Structural design and mechanical behavior of alligator (Alligator

mississippiensis) osteoderms
Chang-Yu Sun, Po-Yu Chen

Department of Materials Science and Engineering, National Tsing Hua University, 101, Sec. 2, Kuang-Fu Rd, Hsinchu 30013, Taiwan
a r t i c l e i n f o
Article history:
Received 9 April 2013
Received in revised form 12 July 2013
Accepted 16 July 2013
Available online 24 July 2013
Keywords:
Osteoderm
Armor
Composites
Mechanical property
Toughening mechanisms
a b s t r a c t
Alligator is a well-adapted living fossil covered with dorsal armor. This dermal shield consists of bony
plates, called osteoderms, interconnected by sutures and non-mineralized collagen bers, providing a
dual function of protection and exibility. Osteoderm features a sandwich structure, combining an inner
porous core and an outer dense cortex, to offer enhancements for stiffness and energy absorbance. In this
study, we investigated the multi-scale structure and mechanical behaviors of the American alligator (Alli-
gator mississippiensis) osteoderm. Microcomputed tomography was applied to reveal the complex neuro-
vascular network. Through the observation under optical and scanning electron microscopes, the
osteoderm was found to consist of woven bone in the dorsal region and lamellar-zonal bone in the ven-
tral region. Nanoindentation and compressive tests were performed to evaluate the mechanical proper-
ties of osteoderms. The varying mineral contents and porosity result in a graded mechanical property: a
hard and stiff dorsal cortex gradually transform to a more compliant ventral base. Three protective mech-
anisms optimized for alligator osteoderms were proposed and elucidated.
2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Many structural biological materials have been extensively
investigated in recent years due to their superior mechanical prop-
erties, considering the weak building blocks of which they are
composed [13]. Currently, exible and lightweight dermal armors
have aroused increasing interest due to their intriguing designs for
protection [4,5], including sh scales [612], turtle shells [1316]
and armadillo carapaces [17,18]. Fish scales, such as P. senegalus
[7], A. gigas [8] and A. spatula [10,12], have been widely studied la-
tely. Despite the differences in material compositions, mineral con-
tent and thicknesses, they all applied a similar strategy of
combining a stiffer and harder external region with a softer inter-
nal base. The scales of these marine species exhibit exibility
through interlocking and overlapping [4,5].
On the other hand, armadillo carapace [17,18] and turtle shell
[1316] utilize rather distinct strategies from sh scales. These ar-
mors share many similar structural features: (1) the main constit-
uents of these mineralized tissues are bone, consisting of collagen
bers and hydroxyapatite minerals; (2) the bony plates are con-
nected by soft tissues or joints; (3) they are covered by keratinous
layers on the outer surface; (4) they are both sandwich composites
with a dense cortex and a porous core. Chen et al. [17] found that
the non-mineralized collagen bers are responsible for the
macroscopic mechanical responses of the armadillo carapace. The
stretching of these connective bers between hexagonal plates is
the major contribution to tensile and shear strengths [17]. In turtle
shells, the bony segments are juxtaposed with zigzag joints inter-
locking in between, called sutures. The sutures are three-dimen-
sional (3-D) and complicated structures with organic tissues,
giving rise to effortless deformation under small loads and trans-
ferring to stiffer responses after locking under higher degrees of
movement [15]. Rhee et al. [13] reported that the porous core of
the turtle shell is made of closed-cell foam, causing the sandwich
structure to undergo a nonlinear deformation, which leads to a
higher specic energy absorption compared with the dense cortex
alone. Recent investigations conducted by Achrai and Wagner [14]
revealed that the dorsal and ventral cortices of the sandwich struc-
ture own various mechanical properties as a result of different -
ber arrangements. The randomly oriented brillar network in the
dorsal cortex can sustain sharp impact isotropically, while the ply-
wood arrangement of bers in the ventral cortex possesses aniso-
tropic mechanical properties and is benecial for structural
support [14]. The turtle carapace also appears to be a functionally
graded material (FGM) in terms of composition, porosity and
mechanical properties.
Crocodilian osteoderm is another interesting topic in natural
exible dermal armors. These ancient reptiles have long been con-
sidered as erce carnivorous tetrapods with heavily armored skins.
Although they seldom encounter predators, territorial ghts
among the same species can often be deadly because of their
1742-7061/$ - see front matter 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.actbio.2013.07.016

Corresponding author. Tel.: +886 3 571 5131x33889.

E-mail address: poyuchen@mx.nthu.edu.tw (P.-Y. Chen).
Acta Biomaterialia 9 (2013) 90499064
Contents lists available at ScienceDirect
Acta Biomaterialia
j our nal homepage: www. el sevi er . com/ l ocat e/ act abi omat
extremely high bite force, reaching 10 kN, the highest value that
has been reported for living animals to date [19]. Thus, well-devel-
oped armor designs for excellent mechanical performances are de-
manded, along with some exibility for speedy and agile
movements in order to capture preys. As a matter of fact, the dorsal
sheaths of crocodilians have been used as armor suits for ancient
warriors since they are found to repel knives and arrows, and are
even bulletproof under certain conditions, as discovered recently
[20]. However, the microstructure, mechanical properties and
deformation mechanisms have not been thoroughly investigated.
In this study, we investigated the osteoderm of American alliga-
tor (Alligator mississippiensis) by multi-scale structural character-
izations using materials science approaches. Mechanical
behaviors were evaluated and related to the structure features at
varying length scales, and the deformation as well as toughening
mechanisms of this biological armor when subjected to external
forces were proposed. We hope this study can provide further
understanding of biological defensive designs, and offer inspiration
for novel synthetic armors and advanced composites.
2. Background
Reptiles are cold-blooded animals featuring scales that cover
their whole body. Among them, crocodilians, including crocodiles,
alligators and caimans, are amazing living fossils which appeared
150 million years ago and have evolved into one of the most adap-
tive modern animals on the planet. These large tetrapods possess
not only keratinous scales on their external surfaces, but also un-
ique bony plates underneath the keratinous scales for reinforced
protection, called osteoderms. Crocodilian osteoderms are found
mainly on the dorsal dermis (also on the abdomen for some spe-
cies, such as most caimans), sheltering areas from the nuchal to
the caudal region. These natural armors are composed of mineral-
ized bony plates which are connected by brous tissues, similar to
armadillo and turtle carapaces. The hierarchical structure of the
dermal armor of Alligator mississippiensis is schematically pre-
sented in Fig. 1. The whole armor includes about 70 pieces of bony
plates. Each plate has a longitudinal keel in the middle. Through a
transverse cross-section, various structural features are demon-
strated. The external surface of the bony plates is covered by a thin
layer of keratinous scutes. These scutes, or scales, cover the dorsal
armor of alligator as well as all other parts of its body, and may
vary in shape, composition and formation mechanism. They result
from morphological transitions through differentiation and
keratinization of the crocodilians epidermis [21]. Harder and
tougher b-keratin outer layer coats the osteoderms to provide wear
resistance, while the a-keratin forms mainly the matrix and hinge
regions, acting as a barrier to water and electrolyte exchange
[21,22]. In addition, connective bers are found at the junction of
laterally neighboring bony plates.
Osteoderm is not an original element in evolutionary for croc-
odilians. This type of integumentary skeleton is a plesiomorphic
trait for tetrapods [23,24], and has been well demonstrated in
many dinosaurs, such as the renowned stegosaurs [25], ankylo-
saurs [20], and other extinct relatives such as squamates [26]
and archosaurs [27]. The osteoderms in various species differ in
size, shape, ornamentation and functions. In addition to protection
from claws and teeth of predators, other functions of osteoderms,
including heat transfer [25,28], mineral storage [29] and locomo-
tion aid [30], have been suggested. Since these integumentary skel-
etons of reptiles are not subjected to external forces and do not
likely undergo bone remodeling, the ages of these animals can be
estimated by counting the growth marks in the osteoderms corre-
sponding to the seasonal changes in growth rate [31], which is an
equally valid yet much more convenient method to obtain ages
than counting or observing the growth marks in the interior bones
(e.g. long bones) from living or preserved species [32].
