Three fold cross-links strengthen aligned fiber

bundles by resisting premature rupture of covalent

backbones
H. Shabbir1 , C. Dellago 1
and M. A. Hartmann2
1
Faculty of Physics, University of Vienna, Boltzmanngasse 5, 1090 Vienna, Austria
2
Ludwig Boltzmann Institute of Osteology at Hanusch Hospital of WGKK and
AUVA Trauma Centre, Meidling, Vienna, Austria
E-mail: [email protected], [email protected]

Abstract.
The influence of the coordination of (reversible) cross-links on the mechanical
properties of aligned fiber bundles is investigated. Two polymeric systems containing
cross-links of different coordination (two- and three fold coordination) but having the
same binding energy are investigated. In particular, the response to loading of these
systems is compared. Mechanical parameters (strength, stiffness and work to fracture)
are obtained by computational loading tests. The influence of coordination is studied
for simple test systems with pre-defined topologies that maximize strength as well as
for more realistic fiber bundles containing nine chains. The results show that a higher
coordination of cross-links has a beneficial effect on the strength and the stiffness of the
systems, while the work to fracture was found larger for the system having a smaller
coordination of cross-links. It can be concluded that controlling the coordination of
cross-links is a versatile tool to specifically tailor the mechanical properties of polymeric
structures.

PACS numbers: ***

Keywords: ***

Submitted to:
Influence of cross-link coordination on the mechanical behavior of polymers 2

1. Introduction

Structural properties of polymers are rather independent from the properties of their
backbones. A common strategy to specifically tailor the mechanical performance of
natural and synthetic polymers is to (reversibly) cross-link the polymer [1]. A cross-link
is an additional bond connecting distant parts of one polymer or two distinct polymers.
One prominent technological example is the vulcanization of rubber. In this process,
natural rubber is permanently cross-linked with sulfur bridges, thereby dramatically
enhancing its stiffness, strength and wear properties [2, 3].
In biological materials, cross-links are often considerably weaker than the covalent
backbone that holds the structure together. Consequently, upon loading, these bonds
rupture before the covalent backbone fails. The rupture of cross-links unveils the
segments of the structure that were previously shielded from loading, unraveling what
is picturesquely called the hidden length [4]. Cross-link rupture and hidden length
unraveling are highly efficient mechanisms to dissipate large amounts of energy and
obtain enhanced mechanical properties [5, 6]. Furthermore, these bonds are often
reversible, meaning that cross-links may reform after the load is released. As a
consequence, these biological materials show self healing properties even in the absence
of living cells [7]. In the biological context, such cross-links are often referred to as
sacrificial bonds [4] and can be found in numerous materials such as mussel byssus
threads [8, 9], bone [10, 11, 12], wood [13] or spider silk [14, 15]. The nature of these
bonds ranges from weak hydrogen bonds in spider silk [16, 17, 18] to relatively strong
metal coordination bonds in the mussel byssus that have binding energies approximately
a quarter the strength of the backbone [19, 20].
These special mechanical properties of cross-links have aroused a considerable
amount of theoretical and modeling effort to describe, understand and predict their
structural and mechanical effects. These studies include the investigation of stiff
networks [21], the unbinding transition of semiflexible polymer bundles [22], the influence
of network connectivity and entanglement on network mechanics [23], as well as
calculating the force-extension relation of cross-linked polymers [24, 25]. Also, the
effect of hidden length opening on the mechanical properties of biological materials
was explicitly studied [26, 27, 28].
In all these aforementioned studies, a cross-link is defined as an additional bond
between either two parts of the same polymer (intra-chain cross-link) or between two
different polymers (inter-chain cross-link). Microscopically, this means that cross-links
are modeled to form between two monomers. While this situation holds true in many
materials, there are also important other classes of cross-links that connect more than
two monomers. Typical examples of such cross-links are metal-coordination complexes,
like the Fe-DOPA or Zn-histidine complex found in the mussel byssus [29, 8, 30]. Such
a coordination complex consists of one metal ion bound to several organic ligands
[8, 31, 32, 33]. These metal complexes form fast and reversibly, have a high strength and
provide the material with remarkable mechanical properties including high toughness
Influence of cross-link coordination on the mechanical behavior of polymers 3

and stiffness, enormous extensibility up to 100%, self healing and exceptional adherence
in wet conditions [7, 9].
There are only few studies that explicitly investigate the influence of the
coordination of cross-links on the mechanical behavior. In the present paper, we
generalize a model presented in reference [34] to study the effect of cross-linking a single
chains to the important case of aligned fiber bundles. We refer to the coordination of
cross-links as the number of ligands bound to a metal ion in a coordination complex,
i.e, the number of monomers involved in one cross-link. We call a two-fold coordinated
cross-link a dimer or bis-complex and a three-fold coordinated cross-link a trimer or
tris-complex. The deformation behavior of aligned fiber bundles that are cross-linked by
three-fold or two-fold coordinated cross-links is compared. The physical properties of
the cross-links are modeled using a generic approach given by the framework provided
by the reactive empirical bond order (REBO) potential [35, 36]. REBO-like potentials
give the possibility to control the coordination of cross-links.
The obtained results indicate that inter-chain cross-links play a crucial role in the
mechanical behavior of aligned, interacting chain bundles. As was shown previously, the
presence of cross-links may deteriorate the strength of aligned fiber bundles compared
to the non cross-linked case [37]. This effect is further analyzed in the current study
and the results show that a higher coordination of cross-links may partially restore the
strength.
The paper is structured as follows. First, the model is briefly explained and the
mechanics of single and double polymer chains with pre-defined cross-link topologies are
discussed. Then, the mechanical performance of aligned fiber bundles containing three-
fold coordinated cross-links is comprehensively investigated and compared to the case
of two-fold coordination only. The influence of cross-link density and grafting density
is discussed. Finally, the results are summarized and possible future research directions
are discussed.

