Journal of Thermoplastic

Composite Materials
http://jtc.sagepub.com/

Joining Composite Materials and Structures: Some Thought-Provoking

Possibilities
Robert W. Messler, Jr
Journal of Thermoplastic Composite Materials 2004 17: 51
DOI: 10.1177/0892705704033336

The online version of this article can be found at:

http://jtc.sagepub.com/content/17/1/51

Published by:

http://www.sagepublications.com

Additional services and information for Journal of Thermoplastic Composite Materials can be found
at:

Email Alerts: http://jtc.sagepub.com/cgi/alerts

Subscriptions: http://jtc.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://jtc.sagepub.com/content/17/1/51.refs.html

>> Version of Record - Jan 1, 2004

What is This?

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

 Joining Composite Materials and
Structures: Some Thought-provoking
Possibilities

ROBERT W. MESSLER, Jr*

Rensselaer Polytechnic Institute
Troy, NY USA

ABSTRACT: Custom-engineered functionally-specific properties leading to greatly

enhanced design versatility have drawn designers to composites, but will make these
materials-of-preference only if better ways can be found for producing better joints.
The very micro- and macrostructure that gives rise to the unparalleled properties
possible with composites exacerbates the inherent weaknesses virtually always
associated with joints. Prevailing methods of mechanical fastening and adhesive or
thermal bonding (welding) are nonoptimal, if not inappropriate, extensions of
methods used for joining conventional metallic or polymeric structures with little real
change in over two millennia. This paper challenges the old paradigm of joining
being a pragmatic process performed as a last step in manufacture and proposes that
it must increasingly become an enabling technology that is integrated with material
and structure synthesis. Four new ways to think about joining composites are
proposed, and a possibility for each approach is presented to provoke thinking
among design and process engineers.

KEY WORDS: composites, joining, mechanical fastening, adhesive bonding,

thermal bonding, welding.

INTRODUCTION

IKE ANY REVOLUTION, the revolution in materials that has been taking
L place over the past few decades, and continues to be taking place, is a
two-edged sword. On the one hand, diverse, numerous, and dramatic
advancements in material types, compositions, forms and properties, as
well as better fundamental understanding of microstructure–property

*Dr. Robert W. Messler, Jr., Fellow of the ASM International and Fellow of the American Welding Society,
is Professor of Materials Science & Engineering and Director of Materials Joining at Rensselaer Polytechnic
Institute, Troy, NY USA.

Journal of THERMOPLASTIC COMPOSITE MATERIALS, Vol. 17—January 2004 51

0892-7057/04/01 0051–25 $10.00/0 DOI: 10.1177/0892705704033336
ß 2004 Sage Publications

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

52 R. W. MESSLER, JR

relationships, the behavior and demands of materials in complex engineer-

ing structures, and ever-better computer-based modeling have stimulated
and liberated designers’ minds even if not their practice. The possibilities for
greater creativity, greater innovation, better performance, lower cost, and
greater reliability in designs employing new materials is unparalleled in the
history of humankind.
On the other hand, the use of materials that are inherently more brittle
(e.g., ceramics and intermetallic compounds), anisotropic (e.g., directionally-
reinforced composites), or possess the ability to sense and actively (rather
than passively) respond to their environment (e.g., ‘‘smart’’ materials and
composites) requires these same designers to move away from established
wisdom, phenomenological experience, and areas of practice that were
relevant to more conventional, more tolerant, and more homogeneous
engineering materials [24]. Furthermore, the implementation of new
materials inevitably demands development of new and more suitable
process technologies [5,7,18]. Indeed, the evolution of civilization with
progression from the Stone Age to the Bronze Age to the Iron Age was
paced by the ability to process these materials into useful products, not just
by their discovery. Processes and materials must evolve – or undergo a
revolution – together. If we are in or are entering an ‘‘age of composites’’,
the full blossoming of that age has been, is, and will continue to be paced by
our ability to process these interesting and promising materials into
products and structures.
Figure 1 schematically portrays the evolution of materials that has taken
place over recorded history to the present time and projected into the future

Figure 1. Schematic portrayal of the evolution of materials over history, and the growing
importance of composite materials in the future (from Ashby, [2]).

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 53

[2]. The increasing presence of composite materials is clear, and may be

understated if the right advances occur in processing including joining.

The Need for Composites in Modern Structural Design

Early in the history of humankind, it was recognized that, for some uses,
combinations of materials produced properties in the mixed material or
‘‘composite’’ that were superior to those of the component materials
themselves; an ancient example being mud bricks reinforced with straw; a
modern example being graphite fiber-reinforced thermoplastic or thermo-
setting polymers. A useful ‘‘working’’ definition of a composite is that it is a
combination of a reinforcing material within a matrix material that act
together to provide enhanced properties unobtainable in either single
material. Often the matrix provides protection from damage and gives
mechanical stability to the reinforcement under loading; at the same time
the reinforcement can enhance the strength, hardness, stiffness or other
properties of the matrix.
By their nature, synthetic (as opposed to natural) composites are
‘‘engineered’’ to provide functionally-specific properties; whether mechan-
ical, electrical, magnetic, thermal, or optical, or combinations of these. If
they are directionally aligned, reinforcements give rise to properties that are
anisotropic, and as such lead to the possibility of tailorable anisotropy.
Composites inherently exhibit material damage tolerance and can impart
structural damage tolerance; arresting propagating damage either within the
material itself or at interfaces between materials comprising different
structural elements, respectively. Finally, and most recently, ‘‘smart’’
behavior can be imparted to materials in the form of composites within
which sensors and/or actuators (e.g., piezoelectrics) are embedded.
Table 1 lists some of the functionally-specific properties offered by
composite materials.

