
scale orientation), after which the stress drops significantly.
Using optical observation, we could associate this sudden
decrease in stress to the sudden cracking of the bony layer.
Subsequently, the collagen layer progressively detaches from
the stiff bony layer, while collagen plies tear one after the
other, yielding step-like patterns on the stress–strain curve up
to total failure at about 40% strain.
In order to assess the mechanical response of the collagen
layer alone, we performed additional tensile tests on scales
with the bony layer removed. The collagen layer was carefully
peeled out of the scale which rested on a flat and hard surface.
Fig. 2. Arrangement of the collagen fibrils in a striped bass scale. (a) Top view of the The layered structure of the fish scale makes it easy to
whole scale showing surface features (optical micrograph); (b) bottom view showing
collagen fibers in the collagen layer; (c) removal of the bony layer reveals the radial– ‘‘delaminate’’ with a minimum of force. While the collagen
circumferential (R–C) pattern of the collagen fibrils; (d) locally, the fibers are orthogonal material at the separation site might be partially damaged, the
from one layer to the next; (e) schematics of the R–C pattern.
peeling force was deemed insufficient to damage the rest of
the collagen layer and to alter its overall mechanical proper-
ties. The remaining 0.05 mm thick collagenous material was
scales, which occurs by deposition of collagen at the periphery tested in tension along the 0, 45, and 908 directions.
of the scale.[35,36] The stress–strain response (Figure 3a) displayed a linear
regime with a modulus of about 450 MPa, followed by a
2. Tensile Testing on Individual Scales progressive failure after a peak stress of 65 MPa. The ultimate
strain was the same with and without bony layer. The
In order to assess the mechanical response of individual behavior of the collagen cross-ply is consistent with
scales, we performed tensile tests in hydrated conditions on the behavior of single collagen type I fibrils.[38] Assuming
small tensile samples prepared from individual scales. Whole, that the fibers do not carry any stress if they are perpendicular
fresh striped bass (Morone saxatilis) were acquired from a fish to the loading direction, only half of the material actually
supplier (Nature’s Catch, Inc., Clarksdale, MS, USA) and kept carries stress in the collagen layer. Since individual collagen
on ice. Scales were plucked using tweezers and stored in a fibrils have a Young’s modulus of about 1 GPa, a tensile
freezer at 20 8C until tested. Before the test, the scales were strength of 200 MPa and a strain at failure of 30%,[38] the
removed from the freezer and put in a water bath for about modulus of the 908 ply laminate can be estimated at 500 MPa,
5 min for thawing, and then cut into small dog-bone-shaped its strength at 100 MPa and its strain at failure at 30%. This
specimens using a multi-tube rotary hole-punch and dissect- compares well with our experimental results, showing that the
ing scissors. The resulting samples had a gage length of 4 mm, tensile behavior of the collagen layer is largely controlled by
a gage width of 1.5 mm and an average thickness of about the stretching of straight, individual collagen fibrils.

ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4 ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com B187
 D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .
RESEARCH ARTICLE

Fig. 3. (a) Tensile stress–strain curves for fish scales along 0, 45, and 908 from the longitudinal axis of the fish. Summary of results for (b) Young’s modulus; (c) strength. The
error bars indicate standard deviations.

While it was not possible to isolate the bony layer for given by:
testing, its properties were inferred from the whole scale
1
and collagen only tensile test results. In the elastic regime, the ES ¼ ðEC þ EB Þ (1)
2
whole scale behaves like a two-layer, constant strain
composite. Since the thickness of the bony and where EC and EB are the Young’s moduli of collagen and bony
collagen layers is similar, the modulus of the scale is layers, respectively. The modulus of bony layer can then be

B188 http://www.aem-journal.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4
D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .

