Smart Materials and Structures

Related content
- 3D flexible NiTi-braided elastomer
Velcro-like fasteners based on NiTi micro-hook composites for smart structure applications
L Heller, D Vokoun, P Šittner et al.
arrays - Impact damage resistance and damage
suppression properties of shape memory
alloys inhybrid composites—a review
To cite this article: D Vokoun et al 2011 Smart Mater. Struct. 20 085027 S L Angioni, M Meo and A Foreman

- Numerical simulation of superelastic shape

memory alloys subjected to dynamicloads
Corneliu Cismaiu and Filipe P Amarante
dos Santos
View the article online for updates and enhancements.

Recent citations
- The rotation toughening mechanism of
barb–barbule joint in the barb delamination
of feathers
Qiang Chen et al

- A NiTi alloy weft knitted fabric for smart

firefighting clothing
Alenka Šalej Lah et al

- Hadha Afrisal et al

This content was downloaded from IP address 193.137.92.107 on 29/08/2020 at 16:18

IOP PUBLISHING SMART MATERIALS AND STRUCTURES
Smart Mater. Struct. 20 (2011) 085027 (13pp) doi:10.1088/0964-1726/20/8/085027

Velcro-like fasteners based on NiTi

micro-hook arrays
D Vokoun1 , P Sedlák2, M Frost2 , J Pilch1 , D Majtás1 and P Šittner1
1
Institute of Physics ASCR, v.v.i., Na Slovance 2, 182 00 Prague 8, Czech Republic
2
Institute of Thermomechanics ASCR, v.v.i., Dolejškova 1402/5, 182 00 Prague 8,
Czech Republic

Received 15 March 2011, in final form 14 June 2011

Published 28 July 2011
Online at stacks.iop.org/SMS/20/085027

Abstract
A recently developed Velcro-like fastener utilizes superelastic deformation of two interlocked
NiTi hooks when pulled apart. This work focuses on experimental analysis (evaluation of
normal detachment force at different temperatures) and modeling (simulation by a finite
element implemented SMA model) of the unhooking process. It is claimed that nonlinear
superelastic deformation of NiTi leads to unique properties of the NiTi hook fasteners such as
high strength (∼15 000 kg m−2 ), a significant increase of strength with increasing temperature,
absorption of impact loads, damping of mechanical vibrations, forceless contact or silent release
and better functioning in dirty environments compared to conventional Velcro fasteners.
(Some figures in this article are in colour only in the electronic version)

1. Introduction

Velcro-type releasable fasteners are typical examples of man-

made bio-inspired engineering applications. The inventor
of the first hook-and-loop fastener, Swiss engineer George
de Mestral, proposed the Velcro fastening mechanism while
observing sticking burdock seeds. Since then, other examples
of hook-and-loop fasteners have been found in Nature, e.g. [1].
Recently, a number of innovative Velcro-type fasteners made of
hard plastic [2] or steel [3] have been reported in the literature.
These innovative fasteners still utilize interlocked hooks/loops
of various shapes, sizes, arrangements and materials [4, 5]
released by a peel-off mechanism. Mechanical models able
to simulate the hook detachment force of conventional Velcro
fasteners have been published (e.g. [6]). It has already been
suggested that shape memory alloy (SMA) hooks may be
utilized in thermally releasable Velcro-type fasteners [7, 8]
exploiting the shape memory effect ( Af ∼ 70 ◦ C) on heating.
Recently, we have developed a Velcro-like fastener [9, 10]
Figure 1. (a) A photograph of a single NiTi hook, (b) a scheme of a
made of an array of superelastic NiTi hooks attached to fixture for shape setting of the SMA wire and cutting the wire into
a flexible substrate which is released via the conventional individual NiTi hooks.
peel-off mechanism. Originally, this research work was
motivated by the idea to develop a Velcro-like fastener for
silent use [11]. The original idea also involved the use unhooking process is rather different. It appeared that NiTi
of permanent magnets [12]. Later on we realized that, hook fasteners exhibit additional unique features originating
although these novel superelastic fasteners work in a similar from the superelastic bending deformation of NiTi hooks. We
way to conventional Velcro fasteners, the mechanics of the thus decided to investigate the mechanics of the unhooking

0964-1726/11/085027+13$33.00 1 © 2011 IOP Publishing Ltd Printed in the UK & the USA
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Figure 2. Tensile stress–strain curves of NiTi wire FWM #1 annealed at 450 ◦ C for 30 min: (a) a monotonic test until rupture at 20 ◦ C,
(b) cyclic tests at various test temperatures.