3. Materials and methods
3.1. Sample preparation
A complete dorsal armor of an American alligator was obtained
from Jernigans Taxidermy (Waco, Texas, USA). The alligator armor
was prepared using a relatively harmless and natural method
without using strong chemicals, which may alter the natural state
of the samples. The longitudinal length of the armor is 0.85 m,
indicating the animal may have had a body length of 1.61.8 m
since the armored part, excluding the head and about half the tail,
accounts for 50% of the length of the entire animal. It it is likely to
be a mature alligator, 810 years old, depending on the gender. It
has been reported that male alligators in Texas can be 8 ft long
at the age of 10 and female alligators 67 ft long at the same age
[33].
Keratin coverings on the dorsal surface and dried dermis on the
ventral surface were removed in order to observe the structural
Fig. 1. Hierarchical structure of alligator osteoderm from macro-, meso- and micro-, to nanometer-length scales.
9050 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064
features of individual osteoderms. All samples prepared for micro-
structural characterization and mechanical testing were taken
from the central and caudal parts of the whole armor to maintain
consistency, since the cervical and transverse terminal osteoderms
possess rather irregular shapes and non-uniform keel heights. It
should be noted that the experimental samples are taken from a
single alligator, and may not be representative of the entire
species.
3.2. Elemental analysis
3.2.1. Ash content measurement
17 samples sectioned from ve mid-dorsal osteoderms (34
samples from each) with regions varied from the keel to the edge
were used to determine the average mineral content of the osteo-
derm by the ash-weight method [17]. Since the samples for the
measurement were all from the mid-dorsal osteoderms, the results
may stand for the major portion of the alligator armor. However, it
should be noted that the mineral content may change for osteo-
derms at different locations on the body. Samples were dried on
a hot plate at 105 C for 12 h and the dry weights were measured
using an electronic balance. Samples were then ashed at 600 C
for 24 h and ash weights were measured. The water content and
ash content (in wt.%) were calculated.
3.2.2. X-ray diffraction
X-ray diffraction (XRD) was carried out on powders of ground
alligator osteoderms utilizing a powder X-ray diffractometer
(XRD-6000, Shimadzu Co., Kyoto, Japan). A continuous scan using
Cu K
a1
(k = 0.154 nm) as the radiation source was performed in a
h2h mode scanning from 2h = 20 to 60, with a step size of
0.02 at a rate of 2 min
1
.
3.2.3. Electron probe microanalysis
The localized elemental compositions were analyzed by eld-
emission electron probe microanalysis (FE-EPMA) with a JEOL
JXA-8500F EPMA (JEOL Ltd., Tokyo, Japan). Three cross-sectional
samples (10 5 5 mm
3
) with both dorsal and ventral regions
of the osteoderm were sectioned and embedded in epoxy followed
by grinding and polishing. The samples were coated with a thin
layer of carbon instead of other common conductive coatings such
as gold or platinum because these heavy metal coatings can se-
verely suppress the emission of X-rays induced by the incident
electron, serving as a barrier layer for the signals to come out
and be detected. Five quantitative measurements were taken from
the dorsal and ventral region, respectively, for each specimen, and
the results were then averaged to compare the compositional dif-
ferences between the two regions.
3.2.4. Energy-dispersive spectroscopy
Elemental mapping at interfacial regions was achieved by an
energy-dispersive spectroscope (EDS) within a eld-emission scan-
ning electron microscope (FE-SEM) (JSM-7600F, JEOL Ltd.). Sam-
ples were sectioned and ground from the edge of an osteoderm
and coated with a thin layer of carbon (10 nm) to prevent elec-
tron charging. X-rays were produced from electron bombardments
under an accelerating voltage of 10 kV and a working distance of
15 mm. A silicon-drift detector (SDD) (X-Max SDD, Oxford Instru-
ments, Abingdon, Oxfordshire, UK) was used to collect the charac-
teristic X-rays from the sample, while the software AZtec (Oxford
Instruments) was applied to analyze and map the elemental
distributions.
3.3. Structural characterization
3.3.1. Macroscopic observation
External shape and morphology of osteoderms were taken from
central and edge regions of the whole armor. For cross-sectional
observation, a sample was cut by a hand saw and through the keel
region followed by grinding and polishing. Photographs of top, bot-
tom and cross-sectional views were taken by a digital camera.
3.3.2. Microcomputed tomography (l-CT)
l-CT was accomplished by unmonochromatized synchrotron
hard X-rays with energy ranging from 5 keV to 35 keV at the Na-
tional Synchrotron Radiation Research Center (NSRRC) in Hsinchu,
Taiwan [34]. l-CT images were obtained from a CCD camera (mod-
el 211, Diagnostic instruments, 1600 1200 pixels) after convert-
ing the X-rays into visible lights by a scintillator. The resolution of
the corresponding images (in pixel size) of a 2 lens was
2.8 lm 2.8 lm. In order to reconstruct 3-D tomographic mod-
els, images were collected with a regular step size of 0.3 over a to-
tal 180 rotation of the specimen stage. Xradia software (Xradia
Inc., Pleasanton, CA, USA) was applied to reconstruct the raw data,
which was then visualized by Amira software (Visualization Sci-
ence Group, a FEI company, Burlington, MA, USA).
3.3.3. Optical microscopy and stereoscopy
Cross-sectional samples through the keel region were ground
and polished before observing under an optical microscope
(BX51M, Olympus Co., Tokyo, Japan) equipped with a 0.8 mega-
pixel digital camera (DP12, Olympus Co., Tokyo, Japan). An inte-
grated view of the keel cross-section was achieved by combining
17 micrographs of consecutive regions.
Stereoscopic images were taken from an Olympus SZX7 Zoom
Stereomicroscope (Olympus Co., Tokyo, Japan) with a 2.0 mega-
pixel CCD camera (Innity 1, Lumenera Co., Ontario, Canada). The
magnication of the stereoscope ranged from 8 to 56.
3.3.4. Scanning electron microscopy
Microstructural characterization of the fracture surfaces were
observed by a FE-SEM (JSM-7600F, JEOL Ltd.). Fracture surfaces
were created by exerting a bending force through a clamp and a
wrench. The specimens were coated with a thin layer of platinum
to enhance electron conductivity on the surface. Secondary elec-
tron images (SEIs) were taken with an accelerating voltage of
10 kV and a working distance of 10 mm.
3.4. Mechanical testing
Schematic representations of specimens prepared for mechani-
cal testing are shown in Fig. 2. The system of coordinates we
adopted throughout the text is illustrated in Fig. 2a. Longitudinal
is dened as the direction along the keel long axis, transverse is re-
ferred to the direction perpendicular to the keel and vertical is the
direction through the thickness of the osteoderm. Longitudinal and
transverse are both included when the term horizontal is
referred.
3.4.1. Nanoindentation
A sample for nanoindentation was taken from the keel region of
a caudal osteoderm. The sample was mounted with the longitudi-
nal cross-sectional area (Fig. 2a) revealed, followed by grinding and
ne polishing with Al
2
O
3
suspensions from particle size of 1 lm,
0.3 lm and nally 0.05 lm. The average surface roughness of the
nal sample was 20 nm, measured with atomic force microscopy
(Dimension Icon, Bruker Corp., Billerica, MA, USA). Nanoindenta-
tion tests were conducted by using a Hysitron TI900 TriboIndenter
(Hysitron Inc., Eden Prairie, MN, USA) with a Hysitron TI-0039
C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064 9051
Berkovich diamond tip under a load-controlled mode. The radius of
the tip was 100 nm. The load function can be separated into three
stages: linear loading, holding and linear unloading, and the dura-
tion for each stage was 5 s, with the peak load set to be 1000 lN.