pelagic fishes and sharks.[39] Juveniles striped bass are known

RESEARCH ARTICLE
estimated using:
to have several sources of predation including bluefish
EB ¼ 2ES EC (2) (Pomatomus saltatrix) and cannibalism.[39] The main function
of the scales is mechanical protection against predators[3,7,8]
The strength of bony layer can be evaluated with a similar and, in particular, the scale must be capable of preventing
approach. In the linear regime, with the uniform strain sharp objects such as teeth[40] from completely penetrating the
assumption, the stresses in the bony and collagen layers are skin and reaching the softer and more fragile underlying
proportional to their stiffnesses. From the whole scale test, the tissues. A sharp tooth concentrates the biting force over a very
stress sS at which the bony layer fails is known. Just prior to small area on the scale, which leads to severe contact stresses.
failure the stress in the bony layer is then given by: We have assessed the resistance to penetration of single
EB striped bass scales using an experimental setup that simulates
sB ¼ sS (3)
ES a predator’s bite using the same miniature loading stage
described above in the tensile tests. A sharp steel needle (tip
This model assumes that both materials are in the linear radius ¼ 25 mm) was used to simulate a predator’s tooth
elastic range up to the failure of the bony layer. In reality (bluefish, for example, have teeth of a shape similar to a needle
Figure 3a shows that the scale softens slightly when loaded in and a radius smaller than 250 mm). The needle was driven
tension, probably due to damage accumulation in the bony layer through a scale resting on a silicone rubber substrate
(collagen behaves linearly over this range of strain). Equation 3 (E  1.8 MPa measured by spherical indentation) used to
therefore slightly overestimates the actual strength of the bony simulate the soft dermis and tissues underlying the scale, at a
layer. We used Equations (2) and (3) to estimate the modulus speed of 0.005 mm  s1 while the force was recorded
and strength of the bony layer. The results show that the bony (Figure 4a). The load-deflection curves resulting from these
layer is about twice as stiff as the collagen layer, with about the tests were highly repeatable throughout all tested samples,
same tensile strength (Figure 3b,c). The bony layer is, however, with a slight force drop at about 2 N and a maximum
more brittle, failing at about 10% strain while the collagen layer penetration force of 3–3.5 N (Figure 4b). For comparison, we
fails at strains in excess of 40%. Interestingly, we also found that also performed puncture tests on thin polystyrene (PS) and
the whole scale displays in-plane anisotropic properties, but polycarbonate (PC), which are modern engineering polymers
only because of the bony layer; the collagen layer is isotropic in typically used when light weight, stiffness, strength, optical
plane in terms of both modulus and strength. This set of translucence, and impact resistance are required (their
experiments highlights the main traits of the scale’s compo- applications include CD cases, biomedical equipment such
nents: the bony layer is stiff, hard and brittle because of its high as petri dishes, and protective gear such as safety glasses or
mineral content, while the underlying collagen cross-ply is squash goggles). For proper comparison we prepared 10 mm
softer and more deformable, with larger strains at failure. diameter disks of these polymers, and we adjusted the
thickness so the areal density (the mass per unit area of
3. Resistance to Sharp Penetration: Puncture of Individual protective material) of the protective layer was then same for
Scales fish scale, PC and PS. Remarkably, the fish scale provided a
Adult striped bass have a few natural predators, including significantly higher resistance to puncture compared to these
various aquatic birds, marine mammals, and potentially large high performance engineering polymers (Figure 4b).

Fig. 4. (a) Experimental setup for the puncture test; (b) typical results for striped bass scales with and without bony layer. Results for polystyrene and polycarbonate are also
shown for comparison.

ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4 ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com B189
 D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .
RESEARCH ARTICLE

Finally, we performed additional puncture tests on scales have the same thickness, but since the bony layer is stiffer, the
with the bony layer removed. The isolated collagen layer was neutral plane of the scale lies within the bony layer. As a
easily penetrated, providing only half of puncture resistance result, flexural deformations generate tensile stresses in the
in terms of force. The bony layer is therefore an important lower side of the bony layer. Once these stresses reach
component of the system, operating in synergy with the the tensile strength of the bony layer, cracks initiate at the
collagen layer to increase the performance of individual collagen/bone interface and rapidly propagate towards
scales. The fish scale followed a sequence of mechanisms that the surface of the bony layer. Interestingly, the patterns of
was highly repeatable from scale to scale as well as across the flexural cracks always followed a cross pattern, whose
locations on a given scale. The penetration curves consist of orientation invariably followed the orientation of the local
three distinct stages that we investigated in detail by imaging radii and circuli (Figure 5c) and underlying collagen fibrils.
of the puncture site at different points on the penetration curve The microstructure of the bony layer therefore induces the
(Figure 5). Stage I is the initial linear region, which is failure of the bony layer along specific directions. Upon
dominated by flexion of the entire scale and by damage and cracking of the bony layer, four ‘‘flaps’’ of bony material
indentation of the surface of the bony layer. At a force of about immediately deflect downwards, generating circumferential
2 N, the force drops slightly, which we associated to the cracks. At this point, the underlying collagen layer, while
sudden cracking of the bony layer. Bony and collagen layers remaining intact, detaches from the bony layer over a ring-like

