Downloaded from rsif.royalsocietypublishing.

org

Life extension of self-healing polymers with rapidly growing

fatigue cracks
A.S Jones, J.D Rule, J.S Moore, N.R Sottos and S.R White
J. R. Soc. Interface 2007 4, 395-403
doi: 10.1098/rsif.2006.0199

References This article cites 41 articles

http://rsif.royalsocietypublishing.org/content/4/13/395.full.html#ref-list-1

Article cited in:

http://rsif.royalsocietypublishing.org/content/4/13/395.full.html#related-urls
Email alerting service Receive free email alerts when new articles cite this article - sign up in the box at the top
right-hand corner of the article or click here

To subscribe to J. R. Soc. Interface go to: http://rsif.royalsocietypublishing.org/subscriptions

This journal is © 2007 The Royal Society

 Downloaded from rsif.royalsocietypublishing.org

J. R. Soc. Interface (2007) 4, 395–403

doi:10.1098/rsif.2006.0199
Published online 19 December 2006

Life extension of self-healing polymers

with rapidly growing fatigue cracks
A. S. Jones1,*, J. D. Rule1,2, J. S. Moore1,2, N. R. Sottos1,3 and S. R. White1,4
1
Beckman Institute, University of Illinois, Urbana, IL 61801, USA
2
Department of Chemistry, 3Department of Materials Science and Engineering, and
4
Department of Aerospace Engineering, University of Illinois at Urbana-Champaign, Urbana,
IL 61801, USA

Self-healing polymers, based on microencapsulated dicyclopentadiene and Grubbs’ catalyst

embedded in the polymer matrix, are capable of responding to propagating fatigue cracks by
autonomic processes that lead to higher endurance limits and life extension, or even the
complete arrest of the crack growth. The amount of fatigue-life extension depends on the
relative magnitude of the mechanical kinetics of crack propagation and the chemical kinetics
of healing. As the healing kinetics are accelerated, greater fatigue life extension is achieved.
The use of wax-protected, recrystallized Grubbs’ catalyst leads to a fourfold increase in the
rate of polymerization of bulk dicyclopentadiene and extends the fatigue life of a polymer
specimen over 30 times longer than a comparable non-healing specimen. The fatigue life of
polymers under extremely fast fatigue crack growth can be extended through the
incorporation of periodic rest periods, effectively training the self-healing polymeric material
to achieve higher endurance limits.

Keywords: autonomic materials; self-healing polymers; fatigue

1. INTRODUCTION healing reaction progresses, the crack plane fills with

Fatigue and the associated slow accumulation of polymerized dicyclopentadiene and provides a crack-tip
damage and deterioration in performance plague all shielding effect (Elber 1970; Ur–Rehman & Thomason
materials, particularly those that fail in a brittle 1993; Brown et al. 2005a,b) leading to the retardation or
manner. Inspired by biological systems that continu- permanent arrest of further fatigue crack propagation.
ously adapt and remodel in response to continuous (or Brown et al. (2002) have demonstrated that this
periodic) fatigue cycles, the self-healing polymer materials system recovers up to 90% of its original
reported here responds to fatigue loading by autonomic fracture toughness after a quasi-static Mode I fracture.
processes that lead to higher endurance limits and life A pin-loaded, tapered double-cantilever-beam specimen
extension, or even complete arrest of crack growth was subjected to a constant displacement rate of
(infinite life). 5 mm sK1 until unstable crack growth occurred. After
The materials system consists of a self-healing fracture, the polymer halves were placed back together
polymer that incorporates a microencapsulated healing and allowed to rest at room temperature for 48 h.
agent (dicyclopentadiene) and bis-tricyclohexylpho- Subsequent testing of the same specimen resulted in a
sphine benzylidene ruthenium (IV) dichloride, a solid 90% recovery of the original fracture toughness. In
chemical catalyst known as Grubbs’ catalyst, in a addition, the inclusion of the self-healing materials
polymer matrix (EPON 828, Miller-Stephenson (microcapsules and catalyst) in the polymer matrix
Chemical Co.; White et al. 2001; Brown et al. 2002). significantly increased the inherent fracture toughness of
As shown in figure 1, a crack is formed in the polymer the virgin polymer specimen (Brown et al. 2004). The
and then grown under cyclic loading in tension. quasi-static fracture of a neat polymer specimen resulted
Embedded microcapsules are ruptured by this crack in a predominantly mirror-like smooth fracture surface
growth, which then release the dicyclopentadiene into typical of cleavage-like brittle fracture. On the other
the crack plane through capillary action. Poly- hand, the quasi-static fracture of the self-healing system
merization of the dicyclopentadiene is triggered by resulted in fracture surfaces that are predominantly
contact with the suspended catalyst phase. As the covered with hackle markings (a surface morphology
resulting from small-scale secondary crack formation
parallel to the fracture plane; Rabinovith et al. 2000),
*Author and address for correspondence: Indiana University-Purdue
University Indianapolis, 723 West Michigan Street, Indianapolis, which leads to increased fracture toughness.
IN 46202-5132, USA ([email protected]). Most polymeric materials suffer from poor fatigue
One contribution of 9 to a themed supplement ‘Self-healing polymers resistance and will fail at stress levels much lower than
and composites’. they can withstand under monotonic loading conditions

Received 18 October 2006

Accepted 24 November 2006 395 This journal is q 2006 The Royal Society
 Downloaded from rsif.royalsocietypublishing.org

396 Life extension of self-healing polymers A. S. Jones et al.

Figure 1. Details of the self-healing process during fatigue.