2. The Model

The inspiration of this work stems from metal coordination bonds, where several
(organic) ligands bind to a single metal cation [29]. Due to the long range nature
of the Coulomb force, a first principles treatment of such bonds is complicated [38, 39]
and a mechanical assessment of these bonds is even more scarce [20]. This is why we
decided to choose a generic, coarse-grained approach to describe such bonds. While the
used approach is tractable and simple enough to study rather large systems, it gives
the possibility to explicitly study the influence of the coordination of cross-links on the
mechanical behavior of polymeric systems which is the main objective of this work. As
the model was already described elsewhere [34], here we only shortly recapitulate its
main feature and describe the additions necessary to model chain bundles.
The systems investigated in this paper are either composed of one, two or nine
polymer chains each composed of NC = 50 beads. This number was chosen because
Influence of cross-link coordination on the mechanical behavior of polymers 4

typical protein segments containing metal coordination bonds are relatively short
consisting of only 30 to 80 amino acids [40]. The contour length of the polymer is
given by Lc = (NC − 1)r0 where r0 is the equilibrium bond length. The first and last
bead of each chain are permanently grafted to a substrate and are held fixed during the
simulations. For the system consisting of nine chains, the grafting points are located on
a square lattice with lattice constant d defining the grafting density. In the directions
perpendicular to the chain axis, periodic boundary conditions are used (figure 1).
The polymers are modeled using a bead-spring model. Hard sphere radius R is
assigned to each bead to impede physical overlapping of beads. The covalent interactions
along the backbone are described using a Morse potential. The excluded volume and
covalent interactions are as(follow
0 rij > R
UEV (rij ) = (1)
∞ else
Uij (rij ) = E0 (1 − e−α(rij −r0 ) )2 − 1
 
(2)
The cross-links are introduced by declaring Ns of the beads as sticky. The sticky beads
refer to a possible ligand of a metal ion, thus only these sites can form cross-links.
ρs = Ns /NT is the cross-link number density in the system (NT denotes the total
number of particles in the system). The preferred coordination of cross-links is controlled
using the framework provided by the REBO (Reactive Empirical Bond Order) potential
[36, 35]. The main feature of REBO-like potentials is that the strength of a bond is
dependent on the local coordination, i.e., the number of neighbors of a given particle.
The relevant equations read
1 X
URCL = [fR (rij ) + bij fA (rij )] fc (rij ) (3)
2 <i6=j>

fR = Ae−λrij , fA = −Be−µrij , (4)

where the sum runs over all pairs of sticky sites. fR and fA are pair-wise repulsive and
attractive contributions, respectively. fc (rij ) is a continuous cut-off function

 1
 h i rij < R0
1 rij −R0
fc (rij ) = 2
1 + cos π ∆ R0 < rij < R0 + ∆ (5)

0 rij > R0 + ∆


and the bond-order parameter bij depends on the local environment of each sticky site
(i.e., the bond angle θijk between three sites)
1
bij = (1 + β n ξ n )− 2n
X
ξ = g(θijk )fc (rik )
k6=i,j

c2 c2
g(θijk ) = 1 + − (6)
d2 d2 + (h − cos θijk )2
The sum in ξ runs over all sticky sites k (with the exception of i and j) inside the
cut-off region of site i. In particular, REBO potentials are reactive, meaning that
Influence of cross-link coordination on the mechanical behavior of polymers 5

Table 1. Simulation parameters describing covalent bonds and reversible cross-links.

The unit of length is set by the hard sphere radius R, the energy scale is given by kB T .
The two parameter sets given correspond to a system that can only form dimers (I)
and where the most stable cross-link is a trimer (II). The parameters are chosen such
that the binding energy of one complex is a quarter of a covalent bond.

Covalent Bonds Parameter Set I Parameter Set II

(Dimers only) (Trimers & Dimers)

E0 /kB T 200.0 A/kB T 1101324 367108

R - B/kB T 14840 5396
r0 /R 3.0 λ·R 3.33 3.33
α·R 2.0 µ·R 1.66 1.66
n 0 4
β - 1.0
h = cos θ0 - 0.5
c - 8.0
d - 2.0
R0 /R 4.2 4.2
∆/R 1.2 1.2

bond formation, rupture and reformation are explicitly taken into account without any
external input. The values of the parameters used in the simulations are summarized
in table 1.
In the chosen approach, metal-coordination complexes are described in a coarse-
grained manner via effective interactions between the ligands (the sticky sites). In this
description the metal ion itself is not explicitly considered in the description (see the
top part of figure 1). It would be desirable to obtain the effective interaction parameters
needed for the REBO potential via detailed ab initio calculations of the coordination
complexes. Unfortunately, such calculations are difficult and only a limited number of
investigations on the mechanics of metal coordination complexes using first principles
approaches exist [41, 42, 20]. Thus, the parameters used in this work were chosen
empirically to fulfill known characteristics of these bonds. First, the binding energy
of a cross-link complex is approximately a quarter of a covalent bond. Second, a tris-
complex is energetically the most stable complex in the system. Third, the equilibrium
bond length of the cross-links and covalent bonds are identical. The numerical values of
the parameters fulfilling these constraints and used in the simulations are summarized
in table 1.
The main focus of this work is to elucidate the influence of the coordination of cross-
links on the mechanical behavior of polymeric systems. This is done by comparing the
mechanical behavior of two systems having the same number and distribution of sticky
sites, but differing in the coordination of the formed cross-links. In the first system, a
Influence of cross-link coordination on the mechanical behavior of polymers 6

Figure 1. Top: Schematic of the model used. A linear polymer is described as

a collection of covalently bound hard spheres, (black lines indicate the backbone
of the polymer). Non-sticky, end and sticky beads are shown in black, red and
grey, respectively. Sticky sites have the ability to form cross-links via coordination
with metal ions. Cross-links are described in a coarse-grained manner via effective
interactions between the sticky sites themselves (grey bars). Consequently, the metal
ion and its interactions are not modeled explicitly (dashed green lines). Bottom:
snapshots of the system consisting of nine chains for two different types of cross-links:
two-fold coordinated cross-links (left) and three-fold coordinated cross-links (right). In
the image non-sticky beads are removed for better visualization, end and sticky beads
are shown. The green rectangular box represents the the system size.
Influence of cross-link coordination on the mechanical behavior of polymers 7

two-fold coordinated dimer is the only stable cross-link that may form (parameter set
I). In the second system, the energetically most stable form of cross-links is a three-fold
coordinated trimer, but also a two-fold coordinated dimer has a nonzero binding energy
(parameter set II). The two parameter sets used for modeling the cross-links in these
two systems are listed in table 1. The parameters are chosen such that the energy of
one complex as well as the equilibrium bond length are the same in both systems.
In the current work, we present load-displacement curves obtained in computational
displacement controlled loading tests. Using Monte Carlo (MC) simulations, the
polymer was stretched gradually from an initial compressed state characterized by a
small end-to-end distance L/Lc = 0.15 until the system failed via rupture of all chains
above the contour length L/Lc = 1. During each elongation step of ∆L/Lc = 0.0002, the
load is computed by averaging 5000 independent configurations. A standard Metropolis
algorithm was used to obtain these configurations [43].

