Experimental investigations of spontaneous bimaterial interfacial

fractures



Kaiwen Xia
1
, Carl-Ernst Rousseau
2
and Ares J. Rosakis
3

1. Department of Civil Engineering and Lassonde Institute
University of Toronto
Toronto, Ontario M5S 1A4, Canada
2. Department of Mechanical Engineering
University of Rhode Island
Kingston, RI 02881
3. Graduate Aeronautical Laboratories, 105-50, Caltech
California Institute of Technology
Pasadena CA 91125, USA



Abstract
Following our earlier novel spontaneous fracture designs of frictional fracture (compression and shear) and
mixed-mode fracture (tension and shear), we set up a laboratory model for mixed-mode spontaneous fracture
along bimaterial interfaces created by bonding two photoelastic polymer plates (Homalite-100 and
polycarbonate). To achieve spontaneous propagation, we applied static tensile loading and initiated dynamic
crack propagation by creating a small incipient crack within the interface and using an exploding wire
technique. The fracture propagation thus obtained is controlled by the far-field static loading, the local
material properties, and bonding conditions. Using photoelasticity diagnostics combined with high speed
photography, we are able to directly measure important dynamic fracture parameters such as crack speed,
stress intensity factor, and energy release rate. A rich spectrum of phenomena associated with the material
contrast was observed. These observations will have a profound influence on engineering practice involved
the use of materials and structures with bimaterial interfaces.


Introduction
Interfacial fracture has long been identified as one of the primary causes of failure in layered materials and
adhesive joints, and has therefore received much scrutiny [1]. The interfacial fracture spans in the length
scale of sub-microns in multi-phase nano-materials to thousands of kilometers in geological faults. Among
various scenarios, two situations can be commonly recognized under dominant tensile loading conditions [2].
First, when an interface is stronger than the weaker constituent material, cracks tend to kink into the weak
material and grow nearly parallel to the interface [3]. Thus, failure is controlled by the properties of the weak
material. The second scenario appears when the interface is the weakest link of the system, causing crack
initiation and propagation to occur along the interface. In the latter case, which will be elaborated in the
paper, the weakness pertinent to such multi-material systems arises from the inherent disadvantage present
at interfaces that can be no stronger than the weaker of the two adjacent materials. The physical nature of
such composites is further complicated by the discrepancy in elastic property that gives rise to inbuilt mixity in
stress states.


Although the resulting cracks can initiate in either the quasi-static or the dynamic regime, the latter is more
catastrophic, and is also more difficult to evaluate experimentally and theoretically. Consequently, direct
measurements of dynamic stress states of bimaterial cracks were initiated only a decade ago, utilizing optical
methods combined with high speed photography. Pioneering experimental investigation in dynamic
bimaterial fracture include the works of Tippur and Rosakis [4], Lambros and Rosakis [5], and Singh and
Shukla [6].


Modeling of dynamic interfacial fracture, however, initiated earlier, dating back to the 1967 work by Goldstein
[7]. This work predicted the possibility of attaining wave speeds in excess of the lower of the two Rayleigh
wave speeds of the material constituents. This was subsequently confirmed by Lambros and Rosakis [5],
through optical measurement. Several new measurements have since been conducted to further reinforce
these findings.


Analytical developments have also emerged to assist in the interpretation of these experimental
developments. Yang et al. [8] have formulated expressions that describe the crack tip field for steadily
growing interfacial cracks between isotropic materials. Through their work, the velocity dependence of the
universal bimaterial oscillatory index was for the first time ascertained. Further, Deng [9] performed a steady-
state asymptotic analysis for interfacial cracks that resulted in the establishment of higher order terms
associated with the bimaterial crack tip stress field. Similar results for non-steady state crack growth soon
followed. Indeed, Liu et al. [10] uncovered the requisite analytical expressions along with the associated
higher order terms. The work was also accompanied by experimental investigation.