The area function of the tip was calibrated before conducting the
tests with a fused quartz bulk specimen as a standard material
due to its low elastic-modulus-to-hardness ratio [35]. A series of
indentations was performed vertically across the cross-section
(Fig. 2b) with an interval of 300 lm and a total number of 53
groups. Each group contained eight indent points forming a 2 4
rectangular region with a lateral space of 15 lm between neigh-
boring indents, which is a signicantly large distance compared
with the indent size to avoid effects of adjacent indents. The hard-
ness and reduced modulus values of each group were then aver-
aged. The area function was calculated again after the test and
showed no signicant variations (0.3%) on the tip geometry, sug-
gesting that the results were reliable even after a large amount of
indentations were performed.
3.4.2. Compressive testing
Samples for compressive testing were cut into 2.3
2.3 4.5 mm
3
rectangular pieces by a rotating diamond blade.
The dimensions were chosen to prevent buckling by the Eulers cri-
teria. Each facet of the samples was then ground carefully using a
clamp to ensure that the two surfaces in contact with the upper
and lower load cells are parallel to each other, while keeping the
side surfaces perpendicular to the ends as precise as possible to
eliminate eccentric loading. 80 vertical and 80 horizontal samples
were prepared, as shown in Fig. 2c. The horizontal samples were
taken from at (non-keel) regions, since the large deviation on
the vertical direction of the keel cross-section may have signicant
inuences on experimental results. However, for vertical samples,
the height 4.5 mm cannot be satised at locations far away from
the keel due to the decrease in thickness from the keel toward
the edge. Therefore, vertical samples were taken from non-keel re-
gions near the keels. The difference in regional distribution be-
tween the horizontal and vertical samples led to different
portions of dorsal and ventral regions, which is clearly demon-
strated in Fig. 2c: vertical samples contain more dorsal region than
the horizontal samples. 40 samples from each group (longitudinal
and vertical directions) were immersed in Hanks balanced saline
solution (HBSS) (H2387, SigmaAldrich Co., St Louis, MO) for
24 h before mechanical testing. The rehydrated samples were
tested immediately after taking out of HBSS in order to prevent fur-
ther drying. The other set of samples (40 each direction) were
tested in ambient dry condition. Compressive tests were conducted
by using a universal testing machine (Instron 3343 Single Column
Testing System, Norwood, MA, USA) with a 1 kN load cell at a strain
rate of 1 10
3
s
1
.
3.4.3. Flexibility demonstration
Two adjacent osteoderms were taken from the mid-dorsal re-
gion of the alligator armor to demonstrate the exibility of the
joints. The bony plates were sectioned transversely to better dis-
play the angles bent by bare hands.
3.4.4. Whole osteoderm compression
A transversely cross-sectioned large osteoderm sample with
keel height of 15 mm was used to demonstrate the deformation
of the sandwich structure under compression. The sample was
ground at the bottom to create a at contact surface, followed by
immersion in HBSS for 24 h before the test. The large-scale
compressive tests were carried out by a universal testing machine
Fig. 2. Schematic illustrations of sample preparations for mechanical tests. The shades of the color in (b) and (c) denote different regions, i.e. dorsal and ventral region. (a) The
headtail direction along which the keel is oriented is dened as the longitudinal (or parasagittal) direction. The direction along the lateral row of osteoderm is dened as the
transverse direction. Both directions are included when horizontal direction is referred. The horizontal direction is perpendicular to the vertical direction. (b) An illustration
of the direction where nanoindentation tests were performed. (c) An illustration of the locations where compression samples of different orientations were taken. Due to the
geometrical limitations, vertical samples are taken near the keel and possess more dorsal region, while the horizontal samples are taken mainly within ventral regions.
9052 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064
(Instron 4468, Double Column Testing System, Norwood, MA, USA)
with a 50 kN load cell at a strain rate of 1 10
3
s
1
.
4. Results and discussion
4.1. Macroscopic observation
An osteoderm from the mid-dorsal region appears in a quadrate
shape 5 cm in length and width with a parasagittally aligned keel
of 1.5 cm in height (Fig. 3a). Osteoderms taken from different
locations show distinct appearances, in both shape and keel height.
An osteoderm from the transverse terminal is shown in Fig. 3b for
comparison. The large concave regions on the external surface con-
tain small cavities that connect to the vascular channels, which
were proposed to be the evidence for the thermoregulation func-
tion of alligator osteoderms [30]. The neurovascular foramina enter
the bony plates from the ventral surface, as shown in Fig. 3c, where
the grooves are traces of bifurcated dorsal median arteries circulat-
ing across the surface [30]. The transverse edge of an alligator
osteoderm contains connective collagenous bers (Fig. 3d) be-
tween two adjacent plates, similar to the armadillo carapace
Fig. 3. Photographs showing the top views of mid-dorsal (a) and transverse terminal (b) osteoderms show different shapes and features. The ventral surface (c) of a mid-
dorsal osteoderm shows that arterial grooves were used to hold vessel branches. The arrows indicate pits where the artery bifurcates into nutrient foramina entering the
osteoderm. The edge of the osteoderm between laterally neighboring plates is shown in (d), which was covered by non-mineralized connective bers. By removing the non-
mineralized connective bers, the sutures can be observed in (e). A SEM image in (f) shows the 3-D feature of the sutures, which contain pits that are connected to the
neurovascular system within the plates.
C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064 9053
[17]. These bers account for connectivity and exibility enhance-
ments. By removing the connective soft bers, serrated sutures
(Fig. 3e) are revealed, which share the same functional design for
interlocking as the turtle shell [15]. From SEM, we can clearly ob-
serve the 3-D characteristic of the sutures (Fig. 3f). Numerous tiny
pits are also shown in Fig. 3f, which are connected to the neurovas-
cular foramina, presumably for vascularization, sensing and nutri-
ent transportation.
4.2. Mineral content measurement and elemental analysis
From the ash-weight measurement, the water content is
10.70 0.58 wt.% and the mineral content of the dried specimen
is 65.77 2.26 wt.%, giving 34.23 2.26 wt.% of the dried alligator
osteoderm to be organic components. The osteoderm of American
alligator possesses a similar mineral content to that of armadillo
carapace (65 wt.%) [17] and bovine femur (67 wt.%) [36], which
is higher than other natural armors such as tortoise shell
(53 wt.%) [37] and sh scale (46 wt.%) [6], as well as that of
some mammalian compact bones, for instance, elk antler
(57 wt.%) [38]. It is also found that the mineral content decreased
gradually from the keel (67.13 0.58 wt.%) through the transition
region (65.54 0.56 wt.%) to the edge (62.89 1.60 wt.%) as shown
in Fig. 4a, which is functionally graded since the edge with sutures
and connective bers serve for exibility enhancements, and may
be a result of evolutionary convergence with the turtle shell [14].
To conrm the mineral constituents, the crystalline phase was
detected by XRD. The resulting pattern in Fig. 4b can be indexed
to JCPDS 09-0432, revealing hydroxyapatite as the main compo-
nent, which is the same as the minerals in bone and other bony tis-
sues [6,17,38]. Furthermore, localized elemental analysis measured
by FE-EPMA shows that the dorsal region contains more calcium
(29.58 1.15 wt.%) and phosphorus (13.13 0.68 wt.%) than the
ventral region (Ca: 24.53 1.62 wt.%, P: 10.38 0.75 wt.%). The re-
sults indicate that the dorsal region contained more minerals,
which are primarily non-stoichiometric hydroxyapatite, while the
ventral region appeared to be less-mineralized and had more or-
ganic constituents.