3. Results and Discussion

3.1. Single chain

To gain a first understanding of the influence of the topology of cross-links on the
mechanical performance of cross-linked fibers, first a single chain with a predefined
topology of cross-links was studied. In the upper part of figure 2, the initial topology of
the investigated two- and three-fold coordinated structures are shown. In the figure, the
sticky sites are numbered consecutively and cross-links are indicated by grey bars, black
lines depict the covalent backbone. Following the definition given in reference [4], we
refer to the hidden length (indicated by the blue color) as the part of the backbone that
is shielded from loading by the cross-links. In both systems (parameter set I and II) the
cross-link complexes carry the same amount of energy by construction. This means that
although the systems differ in sticky site density, their ground state energy is the same.
The topology of cross-links is chosen such that the complexes are loaded in parallel
leading to an effective shear force on the backbone. Such a topology distributes the
applied load over all cross-links, meaning that the strength of the system is correlated
with the number of cross-links [44, 28]. This cross-link topology resembles a minimal
(parallel) beta-sheet structure and can be found in many mechanically strong biological
materials such as spider silk, amyloid fibers and muscle protein titin [45]. Corresponding
load-displacement curves are shown in the lower part of figure 2. The relative force (i.e.,
the force relative to the strength of the non cross-linked polymer F/F0max ) is plotted as
a function of the reduced length (i.e., the end-to-end distance over the contour length
L/Lc ). (Ultimate) strength is defined as the maximum load a structure can take. In
particular, the strength is the maximum load observed in the force-extension curve.
The obtained load-displacement curves show that in the case of parameter set I
five cross-links forming the shown topology are sufficient to rupture the backbone of
the polymer before the contour length is reached. In the load-displacement curve (blue
Influence of cross-link coordination on the mechanical behavior of polymers 8

line in figure 2) no cross-link rupture can be observed; the rise of the load around an
elongation of approximately 0.6 corresponds to backbone rupture. In the chosen beta-
sheet like topology the cross-links are loaded in parallel, meaning that they cooperatively
deform and share the load. Because the energy (and strength) of one two-fold complex
is equal to the quarter of one covalent bond, five of these complexes are sufficient to
take more load than the backbone [37]. Furthermore, the strength of the system is
slightly higher than the strength of the non cross-linked polymer (F/F0 )max = 1.06.
This slight increase stems from the fact that the system fails at extensions below one.
This effectively hinders large fluctuations of the backbone that are known to deteriorate
the mechanical behavior [46, 47].
The load-displacement curve of the system simulated with parameter set II shows a
more complex mechanical behavior than the classical two-fold coordinated system (red
curve in figure 2). Although, the binding energy of a single cross-link complex is the
same in both systems, no premature rupture of the covalent backbone (i.e., failure of the
backbone before the contour length) is observed in the three-fold coordinated system.
Comparison with the non-cross-linked reference system shows that the system recovers
its extensibility, while maintaining its strength. Furthermore, the systems shows a
dramatic increase in the work to fracture W compared to the other two systems. The
work to fracture is a measure of the toughness of the system and can be obtained
by calculating the area under the load-displacement curve. The work to fracture is
normalized with the work to fracture of the non cross-linked reference system W0 . It
is found that W/W0 increases from 0.61 to 1.67 from the two-fold coordinated system
(parameter set I) to the three-fold coordinated system (parameter set II).
The load-displacement curve obtained for parameter set II can be roughly separated
in two distinct regions: region 1 ranges approximately from L/Lc = 0.3 to 0.7 and
consists of several peaks of different height. Region 2 starts at extensions larger than 0.7
and shows a much more continuous loading, where no single peaks can be observed (see
figure 2). Each force peak observed in region 1 corresponds to the (partial) rupture of
one cross-link complex. Figure 3 shows the mean strain in the structure at peak load (the
numbering of the peaks corresponds to the numbering given in the load-displacement
curve in figure 2). The graphs corresponding to the first two peaks are enlarged to show
better visualize the course of the backbone (solid arrows in the graph corresponding to
peak 1) and the flow of the strain through the structure (dashed arrows in the graph
corresponding to peak 2). For each snapshot also the corresponding extension L/Lc and
the number of intact and broken bonds are given. Comparison of the backbone and of the
force flow shows that the investigated topology shields a large portion of the backbone
from being loaded (the hidden length). Following the subsequent rupture events, it can
be seen how the three-fold coordinated complexes gradually decompose into two-fold
coordinated complexes. It needs 5 rupture events until this process is completed and
the initially three-fold coordinated complexes are completely transformed into two-fold
coordinated complexes. Then, the two-fold coordinated cross-links rupture one by one
until the hidden length is completely unraveled (at L/Lc ≈ 0.65). The unfolding of the
Influence of cross-link coordination on the mechanical behavior of polymers 9

(F/F0 )max (L/Lc )max W/W0

Parameter set I 1.06 0.61 0.61
Parameter set II 0.99 1.06 1.67
Non cross-linked 1.0 1.056 1.0

Figure 2. Initial configurations (top) and corresponding load-displacement curves

(bottom) of systems containing five cross-link complexes of different coordination. The
topologies are designed such that they resemble a beta-sheet structure and that the
cross-links are loaded in parallel. The chains have a total number of 50 monomers. In
the load-displacement curves, the blue and red lines correspond to the two (parameter
set I) and three-fold coordinated system (parameter set II), respectively. The grey
line denotes the deformation of a reference system without any cross-links. The table
shows the strength, maximum extension and work to fracture of the different systems.
Strength is normalized with the theoretical strength of a non cross-linked system, work
to fracture is normalized with the work to fracture of the non cross-linked system.
Influence of cross-link coordination on the mechanical behavior of polymers 10

hidden length gives the possibility to some of the sticky monomers to reform. Most
of the reformed cross-links are of independent topology [46]. The second region in the
load-displacement curve corresponds to the rupture of these independent cross-links.
Important remark regarding the normalized strain in figure 3, sacrificial and covalent
bonds are deformed distinctly because these bonds have different stiffness. Therefore,
strain in sacrificial and covalent bonds are normalized by their respective maximum
value in each topology.
The load displacement curve as well as the corresponding mechanical parameters
(figure 2) show clearly that the deformation behavior of three- and two-fold coordinated
cross-link complexes is very different, although both carry the same amount of binding
energy. From a topological perspective, this difference stems from a different geometric
arrangement of bonds. While the two-fold coordinated complexes may form a topology
that is loaded in pure shear via a cooperative deformation of all bonds, three-fold
complexes consist of three bonds, two of which loaded cooperatively and one loaded
in direction of the backbone. As an example, consider the first three-fold complex
consisting of sticky beads 1, 2 and 9 in figure 2. Effective bonds 1-9 and 2-9 are loaded
cooperatively, while bond 1-2 is loaded along the backbone. The same observation holds
true for each three-fold coordinated complex—out of 3 effective bonds, 2 are loaded in
parallel, 1 along the backbone. It is these bonds along the backbone that are loaded non-
cooperatively and rupture first, thus transforming the three- into two-fold coordinated
complexes. These bonds rupture independently, thus preventing the structure from
backbone failure.