Although the historical context of this problem has focused on material decohesion, the associated concepts
are applicable also to the modeling of earthquakes [11, 12] and to the spontaneous rupture of un-bonded, but
otherwise joined materials in relatively firm contact. The latter case, also called laboratory earthquake, is
intent on understanding earthquakes resulting from disturbances to natural faults, and was explored by Xia et
al. [13]. Recently, the idea of spontaneous fracture experiments was extended to the study of coherent
interfacial fractures between identical and dissimilar materials under far-field mixed-mode [14]. As will be
discussed later, there are several advantages to spontaneous experimental fracture set-ups over traditional
interfacial experimental designs inherited from the standard static fracture experiments, such as three-point
bent, compact tension, and wedge driving crack. The present work expands upon the preliminary
observations of interfacial fracture in bimaterial systems under mixed-mode far-field loading [14] to establish a
comprehensive and more complete understanding of dynamic interfacial failure in bimaterial systems.


Experimental Design of Spontaneous Interfacial Fracture
Two different birefringent materials, Homalite-100 and polycarbonate (9.5 mm in thickness) are used in this
study. The material properties are listed in Table 1. The wave speeds are measured using ultrasonic method
while the remaining properties are obtained from the literature [15]. The combined dimension of the two
equally-sized plates that form a full specimen is 150 mm 150 mm. Before bonding the two halves of the
specimen, a small groove, 0.15 mm in depth, is machined on the interface of one of the two half plates
through its thickness. A thin metal wire of diameter 0.1 mm is placed in the groove and the plates are bonded
using an adhesive (HARDMAN five minutes epoxy). An average interface thickness of 0.1 mm is achieved.
The static mode-I fracture toughness of the adhesive sandwiched between the two materials is 0.46 MPa-m
1/2

[14].

Table 1 Summary of optical and dynamic mechanical properties of photoelastic materials
Material Property Homalite-100 Polycarbonate
Youngs Modulus E (MPa) 3860 2480
Poissons Ratio 0.35 0.38
Stress fringe value f

(kN/m) 23.6 7.0
P Wave Speed C
P
(km/s) 2.104 1.724
S Wave Speed C
S
(km/s) 1.200 0.960
Density (kg/m
3
)
1230 1192


Figure 1 describes the configuration used in this study. The explosion box provides the electric energy to
transform the metal wire into a high pressure, high temperature plasma, which constitutes the most critical
aspect of this experiment. The details of the explosion box were discussed in a previous paper [16]. This
explosion creates a small incipient crack that propagates along the interface because the toughness of the
adhesive is lower than that of the either material. If the far-field loading is large enough, this initial small
crack becomes unstable and propagates dynamically. The interface is inclined at an angle to the horizontal
direction. Therefore, the nature of the far-field loading, provided through the uniaxial tension in the vertical
direction, is mixed-mode. Upon sending the ultra-high electric current to the exploding wire, the explosion box
simultaneously provides a low level triggering signal to the high speed camera. The camera system (Cordin
220) is able to capture the images at a framing rate of 100 million frames per second with exposure times as
low as 10 nanoseconds. In this specific study, the high speed camera is operated at a much slower speed,
around 0.2-0.4 million frames per second. A specially designed tensile fixture is used to obtain static far-field
load in the vertical direction.


Figure 1. Specimen geometry of spontaneous fracture along bimaterial interface with the triggering
mechanism.


The spontaneous fracture set-up generally has some advantages over traditional means of dynamic fracture
experiments. First, the flawless sample is loaded statically before the dynamic initiation and propagation of
the fracture. This results in increased accuracy in the foreknowledge of the stress field, which is especially
important for exaction of stress intensity factors with the full-field optical diagnostics. Second, the load is time
independent during the crack propagation. Thus, theoretically the dynamic stress intensity factor is related to
the equivalent static stress intensity factor through a universal function of crack speed [8, 17]. Third, the
sudden initiation of the small incipient crack in the center of the specimen enables treatment of the problem,
dynamically, as that of a finite crack extending in an infinite domain before the waves reflect from the
boundaries of the sample. This way, the edge effects that are unavoidable in other designs of dynamic
fracture experiments is not longer a problem here and the exact analytical solutions are available. Finally, the
initiation methodology provides an easy and reliable way to synchronize the diagnostic system (optical
method combined with high speed photography) with the fracture process [16].