process in more detail (experimental characterization of the unhooking process, mainly at the curved surface of the NiTi
detachment of hook array prototypes and numerical modeling hooks, at places with the highest stress concentration. We
of the unhooking of two NiTi hooks). The results, including a have already tested NiTi hooks with the maximum curvature
brief summary of the novel features of NiTi hook fasteners, are of 2.1 mm−1 and wire diameter of 0.15 mm. However, these
reported in this work. geometrical parameters were not suitable for hook arrays.
The procedure employed to manufacture prototypes of Hence we have chosen a commercial superelastic NiTi wire
superelastic NiTi hook fasteners is briefly introduced in (FWM #1, chemical composition Ni 56.0 wt%, C 330 ppm,
section 2. Following that, results of mechanical tests aimed O2 280 ppm, H2 25 ppm and Ti balance) with a smaller
mainly at the evaluation of the detachment force in the normal diameter of 0.1 mm and slightly increased the hook curvature
direction at various temperatures are presented. The tests to compensate for the NiTi wire diameter decrease. Heat
were performed on single NiTi hooks, various NiTi hook treatment at 450 ◦ C for 30 min gives superelastic properties
arrays as well as conventional Velcro fasteners. Section 3 similar to straight annealed superelastic wire provided by the
deals with the simulation of the unhooking process using a producer.
numerically implemented SMA model developed especially In order to get a basic idea about the thermomechanical
for thin wire NiTi structures [13]. Finally, the novel features of superelastic properties of the shape-set hooks and determine
the superelastic NiTi hook fasteners are discussed in section 4 the material parameters for the modeling, we evaluated cyclic
based on the confrontation of the simulation and experimental tensile stress–strain responses of straight wire (figure 2(b))
results. given the same heat treatment. The results show that the
tensile superelastic responses of the wire are temperature
2. Experimental details dependent and reasonably stable upon tensile cycling in
the 10–70 ◦ C temperature range for which the fastener is
2.1. NiTi hooks developed. Although the yield stress for plastic deformation
Generally, NiTi wires can be shape-set into hook shapes reaches ∼1500 MPa (figure 2(a)), one needs to count with
(figure 1(a)) by heat treatment at temperatures of 300 to 500 ◦ C ∼0.5% irreversible strain upon cyclic tensile loading up to
while being constrained in a required shape. The aim of 700 MPa (figure 2(b)) which is not recoverable upon heating
the shape setting is twofold—to give the wire the required above Af . The large non-recovered strains measured at
hook shape (geometrical parameters) and to set its superelastic low test temperature below 0 ◦ C are fully recoverable upon
properties (material parameters) as desired. As will become heating above the Af temperature. Material parameters such
evident from the analysis performed in section 3, both are as the Young modulus, transformation strain, temperature
equally important for the fastener’s functionality. When dependence of transformation stress, etc utilized in modeling
pulling apart two interlocked NiTi hooks, the detachment force of the hook superelasticity were extracted from the stress–
depends mainly on the hook geometry, wire thickness, material strain responses in figure 2. More information about the
parameters characterizing the superelastic thermomechanical thermomechanical properties of FWM #1 wires including a list
properties (figure 2) and the environmental temperature. of material parameters can be found on the dedicated website
In our earlier work (a part of it is described in [9]), of SMA round-robin modeling activity [14]—an initiative
we tested the detachment forces of NiTi hooks of various aimed at the analysis of the performance of various SMA
curvatures and wire thicknesses. Our concern was to adjust models.
the normal detachment force (to be at least as large that NiTi wire in a cold-drawn state was shape-set in a furnace
of conventional Velcro fasteners) and to prevent large non- using a fixture in a form of thin steel plate (0.6 mm thick,
recoverable deformations at room temperature during the 6 mm wide and more than 100 mm long—see figure 1(b)).

2
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Figure 3. The scheme of the manufacturing procedure for NiTi hook arrays. The hooks are first inserted into a steel fixture (a plate with
drilled holes) with random orientation of the hook heads (a); the plate with hooks is then put into a container which is subsequently filled with
a two-component silicon based elastomer (b); the steel plate is removed after polymerization (c); and, finally, the empty space with the hook
ends is filled in by a resin or a different material which becomes the substrate of the array. A photo of a hook array is shown in (e).

The geometry of the shape setting tool, particularly the empty space left was filled with a connective material to
plate thickness and curvature at the edges, is crucial for the connect all the hooks’ ends and to become a substrate after
adjustment of the hook’s shape. The as-drawn wire was tightly polymerization or solidification (figures 3(a)–(d)). If we
wound around the steel plate. More than 1000 wire loops used epoxy resin as a connective material, after curing at a
were wound under a constant wire tension. Both the wire temperature of 80 ◦ C for 2 h, the resin became solid, forming
ends were fixed with screws to prevent loosening of the wire a rigid substrate. The rubber layer covering the hook heads
during the shape setting treatment in the furnace at 450 ◦ C was finally peeled off. The hooks were temporarily deformed
for 30 min. The transformation temperatures of the shape-set while this was done but they recovered back to their original
wire were determined by differential scanning calorimetry as shape due to their superelastic effect. Figure 3(e) shows the
Ms = −160 ◦ C, Af = 0 ◦ C. Individual NiTi hooks (figure 1(a)) final NiTi hook sample with the epoxy resin substrate. In
were cut from the shape-set superelastic wire structure. The order to evaluate the peel-off force, the same hook array was
hooks were further integrated into SMA hook arrays using a attached to a flexible substrate. To do that we followed the
manufacturing procedure described in section 2.2. Although same procedure and instead of resin we used thermoplastic
the hooks’ heights reach several millimeters, we call them elastomers containing polyolefins heated up to a temperature
micro-hooks because of their small diameter. of around 100 ◦ C. After cooling down, the substrate became
solid and flexible.
2.2. NiTi hook fasteners Five NiTi hook arrays (table 1) with various NiTi hook
densities, various hook areas and various substrates were
The functionality of a Velcro fastener is mainly characterized prepared for mechanical experiments. Sample E with a smaller
by the detachment force in the normal direction and the area was used for evaluation of the normal detachment force
peel-off force. Prototypes of various NiTi hook arrays on at various test temperatures. It was made with a smaller hook
rigid and flexible substrates were prepared and the normal area to fit the dimensions of the thermal Peltier tubular chamber
detachment force and peel-off force were evaluated on them (figure 4(a)) installed on the loading machine. Sample F was
experimentally. The fastener prototypes were manufactured prepared from a conventional Velcro fastener firmly glued to a
as hook arrays attached to a flat substrate with individual rigid substrate for comparison tests.
hooks oriented perpendicular to the substrate. All the hooks’
heads should be at the same height above the substrate surface.
2.3. Characterization of detachment and peel-off forces
Furthermore, we require that all the hooks are spread regularly
at the nodes of an imaginary square grid and the hooks’ heads 2.3.1. Hook arrays. The experimental setups used for
are oriented randomly. In order to meet the above-mentioned measurement of the detachment force on samples with
requirements, steel plates with the dimensions 22 mm × rigid substrates (figure 4(a)) and the peel-off force on
22 mm × 2 mm and an array of microchannels of diameter samples with flexible substrates (figure 4(b)) at room
0.2 mm laser-drilled into the plates were manufactured. The temperature were installed on the tensile deformation
holes were arranged in a regular square grid in an area of machine Walterbai dedicated to SMA wire testing and on a
4 cm2 . Samples with three different microchannel densities tensile deformation machine designed and assembled in the
were prepared using 121, 441, and 961 holes per plate. laboratory, respectively. Series of detachment force tests
The hooks in the array were aligned with their heads (about 10 measurements) were performed on each sample
about 3 mm above the plate’s surface. Then, the hook heads A–F at room temperature (figures 5–7). In the case of
were covered by a liquid layer of two-component silicon sample D (flexible substrate), only the peel-off force was
based elastomer (Sylgard® ). After polymerization at room measured (figure 6). Samples E and F were used for normal
temperature all the hook heads were embedded in the rubber- detachment force measurements in the temperature range 0–
like elastomer layer and the metal matrix was removed. The 80 ◦ C (figures 8 and 9).