Table 1. List of some of the more important and commonplace functionally-

specific properties possible with composite materials.
1. Higher specific strength
2. Higher specific modulus of elasticity (stiffness)
3. Enhanced damage tolerance
4. Tailorable anisotropy
5. Engineered- or tailorable coefficient of thermal expansion
6. Greater resistance to wear (by adhesion, abrasion, fretting, erosion, etc.)
7. Self lubricity (embedded lubricant)
8. Tailorable thermal conductivity
9. Tailorable electric conductivity or resistivity
10. Embedded sensing and/or actuation

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

54 R. W. MESSLER, JR

The Need for and Challenge of Joining Composites

Rarely can a product or structure be fabricated as one piece. Usually, it is

necessary to build up products or structures by joining detail parts, compo-
nents, or structural elements. The reasons for joining are multifold, and
include: (1) achieving large size, especially for fabrication or erection of stru-
ctures on site; (2) achieving geometric complexity to obtain function unobtai-
nable by casting, forging, forming or other primary processing methods; (3)
achieving structural efficiency (by adding material through joining only where
it is needed): (4) optimizing material utilization (by adding rather than remov-
ing material); and (5) creating hybrid structures composed of dissimilar
materials (using the right material in just the right location).
Joining is needed for and used with materials of every type, e.g., metals,
ceramics, glasses, plastics, and composites of all types, for one or more of
the reasons given above; each material type posing its own challenges based
on its inherent properties [17]. Metals are relatively easy to join by a number
of different methods because of their inherent strength with ductility, ability
to be melted and solidified, and ability to be plastically deformed. Ceramics,
on the other hand, being refractory, mechanically brittle, hard, and prone to
thermal shock pose problems in overcoming limitations on size and geometric
complexity with slip casting, machining, or powder processing. Composites
pose particular challenges to joining because of their inherent nature too.
The performance of a structure is critically dependent on the behavior of
any joints it contains, and, as should be clear by now, most structures
contain joints. All too often, a weight or strength or stiffness advantage
brought about by clever design or optimal use of materials is lost because
the characteristics of the associated joints were not properly understood.
Nowhere is proper joining more important than in structures comprised of
composites, where the composite is almost certainly being used to enable the
attainment of some functionally-specific property(ies) beyond that(those)
attainable from monolithic materials. But, how does one join a material that
relies on reinforcement and achieve properties anywhere close to those of
the parent composite since the integrity or continuity of the reinforcement
across the joint is difficult or impossible to retain or reestablish?
As currently practiced, the problems of producing joints in composites are
two-fold: first to deal with the matrix(ices) to obtain chemical, physical and
mechanical compatibility from joint element to joint element and, second, to
attain (extend or add) or retain (preserve) the integrity of reinforcement(s)
across the joint; both with the goal of obtaining high joint efficiency1.

1
‘‘Joint efficiency’’ is the ratio (usually expressed in percent) of the relevant property in the joint to the
relevant property in the joint element(s).

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 55

Let us look briefly at how composites have been joined in the past to see
what is lacking and what might be done in the future.

PRESENT OPTIONS FOR JOINING COMPOSITE MATERIALS

Permanent versus Nonpermanent Joints

Joints in structures and assemblies can be designed and produced to

be either permanent or nonpermanent. Permanent joints, as the name
implies, are intended to keep a structure or assembly together forever once
the joint is produced. This means, producing the joint in such a way that it
cannot accidentally disassemble. Unfortunately, in reality, this usually
means producing the joint in such a way that it will not be able to be
intentionally disassembled either. The reason is, as any experienced engineer
will tell you, if a joint can be disassembled on purpose, it can disassemble
accidentally. Nonpermanent joints are those that are not permanent;
the possible reasons being multifold. A joint might be intended to be
temporary, until a permanent joint can or needs to be made. It might be
necessary for a joint to be able to be disassembled to allow maintenance,
repair, modification, or planned removal and ultimate disposal of the
structure in an increasingly environmentally-conscious world. A joint might
also have a planned or unplanned life, with sealed envelope flaps being
an example of the former and glued joints in wood furniture being an
example of the latter. Here too, the joint is nonpermanent; the former by
design, the latter by circumstances.
One could reasonably ask: ‘‘Do we ever want anything to last forever;
to be permanent? Or, do we only want something while we need it or want
it?’’ The special challenge associated with producing permanent joints is
their need for assured permanence. Once assembled, permanently joined
structures must not come apart; accidentally or on purpose. That’s the
good news and the bad news.

Fundamental Bases of Joining

At the most fundamental level, there are only three ways to make a
joint between materials and parts composed thereof – by using and relying
only upon mechanical forces, by using and relying upon chemical forces, or
by using and relying upon physical forces. When mechanical forces are
relied upon, materials (uniquely) remain separate and distinct at the atomic
or molecular level. No chemical bonds are formed, intentionally or
incidentally. One part is held to another by the physical interlocking or

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

56 R. W. MESSLER, JR

Table 2. Fundamental bases for joining materials.

Summary of Options for Joining

Mechanical joining using strictly mechanical forces to create
physical interference and/or interlocking by
fasteners attachments
Adhesive bonding using chemical forces to cause atomic-level
bonding using chemical agents
Welding relying on natural tendency of atoms to bond
using physical forces with their origin in
electromagnetic effects using heat and/or pressure

interference of either macroscopic or microscopic features, whether

naturally present or designed or formed in. Mechanical forces are the
basis for mechanical fastening and mechanical attachment. When chemical
forces are relied upon, materials and the parts they comprise are held
together by the formation of chemical bonds, usually as the result of a
chemical reaction, while when physical forces are relied upon, the natural
tendency of atoms, ions, or even molecules to attract one another is what
creates the bond and produces the joint, with no chemical reaction being
necessary. Such chemical forces are the basis for adhesive bonding, while
such physical forces are the basis for welding2.
When only mechanical forces are used to produce a joint, the joint can
virtually always be disassembled at will without damaging the parts. Hence,
mechanical joints are never truly permanent, although they can sometimes
be considered permanent for all practical purposes. When chemical forces
are used to produce a joint, the joint can rarely be disassembled at will
without damaging the parts. Hence, such joints are generally permanent,
although there are exceptions. (Just think about opening an overnight
express mailing envelope that you forgot to enclose something in!) When
physical forces are used to produce a joint, the joint can virtually never be
disassembled without severely damaging the parts.
Table 2 summarizes the fundamental bases for joining materials.