3.2. Two chain system

In a next step, we investigated the behavior of a system consisting of two parallel linear
chains. In contrast to a single chain, a multi-chain system allows also for the presence of
inter-chain cross-links, which have been shown to play a decisive role in the mechanics
of aligned chain bundles [37]. Figure 4 shows the investigated topology of two- and
three-fold coordinated complexes. This topology is similar to the one investigated for
the single chain (figure 2), but is formed between two chains instead of only one. Each
chain has a length of 23 beads and contains five cross-link complexes. The topology
is designed such that most of the cross-links are inter-chain (the two-fold coordinated
case consists solely of inter-chain cross-links). The lateral distance between the chains
was set equal to one covalent bond length. Corresponding load-displacement curves
are shown in the lower part of figure 4. Blue and red lines represent the mechanical
response of the systems simulated with parameter set I and II, respectively, and the
grey line corresponds to the bare non cross-linked polymer as reference.
The bottom part of figure 4 shows the obtained load-displacement curves by plotting
the relative force F/F0max versus the reduced length L/L0 (as before F0max is the strength
of the non cross-linked system). The system containing two-fold coordinated cross-links
only (blue line) shows the same behavior as was already reported in reference [37].
Influence of cross-link coordination on the mechanical behavior of polymers 11

Figure 3. Polymer topology and strain distribution in each bond close to bond
rupture for parameter set II. Sacrificial and covalent bonds are deformed distinctly
due to the difference in the stiffness of these bonds. Therefore, strain in sacrificial
and covalent bonds are normalized by their respective maximum values in each figure.
The numbering of peaks corresponds to the numbering given in figure 2. The first two
images are enlarged to better visualize the course of the backbone (solid arrows) and
the flow of the strain (dashed arrows). It is evident that the investigated cross-link
topology shields a large portion of the backbone from loading (hidden length).
Influence of cross-link coordination on the mechanical behavior of polymers 12

(F/F0 )max (L/Lc )max W/W0

Parameter set I 0.53 1.06 0.91
Parameter set II 0.87 1.057 1.01
Non cross-linked 1.0 1.06 1.0

Figure 4. Starting topology (top) and load-displacement curves (bottom) of systems

consisting of two chains and simulated with parameter sets I and II, respectively. Both
systems show a similar extensibility as the non cross-linked system, while the strength
of the three-fold coordinated system (parameter set II) and the reference system
is significantly higher compared to the system forming two-fold coordinated cross-
links only (parameter set I). The table shows the corresponding strength, maximum
extension and work to fracture.
Influence of cross-link coordination on the mechanical behavior of polymers 13

The strength of five inter-chain cross-links loaded cooperatively exceeds the strength of
the backbone. The two large peaks observed at L/Lc ≈ 0.88 and 1.06 correspond to
the permanent failure of the backbone of the two chains. Because the two chains fail
independently, the strength of the system is reduced approximately by a factor of two
compared to the non cross-linked system, where both chains take the load cooperatively
and fail simultaneously. Thus, the presence of cross-links effectively weakens the system.
Nevertheless, this effect can be partially reversed, when some disorder is introduced in
the system. Cooperative loading of the cross-links is possible only, if the cross-links are
spaced regularly. Whenever some disorder is introduced in their spatial distribution,
the load experienced by the cross-links at a given extension will be different and they
will fail sequentially not exceeding the strength of the backbone.
The mechanical behavior of the three-fold coordinated system is different (red
line). The topology is designed such that most of the cross-links form (effective) inter-
chain bonds. Nevertheless, having two chains only, it is not possible to prevent the
formation of intra-chain cross-links when all possible bonds should be closed. The
main difference in the mechanical behavior compared to the system simulated with
parameter set I is that the polymer backbone does not rupture prior to the contour
length. This improves the mechanical properties like strength and work to fracture
of the system containing three-fold coordinated cross-link compared to the dimer only
case. Values for the mechanical parameters are shown in the table in figure 4. The
strength increases from (F/F0 )max = 0.53 to 0.87. While, the presence of cross-links
weakens a multi-chain system in general due to presence of inter-chain cross-links that
are loaded cooperatively, this weakening is strongly reduced for the three-fold compared
to the two-fold coordinated system. In contrast to the single chain system discussed
in the previous section, the work to fracture scales only slightly with the coordination,
increasing roughly only 10 % from 0.91 to 1.01.
While the load-displacement curve of the system simulated with parameter set I
consists of two large peaks that correspond to the backbone rupture of the two involved
chains, the corresponding curve of the system containing three-fold coordinated cross-
links is more structured. Before the structure finally fails, several smaller peaks (labelled
from 1 to 4 in figure 4) can be observed. These correspond to the rupture and reformation
of cross-links. Figure 5 shows the strain distribution in each bond before the rupture
event directly at peak load. The load is distributed such that the intra-chain cross-links
rupture first. This leads to the formation of a purely two-fold coordinated topology
consisting solely of inter-chain cross-links. The transformation of three-fold to two-fold
topology is completed when peak 4 is reached. In total 7 two-fold coordinated complexes
have formed and peak 4 corresponds to their rupture. Exactly at L/Lc = 1, the load is
shifted to the backbone. From now on covalent bonds start elongating until the structure
fails.
Influence of cross-link coordination on the mechanical behavior of polymers 14

Figure 5. Topology and strain distribution for the system consisting of two chains.
Sacrificial and covalent bonds are deformed distinctly due to the difference in the
stiffness of these bonds. Therefore, strain in sacrificial and covalent bonds are
normalized by their respective maximum values in each figure. The numbering of
peaks corresponds to the numbering given in figure 4. As in the single chain case the
deformation behavior is such that three-fold coordinated complexes first transform into
two-fold coordinated ones, until the system is purely two-fold coordinated at peak 4.