(Sauer & Richardson 1980). This behaviour leads to with 20 wt% of dicyclopentadiene-filled microcapsules
fatigue crack propagation at relatively low stress levels and 2.5 wt% of catalyst depends on the fatigue stress
and represents a critical failure mode in polymers intensity range (DK ). At low stress intensity ranges
(Sutton 1974). One method of improving the fatigue (slow mechanical kinetics of crack growth), the fatigue
performance of brittle polymeric materials is to increase crack is arrested, while at high stress intensity ranges
the inherent fracture toughness of the polymer. (fast mechanical kinetics of crack growth), the fatigue
Methods of increasing the fracture toughness of the crack is unaffected by the self-healing system unless a
polymer include incorporating a rubbery second phase rest period is incorporated during the fatigue loading.
(Bascom et al. 1981; Kinloch et al. 1983; Becu et al. Even with a rest period, only a moderate life extension
1997; Rey et al. 1999; Hayes & Seferis 2001; Nobelen was obtained.
et al. 2003), incorporating solid particles (Azimi et al. The effectiveness of the healing system after a
1995; Sautereau et al. 1995; McMurray & Amagi 1999) fracture event depends on the amount of time the
or the addition of microcapsules (Azimi et al. 1996; specimen has been allowed to heal (Brown et al. 2002;
Brown et al. 2006) to the polymer matrix. Additional i.e. the duration of polymerization of the dicyclopenta-
methods of reducing the fatigue crack propagation rate diene). During fatigue loading, the mechanical damage
include crack-tip shielding mechanisms such as crack is continuously accumulating while, simultaneously,
closure (Elber 1970, 1971; Ur–Rehman & Thomason healing takes place. Improved healing and faster
1993; Sharp et al. 1997; Shin et al. 1998; Song et al. chemical kinetics are needed for further enhancement
1998), which introduces a wedge in the crack plane of the fatigue crack resistance of polymers. One obvious
and reduces the effective stress intensity at the crack method of accelerating the healing kinetics is to use
tip, and hydrodynamic pressure crack-tip shielding faster curing healing agents. Converting endo-dicyclo-
(experiments and theory developed mainly for metals), pentadiene isomer into exo-dicyclopentadiene results in
which reduces fatigue crack growth due to the viscous significantly faster polymerization of the monomer
flow between the crack faces (Galvin & Naylor 1965; (Rule & Moore 2002). Using faster reacting monomers
Endo et al. 1972; Polk et al. 1975; Plumbridge 1977; such as 5-ethylidene-2-norbornene, or blends of
Plumbridge et al. 1985; Tzou et al. 1985a,b; Davis & different monomers (Liu et al. 2006), can result in
Ellison 1989; Yi et al. 1999; Brown et al. 2005a). extremely fast polymerization of the healing agent.
Self-healing polymers based on microencapsulated In addition to modifying the healing agent, greater
healing agents benefit from all of the mechanisms listed concentrations of catalyst dissolved into the dicyclo-
previously for improving fatigue performance. The pentadiene result in faster curing times (Kessler &
fracture toughness of the polymer is increased by the White 2002).
addition of the microcapsules (Brown et al. 2006) and Rule et al. (2005) have shown that encapsulating the
once a microcapsule is fractured, release of the healing catalyst in wax microspheres can prevent deactivation
agent into the crack plane initiates hydrodynamic of the catalyst by the amine-based epoxy curing agents
crack-tip shielding. After the liquid dicyclopentadiene used in self-healing epoxies and dramatically improve
contacts an exposed catalyst particle, polymerization the dispersion of the catalyst throughout the polymer
takes place and the viscosity of the dicyclopentadiene matrix. Less catalyst is needed for effective healing
increases until a solid wedge of polydicyclopentadiene is since it is more reactive and more uniformly dispersed.
formed in the crack plane. Brown et al. (2005b) showed Jones et al. (2006) have shown that for more rapid
that the fatigue performance of self-healing polymers healing not only does the kinetics of polymerization