Elemental mapping by EDS under FE-SEM was conducted to
analyze the compositional difference between the bony plate
and the connective bers (Fig. 5). As the secondary electron im-
age (SEI) in Fig. 5a shows, the area of interest was the interface
between sutures, which is still a part of the bony plate, and the
connective bers at the edge of an osteoderm. Carbon mapping
in Fig. 5b indicated that both regions contain organic contents,
which is collagen [24], and that the amount of organic contents
is obviously much richer in the connective bers. On the other
hand, distribution of Ca (Fig. 5c) and P (Fig. 5d) clearly illus-
trated that hydroxyapatite (Ca
10
(PO
4
)
6
(OH)
2
), as the major min-
eral content, is conned in the suture. Therefore, it was
conrmed that the bers connecting bony plates together are
not mineralized.
4.3. l-CT imaging
A cross-sectional view in the longitudinal direction of an osteo-
derm keel reveals a sandwich structure (Fig. 6a), where the porous
interior is surrounded by compact cortex. A sandwich structure
also appears in various light-weight designs, such as leaves [39],
bird beaks [40] and feathers [41], as well as many defensive de-
vices against impact and bending, including human skull [39], tur-
tle shell [14], armadillo carapace [17], sh armor [11] and
horseshoe crab exoskeleton [1]. The main advantage of sandwich
structures is to provide high bending stiffness with minimum
weight. Defensive designs applying this principle also serve the
function of energy absorbance under impact loads by deformations
of the cellular core through elastic bending, brittle fracture or plas-
tic buckling of the walls or trabeculae before undergoing densica-
tion [39]. The reduced burden of these lightweight armors can thus
enhance locomotion along with improvements in mechanical
properties.
Since cross-sections can only show porosities and incomplete
channels, a computed tomographic technique was applied to gain
3-D information of the complex neurovascular network in the
bony plates. By collecting the transmitted X-ray signals and the
corresponding intensities, we can identify the porosity since the
absorbance along the X-ray path is different for materials with
porosity and without porosity. Through rotation during scanning,
the information of the whole specimen can be obtained and 3-D
models can be further reconstructed. A sectioned image of the
reconstructed 3-D model from l-CT scans is shown in Fig. 6b.
The neurovascular channels form an intricate 3-D network, where
the major cavity in the center branches out toward the dorsal and
ventral regions with much smaller pipes. The bifurcation within
the dorsal region tends to be more complex compared with the
ventral region, whereas both regions contain evidences of sea-
sonal and/or annual growth, indicated by the white arrows in
Fig. 6b.
Fig. 4. Minerals in alligator osteoderm. (a) The amount of mineral content within
the osteoderm was measured by ash-weight method, where the results showed
decrease from the keel to the edge (keel: 67.13 0.58 wt.%, transition:
65.54 0.56 wt.%, edge: 62.89 1.60 wt.%). (b) X-ray diffraction (XRD) pattern of
the alligator osteoderm conrmed that hydroxyapatite is the main mineral
component.
9054 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064
4.4. Microstructural characterization
4.4.1. Optical microscopy
Four different regions with different microstructural morpholo-
gies of bone can be distinguished from the optical micrograph, as
shown in Fig. 7a. At the outer sheath, randomly oriented woven
bone can be observed along with extensive vascularization. The
vascular channels connect to small pits on the external surface,
mainly located within the large concave regions, as previously
shown in Fig. 3a and b. These pits and vascular systems are sug-
gested to be the major evidence of the role of osteoderm in ther-
moregulation of the body [30]. The second region beneath the
outer sheath is composed of dense woven bone and scattered
lamellar rings deposited around the neurovascular channels. An
SEM image taken from this region is shown in Fig. 7b. The patterns
on the surface indicate collagen ber bundles being ruptured and
pulled out, where no preferred orientation can be observed, illus-
trating bundles entangled in a randomly woven manner. Scattered
lamellar rings can be recognized in this region. These concentric
lamellar structures are not likely to be secondary osteons since
no canaliculi and well-developed vascular systems (Haversian
and Volkmanns channels) are observed. Moreover, osteoderms
are not constantly subjected to external loading and bone remod-
eling may be restricted and limited. The third region contains por-
ous woven bone with large neurovascular channels. The woven
bone in this region is similar to that of the outer sheath. The major
neurovascular foramen and branches mainly locate in this region
and the large cavities can reach up to hundreds of micrometers
or several millimeters in diameter. Lamellar bone is also found
around the channels, forming circular rings. At the bottom, the ba-
sal region consists of lamellar-zonal bone [42]. This type of bone is
commonly seen in reptiles, and is related to the poor vasculariza-
tion, which derives from low metabolic rate [42]. The lamellar-zo-
nal bones correspond to seasonal or annual growth, where bone
growth stops or slows down in winters, leaving lines of arrested
growth (LAGs). The lamellae in this region are not the same as
the lamellar bone, but rather a parallel-bered bone, which is con-
structed by woven collagen bers with a preferred orientation
(Fig. 7c). Furthermore, Sharpey-bered bone is also found mostly
in the ventral region and sometimes on the edge of dorsal region,
extending in oblique directions from the margin to the core with
wavy or zigzag structures [20,23,24]. This type of bone is derived
from non-mineralized Sharpeys bers functioning for connection,
which fuse into the osteoderm and anchor the bony plates to the
epidermis. The Sharpey-bered bone in the ventral region indi-
cated locations where the osteoderm contact the epidermis, while
those in the dorsal region implied that the genesis of osteoderm
took place within the skin. It is thus discovered that the alligator
osteoderm consists of various types of bone, resulting in a hetero-
geneous composition. The differences in ber orientation at differ-
ent locations are schematically presented in Fig. 1. It should be
noted that the structural variations between adjacent regions un-
dergo a gradual change.
Thin sections of the osteoderm were observed under the stereo-
scope (Fig. 8). LAGs are clearly observed in the ventral region, while
only vague annuli can be recognized at the dorsal region because
no complete halt occurred during growth. The growth rate is also
different in the two regions. Although this is not a precise quanti-
tative evaluation due to the limited resolution, we can still distin-
guish that the interval between annual growth marks in the ventral
region (0.5 mm) is much narrower compared to those in the dor-
sal region (1.1 mm). Hence, we proposed a growth model, as
shown in the schematic illustration (Fig. 8). The growth rate in
the keel region is higher than the basal region, leading to a unique
shape of a ridged keel. In addition, according to Vickaryous and
Hall [24], the mineralization process of the osteoderm initiates
from the keel, and then extends radially across the whole plate.
Therefore, combining the non-simultaneity in calcication and
Fig. 5. Elemental mapping by energy-dispersive spectroscopy (EDS) at the edge of the osteoderm. (a) An SEI image of the area being analyzed, which shows no signicant
distinguishing structural features of the two regions. (b) Carbon mapping indicates that the carbon content is much richer in the connective bers. (c, d) Calcium and
phosphate mapping clearly demonstrates the lack of minerals, which is mainly hydroxyapatite, within the connective bers.
C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064 9055
the anisotropy in growth rate, it is implied that the degree of min-
eralization may not be uniform within the entire osteoderm, espe-
cially along the vertical direction.