Extraction of Stress Intensity Factor
A. Crack tip stress field at a bimaterial interface
Consider a crack propagating spontaneously and non-uniformly along a line defining the interface of two
homogeneous, isotropic, and linearly elastic solids, at speeds bounded by the lowest Rayleigh wave speed
pertaining to either material. The state of stress describing the ambient conditions surroundings such a two
dimensional crack has been established by Liu, et al. [10], via asymptotic higher order analysis. Relations
necessary to the immediate and complete understanding of this experimental effort are reproduced below.
Thus, individual stress components are given by:

( ) ( ) ( )
{ }
( ) ( ) ( )
{ }
( ) ( ) ( )
{ }
( ) 2 2 '' ''
11
( ) 2 '' ''
22
( ) '' 2 ''
12
1 2 ; 2 ; ,
1 ; 2 ; ,
2 ; 1 ; .
m
l s m l s m s
m
s m l s m s
m
l m l s m s
F z t G z t
F z t G z t
F z t G z t



= + +
= + +
= + +
(1)
Each prime denotes a derivative of the corresponding complex arguments. The variable is the shear
modulus of material 1 (see Fig. 2).



Figure 2. Schematic of dynamic growth of a crack along a bimaterial interface.

Designating by c
l
, and c
s
the longitudinal and shear wave speeds of material 1, respectively, and by v, the
instantaneous crack speed, the following definitions arise:
( )
( )
2
2
, 2
,
1
l s
l s
v t
t
c
= . (2)
Further, a modified coordinate system was introduced, such that z
l,s
=
1
+ i
l,s

2
, where
i
= (
i
/ , i {1, 2},
and is a small arbitrary positive number. The complex functions F(z
l
; t), and G(z
s
; t) are defined as follows:

( )
( )
( ) ( )
( )
( )
( ) ( )
( )
( )
( )
( )
2
3 2
2
3 2
2
2
2
1 2
; ;
cosh
1 2
;
cosh
1 2 1
;
1 1
1 2
;
1 1
s s
i
m l l m l
s s
i
m
l l
s s
m l
l s
s s
m
l l
l s
e
F z t z A z t
D v
e
z A z t
D v
B z t
D v
B z t z

(
+


=

(
+ +

+

`
| | +

|
+ +
\ .

| | +

+
` |

+ +
\ . ) )
, (3)
and

( )
( )
( ) ( )
( )
( )
( ) ( )
( )
( )
( )
( )
2
3 2
2
3 2
2
2
2
2 1
; ;
cosh
2 1
;
cosh
2 1 1
;
1 1
2 1
;
1 1
l s
i
m s s m s
l s
i
m
s s
l s
m s
l s
l s
m
l s
l s
e
G z t z A z t
D v
e
z A z t
D v
B z t
D v
B z t z

(
+


=

(
+ +

+

`
| | +
+
|
+ +
\ .

| | +


` |

+ +
\ . ) )
. (4)
In the above, several new variables are introduced. However, they are all universal to the field of linear
elastic fracture mechanics and their complete definition can be found in Liu, et al., [10] and in Freund [17].
A
m
(z, t) and B
m
(z, t) are entire function to be expanded into Taylor series, such that:

( )
( )
( )
( )
( )
( )
0
0
, ,
, .
n n
m m
n
n n
m m
n
A z t A t z
B z t B t z

=
=
=

(5)
In the present work, only the terms for which m = 0 are used in the expansions, for which case, the dynamic
stress intensity factor is readily derived from the complex expression:

( )
( )
( )
( )( )
0
0
1
3 2 1 2 2 2
d
K t
A t
i i
=
+ +
. (6)
Evaluation of the energy release rate follows directly from the stress intensity factor:

( )
( )
( )
( )
( )
2 2
2 2
1 2
2
1 2
1 1
4cosh
l s l s K K
G
D v D v

(


+
(
= +
` `
(
(

) )

. (7)
Finally, the mode mixity inherent to all bimaterial systems is derived from the arc tangent of the ratio of the
imaginary to the real arguments of the stress intensity factor.