J. R. Soc. Interface (2007)

 Downloaded from rsif.royalsocietypublishing.org

Life extension of self-healing polymers A. S. Jones et al. 397

need to be accelerated, but also the dissolution kinetics (a)

of the catalyst. If the polymerization kinetics of the
healing agent is rapid, while the catalyst dissolution is
slow, the catalyst may not have a chance to fully
dissolve into the healing agent. In this case, healing is
heterogeneous and isolated in locations around catalyst
particles, leaving the majority of the fracture plane
unhealed and dramatically reducing the healing effi-
ciency of the system.
Self-healing polymers based on embedded micro-
capsules of dicyclopentadiene and catalyst in a polymer
matrix have been demonstrated (Brown et al. 2005b) to 50 µm
be effective in preventing fatigue crack growth under
(b)
conditions of slow mechanical kinetics and retarding
crack growth under moderate conditions. Improve-
ments to the chemical kinetics of healing or incorporat-
ing prescribed rest periods need to be used to effectively
prevent fatigue crack growth under higher stress
intensity ranges (fast mechanical kinetics). In the
current work, the role of accelerated chemical kinetics
via catalyst tailoring, as well as periodic rest periods, on
the fatigue response of self-healing polymers is
investigated.
10 µm

2. EXPERIMENTAL PROCEDURES (c)

2.1. Materials and specimen fabrication

Dicyclopentadiene-filled, urea-formaldehyde microcap-
sules were fabricated by in situ polymerization in an oil-
in-water emulsion. Details of the microencapsulation
process can be found in Brown et al. (2003). Wax
microspheres containing Grubbs’ catalyst were fabri-
cated by rapidly stirring a 708C aqueous solution of
melted wax, catalyst and poly(ethylene-co-maleic
anhydride) with an overhead digital mixer. Cold 10 µm
water was added to the solution to solidify the wax.
The wax microspheres were then filtered, dried and Figure 2. Morphologies of Grubbs’ catalyst: (a) as-received,
sifted before use. Wax microspheres excluding catalyst (b) recrystallized by non-solvent addition and (c) recrystal-
were fabricated in an identical fashion. lized by freeze-drying.
Different morphologies of Grubbs’ catalyst were
used in this study. The as-received catalyst (Sigma- self-healing section
Aldrich) consisted of large (approx. 150 mm long by (a) (b)
40!50 mm in cross-section) parallelepiped crystals.
The other morphologies, achieved by recrystallizing
by non-solvent addition and freeze-drying the as-
received catalyst, consisted of a rod-like morphology 92
(5 mm in diameter and up to 200 mm long) and a platelet
morphology (1 mm thick and 5 mm in diameter). Details
of the recrystallization processes can be found in Jones
et al. (2006). Scanning electron microscope images of 6.4
different catalyst morphologies are shown in figure 2. 76.2
Self-healing, tapered, double-cantilever-beam
Figure 3. Geometries during the moulding process: (a)
(TDCB) specimens (Mostovoy et al. 1967) were polymer shell and (b) complete specimen (grey area indicates
fabricated using a two-step moulding process resulting the self-healing polymer material), dimensions in millimetre.
in a sample geometry shown in figure 3. First, a neat
polymer shell with a void at the core was fabricated by DETA and the self-healing constituents. The specimen
pouring degassed EPON 828 epoxy resin (DGEBA) and was then allowed to cure for 24 h at room temperature
12 parts per hundred (pph) Ancamine DETA (diethy- followed by a 24 h cure at 308C. In all cases (where that
lenetriamine) curing agent into a mould with an component is used and except as noted), the quantity of
appropriate insert and allowing it to cure for 24 h at each component was as follows: 20 pph of 180 mm
308C. The insert was removed and the core was diameter dicyclopentadiene-filled microcapsules; 5 pph
subsequently filled with a degassed mixture of epoxy, of wax microspheres; and 5 pph of wax microspheres