Fig. 5. Detailed sequence of a puncture test. (a) Load–displacement curve showing three distinct stages; (b) associated mechanisms; and (c) imaging. The images of the bony side and
the last one of the collagen side were obtained by scanning electron microscopy (SEM), while the first four images of the collagen side were taken with an optical microscope.

B190 http://www.aem-journal.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4
D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .

RESEARCH ARTICLE
area observable with an optical microscope (the scale, while Figure 6a,b. Each of the fours flaps was assumed to be rigid
opaque to electrons, is transparent to visible light). The and hinged along a straight line at the bone/collagen
cracking of the bony layer marks the beginning of stage II, interface. The force from the needle was assumed to be
dominated by further flexion of the scale, opening of the cross evenly distributed between the tips of the 4 flaps. Only two
cracks as the four ‘‘flaps’’ of bony material are bent towards mechanical loads resist the deflection of the flaps by balancing
the collagen layer, radial propagation of the cross cracks, and the force from the needle: i) the bending moment transmitted
further delamination between collagen and bony layers. through the remaining ligament of the bony layer, and ii) the
Eventually the deflection and opening of the flaps are intact collagen layer, which acts as a ‘‘retaining membrane’’
sufficient to let the needle reach the collagen layer and for the flaps. Interestingly, in this configuration the collagen
completely puncture it (stage III). The initial failure of the layer is in a state of biaxial tension at the penetration site, and
collagen layer indicates the beginning of stage III, and the the purpose and advantage of its cross-ply structure becomes
sharp drop in force at this point suggests that the failure is evident.
rapid, possibly because the collagen layer is stretched. The bending moment transmitted at the ligament was
Throughout the rest of stage III, the scale is deflected by evaluated by assuming perfect plasticity in the bony layer
the needle, the delamination between collagen and bony with sY ¼ 60 MPa (evaluated from our tensile test). In order to
layers propagates more extensively, and the radial cracks balance this moment, the force applied by the needle was
continue to grow. estimated at 1.2 N (see Appendix for calculation details),
We further investigated the mechanisms operating at stage which is actually below the force at which the bony layer
II, since they are powerful enough to increase the resistance of fractures. This value represents an upper bound estimate,
the scale by an additional 1 N (representing an additional 50% since in reality the bony layer probably cracks before the full
of the load at failure of the bony layer). In particular, we plastic state can be reached. This prediction shows that in
examined the controlled deflection of the four ‘‘bony flaps’’ stage II, there is no bending moment transmitted at the bony
using an idealized three-dimensional geometry shown in flaps, and that the flaps can be assumed to rotate about

Fig. 6. (a) Top view of the idealized puncture configuration; (b) three-dimensional view of single bony flaps. The force applied by the needle (F/4) and the forces generated by the
tension in the collagen layer are shown. (c) Effect of collagen layer thickness and (d) collagen resilience on the normalized penetration force.

ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4 ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com B191
 D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .
RESEARCH ARTICLE