Mechanical Fastening and Mechanical Attachment

The approach of producing joints using only mechanical forces is, or

should be, generically known as mechanical joining. Within mechanical
joining there are two major subdivisions: (1) the well-known and widely
used mechanical fastening, and (2) the less recognized but equally widely

2
There are times (processes) where it is impossible to tell the difference between chemically-based and
physically-based joining, as in solvent cementing and thermal bonding of thermoplastics [17].

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 57

used (and growing!) mechanical attachment [17]. Both rely exclusively on

the use of physical interference or interlocking at a macroscopic or
microscopic level, or both, to hold parts together. In mechanical fastening, a
supplemental device is used to cause the interference between parts, with
well recognized examples being rivets, bolts, screws, nails, and staples, and
less well recognized but well-known examples being stitches, laces, lashings,
wire-wraps, as well as buttons and zippers. In mechanical attachment, no
supplemental devices (or fasteners) are needed. Joining is accomplished by
interference or interlocking between designed-in or fabricated-in geometric
features integral to the parts being joined. Examples include dove-tails,
tongues-and-grooves, T-slots, wedges, and integral snap-fit features molded
into plastic parts, as well as elastic thermal shrink-fits, elastic or plastic
mechanical interference fits, and plastic staking, crimping, folding and
hemming all also found in plastic (especially thermoplastic) parts.
With the exception of the plastic deformation methods listed above,
all mechanical attachments and all mechanical fasteners can be removed
at will without damaging parts or materials. As a result, neither
mechanical fastening nor mechanical attachment can be considered to
be truly permanent. Thus, the two greatest advantages of mechanical
joining are that it can be accomplished without changing the composition
or structure of material comprising parts and it allows intentional
disassembly. Ironically, the greatest disadvantage is that it makes possible
accidental disassembly unless very special precautions are taken at the
design stage.
Mechanical fastening has been used for joining polymer–matrix
composites (PMCs), in particular [4,11,13,16]. Some of these joints are
intended to be permanent, and some are not. The problems with
mechanically fastening PMCs arise from the inherent nature of these
materials, and include: (1) tendency for viscoelastic deformation of the
polymer–matrix under load leading to (a) loss of the preload (or clamping
force) critical to the operation of tension-loaded fasteners (e.g., bolts) due to
joint element relaxation or (b) elongation of the fastener hole due to bearing;
(2) severe concentration of stress due to the anisotropic elastic behavior of
unidirectionally-reinforced composites; (3) loss of reinforcement integrity at
drilled holes; and (4) susceptibility to galvanic corrosive attack of metal
fasteners when coupled to highly cathodic reinforcing fibers (e.g., graphite).

Adhesive Bonding

Adhesive bonding is a process of joining materials with the aid of

a substance able to hold those materials together by surface attachment.
The materials being joined are called adherends, while the bonding

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

58 R. W. MESSLER, JR

substance (or chemical agent) is called the adhesive. The forces that enable
the surface attachment are a combination of substantial secondary (and,
occasionally, primary) chemical bonding, often the result of some chemical
reaction, and, at least for rough-surfaced or porous adherends, some
mechanical locking at the microscopic level between the adhesive and the
adherends. Thus, the adhesive bonding process is fundamentally a chemical
bonding process.
The principal advantages of adhesive bonding are minimal altera-
tion of the chemical composition and microstructure of the adherends,
suitability to joining dissimilar materials, ability to produce large load-
bearing areas, tendency to distribute loading, and capability for sealing,
insulating, damping, and resisting fatigue [14]. The greatest short-
coming of adhesive bonding is susceptibility to degradation by a number
of environmental factors, including extremes of temperature, extremes of
moisture, and attack by solvents, ultraviolet radiation, fungus and vermin.
Adhesive bonding offers the potential for producing permanent joints,
particularly when polymeric adhesives are used to bond polymeric adherends
or composites thereof. The problems associated with the permanent joining
of composites with adhesives include: (1) susceptibility to environmental
degradation by high or low temperatures, water, humidity, salt or salt
spray, solvents, microbes, radiation, vacuum, ultraviolet light, fungus, and
vermin; (2) difficult inspection, except by indirect, nondestructive means;
and (3) difficult repair of processing- or service-induced defects [17].
Table 3 summarizes the relative advantages and disadvantages of
mechanical fastening and adhesive bonding for joining PMCs.

Welding

Without question, the most permanent of all joints are made by welding,
where welding is the process in which materials of the same type or class
(e.g., metals, or ceramics, or polymers) are joined together through the
formation of primary (or, occasionally, secondary) chemical bonds under
the combined action of heat and pressure. As opposed to adhesive bonding,
welding usually does not involve or require any chemical reaction; relying
solely on the natural tendency of atoms, ions, or molecules to attract one
another and bond3. The types of bonds formed across the joint are the same
as the types of bonds found in the materials being joined. The principal
advantage of welding is the production of joints of high structural integrity
and efficiency and permanence. Shortcomings of the process are alteration

3
The evolving process of combustion synthesis-based welding often does, in fact, produce a joint (or ‘‘weld’’)
by a reaction between the reactants and the substrate [22].

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 59

Table 3. Summary of the relative advantages and disadvantages of

mechanical fastening and adhesive bonding for joining polymer–matrix
composites.
Mechanically-fastened Joints
Advantages Disadvantages
1. No special surface 1. Unavoidable stress concentration
preparation needed from holes for fasteners
2. Less critical fit-up requirements 2. Possible leaks at joints or holes
3. Disassembly is possible 3. Weight penalty due to fasteners,
without damage inserts, or doublers
4. Inspection possible 4. Possible galvanic corrosion of
by direct observation fastener by composite.’
5. Greater resistance to out-of-plane
(peel or cleavage) loading
6. Less sensitivity to
environmental degradation
Adhesively-bonded Joints:
Advantages Disadvantages
1. Little undesirable stress 1. Difficult or impossible disassembly
concentration but, rather, without damage to components
favorable load spreading 2. Susceptibility to environmental
2. Little or no weight penalty degradation
3. Smooth external surfaces 3. Requirement for special surface
for aerodynamics preparation to activate for bonding
4. Freedom from galvanic 4. Difficult to inspect except by indirect
corrosion in assemblies of nondestructive means
dissimilar materials 5. Difficult to repair processing- or
5. Facility with thin sections service-induced defects.
6. Lower sensitivity to cycling
loading (fatigue) due to lower
stress concentration and better
strain accommodation
7. Freedom from fastener
bearing problems.