J. R. Soc. Interface (2007)

 Downloaded from rsif.royalsocietypublishing.org

398 Life extension of self-healing polymers A. S. Jones et al.

Kmax 3. RESULTS AND DISCUSSION

3.1. Effect of catalyst morphology and wax
protection
K (MPa m 1/2 )

Recrystallized catalyst dissolves into dicyclopentadiene

∆KI
faster than the as-received catalyst ( Jones et al. 2006)
and accelerates the overall polymerization kinetics.
Figure 5 shows the gel times for the polymerization of
dicyclopentadiene using different concentrations of
as-received and recrystallized catalyst. The recrystal-
Kmin
lized catalyst gels the dicyclopentadiene approximately
time (s) four times faster than the as-received catalyst. Unfor-
tunately, the faster dissolving morphology is also more
Figure 4. Schematic of fatigue loading parameters, frequency susceptible to deactivation resulting from the exposure
is 5 Hz, RZ0.1.
to amine-based epoxy curing agents. Therefore, wax
encapsulation of the recrystallized catalyst is needed to
containing 5 wt% Grubbs’ catalyst. This process prevent significant deactivation of the catalyst during
created TDCB specimens with self-healing components the fabrication of the polymer specimen. The improved
along the path of the crack and neat epoxy material reactivity and dispersion of the protected catalyst
everywhere else. All polymer control specimens were allows a significant reduction in the amount required to
created in an identical manner (initially creating the create effective self-healing polymers. Retardation of
epoxy shell, then filling the core with neat epoxy, etc.) fatigue crack growth can be accomplished with catalyst
to ensure that any residual stresses due to the two-step ratios as low as 0.25 pph versus previous (Brown et al.
fabrication process would be identical in all cases. 2005b) catalyst loadings of 2.5 pph.
The fatigue performance of three self-healing
materials systems and two control specimens is
2.2. Experimental processes compared in figure 6. The fatigue response of a polymer
Fatigue fracture testing was performed by pre-cracking specimen without any healing constituents (curve A)
the TDCB specimens with a razor blade and immedi- serves as a baseline control. As a second control, fatigue
ately cyclically loading the specimen under load-control data for a specimen that contains dicylcopentadiene-
conditions on an Instron DynoMight 8841 low-load filled microcapsules but no catalyst (curve B) is
frame with a 250 N load cell. The specimen was pin- included in figure 6. The improvement in fatigue life
loaded and a 5 Hz triangular waveform was applied for control sample B with respect to A is due to the
with a stress ratio (R) of 0.1 and a maximum stress toughening effect of the microcapsules and the hydro-
intensity of 0.676 MPa m1/2, as shown in figure 4. Mode dynamic crack-tip shielding after the release of the
I fatigue cracks were constrained along the centreline of dicyclopentadiene into the crack plane. The third curve
the specimen by use of side grooves moulded in the (C) is from a self-healing specimen that comprised a
specimen. The crack tip was measured optically in four wax-protected, as-received catalyst along with micro-
to five locations along the length of the specimen during capsules of dicyclopentadiene. The chemical kinetics of
the experiment. Due to the taper on the double- healing are relatively slow due to the low catalyst
cantilever-beam specimen, there is a linear relationship dissolution rate and, consequently, low concentration of
between the compliance of the specimen and the crack the catalyst in the dicyclopentadiene.
length. During the course of the experiment, the crack- Curve D in figure 6 corresponds to a self-healing
tip position was determined from the specimen com- materials system that comprised dicyclopentadiene
pliance along with optical measurements. microcapsules and recrystallized catalyst in wax micro-
The performance of a self-healing system under spheres. Here, the faster dissolving catalyst is protected
fatigue crack growth was evaluated using the fatigue from deactivation by the wax and results in faster
life extension ratio (l). This ratio was defined as the healing kinetics, and thus greater life extension.
difference between the number of cycles to failure of a Experiments on the quickly dissolving catalyst
self-healing specimen (Nhealed) versus the number of morphology formed by freeze-drying the as-received
cycles to failure of a control (without healing) specimen catalyst (curve E) show even greater fatigue life
(Ncontrol), as shown in equation (2.1), extension. In comparison with other materials systems,
Nhealed KNcontrol more significant retardation of crack growth is achieved
lZ : ð2:1Þ earlier in the test due to faster healing kinetics. The
Ncontrol non-uniformity of the crack growth curve correlates to
The gel time measurements were made by placing local variations in the availability of catalyst on the
either the recrystallized catalyst or the as-received fracture plane. A careful examination of the fatigue
catalyst (3 mg) into three 4 ml glass vials and adding fracture surfaces as shown in figure 7 reveals that in
1.0 ml of endo-dicyclopentadiene to the vial via syringe. regions of the crack plane where there is reduced
Each vial was shaken for 1 min and was periodically catalyst concentration (due to the non-uniform dis-
observed and shaken. The gel time was recorded when persion of the catalyst), the crack growth rate is
the monomer solution stopped flowing as the vial was accelerated, and in areas where there is high catalyst
rotated. concentration, the surface is covered with poly-DCPD,

J. R. Soc. Interface (2007)

 Downloaded from rsif.royalsocietypublishing.org

Life extension of self-healing polymers A. S. Jones et al. 399

30

gel time (min)

22.5

as-received 50 µm

15

7.5
recrystallized
50 µm

0 12.5 25 37.5 50
concentration (mg ml−1)
Figure 5. Gel time of DCPD with different concentrations of as-received catalyst and recrystallized Grubbs’ catalyst by
non-solvent addition.

the crack growth is retarded and, in some regions, 30 A B C D E

practically arrested.
The inclusion of wax microspheres into the polymer

crack growth (mm)