frictionless hinges. The experiments show circumferential interface weak enough to delaminate and allow the collagen
cracks in the region of the hinges, confirming that little or no layer to stretch under the bony flaps. While this first study on
bending moment can be transmitted though the bony hinge. the puncture mechanics of fish scale does not consider viscous
The second mechanism examined was associated with the effects, it is likely that viscoelastic and viscoplastic effects also
collagen, which acts as a retaining membrane for the flaps. contribute to the energy dissipation capability of the scale. The
In the model, the collagen layer was assumed to have actual skin of the fish is of course covered with a large number
completely delaminated from the overlying surface of the of overlapping scales and, for striped bass, we counted that
bony layer, which is consistent with experimental observa- any given point on the surface of the body is covered with 3 or
tions towards the end of stage II. The collagen layer then acts 4 layers of scales. The resulting multilayer system alternates
as a retaining membrane, with biaxial tension as the dominant hard and soft layers in an arrangement reminiscent of the
stress. Based on the idealized geometry, the moment balance design of bulletproof glass. In addition, overlapping scales
of a single bony flap about the hinge led to a simple expression ensures compliance and breathability, two highly desirable
for the penetration force as a function of bony and collagen properties for personal armors. A biomimetic design at the
layer thickness (tB, tC), length of the flaps (L), and stiffness (EC) individual scale level could therefore be combined with a
and strength (sC) of the collagen layer. Full penetration at the clever arrangement of the scales at the macroscale to yield a
end of stage II was assumed to be reached when the collagen hierarchical protective system with attractive properties.
failed in tension (at a stress of 65 MPa according to our tensile
tests). Our model predicted a penetration force of F ¼ 3 N Appendix: Analytical ‘‘Four Flaps’’ Model
based on the properties of the collagen layer (EC ¼ 500 MPa,
A simplified analytical model was derived to capture the
sC ¼ 65 MPa) and optical observation (tB ¼ tC ¼ 100 mm,
progressive deflection of the four flaps and the retaining
L ¼ 200 mm). This prediction is remarkably close to penetra-
membrane effect provided by the underlying collagen layer
tion force we measured experimentally, which demonstrates
(Figure A.1).
that the retaining membrane effect dominates stage II and
controls the ultimate penetration resistance of the scale. A.1. Strain
In terms of design, the model reveals that longer flaps are The strains in the collagen were determined by tracking the
desirable, and since L is larger for larger ‘‘teeth’’ and higher displacement of the point in the centroid of the section of
forces, the scale provides a greater resistance to penetration for collagen (point D). Upon deflection of the flap point D moves
larger teeth and stronger bites (Figure 6c,d). A thick collagen to D’, so that its displacement vector can be written, in the
layer is also beneficial, although a minimum of bony material Cartesian system xyz (Figure A.1):
is required to form stiff flaps. This finding substantiates 
 L t
 pffiffiffi ðcosu1Þ C sinu
previous discussions on mechanical benefits of thicker  2 2 2
collagen fibrils and plywood organization in the collagen DD~0 ¼  0 (A.1)
 L tC
layer.[8,32,33,41] Finally, a soft and strong material increases  p ffiffiffi sinu ðcosu1Þ
 2
resistance to penetration, although a too soft backing layer 2 2
may lead to excessive deflection that may damage the
Any fiber crossing the cracks will extend by a distance
underlying tissues even before needle penetration. A cross-ply ~0 on the unit vector
(using scalar product to project DD
of collagen is therefore ideal for this function, and the function
collinear to T):
of the harder bony layer is to protect the collagen layer from
2 32 1
3
direct contact with the tip of the needle, and to mitigate the L tC
pffiffiffi ðcosu1Þ sinu 6  p ffiffi
ffi
stresses transmitted onto the softer collagen layer by 6 2 2
6 2 76
76 27 7 L
6 7:6 1 7 tC
redistributing them over a large area. 6 0 7 6 pffiffiffi 7 ¼ ð1cosuÞ þ pffiffiffi sinu
4 5 7 4 2 2
L tC 4 2 5
 pffiffiffi sinu ðcosu1Þ
2 2 2 0
4. Conclusions
(A.2)
Individual teleost fish scales are therefore high perfor-
mance natural protective systems, offering resistance to The collagen is detached from the bony layer in the
puncture superior to modern engineering polymers typically puncture area, which leads to the assumption that the tensile
used for protective applications. Remarkably, fish scales are strain in the collagen is uniform and equal to:
made of a set of materials that are both softer and weaker than 1 tC
these engineering polymers, which highlights the important "C ¼ ð1cosuÞ þ pffiffiffi sinu (A.3)
2 L 2
role of the structure and architecture of the scale in
‘‘amplifying’’ the properties (as seen in other class of Interestingly, the geometry and kinematics for the system
biological materials[42–44]). The high performance of the scales is such that the strain in the collagen layer is equi-biaxial and
is the result of a fine balance of structure and material uniform in the puncture site. Knowledge of strain in the
properties, and in particular the hardness and stiffness of the collagen layer leads to stress, using the modulus found from
outer layer, the softness and strength of the inner layer, and an the tensile tests (E ¼ 500 MPa).