3.3. Aligned fiber bundle

After having discussed the simplified situations of one and two chains, now the more
realistic situation of an aligned fiber bundle consisting of nine chains shall be discussed.
The starting configurations are prepared by slowly reducing the end-to-end distance of a
system without sticky sites from L/Lc = 1 to 0.15. Then, Ns randomly chosen beads are
declared sticky, while it is ensured that two sticky beads are not direct neighbors along
the chain. The system is equilibrated and cross-links may form between sticky sites that
are spatially close. To obtain load-displacement curves, the system is stretched until all
chains are ruptured at elongations of L/Lc = 1.1. The presented curves are averages
over 48 independent starting configurations. Strength, stiffness, work to fracture and
the elongation at which the first chain rupture occurs are computed as a function of
sticky site density (ρs = Ns /NT ) and grafting density (defined by the lateral distance d
between the chains). The grafting density is strongly influencing the ratio of inter- to
intra-chain cross-links and is, thus, one main factor determining the mechanical behavior
of flexible, aligned fiber bundles. It has been shown that the inter-chain cross-links are
responsible for a dramatic decrease in the strength of two-fold coordinated systems [37].
Figure 6 shows the load-displacement curves of the two investigated systems as
well as the number of inter- and intra-chain cross-links for several grafting densities
(d/R = 7.5, 15.0, 22.5 and 30), respectively. The load-displacement curves show
that the system containing three-fold coordinated cross-links shows in general a higher
strength compared to the system containing solely dimers. Only in cases where the
chains are that much separated that no inter-chain cross-links may form and the chains
Influence of cross-link coordination on the mechanical behavior of polymers 15

deform completely independently (separations larger than d/R = 30.0) the strength
of both systems coincides, approaching the strength of the non cross-linked polymer.
Furthermore, the load-displacement curve of the three-fold coordinated system shows a
much smoother course than the corresponding load-displacement curve of the two-fold
coordinated system. In particular, the rupture of single peaks is clearly visible in the
the latter system, especially in the two cases of lowest grafting density (i.e, d/R = 22.5
and 30).
One observation in the microscopic deformation behavior of the two investigated
system is that the number of intra-chain cross-links decays much faster in the three-fold
coordinated compared to the two-fold coordinated system (see right column of figure
6). This is because in the system containing trimers, it is the trimers that rupture
first as can be seen in the coordination plot shown in figure 7. This figure shows the
relative abundance of cross-links of different coordination as a function of elongation
for a grafting density of d/R = 30.0. At this grafting density, the behavior of intra-
chain cross-links can be directly visualized as due to the large distance of the chains
no inter-chain cross-links are present. The figure shows that the number of trimers
drastically decreases starting from L/Lc ≈ 0.4. The number of dimers shows a much
less pronounced decrease over the entire deformation range. A sharp decrease to zero can
only be seen shortly before the contour length is reached. This confirms that in systems
containing three-fold coordinated cross-links it is mostly the rupture of trimers that is
responsible for the decline in intra-chain cross-links. While a broken trimer removes
two bonds out of the system, in the two-fold coordinated system a rupture of a dimer
reduces the number of bonds only by one. It is this disproportion that is responsible for
the faster decline of intra-chain cross-links in the system containing trimers compared
to the purely two-fold coordinated one.
Figure 8 summarizes the mechanical properties of the investigated systems as a
function of sticky site density and grafting density. Shown are the extension at which
the first backbone rupture occurs (L/Lc )r , the strength (F/F0 )max , the work to fracture
W/W0 and the stiffness Y /Y0 . Blue and red lines correspond to the properties of the
systems simulated with parameters sets I and II, respectively. The grafting density can
be used to tailor the behavior of the system from non-interacting independent chains
(d/R = 30) with zero inter-chain cross-links to highly interacting chains characterized
by a large number of inter-chain cross-links. An important observation is that in the
three-fold coordinated system the first polymer backbone rupture occurs above the
contour length for all investigated sticky site and grafting densities, i.e. (L/Lc )r > 1.0.
This behavior is different to the two-fold coordinated system where (L/Lc )r < 1.0
for interacting chains and high sticky site densities. It is this premature rupture of
the backbone that is responsible for the dramatic loss in strength for the two-fold
coordinated system at high sticky site densities as can be seen in the upper, right part
of the figure. For high cross-link densities, the strength in the system simulated with
parameter set I decreases almost by a factor of two compared to the non-interacting
case. Also in systems simulated with parameter set II a decline in the strength can be
Influence of cross-link coordination on the mechanical behavior of polymers 16
1.0
Inter-Chain Cross-links
Intra-Chain Cross-links

Fraction of Cross-links
0.8

0.6

0.4

0.2

0.0
0.2 0.4 0.6 0.8 1.0
L/Lc
1.0

Fraction of Cross-links
0.8

0.6

0.4

0.2

0.0
0.2 0.4 0.6 0.8 1.0
L/Lc
1.0
Fraction of Cross-links

0.8

0.6

0.4

0.2

0.0
0.2 0.4 0.6 0.8 1.0
L/Lc
1.0
Fraction of Cross-links

0.8

0.6

0.4

0.2

0.0
0.2 0.4 0.6 0.8 1.0
L/Lc

Figure 6. Load-displacement curves (left) and fraction of inter- and intra-chain

cross-links (right) of a system consisting of nine aligned chains. The different curves
correspond to different grafting densities, defined by the lateral distance d/R of the
grafting points of the chains: (A) d/R = 7.5, (B) 15, (C) 22.5 and (D) 30. Blue and
red lines correspond to systems simulated with parameter set I and II, respectively.
The black line indicates the response of the non cross-linked reference curve. The
sticky site density was ρs = 0.36 for all curves. The results shown are averages over
48 independent simulation runs.
Influence of cross-link coordination on the mechanical behavior of polymers 17

1.0
Coordination plot
Monomer
Dimer

Relative Abundance
0.8
Trimer

0.6

0.4

0.2

0.0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
L/LC
Figure 7. Coordination plot of the system containing trimers (parameter set II) with
d/R = 30.0. The figure shows the fraction of cross-links of different coordination
in the system. While—at the onset of deformation—most cross-links are three-fold
coordinated, there is also a considerable amount of dimers in the system that also
possess a non-zero binding energy. Only intra-chain cross-links are present, because
at the grafting density chosen the chains are that much separated that no inter-chain
cross-links may form.