matrix influences the fatigue performance of the
polymer significantly. The wax not only acts as an 22.5
additional toughening component, but also dissolves
into the released dicyclopentadiene leading to increased
viscosity of the fluid and additional hydrodynamic 15 iv
crack-tip shielding. Figure 8 summarizes the effect of
adding wax on the fatigue life of materials systems that
contain microcapsules of dicyclopentadiene. The speci- iii
7.5
men corresponding to curve F does not include the
catalyst; hence, even though no chemical healing takes ii
place, the addition of wax alone increases the fatigue i
life. Figure 8 also contains data for a self-healing
0 40 80 120 160
specimen (curve G) that includes recrystallized catalyst
without wax microspheres. This specimen was fabri- time (h)
cated using 2.5 pph of catalyst for comparison with
Figure 6. Fatigue at KmaxZ0.676 MPa m1/2, RZ0.1, fZ5 Hz
the prior study by Brown et al. (2005b) (10 times the of TDCB polymer specimens. A, neat polymer; B, contains
amount used in all other specimens). In this case, the 20 pph microcapsules; C, contains 20 pph microcapsules and
faster polymerization kinetics of the recrystallized 5 pph of wax microspheres with 5 wt% of as-received catalyst;
catalyst results in longer fatigue life, but due to the D, contains 20 pph microcapsules and 5 pph of wax micro-
catalyst being exposed to the amine-based curing spheres with 5 wt% of recrystallized catalyst (non-solvent
agents of the epoxy during the specimen fabrication, addition); E, contains 20 pph microcapsules and 5 pph of wax
its reactivity is significantly reduced. microspheres with 5 wt% of recrystallized catalyst (freeze
Table 1 summarizes the fatigue life extension factors dried). Numerals i, ii, iii and iv on curve E demark the
(equation (2.1)) for all the specimens presented in locations where SEM images of the fracture plane are
figures 6 and 8. The life extension factor has been provided in figure 7.
calculated with respect to both a neat polymer speci-
men to demonstrate the total effect of the self-healing frequency ( f ) and the maximum applied stress
constituents and the microcapsule-only specimen to intensity factor (Kmax). Keeping R and f constant, the
demonstrate how self-healing polymers would perform mechanical kinetics is controlled through Kmax. As
compared with other non-healing toughening Kmax is increased, the crack growth rate (da/dN ) will
mechanisms. At the relatively high loading rate for increase according to the Paris power law,
these experiments, the greatest life extension was da
achieved for the specimen with the fastest poly- Z C ðKmax KKmin Þn : ð3:1Þ
dN
merization kinetics.
There is a limit, however, as Kmax approaches the quasi-
static fracture toughness of the polymer, KIC, unstable
3.2. Interplay between mechanical and chemical
fracture will occur. Data in figures 6 and 8 correspond
kinetics
to a Kmax of 0.676 MPa m1/2, which is 62% of the quasi-
The mechanical kinetics of fatigue crack growth is static fracture toughness, Kmax/KICZ0.62. This ratio is
dependent on the stress ratio (RZKmin/Kmax), the significantly higher than the range of stress intensities

J. R. Soc. Interface (2007)

 Downloaded from rsif.royalsocietypublishing.org

400 Life extension of self-healing polymers A. S. Jones et al.

(i) (ii)

poly-DCPD epoxy microcapsule

(iii) (iv)

Figure 7. SEM of the fracture plane of a self-healing polymer correlating to curve E of figure 6. In all images, the crack is travelling
from left to right. The width of each image represents 1 mm. (i) Fracture surface covered densely with poly-DCPD, (ii) fracture
surface with occasional regions lacking poly-DCPD, (iii) fracture surface with adhesively bonded epoxy from the opposite
fracture plane resulting in the retardation of crack growth rate and (iv) fracture surface showing occasional regions of poly-
DCPD and significant uncovered areas.

30 B F G Table 1. Life extension factor of self-healing polymers with

KmaxZ0.676 MPa m1/2, RZ0.1, fZ5 Hz. l
, life extension
with respect to neat polymer and l

, life extension with
crack growth (mm)

respect to polymer with microcapsules.

22.5
specimen type life (h) l
l

15 A neat 4.8 0 n.a.

B microcapsule only 20.2 3.3 0
C wax with as-received 84.8 16.8 3.2
Grubbs’ catalyst
7.5 D wax with recrystal- 124.2 25.1 5.1
lized Grubbs’
catalyst
0 E wax with freeze-dried 156.4 31.9 6.7
0 40 80 120 160 Grubbs’ catalyst
time (h) F wax microspheres 76.6 15.1 2.8
without Grubbs’
Figure 8. Fatigue at KmaxZ0.676 MPa m1/2, RZ0.1, fZ5 Hz catalyst
of TDCB polymer specimens. B, contains 20 pph microcap- G recrystallized Grubbs’ 102.6 20.6 4.1
sules; F, contains 20 pph microcapsules and 5 pph of wax catalyst without
microspheres; G, contains 20 pph microcapsules and 2.5 pph wax
of recrystallized catalyst (non-solvent addition).