3
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Table 1. NiTi hook array samples.

Mean maximum Mean fastener
Hook density Hook area Substrate normal detachment strength at 20 ◦ C
Sample (hooks cm−2 ) (cm2 ) type force at 20 ◦ C (N) (kg m−2 )
A 30 4 Rigid 0.7 178
B 110 4 Rigid 16.1 4103
C 240 4 Rigid 42.1 10 729
D 110 4 Flexible — —
E 240 1.3 Rigid 21.2 16 616
F Velcro 2 Rigid 4.6 2344

is a force maximum beyond which the force continuously

decreases down to zero—i.e. no sudden sharp force drops
occur during detachment of NiTi hook fasteners with higher
hook densities. The force–stroke curves from individual
tests display significant scatter. The thick black curves in
figures 5(a)–(c) represent the averaged force–stroke curves
over all the tests. The maximum forces scatter around the
mean value of the maximum detachment force. Most likely
this is due to the different numbers of interlocked hooks in
each measurement. The maximum detachment force linearly
increases (figure 5(d)) with increasing density of hooks in
the array. Maximal normal detachment force values were
obtained for samples C and E with 240 hooks cm−2 (table 1).
Figure 6 proves that the peel-off force is less than 20% of the
normal detachment force. The force–stroke curves measured
on sample E and the Velcro fastener (sample F) of the same
area at room temperature are compared in figure 7. Not only
is the strength of the Velcro fastener much lower, but the
character of the force–stroke curve is completely different.
Note also that the force–stroke curves of Velcro display sudden
force drops in spite of the large density of interlocked hooks in
loops.
The results of the detachment tests performed at various
temperatures on NiTi hook array E (figure 8) and the
conventional Velcro fastener (sample F) (figure 9) clearly
show the essential difference. While the maximum normal
detachment force of the NiTi hook array increases with
increasing temperature, it is almost temperature insensitive in
Figure 4. The experimental setup for measurement of the the case of the Velcro fastener. As these are single test data
detachment force in the normal direction (a) and the peel-off
force (b) of NiTi hook arrays with rigid (a) and flexible (b) (not the mean curves), however, the maximum normal forces
substrates. A tubular Peltier furnace can be seen in (a). are subject to statistical errors (see figure 5).

2.3.2. Single hook couple. In order to provide the

In each measurement, the contact was established by experimental data necessary for comparison with the numerical
simply allowing the NiTi hooks to self-interlock on coming simulations (section 3), force–stroke curves of two single
together. It should be pointed out that interlocking can be interlocked NiTi hooks were measured in similar detachment
made almost forcelessly and only a fraction of hooks become tests (figures 10 and 11). These responses represent the basic
interlocked. The interlocked hook arrays were pulled apart master curves of the unhooking superelasticity as will be
perpendicularly either to the plane of the rigid substrates discussed in section 3. Figure 10 shows several closed cycle
(samples A, B, C, E, F) or to the part of the flexible substrate force–stroke curves measured on two interlocked NiTi hooks at
which was firmly connected to the clamps (sample D). The a temperature of 20 ◦ C with progressively increasing maximum
force and stroke were continuously measured until complete stroke. Although there is no transformation plateau as in the
release of all hooks in the array. Figures 5(a)–(c) show tensile tests (figure 2), the force–stroke curves are hysteretic
the force–stroke curves from the series of detachment tests and nearly half of the supplied mechanical energy is dissipated
on samples A–C, respectively. Note that, in the case of in the largest loop. Figure 11 shows the results of a series of
arrays with higher hook densities (figures 5(b) and (c)), there detachment tests performed on two interlocked NiTi hooks in

4
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Figure 5. Normal detachment tests on NiTi hook arrays with variable hook densities ((a) 30, (b) 110 and (c) 240 hooks cm−2 ) at 20 ◦ C. Since
the individual force–stroke curves ((a)–(c)) exhibit significant scatter, the black thick line denoting the averaged response is included. (d) The
maximum detachment force versus hook density.

and there is a demand for silent Velcro fasteners [3] we

decided to perform experimental investigation of the levels of
acoustic noise during the detachment of NiTi hook fasteners
and conventional Velcro fasteners. The measurement was
carried out by a commercial company authorized for sound
level measurements. Figure 12 presents the results, showing
the experimental setup in the inset. The noise levels were
registered with a pre-polarized 1/2 inch free-field microphone
and the signal was analyzed with a sound level meter.
Before the measurements, both devices were calibrated and
background was measured. The detachment was performed
with a mechanical tester. Because of non-negligible levels
of background sound (sound coming from a cooling unit of
Figure 6. Normal (sample B) and peel-off (sample D) detachment the mechanical tester), the microphone was attached close
tests on NiTi hook arrays at 20 ◦ C.
(0.1 m) to the samples. The measurement accuracy was
±1.5 dB. NiTi hook sample B (rigid substrate) and Klettostar®
a wide temperature range from −40 to 80 ◦ C. All the tests (a commercial Velcro-type product made of plastic hooks and
were carried out with one couple of NiTi hooks. Before each
fiber loops firmly attached to the sample holder) with the same
test, the hook couple was heated to 100 ◦ C to reset its shape
fastener area (2 cm × 2 cm) were pulled apart in the normal
and eliminate the effects of previous history. This is important
particularly at temperatures below 0 ◦ C where the loading is direction while the noise was recorded. Among samples A–
not reversible, as documented by the partial loading–unloading C, sample B exhibited the maximum normal detachment force
at the lowest test temperature of −40 ◦ C. closest to that of the Klettostar sample. The results of the
spectral analysis (0–16 kHz) indicate that the time integrated
sound level L Zeq generated by the samples increases with
2.4. Fastener detachment noise
increasing frequency but the level is significantly lower for
Since the detachment of superelastic hook arrays appears to the NiTi hook fastener than for the conventional Klettostar
be relatively quiet compared to conventional Velcro fasteners Velcro fastener.

