Available online at www.sciencedirect.

com

Acta Biomaterialia 4 (2008) 1114–1117

www.elsevier.com/locate/actabiomat

Brief communication

Characterization of the mechanical properties of tough

biopolymer fibres from the mussel byssus of Aulacomya ater
O.P. Troncoso, F.G. Torres *, C.J. Grande
Polymers and Composites Group, Department of Mechanical Engineering, Catholic University of Peru, Lima 32, Peru

Received 14 September 2007; received in revised form 11 December 2007; accepted 25 January 2008
Available online 7 February 2008

Abstract

Byssus fibres are tough biopolymer fibres produced by mussels to attach themselves to rocks. In this communication, we present
the mechanical properties of the byssus from the South American mussel Aulacomya ater which have not been previously reported
in the literature. The mechanical properties of the whole threads were assessed by uniaxial tensile tests of dry and hydrated spec-
imens. Elastoplastic and elastomeric stress–strain curves were found for byssal threads from A. ater in the dry and hydrated state,
respectively. The results obtained from mechanical tests were modelled using linear, power-law-type and Mooney–Rivlin relation-
ships. These methods for dealing with tensile measurements of mussel byssus have the potential to be used with other stretchy
biomaterials.
Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Mussel byssus; Mechanical properties; Biopolymer fibres

1. Introduction proteins [6]. Studies performed on byssal threads from

Byssus fibres are tough biopolymer fibres produced by
mussels to attach themselves to rocks. They are secreted
by the mussel foot in an injection moulding-like process
[1] and need to meet specific mechanical requirements of
stiffness, strength, toughness and extensibility in order to
withstand water velocities of over 10 m s1 and water accel-
eration of around 400 m s2 in the marine intertidal zone
[2,3]. Their high strength and extensibility make the tough-
ness of the distal and proximal regions six times greater
than that of the human tendon collagen and comparable
with that of Kevlar and carbon fibres [4].
Mussel byssus is a complex structure divided in three
regions (Fig. 1): the stem, the thread and the plaque [5].
The stem is the structure that supports each byssal thread,
whereas the plaque is the adhesive pad located at the end
of the thread which contains at least six different types of
*
Corresponding author. Tel.: +51 1 6262000; fax: +51 1 6262461.
E-mail address: fgtorres@pucp.edu.pe (F.G. Torres).
Mytilus edulis mussels [7–9] found three different forms
of collagen, namely byssal precollagen-P (preCol-P), bys-
sal precollagen-D (preCol-D) and precollagen-NG (pre-
Col-NG).
Several studies have reported the use of tensile tests to
assess the mechanical properties of byssus [10–14]. Vac-
caro and Waite [10] and Sun et al. [14] observed the rela-
tionship between the medium and the stiffness of threads,
while Bell and Gosline [12] examined the material and
structural properties of three mussel species and found
similar results for all of them. These studies used meth-
ods that neglect the fact that byssal fibres undergo finite
deformation as they can be stretched by about 100%
before breaking [4].
The aim of this paper is to report the mechanical prop-
erties of dry and hydrated byssus from a mussel species not
previously reported (Aulacomya ater), and also to present a
set of methods for dealing with tensile measurements of
mussel byssus by modelling the results obtained using lin-
ear, power-law and Mooney–Rivlin relationships.

1742-7061/$ - see front matter Ó 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.actbio.2008.01.014
 O.P. Troncoso et al. / Acta Biomaterialia 4 (2008) 1114–1117 1115