B192 http://www.aem-journal.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4
D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .

RESEARCH ARTICLE
Fig. A.1. Schematic diagram of the deformation and force in the four flaps model.

A.2. Deflection of the Flaps get the force retained by the collagen layer as follows:
Similarly, point C at the upper tip of the flaps moves to C’ 0 pffiffiffi 1
tC
upon deflection, and 2tanu þ 2
B LC
F ¼ 2LtC s C @ pffiffiffi tB A (A.7)

 L 1 þ 2 tanu
 pffiffiffi ðcosu1Þ þ tB sinu L
 2
~ ¼ 
CC
0
0 (A.4)
 L
  pffiffiffi sinu þ t ðcosu1Þ Normalized by the square of the scale thickness t and the
 B
2 tensile strength of collagen sC:
0 pffiffiffi 1
The deflection at the loading point is then the z component tC
F L tC B 2 tanu þ 2
~0 :
of CC ¼2 LC (A.8)
@ pffiffiffi tB A
s C t2 t t
1 þ 2 tanu
L L
d ¼ pffiffiffi sinutB ðcosu1Þ (A.5)
2

A.4. Force Generated by Flexural Stresses at the Hinges

Assuming that the entire thickness of the bony layer
A.3. Force undergoes plasticity, the bending moment transmitted is
Under the concentrated force F/4, the rigid bony flap calculated as follows:
rotates along the y axis and the loading location changes from  
pffiffiffi tB 2
C to C’. The force vectors acting on the bony and collagen M ¼ 2L sB (A.9)
2
layers are:
2 3 The force applied by the needle is balanced by this
T
pffiffiffi moment:
2 3 6 27
0 6 7
6 7 L F
6 7
~¼ 6
F 4 0 5
7 ~¼ 6 T 7
T (A.6) pffiffiffi ¼ M (A.10)
6 pffiffiffi 7 24
6 7
F=4 6 2 7
4 5
Combining Equations (A.9) and (A.10), we get the force
0 generated by flexural stresses at the hinges:
~0 and T
~ is equivalent  2
As the cross product of the vectors OD tB
~0
~ then we F¼8 sB (A.11)
to half of the cross product of the vectors OC and F, 2

ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4 ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim http://www.aem-journal.com B193
 D. Zhu et al./A ‘‘Modern’’ Fish Scale . . .
RESEARCH ARTICLE