B. Extraction of fracture parameters
The fundamental relation pertaining to photoelasticity is represented by the stress-optic law:

max
2
Nf
h

= , (8)
where N is an observed the fringe order number, f

is the fringe constant of the photoelastic material, and h

is the material thickness. The relation between these various parameters results in a visual representation of
the contours of constant maximum shear stress in the material. Substitution of the dynamic bimaterial stress
field equations into the above relates the isochromatic fringe patterns surrounding the advancing crack tip to
dynamic stress intensity factor, crack tip speed, and other non-singular higher order terms, through:

( ) ( ) ( )
{
( ) ( ) ( )
}
2
2
2 2 '' ''
0 0
2
'' 2 ''
0 0
1 ; 2 ;
2
2 ; 1 ; .
l l s s
l l s s
Nf
F z t G z t
h
F z t G z t

| |
(
= + +
|

\ .
(
+ + +

(9)
In the present work inference of the dynamic stress intensity factor is conducted through an overdeterministic
method, wherein up to six higher order terms are used. The unknown coefficients A
m
, and B
m
appear as non-
linear terms in Eq. (9). The solution is obtained by first designating the real and imaginary expressions of the
above equation as D, and T, respectively. Subsequently, for each digitized fringe data point, the following
relation is established:

2
2 2
0
2
k
k k k
N f
g D T
h

| |
= + =
|
\ .
. (10)
The set of k non-linear equations is solved simultaneously using a least-squares analysis, based on initial
estimates of A
m
and B
m
. Correction factors are subsequently introduced, and the process is iterated upon
until Eq. (10) is satisfied to within an error of 0.0001. Further comprehensive details of the solution scheme
can be found in the original work of Sanford and Dally [18].


Experimental Results and Discussions
A. Asymmetry of crack speeds and the isochromatic fringe patterns
We conducted experiments of spontaneous interfacial fractures between polycarbonate and Homalite-100
bonded along interfaces of different angles of inclination, and different level of uniaxial tension. Results of
these experiments reveal a common response of the various materials to the loading environment. First, the
bilaterally propagating spontaneous fracture occurs at two different speeds. Second, shortly following fracture
initiation, a constant speed level is established in both directions, with negligible variation. Finally, larger load
levels induce faster crack propagation.

-30 -20 -10 0 10 20 30 40
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
Test # 90, 27.5 degree
G
T
Crack tip (mm)
G

(
k
J
/
m
2
)
0
10
20
30
V
right
=810 m/s
V
left
=675 m/s
T
i
m
e

(

s
)
(A)

-30 -20 -10 0 10 20 30 40
0.00
0.25
0.50
0.75
1.00
1.25
1.50
Test # 82, 17.5 degree
G
T
Crack tip (mm)
G

(
k
J
/
m
2
)
0
10
20
30
40
50
60
V
right
=765 m/s
V
left
=690 m/s
T
i
m
e

(

s
)
(B)

-30 -20 -10 0 10 20 30 40
0.00
0.25
0.50
Test # 86, 27.5 degree
G
T
Crack tip (mm)
G

(
k
J
/
m
2
)
0
10
20
30
V
right
=690 m/s
V
left
=750 m/s
T
i
m
e

(

s
)
(C)
Figure 3. Isochromatic fringe pattern with the plot of energy release rate and crack tip time history.

Figure 3 shows three typical isochromatic fringe patterns displayed with accompanying plots that identify
crack-tip location and energy release rate histories featuring the different levels of uniaxial tension (7.5 MPa
(A) and (B), 5.6 MPa (C)). In Figure 3(A) and 3(C), the inclination angle is 27.5, while in Figure 3(B), the
inclination angle is 17.5. Several observations can be drawn from these figures.


First, the two crack ends propagate essentially at constant speeds with one end faster than the other.
Propagation of the faster side is in the direction of the particle motion of the more compliant constituent
(polycarbonate). For the cases corresponding to Figures 3(A) and 3(B), the lower material is more compliant,
resulting in faster propagation to the right. In Figure 3(C), the upper material is more compliant, causing the
velocity trend to reverse. This asymmetry in speed is attributed to the existence of the material contrast
across the interface and the coupling between shear and tension as a result of this contrast [12, 19]. Similar
phenomena have been predicted theoretically in relation to frictional interfacial crack under combined shear
and compression [20],

and have been observed experimentally in relation to dynamic shear ruptures that
occurs along frictional interfaces separating Homalite-100 and polycarbonate plates

[13].