5
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Figure 7. Normal detachment tests on (a) NiTi hook arrays (sample E) and (b) a Velcro fastener (sample F) at 20 ◦ C. Force is recalculated for
equal substrate area 1 cm2 for easy comparison of strength. The black thick lines denote the averaged responses over about ten force–stroke
curves measured in separate tests.

Figure 8. Normal detachment tests on the NiTi hook array (sample E) at different temperatures: (a) force–stroke curves, (b) the temperature
dependence of the maximum detachment force.

Figure 9. Normal detachment tests on the Velcro fastener (sample F) at different temperatures: (a) force–stroke curves, (b) the temperature
dependence of the maximum detachment force.

3. Simulation of the NiTi hook superelasticity using fastener, exhibits quite different properties. We believe that
an SMA model this is because the mechanics of the unhooking process
in these two fasteners are different. The idea was that
The NiTi hook–hook fastener, although it works in an this is at least partially due to the nonlinear superelastic
apparently similar way to a conventional hook–loop Velcro response of NiTi wires (figure 2). This superelastic response

6
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Figure 10. Unhooking of two interlocked NiTi hooks at 20 ◦ C with

Figure 12. Results of the spectral analysis of the noise generated
partial cycles. Significant energy dissipation and hysteresis appears
during the normal detachment test at 20 ◦ C on a NiTi hook array
only upon loading beyond the knee point on the force–stroke curve.
(sample B) and Klettostar Velcro of the same substrate area and
comparable strength. The inset shows the measurement setup.

leads to temperature dependent nonlinear hysteretic stress–

strain behavior with large reversible strains at high stresses.
various temperatures. The simulation of the unhooking process
This is thought to result in a unique reversible force–stroke
predicts local stresses and strains in various parts of the hook
response of two interlocked NiTi hooks pulled apart (hereafter
during unhooking. Confronting these simulated maxima with
called ‘unhooking superelasticity’). In order to prove that,
the stress and strain reversibility limits (e.g. σmax = 1.5 GPa,
however, it is necessary to simulate the unhooking process by
εmax = 11%, see figure 2), simulation is used to save time
a numerically implemented SMA model (the distribution of
and material consuming experimental work when designing the
stresses and strains in NiTi hooks during unhooking is highly
nonhomogeneous). fastener for stable cyclic performance.
The simulation of the unhooking process has two principal When simulating the unhooking process, two important
goals. The first one is the simulation of force–stroke curves phenomena come into play. The first one is the highly
of two interlocked NiTi hooks for various material and nonlinear temperature sensitive deformation behavior in
geometrical parameters of the hooks and test temperatures. combined bending and tension due to the martensitic phase
Ultimately, assuming the NiTi array as a set of hooks, the transformation. The other phenomenon is the problem of the
simulation can be used to design NiTi hook fasteners for contact of the two hooks during unhooking. The point where
specific requirements—e.g. normal detachment force. The the interlocked hooks come into contact (loading force point)
other goal is to solve the problem of potential overloading and moves during the unhooking process. Frictional forces at the
irrecoverable plastic deformation of the hook when designing contact have a substantial influence on the resultant force–
the superelastic hook fastener. Obviously, plastic deformation stroke response. The notion of unhooking superelasticity thus
of hooks has to be avoided to assure reliable long time covers both the nonhomogeneous deformation of the hook and
performance of the superelastic fastener. In particular, it is the contact problem with friction.
important to determine the values and locations of stress and In order to simulate the thermomechanical behavior of
strain maxima in the hook during the unhooking process at NiTi wire structures as textiles we developed the uniaxial

Figure 11. (a) Unhooking of two interlocked NiTi hooks at various temperatures; (b) the temperature dependence of the maximum
detachment force. The dramatic change of the response in the first stage (stroke 0–0.3 mm) between 20 and 40 ◦ C is related to the absence of
deformation processes related to the R-phase at higher temperatures.

7
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Figure 14. A numerical simulation of the unhooking process: a