considered previously by Brown et al. (2002) in which with slower polymerization kinetics would have
no healing took place at the largest loads (Kmax/KICZ provided a life extension factor l!3.
0.5) unless a rest period was provided. Above this At lower loading levels where the chemical kinetics
value, only unstable crack growth occurred followed by dominate, even under continuous fatigue the polymeric
rapid failure. system autonomically responds by increasing its
The faster chemical kinetics exploited in the endurance limit and permanently arresting crack
current experiments enable significant lifetime exten- growth. At intermediate loading levels, self-healing
sion at these higher load levels where the mechanical systems that are chemically tuned to have faster
kinetics of crack growth dominate. For lower Kmax healing kinetics provide greater life extension (see
values, the crack growth rate is reduced and the figure 6). As the healing kinetics are further acceler-
chemical kinetics dominate and extend the fatigue life ated, even greater fatigue extension can be expected. In
dramatically. Figure 9 shows data for the wax general, if the healing kinetics are sufficiently rapid,
encapsulated catalyst specimens at a K max of fatigue cracks are effectively arrested as the material is
0.50 MPa m1/2 (Kmax/KICZ0.45). At this load level, exercised. This improved resistance to fatigue crack
complete arrest of the crack is achieved with the faster growth with accelerated healing kinetics is summarized
healing system, whereas the previous healing system in figure 10.

J. R. Soc. Interface (2007)

 Downloaded from rsif.royalsocietypublishing.org

Life extension of self-healing polymers A. S. Jones et al. 401

30 30
Kmax = 0.676 MPa m1/2 no rest rest
crack growth (mm)

crack growth (mm)

periods periods

22.5 22.5

15 15

7.5 7.5

Kmax = 0.50 MPa m1/2

0 1.5 3 4.5 6
0 40 80 120 160
time (h) time (h)

Figure 9. Fatigue crack growth at KmaxZ0.50 Mpa m1/2 Figure 11. Fatigue with fast mechanical kinetics. KmaxZ
(Kmax/KICZ0.45) is completely arrested with accelerated 0.801, RZ0.1, fZ5 Hz. Without rest periods, the specimen
healing kinetics. Fatigue crack growth at Kmax Z fractures quickly, with rest periods (indicated by circles)
0.676 MPa m1/2 is included for reference. significant life extension is achieved.

identical specimens were loaded in fatigue, with one

healing
under continuous fatigue conditions and the other
dominates
incorporating seven rest periods. The rest periods,
life extension factor

8
which lasted for 10–12 h each, resulted in a fatigue life
extension (l) of 3, but the life extension of the fatiguing
polymer will increase in proportion to the frequency of
faster healing the rest periods.
5 kinetics

3 4. CONCLUSION
damage Self-healing systems based on encapsulated healing
dominates agents and solid-phase catalysts can significantly
0 extend the fatigue life of polymeric materials. Retar-
dation of fatigue crack growth is achieved by a
0 0.25 0.5 0.75 1
combination of crack-tip shielding mechanisms and
Kmax /KIC increased toughness of the polymer system. At lower
loading levels, the chemical kinetics of healing dom-
Figure 10. Faster healing kinetics result in improved fatigue
crack growth resistance. Loadings that previously caused fast
inate and the materials system autonomically repairs
crack growth (Kmax/KICZ0.45) can be arrested, while under the damage and arrests further fatigue crack growth.
severe loading conditions the crack growth is significantly At intermediate loading levels, self-healing systems
retarded. Filled circles, original data; open triangles, data that are chemically tuned to have faster healing kinetics
from Brown et al. (2005); open circles with dot, complete provide greater life extension.
arrest of fatigue crack achieved. A combination of protecting the catalyst with wax
microspheres and using quickly dissolving catalyst
If the mechanical kinetics of fatigue crack growth are morphologies leads to accelerated healing kinetics and
extremely rapid, the healing system does not have results in greater life extension. Self-healing materials
sufficient time to inhibit the crack propagation. systems under extreme mechanical loading experien-
Endurance training, by periodic exposure to stress, cing rapid fatigue crack growth can have significant life
can lead to significant life extension and is governed by extension by incorporating rest periods into the loading
the interplay between the mechanical kinetics of crack history. Periodic application of rest periods during
propagation and the chemical kinetics of healing in the loading trains the material, effectively increasing its
polymer. At such high loading levels, the mechanical endurance limit. With this self-healing polymer tech-
nology, the incremental deterioration of polymeric
kinetics dominate, and sufficient periods of rest must be
materials can be stopped, or in the case of severe
incorporated into the training cycle for the materials
fatigue loading conditions, significantly slowed allowing
system to adapt. The rest periods reduce the apparent
for up to 30 times the life of a similar, but non-healing
fatigue crack propagation rate and allow time for the
polymer.
polymerization of the dicyclopentadiene. Figure 11
demonstrates the fatigue life extension of the polymeric The authors gratefully acknowledge the support from AFOSR
system at high loading levels (KmaxZ0.801). Two (Award nos. F49620-02-1-0080 and F49620-03-1-0179).