of the composition and microstructure of materials (for most methods)

and, ironically, the inability to disassemble the joint without severely
damaging the parts.
Two broad classifications of specific welding methods are: fusion and
nonfusion. In fusion welding methods, intentional melting of the base
materials, and perhaps a filler material, and intermixing of the resulting
molten materials leads to bond and joint formation. Fusion methods rely on
elevated temperature to cause melting, and pressure simply to hold joint
elements in contact throughout melting and solidification. In nonfusion
methods, bond and joint formation is accomplished without relying on
melting and intermixing and resolidification. Nonfusion methods rely

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

60 R. W. MESSLER, JR

on pressure or friction to cause bonding in the solid state by macroscopic or

microscopic plastic deformation. Interdiffusion in the solid state is also a
key to sound bond and joint formation4.
Welding (commonly referred to as ‘‘thermal bonding’’ by polymer
engineers) is commonly used to successfully join thermoplastic PMCs.

WHAT ABOUT JOINING MUST CHANGE?

It is interesting that scientists and engineers must tread a fine line between
two old proverbs. The first is: ‘‘If at first you don’t succeed, try and try
again!’’ Thomas Alva Edison adhered to this proverb when he tried more
than ten thousand different materials until he found one suitable for the
filament of his incandescent light bulb. The other proverb is: ‘‘Stop barking
up the wrong tree!’’ Generations of scientists and engineers and inventors
have admired Edison because he persevered and prevailed. If he hadn’t
though, we would either not know about him, or we’d laugh at his
stubbornness. There is, after all, a fine line between perseverance and
stubbornness.
So, let me propose, for the reader’s consideration, four ways in which our
entire approach to seeking better ways to join difficult materials into
increasingly more sophisticated structures with ever greater performance
requirements must change to obtain the full potential of traditional
composites, no less even be able to join emerging or still-to-be-conceived
composites.
1. Stop Thinking Narrowly! We must look in places and directions where we
haven’t looked before to find answers to problems we haven’t found
where we’ve been looking. The rationale’ behind this almost obvious
change is: Materials have changed (and will likely continue to change) so
radically that our methods for processing – including joining – them must
change equally radically. We need to match the revolution in materials
with a comparable revolution in our approach to joining.
2. Shift Joining from a ‘‘Secondary’’ to a ‘‘Primary’’ Process! Joining has
almost exclusively been a process performed as a last step (often as an
after-thought) in the fabrication/manufacture/assembly of a device,
product, or structure. We have historically synthesized the material(s),
then processed the material(s) into shaped component(s) or structural
element(s), and then joined one component or structural element to the
other(s). Increasingly, joining will have to occur at the same time the
material and device or structure is being synthesized as an integral aspect
4
In fact, a third major category of non-fusion welding is diffusion welding, in which joining occurs strictly
as the result of extensive solid-state interdiffusion, with no real need for pressure or deformation.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 61

of primary processing5. Joining must change from a pragmatic process to

make do to an enabling technology to make possible [19].
3. Look at and Extend Successes! We have to look at what has worked
somewhere else before, and push it to new levels of scale – up or down –
and performance. We have to go with proven winners. This is the basis
for reverse engineering [23].
4. Accept High Cost for High Value! We have to be willing to pay a penalty
in labor intensity or skill level (or both) for applications where the payoff
will be truly high. This arises from the apparent law of the universe that
‘‘You get nothing for nothing!’’
So, let’s just open up our minds and let our imaginations run wild for
a moment and look at several possible ways in which joining could
change for the future to enable current composites to be joined to yield
better properties and to allow composites of the future to work at all,
no less reach their full potential.

SOME INNOVATIVE JOINING OPPORTUNITIES FOR

COMPOSITES USING NEW APPROACHES

Moving Boundary Curing (MBC) of Thermosetting Composites:

Extending Our View and Looking Elsewhere

For many applications, composites with a thermosetting polymer matrix

are superior to those with a thermoplastic matrix. As a class, thermosets
offer greater tolerance of elevated temperature and solvents and higher
strength and modulus than thermoplastics. Furthermore, such properties
suggest utility for applications where thicker sections (for load carrying)
would be required. Unfortunately, thermosetting PMCs are more difficult to
process, in general, and join, in particular, than thermoplastic types because
welding (by thermal bonding) is impossible; leaving only mechanical
fastening and adhesive bonding as viable options. And, the challenge is
especially great for joining thick sections. The reason is that in conventional
methods for curing of thermosets, the curing process is initiated at the
surface and propagates inward. Release of exothermic heat of reaction
causes the temperature to rise as the reaction front propagates, leading to
overheating at the interior that becomes worse as thickness increases.
A conceptual approach worthy of pursuit might be ‘‘moving boundary
curing’’ (MBC). In this, part joining would be accomplished simultaneously

5
Examples that this is already occurring are superplastic forming/diffusion bonding used so successfully in
military aerospace for more than two decades, and fabrication of microelectronic devices used every day.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

62 R. W. MESSLER, JR

with part layup and curing, and curing would be accomplished from the
interior to the surface in a controlled manner. Figure 2 shows schematically
how the process would work, step-by-step.
In Figure 2(a), two thick-section structures made from laminated,
unidirectionally-reinforced thermosetting matrix composite to be perma-
nently joined are shown before the process of joining begins. Ideally, these
thick-section structures have not been fully cured, i.e., they are B-staged by
having resin infiltrate into a prepreg. Furthermore, whether they are fully
cured or not, their abutting (or faying) surfaces have been treated to expose
the ends of the continuous, unidirectional reinforcing fibers (or layers in
laminate composites), and the exposed ends are caused to overlap (and,
better yet, interlink or entangle) across the joint gap. Besides these
reinforcements or reinforcing elements, ‘‘initiator elements’’ are also
selectively interspersed in the layups; their ultimate purpose being to initiate
the curing reaction (as will be described shortly).
In Figure 2(b), the joint gap with overlapping (or interlinked or
intertangled) exposed reinforcing fiber ends, is filled with a compatible
thermosetting resin, and, depending on the width of the joint gap, additional
initiator elements are added. By activating or ‘‘driving’’ the initiator