tion, previous studies involving uniaxial testing of

biological tissues have used the true stress values [17].
According to Green [18], the strain components (c) of a
simple tensile test (rxx = constant, ryy = rzz = 0) of a unit
cube with sides parallel to the x-, y-, z-axes, deformed into
a cuboid of dimensions k1, k2, k3 parallel to the x-, y-,
z-axes, may be written in matrix form
0 1 01 2 1
c11 c12 c13 2
ðk1  1Þ 0 0
B C B C
@ c21 c22 c23 A ¼ B @ 0 1
2
ðk22  1Þ 0 C:
A
c31 c32 c33 1 2
0 0 2
ðk3  1Þ
Fig. 1. Schematic view of a mussel byssus (left) and a scanning electron ð1Þ
microscope micrograph of byssus from A. ater (right).
Soft tissues are generally considered as imcompressible
[17]. For such materials, the volume of the deformed unit
2. Experimental cube is constant. Thus
k1 k2 k3 ¼ 1: ð2Þ
Specimens of A. ater, a relatively large marine mussel
from the South American west coast that had not previ- And, due to simple tension
ously been studied, were used. The mussels were collected k2 ¼ k3 : ð3Þ
from bedrocks of the Lima shore and brought to the labo-
ratory where the byssal threads were extracted and washed Finally
with distilled water. Some of the threads were then air dried 1
and the others were kept in either distilled water or sea pffiffiffiffiffi ¼ k2 : ð4Þ
k1
water at room temperature (20 °C).
The diameters of the threads were obtained from digital For a solid with initial length different from 1, the exten-
micrographs, using a BRUNEL BMZ trinocular stereomi- sion ratio (k) is defined as
croscope equipped with a graticule in conjunction with lf
image analysis software (ImageJ, National Institutes of k¼ ; ð5Þ
lo
Health, USA).
The overall structure of the byssus threads was studied where lf is the final (deformed) length and lo is the initial
using an RJ Lee scanning electron microscope, using volt- strength. When the solid is stretched along the x-axis, the
ages in the range 10–20 kV. Specimens were gold coated as cross-sectional area is decreased by the ratio k2k3 = 1/k.
described in a previous report from this laboratory [15]. Thus, the true stress (r) of an applied force (F) and initial
A benchtop tensile testing machine, equipped with a 5 N area (Ao) is
(±1%) load-cell, specifically built for this project, was used F
r¼ k: ð6Þ
in the experiments. The deformation of the fibres was Ao
assumed to be equal to the separation of the cross-heads.
Eqs. (1) and (6) are used to calculate the true stress of a
This assumption is reasonable considering that the compli-
simple tension test.
ance of the machine is negligible compared to that of the
Tensile tests results are summarized in Table 1. The
fibres [16]. The ends of threads were glued to thin card-
results of the tests carried out on byssus immersed in dis-
board frames which were held in the grips and then cut
tilled water are not significantly different from those
before starting the tests. The byssus fibres were tested at
obtained from byssus immersed in sea water. By contrast,
three states: dry, immersed in distilled water and immersed
in sea water with a cross-head speed of 10 mm min1. Five
samples were tested per condition. Initial gauge lengths
were about 15 mm. Statistical tests performed include the Table 1
Material properties of mussel byssus under different conditions
analysis of variance (ANOVA) and the Student–New-
man–Keuls method for pairwise comparisons. The null Condition Ultimate Ultimate Ultimate true
engineering stress strain stress (MPa)
hypothesis was rejected at a significance level of 0.05.
(MPa)
Dry 144.71 ± 33.58a 0.27 ± 0.018a 183.73 ± 44.18a
3. Results and discussion
Immersed in 59.47 ± 25.02b 0.84 ± 0.097b 110.20 ± 49.40a
distilled water
As mentioned before, byssal threads undergo finite Immersed in sea 70.16 ± 20.46b 0.79 ± 0.275b 127.78 ± 50.02a
deformation during tensile testing. Thus the engineering water
stress, which refers to the initial undeformed cross-sec- Values are mean ± SD. Letters indicate significantly different groups
tional area, is not accurate. In order to avoid underestima- (Student–Newman–Keuls pairwise comparisons).