4.4.2. Scanning electron microscopy
Microstructural characterization of a fractured osteoderm keel
is shown in Fig. 9, where comparisons between the dorsal, ventral
regions and the non-mineralized collagen bers are made. At a
lower magnication, the dorsal region (Fig. 9a) shows a densely
packed and relatively at fracture surface, indicating a direct fail-
ure and rather brittle behavior upon fracture. The mineralized col-
lagen ber bundles are fused together and the patterns are
obscured. Under a more detailed view (Fig. 9b), granular morphol-
ogies of mineral aggregates on the surface are observed. Individual
bers can hardly be recognized in this region, implying that the -
bers are highly mineralized and form bundles which cannot be eas-
ily separated. In comparison, the fracture surface taken from the
ventral region (Fig. 9c) exhibits a brous feature. These collagen -
bers were stretched and twisted upon breakage, revealing a more
ductile behavior. Individual brils can be easily distinguished at
higher magnication (Fig. 9d), where the surfaces of these brils
appear to be smoother compared with those in Fig. 9b. It is there-
fore suggested that these collagenous bers should possess a lower
degree of mineralization, in contrast to those in the dorsal region,
which corresponds to the growth rate and mineralization process
as previously proposed in Fig. 8. The SEM observation is also in
good agreement with the previous EPMA results, which suggested
mineral content difference in dorsal and ventral regions.
Fig. 9e shows the microstructural features of the non-mineral-
ized connective bers between neighboring bony plates to provide
exibility for the dorsal shield. These organic bundles are ran-
domly oriented, with smooth surface morphologies as shown in
the higher magnication (Fig. 9f). Characteristic patterns of
67 nm periodicity in collagen brils, which derives from the stag-
gered molecular arrangement, can also be observed in Fig. 9f.
In summary, the alligator osteoderm is a complex, heteroge-
neous, hierarchically structured bio-composite with varying de-
grees of mineralization and porosity at different locations.
Through a thorough compositional and structural characterization
at multiple levels with XRD, EPMA, EDS, l-CT, OM (optical micros-
copy) and SEM, the whole dermal armor is revealed to be a hybrid
systemof mineralized bony plates and non-mineralized connective
collagen bers, whereas each bony plate is also a combination of a
heavily-mineralized interwoven dorsal cortex and a less-mineral-
ized parallel-bered ventral base, with complex 3-D neurovascular
channels branching from the core of the osteoderm. These features
are integrated and organized in the schematic illustrations of Fig. 1.
4.5. Mechanical behavior
4.5.1. Nano-mechanical evaluation
Nanoindentation tests along the cross-section of an alligator
osteoderm reveal the difference in mechanical properties for dorsal
and ventral regions. Hardness and reduced modulus of each indent
are calculated from the loaddepth curves according to the Oliver
Pharr method [43]. Fig. 10a shows the hardness variation across
the keel while the photo above the plot corresponds to the inden-
tation position. The average hardness values are 367 94 MPa for
the ventral region and 690 170 MPa for the dorsal region.
Fig. 10b shows the change in reduced modulus through the keel
cross-section, where the trend is the same as the hardness values.
The ventral region was found to possess an average reduced mod-
ulus value of 13.9 2.1 GPa, gradually increasing to 20.3 3.4 GPa
in the dorsal region. Drops at the exterior region for both hardness
and reduced modulus are observed. The relatively high standard
deviation for both hardness and reduced modulus across the whole
sample can be related to the non-uniform porosity within the
osteoderm. The projected area of contact of nanoindentation on
the specimen is 15 lm
2
, and if we consider the deformed sur-
face proles around the indents, the area inuenced by the inden-
tation can reach up to 10 lm
2
. On the other hand, the size of the
porosity in the osteoderm spreads over a wide range from sub-
micrometer to millimeter, where those smaller than tens of
micrometers may cause serious effects on the indentation results.
We have avoided the larger pores at the scales from millimeters
down to several tens of micrometers under the optical microscope
while choosing the indent positions. However, smaller voids be-
yond the limit of light microscopy can still exist, which we were
unable to identify and avoid. Therefore, it is possible that we have
set our indents on or nearby the porosity and affected the results,
Fig. 6. (a) The sandwich structure of an osteoderm keel cross-section. The porosity
is caused by the foramina branching system. (b) 3-D image reconstructed froml-CT
scans by synchrotron X-rays. The white arrows indicate annual growth of the
branches.
9056 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064
giving rise to the high standard deviation. Furthermore, since the
porosity is three-dimensional, it is very likely that there are voids
underneath the indentation surface at any depth, which we cannot
observe but may have signicant effects on the measured values. In
addition, the osteoderm is composed mainly of woven bone, which
is a loosely structured mineralized tissue. Thus, it is also possible
(a)
Woven & lamellar bone
(b)
Lamellar-zonal bone
(c)
Fig. 7. (a) A combined optical micrograph of an osteoderm keel cross-section. Four regions can be distinguished as marked in the Fig.: outer sheath, woven bone and lamellar
bone, woven bone with neurovascular channels and lamellar-zonal bone. (b) An SEM image of the dorsal region showing tangled woven collagen ber bundles. (c) An SEM
image from the ventral region showing parallel-bered bone (woven bone with a preferred orientation).
Fig. 8. A schematic drawing demonstrating the proposed growth mechanism of an alligator osteoderm. The insets are stereoscopic images of thin-sectioned samples from the
dorsal (left) and ventral regions (right). The black arrows indicate growth marks (LAGs or annuli).
C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064 9057
that we have indented onto the spacing between the collagen ber
bundles, resulting in large deviations.
The distribution of localized mechanical properties is in good
agreement with the microstructural features in different regions
previously described in Fig. 7a. The second region (II) from the
external surface contains highly mineralized, relatively dense wo-
ven bone and has the highest hardness and reduced modulus;
whereas less-mineralized lamellar-zonal bone in the basal region
(region IV) possesses the lowest values in both properties. The
exterior region (I) and the porous core (III) both seem to have an
intermediate mechanical property between the dense dorsal re-
gion and the ventral base, corresponding to their similar microcon-
stituents (woven bone with high porosity). Fig. 10c shows the
typical loaddepth curves of the dorsal and ventral regions, illus-
trating the distinct difference in mechanical properties. Under a
constant maximum load, it is apparent that the dorsal region
showed a higher stiffness and hardness with smaller indentation
depth and larger slope of the unloading path, whereas the ventral
region showed an advantage of higher energy dissipation esti-
mated by the area under the curves. The different mechanical
properties between the dorsal and ventral region may relate to
the compositional difference in mineral content and variation in
porosity, as demonstrated in elemental analysis and microstruc-
tural characterization.
The combination of hard external region and soft internal re-
gion is analogous to many biological composites, such as arthropod
Dorsal Region
(a) (b)
Ventral Region
(c) (d)
Non-mineralized
Collagen Fibers
(e) (f)
Fig. 9. SEM fractographs at low and high magnications of the dorsal (a, b) and ventral (c, d) region of an alligator osteoderm keel and the connective bers (e, f). Different
morphologies indicate the varied amount of minerals on the collagen bers: the dorsal region is highly mineralized, the ventral region is less mineralized and the connective
bers are non-mineralized.
9058 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064
exoskeleton [44] and sh scales [7,8]. This is benecial for defen-
sive armors since the stiff and hard exterior can resist penetrations
by sharp claws or teeth, whereas the compliant interior can arrest
crack propagation and provide toughness, acting as a cushion by
absorbing impact energy. However, most of them reported a dis-
continuity (or extremely steep gradations) in mechanical re-
sponses between the stiff and compliant layers. In this study,
hardness and elastic modulus change gradually rather than there
being discrete interfaces, as found in the scales of alligator gar
(Atractosteus spatula), which are composed of two distinct materi-
als (ganoin and bone) [10,12,45]. The alligator osteoderm contains
only bone but undergoes gradual transitions in microstructures
at different locations due to their growth, leading to mechanical
functions. The advantages of comprising a homogeneous material
with mechanical property gradient (i.e. FGM) instead of abrupt
changes include better stress redistribution and enhanced resis-
tance to interfacial failure [7].