Figure 2. Schematic illustration of the ‘‘Moving Boundary Curing’’ process for joining thick-
section thermosetting composites: (a) Before the process begins with joint elements
properly prepared and positioned; (b) with the joint gap filled with a compatible
thermosetting resin and, depending on the gap width, additional initiator elements;
(c) while driving initiator elements to affect curing from the interior to the exterior or
(d) from the left to the right; and (e) after curing has been completed.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 63

elements in a controlled sequence, the curing reaction can be made to take

place in a controlled way. By driving the initiator elements from the mid-
plane outward in the thick section shown in Figure 2(c), for example, the cure
reaction can be made to occur from the inside out. By driving the initiator
elements from left-to-right in Figure 2(d), the reaction front can be made to
sweep from the left joint element to the right joint element. If the driving
signal to the initiator elements is provided by a model (e.g., finite element
model) of the joint elements and joint, the cure reaction can be intelligently
controlled. Exothermic heat of reaction can be accounted for, and the
externally supplied driving force can be reduced as the reaction front (or
boundary) moves, to assure the curing reaction occurs the same way
everywhere. Figure 2(e) shows a fully cured joint. Obviously, very complex
configurations could be joined using MBC and a suitably complex control
model.
As for the initiator elements, they would be designed to provide an
appropriate signal (energy) to trigger (or initiate) the cure reaction, and to
minimally disrupt or interfere with the composite’s properties. Possibilities
include, but are not limited to, resistive metallic, intermetallic or
carbonaceous elements to trigger the cure reaction by joule heating,
radiation absorbing metallic, intermetallic, ceramic or carbonaceous
elements to trigger the reaction with absorbed energy from an irradiating
source (e.g., microwave or induction), or optical, polymeric or glass
filaments to trigger the reaction with light (particularly in the UV range). It
is not out of the question that special chemical-activator elements could be
used with the cure-activating chemical released by some signal to the
initiator elements.
Moving Boundary Curing offers the possibility of controlling the curing
reaction in thermosetting PMCs to be the same everywhere in structures
of any thickness with any shape complexity. The process also offers the
only viable approach for joining really large structures that must be erected
on-site. With some creative thinking, the process could not only optimally
cure (or, really, cocure) the joint elements and joint, but could allow for the
incorporation of reinforcement across the joint to provide much more
continuous-like behavior.

Functionalized Self-joining of Reinforcements: Turning Joining

into a Primary Process

Within the past couple of years, it has become possible to produce

materials that self-join by ‘‘self-healing’’ or, alternatively, ‘‘self-assembling’’.
In a self-healing material, any physical discontinuities (or flaws) that result
during material/structure synthesis or that arises from degradation in service

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

64 R. W. MESSLER, JR

correct themselves by rejoining automatically [3]. There are a number of

different approaches, but one well established approach uses ‘‘chemically
functionalized nanotubes’’. Here, the structure of the nanotube is altered
when it is broken to create chemically active (or ‘‘functionalized’’) sites.
When such functionalized nanotubes encounter other similarly functiona-
lized nanotubes, they automatically chemically bond to effect ‘‘healing’’.
A similar approach using chemically functionalized nanotubes can be used
for joining between newly synthesized, unflawed materials as well.
One interesting and highly relevant approach is shown in Figure 3, where
functionalized molecules built into nanowires link up to form longer,
continuous nanowires for use in next-generation molecular computers [12].
It is not a great stretch of the imagination to see how a similar technique
could be used to link up nanoreinforcements across joints between
nanocomposites.
The other emerging approach is ‘‘self-assembling’’ micro- and, eventually,
nanostructures [6,10]. In this approach, micro- or nanoscale mechanical
elements comprising an assembly are specially designed so that they self-
assemble when they fall or are shaken into the appropriate arrangement,
and not any other arrangement. Another approach uses molecules, that
because of their structure, self-assemble to effect a joint. Figure 4 shows one
exciting recent development in which ‘‘self-assembling’’ peptide-amphiphile
molecules create a cylindrical micelle, in a process that helps dramatically
speed up the healing of broken bones [1]. It is not much of a stretch of the
imagination to see how such self-assembling micelles would allow
reinforcements in a composite to be joined by creating a self-assembling
‘‘splicing’’ sleeve or collar. This approach, as well as the approach described
just above, could have tremendous potential for accomplishing what is

Figure 3. Molecular devices built into nanowires: (a) schematically and (b) in an electron
micrograph of two 50 nm-diameter nanowires with molecules embedded in the joint between
them (Kwok and Ellenbogen [12]).

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 65

Figure 4. A schematic that shows the self-assembly of peptide-amphiphile molecules into a

cylindrical micelle, a process that helps heal broken bones by mimicking the way collagen
forms a scaffold on which new bone cells can form and grow (Kwok and Ellenbogen [12]).

described in subsection D that follows, ‘‘A Totally New Paradigm for

Joining: Accepting High Cost for High Value’’.
While concepts for self-joining might seem far-fetched, they are not. Self-
healing and self-assembly are already more than laboratory curiosities; they
are being used in real materials and in real structures. The only question is,
how long will it be before the technology is extended to joining the
reinforcements in composites?