J. R. Soc. Interface (2007)

 Downloaded from rsif.royalsocietypublishing.org

402 Life extension of self-healing polymers A. S. Jones et al.

REFERENCES Kinloch, A., Shaw, S., Tod, D. & Hunston, D. 1983

Deformation and fracture behaviour of rubber-toughened
Azimi, H. R., Pearson, R. A. & Hertzberg, R. W. 1995 Role of epoxy. 1. Microstructure and fracture studies. Polymer 24,
the crack-tip shielding mechanisms in fatigue of hybrid 1341–1354. (doi:10.1016/0032-3861(83)90070-8)
epoxy composites containing rubber and solid glass Liu, X., Lee, J., Yoon, S. & Kessler, M. 2006 Characterization
spheres. J. Appl. Polym. Sci. 58, 449–463. (doi:10.1002/ of diene monomers as healing agents for autonomic
app.1995.070580223) damage repair. J. Appl. Polym. Sci. 101, 1266–1272.
Azimi, H. R., Pearson, R. A. & Hertzberg, W. 1996 Fatigue of (doi:10.1002/app.23245)
hybrid epoxy composites: epoxies containing rubber and
McMurray, M. K. & Amagi, S. 1999 The effect of time and
hollow glass spheres. Polym. Eng. Sci. 36, 2352–2365.
temperature on flexural creep and fatigue strength of a
(doi:10.1002/pen.10633)
silica particle filled epoxy. J. Mater. Sci. 34, 5927–5936.
Bascom, W., Ting, R., Moulton, R., Riew, C. & Siebert, A.
(doi:10.1023/A:1004720031733)
1981 The fracture of an epoxy polymer containing
Mostovoy, S., Crosley, P. B. & Ripling, E. J. 1967 Use of
elastomeric modifiers. J. Mater. Sci. 16, 2657–2664.
crack-line loaded specimens for measuring plain-strain
(doi:10.1007/BF02402827)
fracture toughness. J. Mater. 2, 661–681.
Becu, L., Maazouz, A., Sautereau, H. & Gerard, J. F. 1997
Nobelen, M., Hayes, B. S. & Seferis, J. C. 2003 Influence of
Fracture behavior of epoxy polymer modified with core-
elastomer distribution on the cryogenic microcracking of
shell rubber particles. J. Appl. Polym. Sci. 65, 2419–2431.
carborn fiber/epoxy composites. J. Appl. Polym. Sci. 90,
(doi:10.1002/(SICI)1097-4628(19970919)65:12!2419::AI
2268–2275. (doi:10.1002/app.12900)
D-APP14O3.0.CO;2-W)
Plumbridge, W. J. 1977 Mechano-enviornmental effects in
Brown, E. N., Sottos, N. R. & White, S. R. 2002 Fracture
fatigue. Mater. Sci. Eng. 27, 197–208. (doi:10.1016/0025-
testing of a self-healing polymer composite. Exp. Mech. 42,
5416(77)90203-8)
372–379. (doi:10.1007/BF02412141)
Plumbridge, W. J., Ross, P. J. & Parry, J. S. C. 1985 Fatigue
Brown, E. N., Kessler, M. R., Sottos, N. R. & White, S. R.
crack growth in liquids under pressure. Mater. Sci. Eng.
2003 In situ poly(urea-formaldehyde) microencapsulation
of dicyclopentadiene. J. Microencapsulation 20, 719–730. 68, 219–232. (doi:10.1016/0025-5416(85)90411-2)
(doi:10.1080/0265204031000154160) Polk, C., Murphy, W. & Rowe, C. 1975 Determining fatigue
Brown, E. N., Sottos, N. R. & White, S. R. 2004 Microcapsule crack propagation rates in lubricating environments
induced toughening in a self-healing polymer composite. through the application of a fracture mechanics technique.
J. Mater. Sci. 39, 1703–1710. (doi:10.1023/B:JMSC. Am. Soc. Lubr. Eng. Trans. 18, 290–298.
0000016173.73733.dc) Rabinovith, A., Belizovsky, G. & Bahat, D. 2000 Origin of
Brown, E. N., White, S. R. & Sottos, N. R. 2005a Retardation mist and hackle patterns in brittle fracture. Phys. Rev. B
and repair of fatigue cracks in microcapsule toughened 61, 14 968–14 974. (doi:10.1103/PhysRevB.61.14968)
epoxy composites—part i: manual infiltration. Compos. Rey, L., Poisson, N., Maazouz, A. & Sautereau, H. 1999
Sci. Technol. 65, 2466–2473. (doi:10.1016/j.compscitech. Enhancement of crack propagation resistance in epoxy
2005.04.020) resins by introducing poly(dimethylsiloxane) particles.
Brown, E. N., White, S. R. & Sottos, N. R. 2005b Retardation J. Mater. Sci. 34, 1775–1781. (doi:10.1023/A:100455500
and repair of fatigue cracks in a microcapsule toughened 8043)
epoxy composite—part ii: in situ self-healing. Compos. Sci. Rule, J. D. & Moore, J. S. 2002 Romp reactivity of endo- and