4.5.2. Compressive mechanical behavior
Fig. 11a and b shows the representative stressstrain curves of
the compressive tests for vertical and horizontal specimens, respec-
tively, under dry and rehydrated conditions. In the dry condition,
vertical samples can sustain higher stresses (142.1 21.4 MPa)
but eventually underwent a direct fracture in a relatively brittle
manner. On the contrary, the horizontal samples showed lower
strengths (117.5 15.9 MPa) with several stages of deformation be-
fore ultimate failure, fractured in a more ductile mode. It should be
noticed that for both orientations, the ultimate strengths occurred
at about the same amount of strain (30%), meaning that both
directions are able to suffer certain amounts of plastic deforma-
tions. However, the horizontal samples yielded a much higher com-
pressive strain (>40%) before totally breaking down due to the
impedance to direct failure, providing a greater toughness. Further-
more, the osteoderm samples of both orientations were also tested
under rehydrated conditions in HBSS to simulate the actual envi-
ronments of the biological system, where both vertical
(124.2 21.8 MPa) and horizontal samples (88.6 12.5 MPa)
showed a decrease in compressive strength. It was discovered that
in both dry and rehydrated conditions, the vertical samples exhib-
ited higher strengths than the horizontal samples (dry vertical:
142.1 MPa vs. dry horizontal: 117.5 MPa, rehydrated vertical:
124.2 MPa vs. rehydrated horizontal: 88.6 MPa); moreover, the ver-
tical samples in both conditions failed similarly in a direct and
rather brittle way, whereas the horizontal samples in both condi-
tions underwent multiple toughening mechanisms upon breakage
(Fig. 11a and b). This implies that the anisotropic mechanical re-
sponse is actually employed by the armor within biological sys-
tems. The compressive mechanical properties are summarized in
Table 1.
Typical fracture samples after compressive tests in both direc-
tions are examined under the stereoscope, as demonstrated in
Fig. 11c and d. Direct crack propagation for vertical samples can
be found in Fig. 11c, which is in accordance with the relatively
brittle failure shown in the ss curve. Horizontal samples, on
Fig. 10. Nanoindentation tests along the cross-section of an osteoderm show difference in (a) hardness and (b) reduced modulus in various regions. Note the gradient
transition between the two regions. (c) Representative loaddepth curves for dorsal and ventral regions show distinct difference in mechanical properties at the maximum
load of 1000 lN.
C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064 9059
the other hand, exhibit wavy and complex fracture paths and re-
veal evidence of lamellae buckling and ber bridging, as shown in
Fig. 11d. These toughening mechanisms can prevent catastrophic
collapse of the whole structure and correspond to the resisting
steps before ultimate fracture observed in the ss curve
(Fig. 11b). The observation further illustrates the anisotropy in
mechanical properties for alligator osteoderm.
To establish a relationship connecting the macroscopic com-
pressive behaviors and the localized nanoindentation results, the
elastic modulus from the two tests were compared. To begin with,
the elastic modulus was rst calculated from the reduced modulus
reported from the nanoindentation test. The relationship between
reduced modulus and elastic modulus can be presented as
1
E
r

1 m
2
s
E
s

1 m
2
i
E
i
where E
r
is the reduced modulus, E
s
is the elastic modulus of the
specimen and E
i
is the elastic modulus of the indenter tip. m
s
repre-
sents the Poissons ratio of specimen and m
i
means the Poissons ra-
tio of the indenter tip. For a Berkovich diamond tip, E
i
is 1140 GPa
and m
i
is 0.07 [35]. Since we are unable to determine the Poissons
ratio of the osteoderm, the Poissons ratio of a typical compact bone,
which is 0.18 for bovine femur [46], is used to calculate the elastic
modulus. From the above equation, the elastic moduli of alligator
osteoderm from nanoindentation tests are 13.6 GPa in the ventral
region and 20.0 GPa in the dorsal region (using the average reduced
Table 2
Elastic modulus measured from nanoindentation on dorsal and ventral regions of alligator osteoderm, with comparison to bovine and human femurs, and the porosity effect on
the elastic modulus calculated by the Boneld and Clark equation.
Nanoindentation modulus (GPa) 10% porosity (GPa) 20% porosity (GPa) 30% porosity (GPa) 40% porosity (GPa)
Alligator Osteoderm (Dorsal) 20.0 16.4 10.7 5.5 2.1
Alligator Osteoderm (Ventral) 13.6 11.2 7.3 3.7 1.4
Human Femur 20.1 [47]
Bovine Femur 23.1 [48]
Fig. 11. Representative compressive stressstrain curves for (a) vertical and (b) horizontal samples under dry and rehydrated conditions. Stereoscopic images showing typical
failure mechanisms in (c) vertical and (d) horizontal samples after compressive deformation.
Table 1
Mechanical properties of dry and rehydrated alligator osteoderm samples in
horizontal and vertical directions (40 samples for each condition).
Direction Hydration
state
Ultimate compressive
strength r
avg
(MPa)
Compressive elastic
modulus E (GPa)
Vertical Dry 142.1 21.4 1.04 0.11
Vertical Rehydrated 124.2 21.8 0.94 0.15
Horizontal Dry 117.5 15.9 1.19 0.22
Horizontal Rehydrated 88.6 12.5 1.07 0.17
9060 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064
Fig. 12. Experimental demonstrations of two proposed deformation mechanisms: (a) Flexibility provided by sutures and connective bers allows limited bending. The
neighboring plates could be bent upward to 10 and downward to 20. (b) The major porosity in the center core of the osteoderm is able to absorb some energy at
deformations lower than 10% before cracks start to propagate in the cortex. Large channels and small voids (circled) were found to be squeezed and distorted from 0% and
4% to 7% deformation.
C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064 9061
moduli), which showed comparable values with the elastic moduli
of human [47] and bovine [48] femur. However, it is signicantly
higher and cannot relate to the compressive modulus. We sug-
gested that it should be a result of porosity effect for macroscopic
mechanical tests. By applying the Boneld and Clark [49] equation,
which is a modied version of the Mackenzie [50] equation, we can
estimate the contribution of porosity to the elastic modulus:
E E
0
1 1:9p 0:9p
2

where E is the elastic modulus with porosity, E

0
is the elastic mod-
ulus of the solid phase only and p represents porosity. Table 2
gives the elastic modulus of bulk material at different porosities
where the elastic modulus from the nanoindentation test is used
as E
0
. As the result showed, the elastic modulus of compressive
tests clearly cannot reach the calculated value, even when an over-
estimated 40% porosity (the average porosity was 13% as esti-
mated from the l-CT scans) is taken into account, implying that
the alligator osteoderm is a complex composite and may incorpo-
rate factors other than porosity upon mechanical responses at the
macroscale.
4.6. Deformation mechanisms
Apart from the localized and global mechanical responses of the
mineralized bony tissues within the osteoderm, two additional
deformation mechanisms were proposed: exibility and sandwich
structure. Thus, simple tests were conducted to demonstrate these
mechanisms, as shown in Fig. 12. The sutures and non-mineralized
connective bers at the lateral edge of osteoderms provide a
Fig. 13. Schematic illustrations of three deformation mechanisms of alligator dermal armor: (a) Sutures and non-mineralized collagen bers provide exibility; (b) sandwich
structure absorbs energy; (c) graded mechanical property from dorsal to ventral regions offers optimization in load redistribution and energy dissipation.
9062 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064
limited exibility that allows bending to some extent. As illus-
trated in Fig. 12a, the connection device of alligator armor is able
to be bent upward to 10, while the bending downward can go
up to 20. This implied that the osteoderm is designed to suffer
more exure downward, corresponding to the main function of
load dissipation for protection. On the other hand, a whole piece
of osteoderm with a cross-sectional surface revealed was subjected
to a large-scale compressive test to demonstrate the function of the
sandwich structure. We focused on the deformation of the major
channel in the center core and the surrounding small voids before
cracks started to propagate in the dense cortex, which is 10%
deformation. The three successive images at the bottom half of
Fig. 12b showed the cross-section at 0%, 4% and 7% strain, where
the major channel was found to be squeezed and distorted, and
were thus able to absorb additional energy. Also, the void at the
lower right became smaller as the bony plate is being compressed,
possibly resulting from plastic buckling and wall collapsing of the
cellular foam. These simple demonstrations illustrated that exi-
bility and sandwich structure are both incorporated in the protec-
tion mechanisms of the dorsal shield of alligator.