observed for high cross-link densities, but this loss in strength is much less pronounced
than in the former system. The system containing three-fold coordinated cross-links
possesses a higher strength than the dimer system for all grafting densities except
for d/R = 30, where the strength becomes independent of sticky site density and
approaches the strength of a non cross-linked polymer. The average number of non
ruptured chains at contour length is a direct measure of the strength. As this number is
higher in the three-fold coordinated system, this system shows a higher strength than the
two-fold coordinated system. Responsible for the premature rupture of the backbone
are inter-chain cross-links that increase in number with increasing sticky-site density
and increasing grafting density (see figure 6). Inter-chain cross-links may effectively
transmit forces from one chain to the other. Many of such cross-links may be loaded
cooperatively leading to an effective shear force on the backbone exceeding the strength
of the backbone (see also figure 4).
The work to fracture is the energy needed to elongate the system until the system
permanently fails. The work to fracture heavily depends on the number of cross-links
present in the system that need to be broken before the system fails. Thus, it is
not surprising that the work to fracture increases with increasing sticky site density
and increasing grafting density in both systems. More surprising is that the two-fold
coordinated system consistently shows a higher work to fracture than the three-fold
coordinated system for all studied sticky site and grafting densities although the energy
of each complex is the same. This, is because, first, at the onset of deformation only
roughly 70 % of the cross-links are three-fold coordinated (see figure 7). The remaining
30 % have a lower coordination and, thus, a lower binding energy. Second, although
Influence of cross-link coordination on the mechanical behavior of polymers 18

the sticky site density is the same in both systems, the number of cross-link complexes
can differ considerably. A sticky site density of 24 % allows for the formation of 54
bis-complexes, but only 36 trimers. Consequently, at the same sticky site density the
ground state energy of the system simulated with parameter set I is lower than the one
for the system described with parameter set II.
The stiffness of a material is defined as its resistance against deformation and it is
measured as the initial slope of the load-displacement curve. Figure 6 shows that cross-
linked systems show in general a higher resistance against deformation and thus, possess
a higher stiffness Y /Y0 . This observation is confirmed by the numerical values of the
stiffness as depicted in the bottom, right graph of figure 8. The stiffness is normalized
with the stiffness of a single cross-link bond Y0 obtained as the second derivative of
the used potential (parameter set I and II) evaluated at equilibrium bond length. The
results indicate that the stiffness shows a strong dependence on the number of cross-links
present, while the grafting density is only of little influence. Furthermore, the system
containing trimers (parameter set II) shows a consistently higher stiffness than the
system containing dimers only (parameter set I). This is because the number of bonds
is higher in the first compared to the latter system. Although—for the same sticky
site density—the maximum number of dimers is larger than the maximum number of
trimers, the number of bonds is higher in the three fold coordinated system. Analysis
of the microscopic details shows that for the investigated structures the ratio of the
number of bonds is relatively constant ranging from 1.5 to 2. Thus, it can be concluded
that the different number of bonds explains the observed differences in stiffness for the
two systems.

4. Summary and Conclusions

The influence of the coordination of cross-links on the mechanical response of aligned

fiber bundles was investigated. In particular, two systems differing in their coordination
of cross-links were compared. In the first system, a classical cross-link of two-fold
coordination is the only cross-link that may form, while in the second system a three-fold
coordinated cross-link is the most stable configuration. The energetics of the cross-links
were chosen such that both cross-link complexes (the dimer as well as the trimer) carry
the same binding energy. To analyze the effect of coordination, simple systems consisting
of one and two chains were studied, prior to the investigation of a more realistic bundle
containing nine chains. The results clearly show that the coordination of cross-links is an
important parameter to understand the mechanical performance of fiber bundles. The
main finding is that a higher coordination of cross-links partially reverts the reduction
in strength that was recently predicted for classical two-fold coordinated cross-links [37].
This is because three-fold coordinated systems effectively prevent premature rupture of
the covalent backbone frequently observed in two-fold coordinated systems.
The work to fracture and stiffness have also been computed for the two systems. The
two-fold coordinated systems shows a higher work to fracture, but a reduced stiffness
Influence of cross-link coordination on the mechanical behavior of polymers 19

1.05 1.0
1.00
0.8
0.95

(F/F0)max
L Lc ) r

0.90 0.6
( /

0.85
0.4
d/R = 7.5
0.80
d/R = 15.0
d/R = 22.5 0.2
0.75
d/R = 30.0
0.70 0.0
0.1 0.2 0.3 0.4 0.1 0.2 0.3 0.4
ρs ρs

2.2
0.0012

2.0
0.0010

1.8 0.0008
W Wo

Y/Yo
/

1.6 0.0006

1.4 0.0004

1.2 0.0002

0.1 0.2 0.3 0.4

ρs ρs
0.1 0.2 0.3 0.4

Figure 8. Mechanical properties of cross-linked fiber bundles for different sticky site
(ρs ) and grafting densities. Blue and red lines correspond to systems simulated with
parameter set 1 and 2, respectively. Shown are the appearance of the first backbone
rupture (L/Lc )r (top left), the strength (F/F0 )max (top right), work to fracture W/W0
(bottom left) and stiffness Y /Y0 (bottom right).

compared to the three-fold coordinated one. The work to fracture mostly depends
on the number of closed cross-link complexes in the system. Having the same sticky
sites density, the possible number of cross-link complexes is higher in the first system
compared to the latter. The main factor determining the stiffness is the number of
bonds present, the grafting density is only of little influence.
Metal coordination bonds often found as cross-linkers in biological materials are
the source of inspiration for the presented work. The results suggest that living
organisms may tune the mechanical properties of some of their load bearing entities by
tailoring the coordination of cross-links and/or their density. The different mechanical
behavior of inter- and intra-chain cross-links indicate that the mechanical performance of
fiber bundles can be most effectively enhanced by incorporating strong intra-chain and
weak inter-chain cross-links. This is because inter-chain cross-links are often loaded
cooperatively causing the premature rupture of the backbone and, thus, degrading
the overall strength of the bundle. Recently, it was shown experimentally that the
mechanical properties of an elastomer could be greatly enhanced by incorporating weaker
Influence of cross-link coordination on the mechanical behavior of polymers 20

hydrogen bonds and stronger metal coordination bonds into a chemically cross-linked
network [48]. The effect of the presence of bonds of different strength will be investigated
in future simulation studies.