Integral Micro-mechanical Interlocks (IMMIs): Looking at and

Extending Successes

Because most types of continuous-fiber-reinforced composite joints,

whether they are adhesively bonded or mechanically fastened, involve some
cutting of the strength-providing fibers, joint configurations require careful
planning to minimize the possibility of overly compromised joint strength or
joint failure. Such planning must consider weakness in in-plane shear,
transverse tension, interlaminar shear, and bearing strength relative to the
primary asset of a lamina, i.e., the strength and stiffness in the fiber direction
[24]. Not surprisingly, the choice for joining such composites that cannot be
welded (e.g., thermosetting matrix types) lies between mechanical methods
and adhesive bonding. To date, this has meant using the hybrid joining
process of rivet-bonding and weld-binding, as well as stitching combined
with bonding [17].
It has been suggested that improved load transfer across joints between
PMCs can be achieved using a new concept known as ‘‘Integral Micro-

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

66 R. W. MESSLER, JR

mechanical Interlocks6’’ or IMMI joints [20]. The use of macroscopic

interlocks was originally suggested by Goldsworthy and Johnson [8] and
Goldsworthy et al. [9], and was revisited by Lee and Hahn [15].
The concept of IMMIs is based upon three different traditional integral
attachment approaches [20]. These three are: (1) integral attachment using
rigid interlocks, as typified by mortise-and-tenon joints, dovetail joints, T-
slots, and flanges or bosses; (2) integral attachment using elastic interlocks,
as typified by integral snap-fit features [21]; and (3) integral attachment
using plastic deformation to cause interlocking, as typified by staking,
crimping and hemming of mating metal or thermoplastic parts. While each
of these approaches achieves a joint force from strictly mechanical means,
due to physical interference or interlocking, there is the possibility – and
design/processing option – of enhancing joining force using supplemental
adhesive in a hybrid interlocking/bonding embodiment.
While it is left to the interested reader to find details elsewhere [20], the
remainder of this section provides an overview.
Rigid IMMIs consist of small (macroscopic or microscopic) 1- or 2-
dimensional protrusions and penetrations as the active components of an
interlocking pair that resist loading in shear only; being exact analogs of
their larger macroscopic cousins (e.g., tongues-and-grooves). Examples of
rigid IMMIs are shown in Figure 5(a), with 1- and 2-dimensional versions;
the former likely preferred for biaxial in-plane loading in shear, the latter for
uniaxial loading in shear. The various dimensions can be adjusted to
optimize load-carrying ability in particular directions. For composites
in which continuous reinforcing fibers are used, it might be possible to
elastically or plastically deform these fibers around the interlocking
protrusions and penetrations to enhance load-carrying continuity either
before the polymeric matrix is applied or while this matrix is soft, and then
having the matrix stabilize the fibers against straightening once the matrix
hardens by cooling (in thermoplastics) or curing (in thermosets).
Two variations of the above concept are also possible: the first is an
analog of pinned joints in conventional fastening (as shown in Figure 5(b));
the second would employ an adhesive compatible with the composite
matrix(ices) to enhance performance, especially against modest tension (as
shown in Figure 5(c)).

6
In the IMMI concept, ‘‘Integral’’ refers to the fact that the joining function is obtained with features that are
continuous with the joint elements, as opposed to being obtained through the use of supplemental devices
(such as fasteners); ‘‘Micro’’ refers to the fact that the features are on a smaller (often much smaller) scale
than traditional integral attachment features; ‘‘Mechanical’’ refers to the fact that the forces that hold the
joint together are mechanical in the origin; and ‘‘Interlocks’’ refers to the fact that the mechanical forces
developed are fundamentally dependent on physical interference or interlocking between features on mating
joint elements.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 67

Figure 5. Examples of rigid IMMIs (a); and rigid IMMIs with implanted posts (b), or enhanced
and made permanent by supplemental adhesive bonding (c) (from Messler and Genc, [20]).

Elastic IMMIs can be of two basic types. The first is a small-scale analog
of conventional snap-fits, i.e., micro-snap-fits, while the second is a small-
scale (perhaps very small scale) analog of hook-and-loop devices (e.g.,
DuPont VelcroTM). In both types, engagement is made possible by elastic
deflection of a hook-like or finger-like feature by a (more rigid) mating catch
or hook or loop, followed by the elastic recovery of this feature and
interlocking of the latch and catch portions of the locking pair of features.
As such, they can be expected to appear in forms most suitable for providing
a retention force in tension (where, like their macroscopic cousins, the
retention force is generally an order of magnitude or more larger than the
insertion force) or resisting shear (although this is generally not advisable
for snap-fits).
Figure 6 shows examples of micro-snap-fits used to provide retention
against tension and to resist shear, and it also shows examples of micro-

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

68 R. W. MESSLER, JR

Figure 6. Examples of elastic IMMIs modeled after integral snap-fit features of various types
(a) and (b) or microhooks-and-loops (c) (from Messler and Genc [20]).

hooks-and-loops. Designs for microsnap-fits are envisioned to be more

(perhaps much more) limited than for their macroscopic cousins, strictly
based on problems associated with scaling down certain features.
Load-carrying capability is enhanced over other more conventional means
by: (1) maximizing the area of shear versus tensile loading; (2) breaking the