Based on the experimental results, we summarized three defor-
mation mechanisms of the alligator armor under external loads
(Fig. 13). Whenanexternal compressive force acts onthe osteoderm,
sutures and non-mineralized collagenous bers connecting adja-
cent bony plates can dissipate loads by providing a limited amount
of exibility (Fig. 13a) [15]. The movements between neighboring
osteoderms can avoid further deformation under small loads. As
the load increases, the sandwich structure of a bony plate functions
as a preliminary route for energy absorbance (Fig. 13b). The porous
interior acts as a cellular foam, undergoing deformations such as
elastic bending, plastic buckling or wall breakage to absorb a certain
amount of energy. This mechanism may prevent cracking and thus
maintain structural integrity of the cortex at small strains. The third
mechanism incorporates the structural and mechanical properties
of the material itself by combining the dorsal and ventral cortex
(Fig. 13c). The hard and highly mineralized dorsal region can sustain
higher vertical stresses as the mechanical testing indicates, and be-
cause of the unique shape of the osteoderm, the loads then transfer
down to the less-mineralized, more compliant ventral region. The
ventral region thus suffers from stresses oriented preferentially in
the horizontal direction. The basal region consists mainly of parallel
oriented collagen brils (lamellar-zonal bone) and possesses better
capability to absorb energy horizontally through various toughen-
ing mechanisms. In addition, the dorsal and ventral region is joined
through a mechanical property gradient, where interfacial tough-
ness is enhanced and stress is better redistributed. The integration
of these three deformationmechanisms may lead to a synergistic ef-
fect and therefore an optimized dermal armor for alligator.
5. Conclusions
Dermal armors developed in reptiles as well as some mammals
and sh are considered to be optimized for both protection and
exibility through millions of years of evolution. In the present
study, we investigated the structure and mechanical behaviors of
the American alligator (Alligator mississippiensis) osteoderms at
multiple length scales. The dermal armor of alligator is a hierarchi-
cally structured composite consisting of mineralized rigid bony
plates connected by non-mineralized collagen bers. Through the
experiments, we established the structuralmechanical property
relationships and deformation mechanisms of the osteoderm, and
eventually proposed how the dorsal shield of alligator protects
against external threats from the whole armor system to individual
plates. The major discoveries of this study are concluded as
follows:
(1) Complex 3-D sutures and the non-mineralized connective
bers between neighboring osteoderms provide exibility
for the whole armor. Strategies such as bridging and stretch-
ing of the collagen bers and interlocking of the sutures are
utilized, which are similar to those observed in armadillo
carapace and turtle shell.
(2) The sandwich structure of the osteoderm shows a compact
cortex surrounding the porous core, enhancing bending stiff-
ness and energy absorption ability with reduced weight. The
intricate 3-D network of the neurovascular systemis respon-
sible for the spongy interior.
(3) The osteoderm is composed of four different bone morphol-
ogies, vertically from the outmost surface to the ventral
region: outer sheath (woven bone with porosity), woven
and lamellar bone, woven bone with major neurovascular
cavities and lamellar-zonal bone. The varying mineral con-
tents and porosity result in different localized hardness
and reduced modulus values across the osteoderm: from a
hard and stiff dorsal cortex gradually transform to a more
compliant ventral base. Similar design strategies have been
applied in various natural armors as well, implying evolu-
tionary convergence for defensive functionality.
(4) Cross-sectional fracture surfaces of osteoderm through keel
region indicate various degrees of mineralization and thus
different microstructures in the dorsal and the ventral
regions: the dorsal region is highly mineralized, showing
granular morphology and at fracture surface while the ven-
tral region is less mineralized, showing exible and twisted
brils. Incorporating the two regions with mild mechanical
gradient leads to the anisotropy in compressive behaviors:
the vertical orientation is able to bear higher loads, while
the horizontal orientation can absorb more energy through
multiple toughening mechanisms including lamellae buck-
ling and ber bridging.
(5) Three deformation mechanisms are proposed for the dermal
armor of alligator: (1) the exibility provided by sutures and
non-mineralized collagen bers can dissipate energy under
small loads; (2) deformations of the cellular foam interior
absorb impact energy without cortex cracking; (3) a combi-
nation of the hard dorsal region and the compliant ventral
region with graded mechanical properties offers optimiza-
tion in load re-distribution and energy absorbance.
Acknowledgements
The authors gratefully thank Yu-ChenChan, Hsien-Wei Chen, Su-
Yueh Tsai and Prof. Jenq-Gong Duh (MSE Department, National
Tsing Hua University) for their support and assistant with the FE-
SEM, FE-EPMA and ash-content measurements. We acknowledge
Chia-Chi Chien, Bai-Hong Ke, Tsung-Tse Lee and Prof. Yeu-Kuang
Hwu (Institute of Physics, Academia Sinica) for help withthe techni-
cal work and advices on l-CT scans and 3-Dimage reconstruction at
the National Synchrotron Radiation Research Center (NSRRC). We
would also thank Hsi-Ming Yang, Li-Chi Hsu and Prof. Jyh-Wei Lee
(MSE Department, Ming Chi University of Technology) for their help
with nanoindentation measurements. This research is supported by
National Science Council, Taiwan (NSC100-2218-E-007-016-MY3
and NSC101-2628-E-007-017-MY3).
Appendix A. Figures with essential colour discrimination
Certain gures in this article, particularly Figs. 113, are dif-
cult to interpret in black and white. The full colour images can
C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064 9063
be found in the on-line version, at doi: http://dx.doi.org/10.1016/
j.actbio.2013.07.016.
References
[1] Meyers MA, Chen P-Y, Lin AY-M, Seki Y. Biological materials: structure and
mechanical properties. Prog Mater Sci 2008;53:1206.
[2] Chen P-Y, McKittrick J, Meyers MA. Biological materials: functional adaptations
and bioinspired designs. Prog Mater Sci 2012;57:1492704.
[3] Dunlop JWC, Fratzl P. Biological composites. Ann Rev Mater Res 2010;40:124.
[4] Yang W, Chen IH, Gludovatz B, Zimmermann EA, Ritchie RO, Meyers MA.
Natural exible dermal armor. Adv Mater 2013;25:3148.
[5] Currey JD. Mechanical properties and adaptations of some less familiar bony
tissues. J Mech Behav Biomed Mater 2010;3:35772.
[6] Ikoma T, Kobayashi H, Tanaka J, Walsh D, Mann S. Microstructure, mechanical,
and biomimetic properties of sh scales from Pagrus major. J Struct Biol
2003;142:32733.
[7] Bruet BJ, Song J, Boyce MC, Ortiz C. Materials design principles of ancient sh
armour. Nat Mater 2008;7:74856.
[8] Lin YS, Wei CT, Olevsky EA, Meyers MA. Mechanical properties and the
laminate structure of Arapaima gigas scales. J Mech Behav Biomed Mater
2011;4:114556.
[9] Marino Cugno Garrano A, La Rosa G, Zhang D, Niu LN, Tay FR, Majd H, et al. On
the mechanical behavior of scales from Cyprinus carpio. J Mech Behav Biomed
Mater 2012;7:1729.
[10] Yang W, Gludovatz B, Zimmermann EA, Bale HA, Ritchie RO, Meyers MA.
Structure and fracture resistance of alligator gar (Atractosteus spatula) armored
sh scales. Acta Biomater 2013;9:587689.