Technol. 65, 2474–2480. (doi:10.1016/j.compscitech.2005. exo-dicyclopentadiene. Macromolecules 35, 7878–7882.
04.053) (doi:10.1021/ma0209489)
Brown, E. N., White, S. R. & Sottos, N. R. 2006 Fatigue crack Rule, J. D., Brown, E. N., Sottos, N. R., White, S. R. &
propagation in microcapsule toughened epoxy. J. Mater. Moore, J. S. 2005 Wax-protected catalyst microspheres for
Sci. 41, 6266–6273. (doi:10.1007/s10853-006-0512-y) efficient self-healing materials. Adv. Mater. 17, 205–208.
Davis, F. H. & Ellison, E. G. 1989 Hydrodynamic pressure (doi:10.1002/adma.200400607)
effects of viscous fluid flow in a fatigue crack. Fatigue Sauer, J. & Richardson, G. 1980 Fatigue of polymers. Int.
Fract. Eng. Mater. Struct. 12, 527–542. J. Fract. 16, 499–532. (doi:10.1007/BF02265215)
Elber, W. 1970 Fatigue crack closure under cyclic tension. Eng. Sautereau, H., Maazouz, A., Gerard, J. F. & Trotignon, J. P.
Fract. Mech. 2, 37–45. (doi:10.1016/0013-7944(70)90028-7) 1995 Fatigue behavior of glass bead filled epoxy. J. Mater.
Elber, W. 1971 The significance of crack closure. Damage Sci. 30, 1715–1718. (doi:10.1007/BF00351600)
Tolerance in Aircraft Struct. ASTM STP 486, 230–242. Sharp, P. K., Clayton, J. Q. & Clark, G. 1997 Retardation and
Endo, K., Okada, T., Komai, K. & Kiyota, M. 1972 Fatigue repair of fatigue cracks by adhesive infiltration. Fatigue
crack propagation of steel in oil. Bull. Jpn Soc. Mech. Eng. Fract. Eng. Mater. Struct. 20, 605–614.
15, 1316–1323. Shin, C. S., Huang, K. C. & Li, Z. 1998 Artificial
Galvin, G. D. & Naylor, H. 1965 Effect of lubricants on the retardation of fatigue crack growth by the infiltration
fatigue of steel and other metals. Inst. Mech. Eng. Proc. of cracks by foreign materials. Fatigue Fract. Eng.
179, 857–875. Mater. Struct. 21, 835–846. (doi:10.1046/j.1460-2695.
Hayes, B. S. & Seferis, J. C. 2001 Modification of thermoset- 1998.00068.x)
ting resins and composites through preformed polymer Song, P. S., Hwang, S. & Shin, C. S. 1998 Effect of artificial
particles: a review. Polym. Compos. 22, 451–467. (doi:10. closure materials on crack growth retardation. Eng.
1002/pc.10551) Fract. Mech. 60, 47–58. (doi:10.1016/S0013-7944(97)00
Jones, A. S., Rule, J. D., Moore, J. S., White, S. R. & Sottos, 180-X)
N. R. 2006 Catalyst morphology and dissolution kinetics Sutton, S. 1974 Fatigue crack propagation in an epoxy
for self-healing polymers. Chem. Mater. 18, 1312–1317. polymer. Eng. Fract. Mech. 6, 587–595. (doi:10.1016/0013-
(doi:10.1021/cm051864s) 7944(74)90015-0)
Kessler, M. R. & White, S. R. 2002 Cure kinetics of the ring- Tzou, J. L., Suresh, S. & Ritchie, R. O. 1985a Fatigue crack
opening metathesis polymerization of dicyclopentadiene. propagation in oil environments: 1. Crack growth
J. Polym. Sci. Part A Polym. Chem. 40, 2373–2383. behavior in silicone and paraffin oils. Acta Metall. 33,
(doi:10.1002/pola.10317) 105–116. (doi:10.1016/0001-6160(85)90224-X)

J. R. Soc. Interface (2007)

 Downloaded from rsif.royalsocietypublishing.org

Life extension of self-healing polymers A. S. Jones et al. 403

Tzou, J. L., Hsueh, C. H., Evans, A. G. & Ritchie, R. O. White, S. R., Sottos, N. R., Geubelle, P. H., Moore, J. S.,
1985b Fatigue crack propagation in oil environments: 2. Kessler, M. R., Sriram, S. R., Brown, E. N. &
A model for crack closure induced by viscous fluids. Viswanathan, S. 2001 Autonomic healing of polymer
Acta Metall. 33, 117–127. (doi:10.1016/0001-6160(85)90 composites. Nature 409, 794–797. (doi:10.1038/35057232)
225-1) Yi, K. S., Cox, B. N. & Dauskardt, R. H. 1999 Fatigue crack-
Ur-Rehman, A. & Thomason, P. F. 1993 The effect of growth behavior of materials in viscous fluid environment.
artificial fatigue crack closure on fatigue-crack growth. J. Mech. Phys. Solids 47, 1843–1871. (doi:10.1016/S0022-
Fatigue Fract. Eng. Mater. Struct. 16, 1081–1090. 5096(99)00002-2)

J. R. Soc. Interface (2007)