1116 O.P. Troncoso et al. / Acta Biomaterialia 4 (2008) 1114–1117
hydrated byssus is more extensible than dry byssus as it
reaches higher maximum strain values. Differences between
true and engineering ultimate stresses are similar for both
dry and hydrated byssus.
Some representative stress–strain curves comparing true
and engineering stresses, are shown in Fig. 2. The true and
the engineering stress are similar only until 5% strain.
Beyond that point, the true stress curves have a steeper
slope than the engineering stress curves. At a strain around
100% true stress values are twice as high as the correspond-
ing engineering stress values. This confirms that using the
engineering stress to describe the mechanical properties
of byssus might lead to underestimation.
In addition, the representative curves (Fig. 2) show that
dry byssus behaves as an elastoplastic polymer while humid
threads immersed in both distilled and sea water display
elastomeric behaviour. This might be due to the fact that
water molecules act as a plastifying agent that interacts
with the long-chain polymer molecules that form the byssal
threads. Similar types of behaviour have been observed in
other protein-based biofibres, such as spider silk [16,19].
Fig. 2a. Representative stress–strain curves of dried byssal threads
showing (a) true stress and (b) engineering stress.
Fig. 2b. Representative stress–strain curves showing (a) true stress and (b)
engineering stress of hydrated byssal threads immersed in sea water; and
(c) true stress and (d) engineering stress of hydrated byssal threads
immersed in distilled water.
The elastoplastic behaviour of dry byssus threads can be
observed in Fig. 2a. The curves have two clearly different
regions. Initially, they follow a linear elastic behaviour
(E = 1746.3 ± 256.4 MPa) at small strains until they reach
a critical strain (0.0364 ± 0.0090) where plastic deforma-
tion begins. Power-law equations are often used to model
this type of behaviour, as follows:
r ¼ K  en ð7Þ
where r and e are the stress and the strain in the plastic
area, respectively, and K and n are constants determined
by fitting the model to the experimental data. The value
of n can be obtained from
d logðrÞ
n¼ : ð8Þ
d logðeÞ
The suitability of the power-law model for the tensile
behaviour of dried byssal threads is clear from Fig. 3 where
a representative stress–strain curve for dry byssus is shown.
The average values of n and K are 0.599 ± 0.034 and
358.92 ± 63.27, respectively.
In contrast to dry byssus, lower stresses were needed to
deform the hydrated byssal threads. As hydrated byssus
displayed an elastomer-like tensile behaviour, rubber elas-
ticity theory was used to model the stress–strain relation-
ships. The Mooney–Rivlin model has been used in the
past to model the deformation of rubber-like materials
and has been applied to the study of soft tissues such as
tendons and ligaments [20–23].
The Mooney–Rivlin model is based on a semiempirical
equation that relates the true stress (r) to the extension
ratio (k) by using two constants (C1, C2). These two con-
stants are material properties and are obtained from exper-
imental results when the reduced stress ([r]) is plotted
against the inverse of the extension ratio (1/k) according to
r 2C 2
½r ¼ ¼ 2C 1 þ : ð9Þ
k  1=k2 k
A representative stress–strain curve for hydrated byssus is
shown in Fig. 4. The experimental points fit well with the
model. The average values of C1 and C2 for byssus im-
mersed in distilled water are 10.54 ± 3.72 and
Fig. 3. Experimental () and theoretical (–) stress–strain curve for dry
byssus.
 O.P. Troncoso et al. / Acta Biomaterialia 4 (2008) 1114–1117
Fig. 4. Experimental () and theoretical (–) stress–strain curve for
hydrated byssus immersed in sea water.
24.24 ± 8.07 MPa, respectively, whereas for byssus im-
mersed in sea water, C1 and C2 are 5.63 ± 4.70 and
18.01 ± 10.10 MPa, respectively.
4. Conclusions
In summary, we present a set of methods for studying
the tensile behaviour of mussel byssus that can be used with
other stretchy biomaterials. In particular, the mechanical