It is noted that fracture energy, and equivalently energy release rate, increases linearly with the crack length
for at both crack ends. The fracture is calculated using Eq. (7), with stress intensity factors obtained by using
the inversion scheme discussed above. A similar phenomenon has been stated for spontaneous mixed-mode
interfacial fracture in identically bonded materials [14]. As is the case for these earlier results, this increase of
fracture energy with crack length is consistent with the observation of constant crack speeds and the nature of
the time independent loading. For a given speed spontaneous fracture, the energy release rate is
proportional to the square of the equivalent static stress intensity factor and thus the crack length. If the
fracture energy remains constant during the crack propagation, the crack will accelerate to its limiting speed,
where the energy release rate tends to zero and the most of the released strain energy will be converted into
kinetic energy. This observation of constant speed crack thus casts doubt on the validity of using the fracture
energy concept as a material parameter to describe crack propagation resistance. Crack propagation
resistance or precisely, the nominal crack propagation resistance might be more appropriately attributed to
loading conditions and geometry. It should also be noted that at the slower propagating end, the energy
release rate is always larger than that of the faster end for the same length crack. This manifests the
decreasing nature of the universal function dependence on the crack speed.


B. Stress intensity factors and variations of mode mixity
Figure 4 presents the crack tip history, energy release rate, stress intensity factors, and the mode mixity for
an experiment featuring 7.5 MPa far-field loading and 27.5 inclination angle. This test has performed under
identical conditions as the one shown in Figure 3(A). Note that the crack speeds of the two experiments are
very similar, which demonstrates the repeatability of this experiment.
-30 -20 -10 0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
Test # 79, 27.5 degree
G
T
Crack tip (mm)
G

(
k
J
/
m
2
)
0
5
10
15
20
25
30
35
V
right
=815 m/s
V
left
=650 m/s
T
i
m
e

(

s
)
(A)
-30 -20 -10 0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
2.5
Test # 79, 27.5 degree
K
1
K
2
Crack tip (mm)
S
I
F

(
M
P
a

m
1
/
2
)
-30
-25
-20
-15
-10
-5
0

M
o
d
e

m
i
x
i
t
y

(
O
)
(B)
Figure 4. Energy release rate, crack tip time history and the evolution of stress intensity factors for a typical
experiment.


As shown in Figure 4(B), the mode mixity of both ends are smaller than the far-field mode mixity of 27.5.
The mode mixity of the slower end is smaller than that of the faster end for the same length of propagation.
The asymmetry is a result of the mode coupling discussed earlier. The fast crack tip has more mode II
component than that of the slow crack tip. This observation is similar to that described in numerical
simulations of mixed-mode propagation of identically bonded materials [21]. This is to be expected since the
limiting speed of mode II cracks is higher than that of mode I cracks [17, 22]. The result also shows that the
debonding process is facilitated by the greater prominence of mode II stress components.


Conclusions
The experimentations conducted herein feature bimaterial systems composed of polycarbonate and Homalite-
100, bonded along an interface. Symmetric and central spontaneous initiation of a dual crack through plasma
explosion resulted in asymmetric propagation, with the faster side proceeding in the direction of the particle
motion of the more compliant constituent. This is the first experiment where two propagation cracks are
generated along the interface in the opposite direction. It is noted that the magnitude of the fracture
parameters are always correlated to crack length. Surprisingly also, energy release rate is shown to be
independent of crack speed, an observation that will be investigated in details in a future publication.