Acknowledgments

The computational results presented have been achieved using the Vienna Scientific
Cluster (VSC). Financial support of the Austrian Science Fund (FWF) within project
P 27882-N27 is gratefully acknowledged.

References

[1] Akihiro Abe, Karel Dusek, and Shiro Kobayashi, editors. Crosslinking in Materials Science,
volume 184 of Advances in Polymer Science. Springer, 2005.
[2] Jan-Michael Y. Carrillo, Fred C. MacKintosh, and Andrey V. Dobrynin. Nonlinear elasticity:
From single chain to networks and gels. Macromolecules, 46(9):3679–3692, 2013.
[3] Paul M. Goldbart, Horacio E. Castillo, and Annette Zippelius. Randomly crosslinked
macromolecular systems: Vulcanization transition to and properties of the amorphous solid
state. Advances in Physics, 45(5):393–468, 1996.
[4] Georg E. Fantner, Emin Oroudjev, Georg Schitter, Laura S. Golde, Philipp Thurner, Marquesa M.
Finch, Patricia Turner, Thomas Gutsmann, Daniel E. Morse, Helen Hansma, and Paul K.
Hansma. Sacrificial Bonds and Hidden Length: Unraveling Molecular Mesostructures in Tough
Materials. Biophys. J., 90:1411, 2006.
[5] Bettye L. Smith, Tilman E. Schäffer, Mario Viani, James B. Thompson, Neil A. Frederick,
Johannes Kindt, Angela Belcher, Galen D. Stucky, Daniel E. Morse, and Paul K. Hansma.
Molecular mechanistic origin of the toughness of natural adhesive, fibres and composites. Nature,
399:761, 1999.
[6] Thomas Gutsmann, Tue Hassenkam, Jacqueline A. Cutroni, and Paul K. Hansma. Sacrificial
bonds in polymer brushes from rat tail tendon functioning as nanoscale velcro. Biophys. J.,
89:536, 2005.
[7] Matthew J. Harrington, Himadri S. Gupta, Peter Fratzl, and J. Herbert Waite. Collagen insulated
from tensile damage by domains that unfold reversibly: In situ X-ray investigation of mechanical
yield and damage repair in the mussel byssus. J. Struct. Biol., 167:47, 2009.
[8] Matthew J. Harrington, Admir Masic, Niels Holten-Andersen, J. Herbert Waite, and Peter Fratzl.
Iron-Clad Fibers: A Metal-Based Biological Strategy for Hard Flexible Coatings. Science,
328:216, 2010.
[9] Clemens N. Z. Schmitt, Yael Politi, Antje Reinecke, and Matthew J. Harrington. Role of
sacrificial protein-metal bond exchange in mussel byssal thread self-healing. Biomacromolecules,
16(9):2852–2861, 2015. PMID: 26296100.
[10] Georg E. Fantner, Tue Hassenkam, Johannes H. Kindt, James C. Weaver, Henrik Birkedal, Leonid
Pechenik, Jacqueline A. Cutroni, Geraldo A. G. Cidade, Galen D. Stucky, Daniel E. Morse, and
Paul K. Hansma. Sacrificial bonds and hidden length dissipate energy as mineralized fibrils
separate during bone fracture. Nat. Mater., 4:612, 2005.
[11] Georg E. Fantner, Jonathan Adams, Patricia Turner, Philipp J. Thurner, Larry W. Fisher, and
Paul K. Hansma. Nanoscale Ion Mediated Networks in Bone: Osteopontin Can Repeatedly
Dissipate Large Amounts of Energy. Nano Lett., 7:2491, 2007.
[12] Himadri S. Gupta, Peter Fratzl, Michael Kerschnitzki, Gunthard Benecke, Wolfgang Wagermaier,
and Helmut O. K. Kirchner. Evidence for an elementary process in bone plasticity with an
activation enthalpy of 1 eV. J. R. Soc. Interface, 4:277, 2007.
Influence of cross-link coordination on the mechanical behavior of polymers 21