simulated force–stroke curve at 20 ◦ C with insets showing the hook’s
shape at the individual stages with denoted areas of high tensile
(red/gray) and compressive (blue/gray) strains due to the stress
Figure 13. Numerical implementation of the SMA model to simulate induced martensitic transformation: (a) the end of pure elastic
the unhooking process as a 2D problem: (a) mesh of the beam loading (minimum strain −1.1%, maximum strain 1.3%), (b) the
elements; (b) a single beam element consists of several truss martensitic phase transformation at the locations with the largest
elements in which the 1D SMA model iRLOOP is applied. deformation is completed (minimum strain −5.4%, maximum strain
8.1%) and (c) the maximum force attained (minimum strain −6.3%,
maximum strain 9.5%).
SMA constitutive model iRLOOP earlier [13, 15]. This model
describes well the thermomechanical hysteretic responses of
NiTi in tension and compression including partial cycles. into a set of truss elements representing parallel slices of the
Detailed comparison of the material model simulations with material (figure 13(b)). The material behavior within a single
experimental stress–strain superelastic curves in tension at truss element is simulated by the iRLOOP constitutive model.
various temperatures can be found in [14]. Without going into The formulation of the problem of unhooking supere-
details, let us recall here only the main properties of this model: lasticity is completed by a specification of the boundary
conditions. The straight end of the hook is fixed, i.e. the
(a) it is derived based on the concept of thermodynamic clamp boundary condition is applied. At the beginning of the
transformation force coupling the effects of stress and computation, the loading force point is located at the most
temperature on the martensitic phase transformation; distant point from the fixation plane. The Hook elongation
(b) it simulates reversible hysteretic stress–strain–temperature (stroke) is basically the distance of the loading force point
responses due to the stress induced martensitic transfor- from the fixation point. During unhooking, the location of the
mation and martensite reorientation; loading force point with respect to the hook is specified by the
(c) it captures the asymmetric stress–strain behavior of SMAs maximum angle between the normal axis of the beam element
in tension and compression; where the contact occurs and normal plane of the loading
(d) it captures partial hysteretic stress–strain–temperature force direction. When the prescribed maximum angle is
responses with a hysteresis algorithm capturing the return exceeded, the loading point moves to the nearest new position
point memory phenomenon; in which the angle is lower than the limit. The prescribed
(e) material parameters (see table 1 in [15]) are obtained from
maximum angle thus represents a kind of friction angle of
tensile experiments on wires including the different Young
the material (angle up to which the friction prevents sliding
moduli of austenite and martensite.
of the two interlocked hooks). This angle can be determined
The model was numerically implemented to treat experimentally. In this way, the simulations include the static
combined tension and bending of the NiTi hooks. A friction in the problem. This was not taken into account in
simple finite element (FE) computational tool for numerical previous work [9].
simulation of unhooking was developed and implemented in Simulation of the force–stroke curve characterizing the
MATLAB commercial software. As the loading force direction unhooking process is introduced in figure 14, illustrating
lies in the same plane as the micro-hook, the unhooking distinct stages of the simulated force–stroke curve together
process is treated as a planar problem. Since the principal with snapshots of the simulated shape and simulated
loading mode is bending and the ratio between the wire radius distribution of strain (tensile in red/gray, compressive in
and the radius of the curved parts of the micro-hook is low blue/gray) within the micro-hook. The general shape of the
enough, it appears reasonable to approximate the geometric hook with variable curvature was taken into account as the
structure of the micro-hook by a system of beam elements input geometry. The loading force point is denoted by an
(figure 13(a)) and assume that the normal lines to the cross arrow. Although the simulated force–stroke curve does not
section areas of the beam remain parallel to the beam axis reproduce the superelastic plateau of the NiTi wire deformed
during the deformation. Each beam element is further divided in tension, there is a change of tangent modulus at the knee

8
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

Figure 15. A comparison of experimental and simulated Figure 16. Numerical simulation of the unhooking process at 20 ◦ C:
force–stroke curves corresponding to unhooking of two interlocked the dissipated mechanical energy during an incomplete
hooks at various temperatures. The response beyond the force loading/unloading force–stroke cycle is partially due to the
maximum is due to the interaction between two hooks which are martensitic transformation (red/dark gray area) and partially due to
deformed slightly differently, which the symmetric simulation is not the friction (forward branch only) of the sliding contact between two
able to treat. interlocked hooks (yellow/light gray area).

point coming from the transition between purely elastic and parameters of the NiTi micro-hook as well as the hook density
superelastic material responses. It is essential to realize that, and size of the hook array to achieve the desired detachment
although much larger recoverable unbending (a change of hook force and elongation of the fastener. Generally, for optimum
radius) is possible with the NiTi hook than would be achievable functionality of NiTi hook fasteners, we expect a significant
with an elastic steel hook of the same geometric parameters, detachment force in the axial direction and strains attained in
the bending deformation is limited. a single micro-hook high enough so that energy dissipation
Simulated and experimental force–stroke curves of two due to the martensitic phase transformation can occur (>2%
micro-hooks at various temperatures are compared in figure 15. at room temperature), but low enough so that the plastic
Although there are differences between the simulated and the deformation is not reached in any hook part (<11% at room
experimental curves, the main trends are reliably captured. temperature). From the manufacturing point of view, a simple
Note, in particular, that the model is able to estimate well geometry consisting of straight parts and arches is favorable.
the knee point as well as the maximum detachment force At the same time, the minimum material consumption for a
and the maximum stroke. A certain disagreement between given wire radius is the cost-effective production requirement.
the calculated and the measured curves arises from the fact In order to find an improved shape of the micro-hook with
that the two real hooks acting in the experiment may not respect to all given requirements, simulations with varied
be completely identical and may affect one another during geometries of the hooks were performed. Based on the
unhooking. Furthermore, the presence of the R-phase likely results of these simulations, a new micro-hook prototype
influences the mechanical behavior of the hooks, particularly was proposed and manufactured and detachment tests were
its dependence on the test temperature, but the model is not performed on these hooks at room temperature for verification
yet able to take the strains due to the B2–R transformation into (figure 17). The experimental results agree quite well with
account. the numerical simulation of the force–stroke curve. Compared
With respect to the high strength and impact energy to the hook simulated earlier (figures 14 and 15) a slightly
absorption achievable with the NiTi hook fastener, it is of larger detachment force and significantly larger elongation are
interest to learn how the mechanical energy is stored and achieved with the newly designed hook while the local stress
dissipated by the unhooking mechanism (see the experimental maxima are about 30% decreased. The lower dissipated energy
incomplete hysteretic force–stroke records in figure 10). compared to that of the previous design reflects the fact that
Figure 16 shows simulated force–stroke curve for the case a smaller amount of the optimized hook volume transformed
where the loading is released before the detachment occurs. martensitically.
The large shaded area corresponds to the dissipated mechanical
energy within the closed force–stroke cycle. While the 4. Discussion
gray area represents the energy dissipated by the phase
transformation, the light gray area corresponds to the A superelastic NiTi hook fastener is released by overcoming
dissipation related to friction (forward branch only). The the characteristic normal detachment force given by the
former is obtained by assuming the friction angle equal to zero resistance to bending deformation of interlocked NiTi hooks.
in the simulation. The responsible deformation process, termed unhooking
The FE implemented model is basically a computational superelasticity, has been introduced above as a structural
tool which can be used to design a NiTi hook fastener, in deformation of the NiTi hooks involving superelastic material
particular to find out the proper material and geometrical deformation mostly in bending and a sliding contact between