References

1. Hutchinson, J.W. and Z. Suo, Mixed-Mode Cracking in Layered Materials, in Advances in Applied
Mechanics, Vol 29. 1992. p. 63-191.
2. Kitey, R. and H.V. Tippur, Dynamic crack growth in particulate bimaterials having discrete and diffuse
interfaces: Role of microstructure. Engineering Fracture Mechanics, 2005. 72(18): p. 2721-2743.
3. Hutchinson, J.W., M.E. Mear, and J.R. Rice, On Crack Paths. Journal of Applied Mechanics-
Transactions of the Asme, 1987. 54: p. 828-832.
4. Tippur, H.V. and A.J. Rosakis, Quasi-Static and Dynamic Crack-Growth Along Bimaterial Interfaces -
a Note on Crack-Tip Field-Measurements Using Coherent Gradient Sensing. Experimental
Mechanics, 1991. 31(3): p. 243-251.
5. Lambros, J. and A.J. Rosakis, Shear Dominated Transonic Interfacial Crack-Growth in a Bimaterial
.1. Experimental-Observations. Journal of the Mechanics and Physics of Solids, 1995. 43(2): p. 169-
188.
6. Singh, R.P. and A. Shukla. Dynamic failure of bimaterial interfaces: A photoelastic investigation. in
Proc. of SEM Spring Conference. 1995. Grand Rapids, MI: SEM.
7. Goldshte.Rv, On Surface Waves in Joined Elastic Materials and Their Relation to Crack Propagation
Along Junction. Journal of Applied Mathematics and Mechanics-Ussr, 1967. 31(3): p. 496-&.
8. Yang, W., Z. Suo, and C.F. Shih, Mechanics of Dynamic Debonding. Proceedings of the Royal
Society of London Series a-Mathematical Physical and Engineering Sciences, 1991. 433(1889): p.
679-697.
9. Deng, X., General Crack-Tip Fields for Stationary and Steadily Growing Interface Cracks in
Anisotropic Bimaterials. Journal of Applied Mechanics-Transactions of the Asme, 1993. 60(1): p. 183-
189.
10. Liu, C., J. Lambros, and A.J. Rosakis, Highly Transient Elastodynamic Crack-Growth in a Bimaterial
Interface - Higher-Order Asymptotic Analysis and Optical Experiments. Journal of the Mechanics and
Physics of Solids, 1993. 41(12): p. 1887-&.
11. Andrews, D.J. and Y. Ben-Zion, Wrinkle-like slip pulse on a fault between different materials. Journal
of Geophysical Research-Solid Earth, 1997. 102(B1): p. 553-571.
12. Cochard, A. and J.R. Rice, Fault rupture between dissimilar materials: Ill-posedness, regularization,
and slip-pulse response. Journal of Geophysical Research-Solid Earth, 2000. 105(B11): p. 25891-
25907.
13. Xia, K.W., et al., Laboratory earthquakes along inhomogeneous faults: Directionality and supershear.
Science, 2005. 308(5722): p. 681-684.
14. Xia, K.W., A.J. Rosakis, and V.B. Chalivendra, Spontaneous mixed-mode fracture in bonded similar
and dissimilar materials. Experimental Mechanics, (In Press).
15. Dally, J.W. and W.F. Riley, Experimental Stress Analysis. 3rd ed. 1991, New York: McGraw-Hill, Inc.
639.
16. Xia, K.W., A.J. Rosakis, and H. Kanamori, Laboratory earthquakes: The sub-Rayleigh-to-supershear
rupture transition. Science, 2004. 303(5665): p. 1859-1861.
17. Freund, L.B., Dynamic Fracture Mechanics. 1990, Cambridge ; New York: Cambridge University
Press. 563.
18. Sanford, R.J. and J.W. Dally, General-Method for Determining Mixed-Mode Stress Intensity Factors
from Isochromatic Fringe Patterns. Engineering Fracture Mechanics, 1979. 11(4): p. 621-&.
19. Weertman, J., Unstable Slippage across a Fault That Separates Elastic Media of Different Elastic-
Constants. Journal of Geophysical Research, 1980. 85(NB3): p. 1455-1461.
20. Qian, W. and C.T. Sun, A frictional interfacial crack under combined shear and compression.
Composites Science and Technology, 1998. 58(11): p. 1753-1761.
21. Geubelle, P.H. and D.V. Kubair, Intersonic crack propagation in homogeneous media under shear-
dominated loading: numerical analysis. Journal of the Mechanics and Physics of Solids, 2001. 49(3):
p. 571-587.
22. Broberg, K.B., Cracks and Fracture. 1999, San Diego: Academic Press. 752.