[13] Jozef Keckes, Ingo Burgert, Klaus Frühmann, Martin Müller, Klaas Kölln, Myles Hamilton,
Manfred Burghammer, Stephan V. Roth, Stefanie Stanzl-Tschegg, and Peter Fratzl. Cell-wall
recovery after irreversible deformation of wood. Nat. Mater., 2:810, 2003.
[14] Nathan Becker, Emin Oroudjev, Stephanie Mutz, Jason P. Cleveland, Paul K. Hansma, Cheryl Y.
Hayashi, Dmitrii E. Makarov, and Helen G. Hansma. Molecular nanosprings in spider capture-
silk threads. Nat. Mater., 2(4):278, April 2003.
[15] Theodor Ackbarow, Xuefeng Chen, Sinan Keten, and Markus J. Buehler. Hierarchies, multiple
energy barriers, and robustness govern the fracture mechanics of alpha-helical and beta-sheet
protein domains. PNAS, 104:16410, 2007.
[16] Sinan Keten and Markus J. Buehler. Strength limit of entropic elasticity in beta-sheet protein
domains. Phys. Rev. E, 78:061913, 2008.
[17] Sinan Keten and Markus J. Buehler. Geometric Confinement Governs the Rupture Strength of
H-bond Assemblies at a Critical Length Scale. Nanoletters, 8:743, 2008.
[18] Sinan Keten, Zhiping Xu, Britni Ihle, and Markus J. Buehler. Nanoconfinement controls stiffness,
strength and mechanical toughness of /beta-sheet crystals in silk. Nat. Mater., 9:359, 2010.
[19] D. S. Barnes and L. D. Pettit. Stereoselectivity in enthalpy changes accompanying the formation
of metal complexes of histidine and other amino-acids. J. Inorg. Nucl. Chem., 33:2177, 1971.
[20] Zhiping Xu. Mechanics of metal-catecholate complexes: The roles of coordination state and metal
types. Sci. Rep., 3:2914, October 2013.
[21] Jan Wilhelm and Erwin Frey. Elasticity of stiff polymer networks. Phys. Rev. Lett., 91:108103,
Sep 2003.
[22] Jan Kierfeld, Torsten Kühne, and Reinhard Lipowsky. Discontinuous Unbinding Transitions of
Filament Bundles. Phys. Rev. Lett., 95:038102, 2005.
[23] Nicholas Agung Kurniawan, Søren Enemark, and Raj Rajagopalan. The role of structure
in the nonlinear mechanics of cross-linked semiflexible polymer networks. J. Chem. Phys.,
136(6):065101, 2012.
[24] Panayotis Benetatos, Stephan Ulrich, and Annette Zippelius. Force-extension relation of cross-
linked anisotropic polymer networks. New J. Phys., 14(11):115011, 2012.
[25] Panayotis Benetatos, Alice von der Heydt, and Annette Zippelius. Tension-induced binding of
semiflexible biopolymers. New J. Phys., 16(11):113037, 2014.
[26] Ahmed E. Elbanna and Jean M. Carlson. Dynamics of polymer molecules with sacrificial bond and
hidden length systems: Towards a physically-based mesoscopic constitutive law. PLoS ONE,
8:e56118, 2013.
[27] Charles K. C. Lieou, Ahmed E. Elbanna, and Jean M. Carlson. Sacrificial bonds and hidden
length in biomaterials: a kinetic constitutive description of strength and toughness in bone.
Phys. Rev. E, 88:012703, 2013.
[28] S. Soran Nabavi, Matthew J. Harrington, Peter Fratzl, and Markus A. Hartmann. Influence of
sacrificial bonds on the mechanical behaviour of polymer chains. Bioinspired, Biomimetic and
Nanobiomaterials, 3:139, 2014.
[29] Elena Degtyar, Matthew J. Harrington, Yael Politi, and Peter Fratzl. The Mechanical Role of
Metal Ions in Biogenic Protein-Based Materials. Angew. Chem. Int. Ed., 53r(45):12026–12044,
2014.
[30] Antje Reinecke, Luca Bertinetti, Peter Fratzl, and Matthew J. Harrington. Cooperative behavior
of a sacrificial bond network and elastic framework in providing self-healing capacity in mussel
byssal threads. Journal of Structural Biology, 196(3):329 – 339, 2016.
[31] Bruce P. Lee, P.B. Messersmith, J.N. Israelachvili, and J.H.Waite. Mussel-Inspired Adhesives and
Coatings. Annu. Rev. Mater. Res., 41:99, 2011.
[32] Niels Holten-Andersen, Matthew J. Harrington, Henrik Birkedal, Bruce P. Lee, Phillip B.
Messersmith, Ka Yee C. Lee, and J. Herbert Waite. pH-induced metal-ligand cross-links inspired
by mussel yield self-healing polymer networks with near-covalent elastic moduli. Proc. Natl.
Acad. Sci. USA, 108:2651, 2011.
Influence of cross-link coordination on the mechanical behavior of polymers 22

[33] Devin G. Barrett, Dominic E. Fullenkamp, Lihong He, Niels Holten-Andersen, Ka Yee C. Lee,
and Phillip B. Messersmith. pH-Based Regulation of Hydrogel Mechanical Properties Through
Mussel-Inspired Chemistry and Processing. Adv. Funct. Mater., 23:1111, 2013.
[34] Huzaifa Shabbir and Markus A Hartmann. Influence of reversible cross-link coordination on the
mechanical behavior of a linear polymer chain. New Journal of Physics, 19:093024, 2017.
[35] J. Tersoff. New Empirical approach for the structure and energy of covalent systems. Phys. Rev.
B, 37:6991, 1988.
[36] J. Tersoff. Empirical Interatomic Potential for Carbon, with Applications to Amorphous Carbon.
Phys. Rev. Lett., 61:2879, 1988.
[37] S. Soran Nabavi and Markus A. Hartmann. Weak reversible cross links may decrease the strength
of aligned fiber bundles. Soft Matter, 12:2047, 2016.
[38] Michael R. Nechay, Crystal E. Valdez, and Anastassia N. Alexandrova. Computational treatment
of metalloproteins. The Journal of Physical Chemistry B, 119(19):5945–5956, 2015. PMID:
25839835.
[39] Crystal E. Valdez, Amanda Morgenstern, Mark E. Eberhart, and Anastassia N. Alexandrova. Pre-
dictive methods for computational metalloenzyme redesign - a test case with carboxypeptidase
a. Phys. Chem. Chem. Phys., 18:31744–31756, 2016.
[40] Matthew J. Harrington and J. Herbert Waite. Holdfast heroics: comparing the molecular and
mechanical properties of Mytilus californianus byssal threads. J. Exp. Biol., 210:4307, 2007.
[41] Shabeer A. Mian, Leton C. Saha, Joonkyung Jang, Lu Wang, Xingfa Gao, and Shigeru Nagase.
Density functional theory study of catechol adhesion on silica surfaces. The Journal of Physical
Chemistry C, 114(48):20793–20800, 2010.
[42] Shabeer Ahmad Mian, Li-Ming Yang, Leton Chandra Saha, E. Ahmed, Muhammad Ajmal, and
Eric Ganz. A fundamental understanding of catechol and water adsorption on a hydrophilic
silica surface: Exploring the underwater adhesion mechanism of mussels on an atomic scale.
Langmuir, 30(23):6906–6914, 2014. PMID: 24835420.
[43] David P. Landau and Kurt Binder. A Guide to Monte-Carlo Simulations in Statistical Physics.
Cambridge University Press, 2009.
[44] Kilho Eom, Pai-Chi Li, Dmitrii E. Makarov, and Gregory J. Rodin. Relationship between the
mechanical properties and topology of cross-linked polymer molecules: Parallel strands maximize
the strength of model polymers and protein domains. J. Phys. Chem. B, 107(34):8730, 2003.
[45] Harold P. Erickson. Stretching Single Protein Molecules: Titin Is a Weird Spring. Science,
276(5315):1090–1092, 1997.
[46] S. Soran Nabavi, Matthew J. Harrington, Oskar Paris, Peter Fratzl, and Markus A. Hartmann.
The role of topology and thermal backbone fluctuations on sacrificial bond efficacy in mechanical
metalloproteins. New J. Phys., 16:013003, 2014.
[47] S. Soran Nabavi, Peter Fratzl, and Markus A. Hartmann. Energy dissipation and recovery in a
simple model with reversible cross-links. Phys. Rev. E, 91:032603, Mar 2015.
[48] Jie Liu, Sheng Wang, Zhenghai Tang, Jing Huang, Baochun Guo, and Guangsu Huang. Bioinspired
Engineering of Two Different Types of Sacrificial Bonds into Chemically Cross-linked cis
polyisoprene toward a High Performance Elastomer. Macromolecules, 49:8593, 2016.