9
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

highly correlated unhooking of individual hooks and relative

insensitivity of the number of interlocked hooks to the way in
which the contact is made. It should be pointed out here that the
strength of the NiTi hook fastener depends significantly on the
manufacturing procedure of the hook arrays. Sample E exhibits
higher strength than sample C (table 1) since individual NiTi
hooks are better aligned in the out of plane direction. Another
important parameter affecting the strength is the number of
interlocked hooks.
The FE simulation (figure 14) shows that during
unhooking the martensitic transformation proceeds massively
in the largest curvature parts of the hooks with maximum
martensite fractions at the surface. The maximum total
Figure 17. A comparison of the experimental and simulated
strains reached prior to unhooking equal 9%. These large
force–stroke curves during unhooking of two interlocked hooks with
optimized shape (inset) at 20 ◦ C. strains are still recoverable in the material model since they
are accommodated by the martensitic transformation yielding
large recoverable strains at high stresses. The maximum
two hooks controlled by friction. The experiments reported in stresses are below 1 GPa which is well below the yield stress
section 2 clearly demonstrated the special properties of NiTi limit (∼1500 MPa) for plastic deformation of these NiTi
hook fasteners such as high strength, well defined strength wires in tension (figure 2(a)). The high strength and shape
and maximum elongation (the stroke value corresponding reversibility of the NiTi hook fastener is thus directly related
to the maximum force), a significant increase of strength to the superelastic deformation.
with increasing temperature or absorption and dissipation of The second unique feature is the temperature dependence
mechanical energy. The simulation work presented in section 3 of the fastener strength (figure 8). The force–stroke curve
related these properties with the mechanics of unhooking characterizing the unhooking process of two interlocked hooks
superelasticity. Let us hence discuss some of these special (figure 15) basically consists of three stages: stage I where
properties of the superelastic hook fastening mechanism in only elastic deformation is expected is characterized by the
detail below. initial linear increase of force, stage II where the martensitic
First of all, the maximum detachment force in the normal transformation accommodates large bending deformation is
direction and the maximum peel-off force measured on sample characterized by a reduced increase of force and stage III
D were 16.1 N and 3.4 N, respectively. For the given sample where the force drops before the detachment occurs. Since the
area, the maximum detachment force in the normal direction martensitic transformation in stage II is strongly temperature
is thus about four times larger than the maximum peel-off dependent, the maximum detachment force significantly
force. This is a desired property of all Velcro-type fasteners. increases with increasing temperature. Although the force–
For comparison, the normal detachment and peel-off force stroke curve of the NiTi hook fastener is different from
densities of Klettostar® (Velcro-like fastener made of polyester that of the two interlocked hooks, the increase of the
and produced by Gottlieb Binder GmbH & Co KG [16]) are maximum detachment force with increasing temperature is
4 N cm−2 and 2.3 N cm−1 , respectively [7]. Due to the large clearly reproduced (figure 8(b)). The significant variation of
difference between the normal detachment force and the peel- strength with temperature is a unique feature of NiTi fasteners
off force (figure 6), the fastener with the flexible substrate can which cannot be found in any conventional Velcro fastener
be used in the same way as a conventional Velcro fastener. (figure 9). Recall that the mechanically reversible superelastic
Nevertheless, there are some differences stemming from the effect is observed only at temperatures above 0 ◦ C. At lower
facts that the two NiTi hook arrays are symmetric (hook–hook temperatures, NiTi hook fasteners still work well, except that
fastener) and that the contact is established by a relatively low the hooks’ original shape has to be restored by heating the
density of metallic hooks (see points (9) and (10) in the list of hooks above the austenite finish temperature Af prior to their
NiTi hook fastener properties below). next use. The strength of NiTi hook fasteners with martensite
If we, however, look more closely at the mechanics of the hooks does not depend on temperature.
unhooking process (as discussed in sections 2 and 3), novel The third feature that is worth mentioning is the large
features of NiTi hook fasteners appear. The first feature is energy absorption provided by the NiTi hook fastener. The
the much higher strength of the NiTi hook fastener compared energy absorption takes place simply due to the ability of
to that of Velcro (figure 7). Even if the estimated strength the martensitic transformation to accommodate large reversible
(20 000 kg m−2 ) is about 2–3 times higher than the normal strains at large stresses (figure 2) without damaging the hooks.
strength (7000 kg m−2 ) claimed for the recently developed Moreover, a significant part of the mechanical energy supplied
high strength steel fastener [2], the NiTi hook fastener can be during the detachment process is in fact dissipated. The
repeatedly released in the normal force test without damaging energy dissipation in various stages of the force–stroke curve
the hooks. The strength as well as the maximum elongation was evaluated by performing partial superelastic loops during
of the NiTi hook fastener are better defined compared to the unhooking of two interlocked NiTi hooks (figure 10). It is clear
Velcro fastener (figure 7). This is assumed to be due to the that the energy dissipation becomes large particularly when