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 69

load-line through the joint, as opposed to having it be straight (to maximize

the energy needed to cause failure); and (3) having some extension of
reinforcements across the joint for hook-and-loop types.
For the case of microhook-and-loop IMMIs, manufacture is more
difficult to envision (and achieve) than design. However, one approach
involves having continuous reinforcements extend out of the matrix as either
single fibers or as doubled-back fibers to form loops, and creating hooks
on the ends of single fibers on the mating joint element. Such hooks could
be produced by controlled heating to cause the ends of the exposed fibers
to droop or by plastic deformation. Getting fibers to extend could be
accomplished by solvent leaching or etching back the matrix to expose fiber
ends. Once microsnap-fit or microhook-and-loop IMMIs are engaged and
locked, the joint can and should probably be back-filled (infiltrated) with
either the matrix material (e.g., thermoplastic or thermosetting resin) or a
compatible adhesive. This would permanently lock the elastic IMMIs, and
would greatly enhance retention (but, perhaps unfortunately, preclude
disassembly).
Plastic IMMIs (such as examples shown in Figure 7) consist of small
macroscopic or microscopic 1- or 2-dimensional protrusions or penetrations
as the active components of one half of a mating joint pair around which
(for protrusions) or into which (for penetrations) matrix material from the
second half of the mating joint pair is comolded. Once cooled or cured,
plastic IMMIs would perform in exactly the same way as rigid IMMIs.
Plastic IMMIs are small-scale (perhaps microscopic) analogs of interlocks
created on or from parts by plastic deformation on a macroscopic scale;
examples of which are used with metals and thermoplastics and include
staking, crimping, and hemming. Like their macroscopic cousins, such
microinterlocks would support both tension and shear, with particular
potential for supporting tension. This gives plastic IMMIs rather unique
capability compared to rigid and elastic types, i.e., they can support tension
well. This is so because, unlike rigid IMMIs which could be envisioned with
undercut penetrations to interlock with overhung protrusions, there would
be no practical problems in creating the interlock through necessary
alignment and sliding of microscopic features. Unlike elastic IMMIs, plastic
IMMIs, once created and cured, would never disengage by elastic deflection,
but, rather, would disengage only when the material comprising the
interlocking features failed in shear.
Figure 8 is a chart that compares rigid, elastic, and plastic IMMIs in 1-D
and 2-D forms, and with added adhesive for enhancement, in terms of load-
carrying capability and suitability for tensile and/or shear loading.
In summary, the concept of IMMIs is intriguing as a means of joining
PMCs, whether they are thermoplastic or thermosetting types.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

70 R. W. MESSLER, JR

Figure 7. Examples of plastic IMMIs employing dove-tail ribs/grooves (a); overhanging posts
(b); or spray-deposited spheres (c) (from Messler and Genc [20]).

Another example of ‘‘looking elsewhere and borrowing from successes’’ is

to employ 3-D weaving techniques across joints between fiber-reinforced
composites of thermoplastic or thermosetting types. The tremendous
mechanical locking effect of such weaves is well known in textiles, and
could prove useful in composites, where processing techniques having their
origin in the textile industry are not uncommon (e.g., part nesting, filament
winding, braiding, sewing, and others). Once the reinforcements are woven
within and across the joint, the joint can be back-filled with the appropriate
resin and allowed to set or caused to cure. Examples of 3-D weaves
are shown in Figure 9.

A Totally New Paradigm for Joining: Accepting High Cost for High Value

While driving along an interstate outside of Chicago a few years ago, the
author passed an area where road workers had excavated the left-most
(high-speed) lane; removing the concrete from a 20-foot long section of the
lane all the way down to the underlying ballast. The ends of the steel

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 71

Figure 8. Comparison of IMMI types: rigid, elastic, and plastic, in 1-D and 2-D forms, without
and with adhesive enhancement, showing load-carrying capability and suitability for tension
and/or shear loads (from Messler and Genc, [20]).

Figure 9. Examples of 3-D weaving used in the textile industry, but readily adaptable to the
composites industry.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

72 R. W. MESSLER, JR

reinforcing bars used in the concrete extended 6–800 out of the chopped edges
of the hole into the hole. When the author returned along the same route
later that evening, the author saw that the construction workers had spliced
new steel reinforcing bars between the exposed ends, sometimes welding the
bridging pieces to the old rebars and sometimes simply splicing them in by
wrapping them with wire (which gripped the diamond-embossed surface of
the rebar). The next morning, when the author passed the site a third time,
workers were back-filling the hole with new cement with mixed-in aggregate;
allowing it to surround and lock onto the restored rebars. Eureka! That was
it! That is how you have to join composites; i.e., by reestablishing continuity
between the reinforcements first, and back-filling with the matrix material as
the last step. This is completely different that how we join composites now;
when we join the matrix material(s) (because we know how to!) and do as
little damage to the reinforcements in the process as we can. That’s not right!
Figure 10 shows the preferred approach (used in on the highways outside of
Chicago, and the author venture to say, everywhere else).
A few weeks after this epiphany, the author happened to watch a cable-
TV documentary on limb reattachment. A team of orthopedic, vascular,
neuro- and plastic surgeons put a severed hand back onto a forearm by (1)
joining the bones (with fasteners and a bioabsorbable cement), (2) joining
the arteries and veins (by sewing and cauterizing to coagulate the protein in

Figure 10. A schematic showing how steel-reinforced concrete roadways are repaired by
re-establishing the integrity of the reinforcement first, and then back-filling with the concrete
matrix. This seems to suggest how joints should be produced in composites – regardless of
their scale – for high-demand applications.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 73

the vessel walls), (3) joining the tendons and ligaments (by sewing), and (4)
joining nerves (by sewing). Finally, when all of these critical embedded
elements of the complex ‘‘composite’’ structure were joined, the plastic
surgeon closed the wound (i.e., ‘‘back-filled the matrix’’).
So, in the future, joining of advanced composites should proceed from
structural reinforcements and elements (analogous to bones and blood
vessels) to actuators (analogous to muscles and tendons and ligaments) to
sensors (analogous to nerves); leaving back-filling of the matrix until last.
While laborious, and clearly not cost-justified for every application, this
approach is the only viable approach for joining composites for the
most sophisticated and demanding applications.

CLOSING THOUGHTS

Given the rate at which technology has been advancing, one can only
imagine what the future will bring. Surely, new and more efficient, safer and
renewable means of generating electric power (e.g., using fusion energy),
faster and more economical means for the transport of people and goods by
road and rail and sea and air, the need to explore and exploit the
undiscovered and untapped mineral resources that lie at the bottom of
the sea, and the quest to explore space as we seek both our past and our
future are a few near-certain examples. Each of these, and more, will
undoubtedly depend on further advances in materials, with composites
likely to lead the way with their tailorable, functionally-specific properties.
And, the continued emergence and application of composites will demand
new and better ways for producing joints, many of which will need to be
permanent.
The drivers for new production joining processes will primarily be new
materials, as new materials often demand new processing methods. Without
a crystal ball, it is impossible to predict what all of the drivers will be,
but some are already clear. New materials will certainly be more highly
engineered, with composites being the most obvious example. Composite
materials will have increasingly engineered compositions and micro- and
macrostructure, demanding more gentle joining methods. Ever-increasing
demands of properties for greater performance, quality for better reliability
and longer life, and ease of manufacture for greater affordability, will
demand processes that are more highly controlled or inherently more robust
to manufacturing-induced variations. And, increasing use of mixed
materials in hybrid structures to optimize properties and minimize cost
will demand greater process diversity and greater process tolerance.
Estimating how far composite joining technology could go is at once the
most difficult yet intriguing challenge of all; because unless we strive for