behaviour of threads from A. ater was modelled using a
power-law equation and the Mooney–Rivlin relation of
rubber elasticity theory. As expected, the ultimate tensile
stress of hydrated byssal threads was lower than the ulti-
mate tensile stress of the dry specimens.
Acknowledgements
The authors would like to thank the International Foun-
dation for Science (IFS, Stockholm, Sweden, RGA F/4194-
1) for financial support. Experimental support with the
SEM from Prof. J. Ruiz at the Materials Laboratory of
PUCP is greatly acknowledged. One of the authors (F.T.)
wants to thank the UNESCO Visiting Professorship of
Patronat Sud-Nord at the University of Valencia for finan-
cial support during part of this research.
References
[1] Waite JH. The formation of mussel byssus: anatomy of a natural
manufacturing process. In: Case ST, editor. Results and problems in
cells differentiation, vol. 19. Berlin: Springer; 1992. p. 27–54.
1117
[2] Denny MW. Wave forces on intertidal organisms: a case study.
Limnol Oceanogr 1985;30:1171–87.
[3] Bell EC, Denny MW. Quantifying ‘‘wave exposure”: a simple device
for recording maximum velocity and results of its use at several field
sites. J Exp Mar Biol Ecol 1994;181:9–29.
[4] Gosline J, Lillie M, Carrington E, Guerette P, Ortlepp C, Savage K.
Elastic proteins: biological roles and mechanical properties. Philos
Trans R Soc B 2002;357:121–32.
[5] Bairati A. The byssus of the mussel Mytilus from the molecules to the
organ: functional performance resides in the ultrastructural assembly.
In: Lanzavecchia G, Valvassori R, editors. Form and function in
zoology. Modena: Mucchi Editore; 1991. p. 163–77.
[6] Waite JH. Adhesion à la moule. Integr Comp Biol 2002;42:1172–80.
[7] Coyne KJ, Qin XX, Waite JH. Extensible collagen in mussel byssus: a
natural block copolymer. Science 1997;277:1830–2.
[8] Qin X, Coyne KJ, Waite JH. Tough tendons: mussel byssus has
collagen with silk-like domains. J Biol Chem 1997;272:32623–7.
[9] Qin XX, Waite JH. A potential mediator of collagenous block
copolymer gradients in mussel byssal threads. Proc Natl Acad Sci
USA 1998;95:10517–22.
[10] Vaccaro E, Waite JH. Yield and post-yield behavior of mussel byssal
thread: a self-healing biomolecular material. Biomacromolecules
2001;2:906–11.
[11] Waite JH, Vaccaro E, Sun C, Lucas JM. Elastomeric gradients: a
hedge against stress concentration in marine holdfasts? Philos Trans
R Soc B 2002;357:143–53.
[12] Bell EC, Gosline JM. Mechanical design of mussel byssus: material
yield enhances attachment strength. J Exp Biol 1996;199:1005–17.
[13] Lucas JM, Vaccaro E, Waite JH. A molecular, morphometric and
mechanical comparison of the structural elements of byssus from
Mytilus edulis and Mytilus galloprovincialis. J Exp Biol 2002;205:
1807–17.
[14] Sun C, Vaccaro E, Waite JH. Oxidative stress and the mechanical
properties of naturally occurring chimeric collagen-containing fibers.
Biophys J 2001;81:3590–5.
[15] Torres FG, Dı́az R. Morphological characterisation of natural fibre
reinforced thermoplastics (NFRTP) processed by extrusion, com-
pression and rotational moulding. Polym Polym Comp 2004;12:
705–18.
[16] Pérez-Rigueiro J, Viney C, Llorca J, Elices M. Silkworm silk as an
engineering material. J Appl Polym Sci 1998;70:2439–47.
[17] Fung YC. Biomechanics. New York: Springer; 1990, p. 242–320.
[18] Green A. Theoretical elasticity. Oxford: Clarendon Press; 1968, p. 80.
[19] Pérez-Rigueiro J, Elices M, Guinea GV. Controlled supercontrac-
tion tailors the tensile behaviour of spider silk. Polymer 2003;44:
3733–6.
[20] DeFrate LE, Li G. The prediction of stress-relaxation of ligaments
and tendons using the quasi-linear viscoelastic model. Biomechan
Model Mechanobiol 2007;6:245–51.
[21] Holzapfel GA. Nonlinear solid mechanics: a continuum approach for
engineering. New York: Wiley; 2000, p. 455.
[22] Johnson GA, Livesay GA, Woo SL, Rajagopal KR. A single integral
finite strain viscoelastic model of ligaments and tendons. J Biomech
Eng 1996;118:221–6.
[23] Weiss JA, Gardiner JC. Computational modeling of ligament
mechanics. Crit Rev Biomed Eng 2001;29:303–71.