10
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

overcoming a certain force (strain) corresponding to the knee Table 2. Number of interlocked hooks and maximum detachment
point on the force–stroke curve. Of course, the large energy force of NiTi hook fasteners.
dissipation would contribute to the damping of mechanical Sample A B C
vibrations of components joined by NiTi hook fasteners ◦
Mechanical work absorbed at 20 C 0.69 10.28 30.74
(mainly low frequency/large amplitude vibrations as common (mJ)
for superelastic dampers [17]). The energy dissipation also Mean number of interlocked couples of 4 63 189
assists in the experimentally observed lower acoustic noise hooks
generated by the NiTi hook fastener since nearly half of the Measured maximum detachment force 0.67 16.13 42.13
(N)
absorbed mechanical energy is dissipated in the NiTi hook
Calculated upper limit on the maximum 1.68 26.46 79.38
fastener by the martensitic transformation (figure 10) before detachment force (N)
the released NiTi hooks start to vibrate at high frequencies
and generate acoustic noise. This is why NiTi hook fasteners
with martensitic wires are quieter than those with superelastic the maximum detachment force between two single hooks
wires. Nevertheless, there might be other factors such as the at 20 ◦ C is 0.42 N (figure 11(a)), we can obtain the upper
low density of NiTi hooks or strongly correlated detachment limit of the maximum detachment force at room temperature
mechanism (one may compare the smooth versus shaky force– using the mean number of interlocked hooks for each sample
stroke curves in figures 7(a) and (b)) which can also make the (table 2). The measured maximum detachment force is,
detachment of the NiTi hook fastener quieter than Velcro. The however, significantly smaller than the calculated upper limit
problem of acoustic noise generation by the NiTi hook fastener, of the maximum detachment force. We believe that the
although it is interesting, is beyond the scope of this paper. main reason for this is that individual hook couples do not
Compared to a conventional Velcro fastener released by deform simultaneously (the force maxima corresponding to
the peel-off mechanism, superelastic NiTi hook fasteners offer individual hook couples are not reached at once) for the
a fastening mechanism characterized by a forceless contact following reasons: (i) the hooks in the samples are not perfectly
closure and a release by overcoming a well defined normal aligned; (ii) when performing the measurement, the rigid
detachment force with several beneficial supplemental features substrates are not perfectly parallel; (iii) some hooks are
discussed above. The strength of a NiTi hook fastener interlocked with two or more hooks, or other constraints are
can be precisely set for specific requirements by a careful present. The detachment of the hook array is thus not simply
selection of the material parameters of the used NiTi alloy, simultaneous detachment of individual inter-hooked couples.
the wire diameter, the hook shape, the hook density and the Further experimental as well as theoretical research on the
size of the array. Manufacturing parameters such as out- coordinated unhooking of hook arrays is currently underway.
of-substrate alignment of hooks, the number of interlocked The aim is to simulate reliably the detachment of NiTi hook
hooks and substrate rigidity also have a strong influence on fasteners. Finally, functional fatigue (variation of strength with
the strength. To facilitate the design of NiTi hook fasteners, the number of detachment cycles) is essential, particularly for
we have developed the SMA modeling based tool and briefly fasteners to be released by a defined force. Fatigue is, however,
demonstrated its use (figures 13–17). an issue even for fasteners which are opened by peeling off,
One of the unknowns in the design of NiTi hook fasteners since individual hooks are loaded equally in the normal and
is the number of interlocked hooks in the arrays in contact. peel-off detachments. Fatigue studies are also in progress now.
Since the spread among the multiple force–stroke curves As the special features of NiTi hook fasteners are expected
measured on a single NiTi hook fastener is much smaller to be exploited in engineering applications, let us summarize
compared to a Velcro fastener, we assume that the number them below for the sake of brevity.
of interlocked hooks does not depend much on the way the
contact is made. The number of interlocked hooks can be (1) High strength—compared to the conventional Velcro
evaluated indirectly as follows. During the sample detachment fastener, a NiTi hook fastener with significantly higher
a certain amount of mechanical work estimated as the area normal detachment force (strength of 30 000 kg m−2
under the detachment force–stroke curve is done. The work evaluated at 80 ◦ C on a rigid surface fastener with a
measured during the detachment of two arrays has to be equal hook density of 240 hooks cm−2 (Sample E)) can be
to the sum of the work of each interlocked hook couple. realized (figure 9). As comes out from the simulations,
Then, assuming the same work for each hook couple, we can the normal detachment force is mainly affected by the
obtain the number of interlocked hooks simply as a ratio of alloy composition ( Ms temperature), the hook diameter
the array fastener work over the single hook couple work. and shape, the hook density, the alignment of hooks in
Taking into account the experimentally measured single hook the out-of-substrate direction, the number of interlocked
couple work of 0.163 mJ and values of work of 0.69 mJ, hooks, the substrate rigidity and the temperature. These
10.28 mJ, and 30.74 mJ measured on samples A, B, and C, parameters can be used to develop optimized high strength
the mean numbers of interlocked hooks were calculated as 4, fasteners for industrial applications.
63, and 189, respectively (table 2). We see that the number (2) Impact absorption—the NiTi hook fastener can accommo-
of interlocked hooks linearly increases with increasing hook date a large normal displacement (∼1 mm) at relatively
density. Consequently, the maximum detachment force also large forces (∼20 N cm−2 ) without releasing the contact
increases with increasing hook density (figure 5(d)). Since and/or damaging the hooks (figure 9). This is a direct