Received: June 21, 2011 [20] M. R. Hebrank, Biol. Bull. 1980, 158, 58.
Final Version: August 10, 2011 [21] L. A. Jawad, J. Nat. Hist. 2005, 39, 2643.
Published online: October 14, 2011 [22] W. T. Liu, Y. Zhang, G. Y. Li, Y. Q. Miao, X. H. Wu, J. Fish
Biol. 2008, 72, 1055.
[23] A. A. Schonborner, G. Boivin, C. A. Baud, Cell Tissue Res.
1979, 202, 203.
[1] L. F. Wang, J. H. Song, C. Ortiz, M. C. Boyce, J. Mater. Res. [24] R. V. Seshaiya, P. Ambujabay, M. Kalyani, Amino acid
2009, 24, 3477. composition of Icthylepedin from fish scales, in: Aspects
[2] F. G. Torres, O. P. Troncoso, J. Nakamatsu, C. J. Grande, of Protein Structure (Ed: G. N. Ramachandran), Academic
C. M. Gomez, Mat. Sci. Eng. C Bio. S 2008, 28, 1276. Press, New York p. 343.
[3] B. J. F. Bruet, J. H. Song, M. C. Boyce, C. Ortiz, Nat. Mater. [25] F. J. Meunier, Am. Zool. 1984, 24, 953.
2008, 7, 748. [26] A. Bigi, M. Burghammer, R. Falconi, M. H. J. Koch,
[4] T. Ikoma, H. Kobayashi, J. Tanaka, D. Walsh, S. Mann, S. Panzavolta, C. Riekel, J. Struct. Biol. 2001, 136, 137.
Int. J. Biol. Macromol. 2003, 32, 199. [27] L. Zylberberg, J. Bereiterhahn, J. Y. Sire, Cell Tissue Res.
[5] T. Ikoma, H. Kobayashi, J. Tanaka, D. Walsh, S. Mann, J. 1988, 253, 597.
Struct. Biol. 2003, 142, 327. [28] F. J. Meunier, J. Castanet, Zool. Scr. 1982, 11, 141.
[6] K. V. Kardong, Vertebrates: Comparative Anatomy, Func- [29] J. Meunier, Tissue Cell 1981, 13, 165.
tion, Evolution, McGraw-Hill, New York 2008. [30] T. M. Scheyer, P. M. Sander, W. G. Joyce, W. Boehme,
[7] L. Zylberberg, J. Geraudie, F. Meunier, J. Sire, Bone 1992, U. Witzel, Org. Divers Evol. 2007, 7, 136.
4, 171. [31] E. C. Bass, F. A. Ashford, M. R. Segal, J. C. Lotz, Ann.
[8] J. Bereiter-Hahn, L. Zylberberg, Comp. Biochem. Phys. A Biomed. Eng. 2004, 32, 1231.
1993, 105, 625. [32] L. Zylberberg, J. Bonaventure, L. Cohensolal,
[9] S. Sudo, K. Tsuyuki, Y. Ito, T. Ikohagi, JSME Int. J. C D. J. Hartmann, J. Bereiterhahn, J. Cell Sci. 1992, 103,
Mech. Sy. 2002, 45, 1100. 273.
[10] J. Y. Sire, J. Fish Biol. 1986, 28, 713. [33] H. S. Youn, T. J. Shin, J. Struct. Biol. 2009, 168, 332.
[11] J. D. Currey, J. Exp. Biol. 1999, 202, 3285. [34] F. J. Meunier, P. M. Brito, Cybium 2004, 28, 225.
[12] J. H. Long, B. Adcock, R. G. Root, Comp. Biochem. Physiol. [35] C. Liu, S. Shen, J. Taiwan Museum 1991, 44, 321.
A Mol. Integr. Physiol. 2002, 133, 911. [36] G. K. Ostrander, The Laboratory Fish, Academic Press,
[13] J. H. Long, M. E. Hale, M. J. McHenry, F M. W. Westneat, Oxford 2000.
J. Exp. Biol. 1996, 199, 2139. [37] F. Barthelat, J. Poissant, Exp. Mech. 2010, 50, 353.
[14] J. H. Long, T. J. Koob, K. Irving, K. Combie, V. Engel, [38] Z. L. Shen, M. R. Dodge, H. Kahn, R. Ballarini,
N. Livingston, A. Lammert, J. Schumacher, J. Exp. Biol. S. J. Eppell, Biophys. J. 2008, 95, 3956.
2006, 209, 4732. [39] K. J. Hartman, Proc. Atlantic Striped Bass Workshop and
[15] M. R. Hebrank, J. H. Hebrank, Biol. Bull. 1986, 171, 236. Roundtable Discussion, 1999, p. 6.
[16] M. A. Meyers, A. Y. M. Lin, Y. Seki, P. Y. Chen, B. K. Kad, [40] W. E. Bemis, A. Giuliano, B. McGuire, Zoology 2005, 108,
S. Bodde, JOM J. Miner. Metals Mater. Soc. 2006, 58, 35. 317.
[17] P. Fratzl, R. Weinkamer, Prog. Mater. Sci. 2007, 52, 1263. [41] H. Onozato, N. Watabe, Cell Tissue Res. 1979, 201, 409.
[18] L. J. Szewciw, D. G. de Kerckhove, G. W. Grime, [42] F. Barthelat, R. Rabiei, J. Mech. Phys. Solids 2011, 59, 829.
D. S. Fudge, P. Roy. Soc. B Biol. Sci. 2010, 277, 2597. [43] M. J. Buehler, Nano Today 2010, 5, 379.
[19] F. J. Vernerey, F. Barthelat, Int. J. Solids Struct. 2010, 47, [44] S. Keten, Z. P. Xu, B. Ihle, M. J. Buehler, Nat. Mater. 2010,
2268. 9, 359.

B194 http://www.aem-journal.com ß 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ADVANCED ENGINEERING MATERIALS 2012, 14, No. 4