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

74 R. W. MESSLER, JR

something revolutionary, we are likely not to find it. Ultimately, and ideally,
composites, like conventional metals and ceramics and intermetallics, will
have to be joined so that the joint never detracts, but hopefully adds, to the
performance of the structure. The path most likely to achieve this goal is to
integrate the process of joining with the process of material and structure
synthesis; i.e., to make joining a primary rather than secondary process.
Current examples are: (1) the co-sintering (and sinter-bonding) of complex
shapes from ceramics; (2) the joining of one ply of a laminated composite
to another during synthesis of the composite; and (3) the simultaneous
superplastic forming and diffusion bonding of titanium structures, including
titanium–matrix composites. In the future, the answer may be found in
the synthesis of materials and structures atom-by-atom, where joining will
be integral with material synthesis and shape production.
Rest assured, however, wherever the evolution or revolution of materials
goes, joining will have to go too!

REFERENCES

1. ASEE PRISM (March 2002). In Briefings, Lucky Break for Bone Fracture Patients, 14.
2. Ashby, M.F. (1992). Materials Selection in Mechanical Design, Pergamon Press, Oxford,
England.
3. Chen, X., Dam, M.A., Ono, K., Mai, A., Shen, H., Nutt, S.R., Sheran, K. and Wudi, F.
(2002). A Thermally Re-mendable Cross-linked Polymeric Material; Science, 295:
1698–1702.
4. Collings, T.A. (1987). Experimentally Determined Strength of Mechanically Fastened
Joints, In: Matthews, F.L. (ed.), Joining Fibre-Reinforced Plastics, Elsevier Applied Science,
pp. 9–64, London, UK.
5. Easterling, K. (1990). Tomorrow’s Materials, 2nd edn, The Institute of Metals, London,
England.
6. Fan, L., Lee, S.S. and Wu, M.C. (1997). Self-assembled Micro-XYZ Stages for Optical
Scanning and Alignment, In: Proceedings of the IEEE LEOS Annual Meeting, November
10–13, San Francisco, CA.
7. Forester, T. (ed.) (1998). The Materials Revolution, The MIT Press, Cambridge, MA, USA.
8. Goldsworthy, W.B. and Johnson, D.W. (1994). Pultruded Joint System and Tower Structure
Made Therewith, U.S. Patent 5,285,613.
9. Goldsworthy, W.B., Johnson, W.B. and Korzeniowski, G. (1994). Bifurcated Column Joint
System for Electrical Transmission Tower, U.S. Patent 5,319,901.
10. Harsh, K.F., Bright, V.M. and Lee, Y.C. (1999). Solder Self-assembly for Three-
dimensional Microelectromechanical Systems Sensors and Actuators A, 77: 237–244.
11. Hart-Smith, L.J. (1987). Design and Empirical Analysis of Bolted or Riveted Joints,
In: Matthews, F.L. (ed.), Joining Fibre-Reinforced Plastics, pp. 227–312, Elsevier Applied
Science, London, UK.
12. Kwok, S.K. and Ellenbogen, J.C. (February 2002). Moletronics: Future Electronics;
Materials Today, 28–37.
13. Landt, R.C. (1986). Mechanical Fasteners for Advanced Composites, In: Buckley, J.D. and
Stein, B.A. (eds), Joining Technologies for the 1990’s: Welding, Brazing, Soldering,

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012

Joining Composite Materials and Structures 75

Mechanical, Explosive, Solid-State, Adhesive, pp. 54–63, Noyes Data Corporation,

Park Ridge, NJ.
14. Landrock, A.H. (1985). Adhesives Technology Handbook, Noyes Publications, Park Ridge,
NJ.
15. Lee, D.E. and Hahn, H.T. (1998). Integral Fit Joint Technologies for Composites, J. of
Thermoplastic Composite Materials, 11(3): 144–169.
16. Matthews, F.L. (1987). Theoretical Stress Analysis of Mechanically Fastened Joints, In:
Matthews, F.L. (ed), Joining Fibre-Reinforced Plastics, pp. 65–104, Elsevier Applied
Science, London, UK.
17. Messler, R.W. Jr. (1993). Joining of Advanced Materials, Butterworth-Heinemann, New
York, NY.
18. Messler, R.W. Jr. (1996). The Challenges for Joining to Keep Pace with Advancing
Materials and Designs, Materials and Design, 16(5): 261–269.
19. Messler, R.W. Jr. (2003). ‘‘Joining Comes of Age: From Pragmatic Process to Enabling
Technology; J. of the Science & Technology of Welding & Joining [in press].
20. Messler, R.W. Jr. and Genc, S. (1997). Integral Micro-mechanical Interlock (IMMI) Joint
for Composite Structures, J. of Thermoplastic Composite Materials, 11(5): 200–215.
21. Messler, R.W. Jr. Genc, S. and Gabriele, G.A. (1997). Integral Attachment Using Snap-fit
Features: a Key to Assembly Automation, Part 1 – Introduction to Integral Attachment
Using Snap-fit Features, J. of Assembly Automation, 17(2): 140–152.
22. Orling, T.T. and Messler, R.W. Jr. (1996). Self-propagating High-temperature Synthesis
(SHS) as a Process for Joining Materials, Welding Journal Res. Supple., 75(3): 93–100.
23. Otto, K. and Wood, K. (2001). Product Design: Techniques in Reverse Engineering and New
Product Development, Prentice Hall, Upper Saddle River, NJ.
24. Schwartz, M.M. (1994). Joining of Composite Matrix Materials, ASM International,
Materials Park, OH.

Downloaded from jtc.sagepub.com at BROWN UNIVERSITY on December 17, 2012