11
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

consequence of the superelastic deformation of the hooks detachment force of the NiTi hook fastener with a
in bending combined with the sliding friction contact. If sufficiently high density of hooks is less sensitive to the
the hooks are made from steel, the reversible displacement way in which the two parts of the fastener are brought into
and force will either be much smaller (thin hook wire) contact. This is essential for the above proposed industrial
or the hooks will be damaged by plastic deformation due fasteners released by defined force and/or stroke.
to excessive force (thick hook wire). This feature can (8) Design of fasteners—compared to conventional Velcro
be used to develop impact tolerant fasteners for industrial fasteners, NiTi hook fasteners can be relatively easily
applications. designed for specific application requirements (the normal
(3) Increase of strength with increasing temperature— detachment force, recoverable stroke, damping capacity,
compared to the conventional Velcro fastener, the normal thermal sensitivity, etc) based on a careful selection of
detachment force of the NiTi hook fastener significantly NiTi alloy, hook geometry, hook density, the number of
increases with increasing temperature (20%/10 ◦ C—see interlocked hooks, etc. To facilitate the design of NiTi
figure 8) since the transformation stress increases with hook fasteners, we have developed a dedicated SMA
increasing temperature. The NiTi hook fastener can modeling based tool and briefly demonstrated its use
be used up to 100 ◦ C; above this temperature plastic (figures 13–17). A new hook shape optimized through the
deformation of the hooks may occur. High temperature modeling offers larger detachment force and significantly
NiTi hook fasteners for industrial applications demanding larger elongation while the local stress maxima appearing
a significant increase of contact strength with increasing in the hook during detachment are decreased by about
temperatures can be developed. 30%.
(4) Decrease of strength with decreasing temperature—taking (9) Forceless contact—two arrays of NiTi hook fasteners
advantage of the temperature sensitivity of the normal can be brought into contact almost forcelessly since they
detachment force (figure 8), simple thermally sensitive are formed by a low density of strong metallic hooks.
releasable fasteners (releasable when the temperature While the normal detachment force of a Velcro fastener
drops below a critical pre-set level) can be developed. depends significantly on the pressure applied when its
Since the Ms temperature can be selected from −100 two parts are contacted, the normal detachment force of
to +100 ◦ C, the release temperature of the fastener can NiTi hook fasteners is rather insensitive to the applied
be adjusted in a wide temperature range. The release contact pressure. Special fasteners can be developed for
of the fastener at low temperature still requires a certain applications where this is an issue.
minimum normal force to be applied. It might be (10) Functioning in dirty environments—one of the serious
necessary to expose the fastener to a temperature above problems of conventional Velcro fasteners is the loss
Af to reset the shapes of the hooks before its next use. of functionality in dirty environments (dust, oil, etc).
(5) Damping of vibrations—taking advantage of the signif- Due to the different hook–hook fastening mechanism
icant energy dissipation (figures 10 and 14), adjustable characterized by a much lower density of metallic
large recoverable displacement and high strength, special hooks, the functionality of NiTi hook fasteners does
fasteners functioning as vibration dampers can be devel- not deteriorate in dirty environments. Special NiTi
oped. The damping capacity increases with increasing hook fasteners for dusty and oily environments can be
stroke amplitude and decreasing temperature, as is typical developed.
for SMA dampers.
(6) Silent release—it is confirmed experimentally that less Other potential issues not investigated yet include reliabil-
noise is generated, particularly at higher frequencies ity (number of attachment–detachment cycles without deterio-
(figure 12), during the normal detachment of a NiTi ration of the functionality), high temperature performance, the
hook fastener when compared to the detachment of response to shear loads or performance in fluids.
a conventional Velcro fastener. The higher the Ms
temperature of the NiTi alloy, the quieter the fastener is. It 5. Conclusions
is assumed that the silent release is mainly due to energy
dissipation by the hysteretic martensitic transformation. It A NiTi Velcro-like fastener utilizing superelastic deformation
seems that the same is true for peel-off detachment, though of two interlocked NiTi hooks is introduced. The procedure
this was not quantitatively evaluated. This feature can be employed for fabrication of fastener prototypes in the form
employed when fasteners with a quiet release are required. of arrays of shape-set NiTi micro-hooks on rigid or flexible
(7) Defined detachment force and stroke—compared to the substrates is described. Basic experimental tests aimed mainly
conventional Velcro fastener (hook–loop), the normal at evaluation of the normal detachment force (strength) at
detachment force and the reversible stroke of the NiTi different temperatures were performed on various fastener
hook–hook fastener are better defined—about the same prototypes as well as on individual hooks.
maximum force and maximum elongation values are It is shown that, compared to the conventional Velcro
obtained repetitively (figure 5). This is a direct fastener, the NiTi hook fastener with a high density of
consequence of the mechanics of superelastic hook– hooks (240 cm−2 ) exhibits several times higher strength
hook fastening. Another supplementary reason is that (∼30 000 kg m−2 ), which is moreover well defined and
the number of interlocked hooks defining the normal strongly temperature dependent.

12
Smart Mater. Struct. 20 (2011) 085027 D Vokoun et al

The mechanics of the superelastic unhooking process was [3] US Patent Specification US5983467 http://www.
investigated using an FE implemented SMA model predicting freepatentsonline.com/5983467.pdf
macroscopic force–stroke curves as well as local strain and [4] Berber S, Kwon Y K and Tomanek D 2003 Phys. Rev. Lett.
91 165503
stress distributions in NiTi hooks evolving during unhooking. [5] Williams J A, Davies S G and Frazer S 2007 Tribol. Lett.
Both the experiment and the simulation confirmed the validity 26 213–22
of the concept of the superelastic NiTi hook fastener. Its [6] Pugno N M 2007 Appl. Phys. Lett. 90 121918
unique features which can be potentially utilized in engineering [7] Schuster A 2001 Dissertation zur Erlangung des Grades
Doktor-Ingenier Der Fakultat fur Maschinenbau der
applications are pointed out.
Ruhr-Universitat Bochum. Bochum
[8] US Patent Specification 7200902B2
Acknowledgments [9] Vokoun D, Majtás D, Frost M, Sedlák P and Šittner P 2009
J. Mater. Eng. Perform. 18 706–10
[10] Czech Patent Specification PV2008-568
DV and DM acknowledge support by a Marie Curie [11] http://department.fzu.cz/ofm/wwwOFM/images/stories/
International Reintegration Grant within the 6th European download/Cerinka estrategies08.pdf
Community Framework Programme. Additionally, the support [12] Vokoun D, Beleggia M, Heller L and Sittner P 2009 J. Magn.
by the Grant Agency of the Czech Republic under Magn. Mater. 321 3758–63
[13] Frost M, Sedlak P, Sippola M and Sittner P 2010 Smart. Mater.
contract nos P108/10/1296, 106/09/1573, AV0Z10100520 and Struct. 19 094010
AV0Z20760514 is acknowledged. [14] SMA roundrobin modeling http://department.fzu.cz/ofm/
roundrobin material parameters of FWM#1 wires
http://department.fzu.cz/ofm/wwwOFM/download/
References FWM No1 wire parameters.pdf
[15] Frost M and Sedlák P 2009 www.esomat.org ESOMAT 2009,
[1] Dejean A, Leroy C, Corbara B, Roux O, Cereghino R, paper No. 08002
Orivel J and Boulay R 2010 PLoS One 5 e11331 [16] http://www.binder.biz/
[2] Metaklet http://www.designboom.com/weblog/cat/16/view/ [17] Heller L, Šittner P, Pilch J and Landa M 2009 J. Mater. Eng.
8949/metallic-velcro-steel-hook-and-loop-fastener.html Perform. 18 603–11

13