[11] Song J, Reichert S, Kallai I, Gazit D, Wund M, Boyce MC, et al. Quantitative
microstructural studies of the armor of the marine threespine stickleback
(Gasterosteus aculeatus). J Struct Biol 2010;171:31831.
[12] Allison PG, Chandler MQ, Rodriguez RI, Williams BA, Moser RD, Weiss Jr CA,
et al. Mechanical properties and structure of the biological multilayered
material system, Atractosteus spatula scales. Acta Biomater 2013;9:528996.
[13] Rhee H, Horstemeyer MF, Hwang Y, Lim H, El Kadiri H, Trim W. A study on the
structure and mechanical behavior of the Terrapene carolina carapace: a
pathway to design bio-inspired synthetic composites. Mater Sci Eng C
2009;29:23339.
[14] Achrai B, Wagner HD. Micro-structure and mechanical properties of the turtle
carapace as a biological composite shield. Acta Biomater 2013;9:5890902.
[15] Krauss S, Monsonego-Ornan E, Zelzer E, Fratzl P, Shahar R. Mechanical function
of a complex three-dimensional suture joining the bony elements in the shell
of the red-eared slider turtle. Adv Mater 2009;21:40712.
[16] Balani K, Patel RR, Keshri AK, Lahiri D, Agarwal A. Multi-scale hierarchy of
Chelydra serpentina: microstructure and mechanical properties of turtle shell. J
Mech Behav Biomed Mater 2011;4:144051.
[17] Chen IH, Kiang JH, Correa V, Lopez MI, Chen PY, McKittrick J, et al. Armadillo
armor: mechanical testing and micro-structural evaluation. J Mech Behav
Biomed Mater 2011;4:71322.
[18] Rhee H, Horstemeyer MF, Ramsay A. A study on the structure and mechanical
behavior of the Dasypus novemcinctus shell. Mater Sci Eng C 2011;31:3639.
[19] Erickson GM, Lappin AK, Vliet KA. The ontogeny of bite-force performance in
American alligator (Alligator mississippiensis). J Zool 2003;260:31727.
[20] Scheyer TM, Sander PM. Histology of ankylosaur osteoderms: implications for
systematics and function. J Vertebr Paleontol 2004;24:87493.
[21] Alibardi L, Thompson MB. Fine structure of the developing epidermis in the
embryo of the American alligator (Alligator mississippiensis, Crocodilia, Reptilia).
J Anat 2001;198:26582.
[22] Grigg G, Gans C. Morphology and physiology of the Crocodylia. In: Glasby CG,
Ross GJB, Beesley PL, editors. Fauna of Australia, vol. 2A. Amphibia and
Reptilia. Canberra: Australian Government Publishing Service; 1993.
[23] Vickaryous MK, Sire JY. The integumentary skeleton of tetrapods: origin,
evolution, and development. J Anat 2009;214:44164.
[24] Vickaryous MK, Hall BK. Development of the dermal skeleton in Alligator
mississippiensis (Archosauria, Crocodylia) with comments on the homology of
osteoderms. J Morphol 2008;269:398422.
[25] Farlow JO, Hayashi S, Tattersall GJ. Internal vascularity of the dermal plates of
Stegosaurus (Ornithischia, Thyreophora). Swiss J Geosci 2010;103:17385.
[26] Buffrenil V, Sire JY, Rage JC. The histological structure of glyptosaurine
osteoderms (Squamata: Anguidae), and the problem of osteoderm
development in squamates. J Morphol 2010;271:72937.
[27] Cerda IA, Desojo JB. Dermal armour histology of aetosaurs (Archosauria:
Pseudosuchia), from the Upper Triassic of Argentina and Brazil. Lethaia
2011;44:41728.
[28] Farlow JO, Thompson CV, Rosner DE. Plates of the dinosaur stegosaurus: forced
convection heat loss ns? Science 1976;192:11235.
[29] Curry Rogers K, DEmic M, Rogers R, Vickaryous M, Cagan A. Sauropod dinosaur
osteoderms from the Late Cretaceous of Madagascar. Nat Commun
2011;2:564.
[30] Seidel MR. The osteoderms of the American alligator and their functional
signicance. Herpetologica 1979;35:37580.
[31] Hutton JM. Age determination of living nile crocodiles from the cortical
stratication of bone. Copeia 1986:33241.
[32] Tucker AD. Validation of skeletochronology to determine age of freshwater
crocodiles (Crocodylus johnstoni). Mar Freshw Res 1997;48:34351.
[33] Saalfeld DT, Webb KK, Conway WC, Calkins GE, Duguay JP. Growth and
condition of American alligators (Alligator mississippiensis) in an Inland
Wetland of East Texas. Southeast Nat 2008;7:54150.
[34] Song YF, Chang CH, Liu CY, Chang SH, Jeng US, Lai YH, et al. X-ray beamlines for
structural studies at the NSRRC superconducting wavelength shifter. J
Synchrotron Radiat 2007;14:3205.
[35] Rho JY, Zioupos P, Currey JD, Pharr GM. Variations in the individual thick
lamellar properties within osteons by nanoindentation. Bone 1999;25:
295300.
[36] Currey JD. Mechanical properties of bone tissues with greatly differing
functions. J Biomech 1979;12:3139.
[37] Kienzle E, Kopsch G, Koelle P, Clauss M. Chemical composition of turtles and
tortoises. J Nutr 2006;136:2053S4S.
[38] Chen PY, Stokes AG, McKittrick J. Comparison of the structure and mechanical
properties of bovine femur bone and antler of the North American elk (Cervus
elaphus canadensis). Acta Biomater 2009;5:693706.
[39] Gibson LJ. Biomechanics of cellular solids. J Biomech 2005;38:37799.
[40] Seki Y, Kad B, Benson D, Meyers MA. The toucan beak: structure and
mechanical response. Mater Sci Eng C 2006;26:141220.
[41] Bodde SG, Meyers MA, McKittrick J. Correlation of the mechanical and
structural properties of cortical rachis keratin of rectrices of the Toco Toucan
(Ramphastos toco). J Mech Behav Biomed Mater 2011;4:72332.
[42] Currey JD. The structure of bone tissue. Bones: structure and
mechanics. Princeton, NJ: Princeton University Press; 2002. pp. 3-26.
[43] Oliver WC, Pharr GM. An improved technique for determining hardness and
elastic modulus using load and displacement sensing indentation
experiments. J Mater Res 1992;7:156483.
[44] Chen PY, Lin AY, McKittrick J, Meyers MA. Structure and mechanical properties
of crab exoskeletons. Acta Biomater 2008;4:58796.
[45] Chen P-Y, Schirer J, Simpson A, Nay R, Lin Y-S, Yang W, et al. Predation versus
protection: sh teeth and scales evaluated by nanoindentation. J Mater Res
2011;27:10012.
[46] Pithioux M, Lasaygues P, Chabrand P. An alternative ultrasonic method for
measuring the elastic properties of cortical bone. J Biomech 2002;35:9618.
[47] Zysset PK, Edward Guo X, Edward Hofer C, Moore KE, Goldstein SA. Elastic
modulus and hardness of cortical and trabecular bone lamellae measured by
nanoindentation in the human femur. J Biomech 1999;32:100512.
[48] Rho JY, Pharr GM. Effects of drying on the mechanical properties of bovine
femur measured by nanoindentation. J Mater Sci Mater Med 1999;10:4858.
[49] Boneld W, Clark EA. Elastic deformation of compact bone. J Mater Sci 1973;
8:15904.
[50] Mackenzie JK. The elastic constants of a solid containing spherical holes. Proc
Phys Soc B 1950;63:211.
9064 C.-Y. Sun, P.-Y. Chen/ Acta Biomaterialia 9 (2013) 90499064