Strength of Materials

https://boilersinfo.
com/
STRENGTH OF MATERIALS
No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or
by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no
expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of information
contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in
rendering legal, medical or any other professional services.
https://boilersinfo.com/
STRENGTH OF MATERIALS
GUSTAVO MENDES
AND
BRUNO LAGO
EDITORS
Nova Science Publishers, Inc.

New York
Copyright © 2009 by Nova Science Publishers, Inc.
All rights reserved. No part of this book may be reproduced, stored in a retrieval system or
transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical
photocopying, recording or otherwise without the written permission of the Publisher.
For permission to use material from this book please contact us:
Telephone 631-231-7269; Fax 631-231-8175
Web Site: http://www.novapublishers.com
NOTICE TO THE READER

The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or
implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of
information contained in this book. The Publisher shall not be liable for any special,
consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or
reliance upon, this material. Any parts of this book based on government reports are so indicated
and copyright is claimed for those parts to the extent applicable to compilations of such works.
Independent verification should be sought for any data, advice or recommendations contained in
this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage
to persons or property arising from any methods, products, instructions, ideas or otherwise
contained in this publication.
This publication is designed to provide accurate and authoritative information with regard to the
subject matter covered herein. It is sold with the clear understanding that the Publisher is not
engaged in rendering legal or any other professional services. If legal or any other expert
assistance is required, the services of a competent person should be sought. FROM A
DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE
AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS.
LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA
Strength of materials / edited by Gustavo Mendes and Bruno Lago.

p. cm.
Includes bibliographical references.
ISBN 978-1-61728-584-4 (E-Book)
1. Strength of materials. 2. Composite materials. I. Mendes, Gustavo. II. Lago, Bruno.
TA405.S765 2009
620.1'12--dc22
2009013535
Published by Nova Science Publishers, Inc. New York
CONTENTS
Preface vii
Chapter 1 High Temperature Mechanical Properties and Microstructure of 1
Sic-Based Fibers under Severe Environments
Jianjun Sha
Chapter 2 Ionomers as Candidates for Structural Materials 61

Daniel J. Klein
Chapter 3 Failure of Layered Composites Subject to Impacts: Constitutive 97
Modeling and Parameter Identification Issues
Stefano Mariani
Chapter 4 Current State of the Art of the Ceramic Composite Material 133
BIOLOX®delta
Meinhard Kuntz, Bernard Masson and Thomas Pandorf
Chapter 5 Particle Modeling and Its Current Success in the Simulations of 157
Dynamics Fragmentation of Solids
G. Wang, A. Al-Ostaz, A.H.D. Cheng and P. Radziszewski
Chapter 6 Non-Oriented Electrical Steels: Materials for Saving Energy and 183
Conserving the Environment
Taisei Nakayama
Chapter 7 Influence of Luting Cement Application Technique on Quartz 217
Fiber Post Regional Bond Strengths
Camillo D’Arcangelo, Francesco De Angelis,
Maurizio D’Amario, Simone Zazzeroni,
Mirco Vadini and Sergio Caputi
Chapter 8 Microstructural Influence on Flexure Strength of a Ceromer 233
Reinforced by Two Types of Fibers (Polyethylene and Glass)
Silvana Marques Miranda Spyrides and Fernando Luiz Bastian
vi Contents
Chapter 9 Influence on Strength Properties of Anisotropy Planes in Slates 247

Samples in the NW of Spain
M.A. Rodríguez-Sastre, M. Gutiérrez-Claverol,
M. Torres-Alonso and L. Calleja
Index 261
PREFACE
The strength of a material refers to the material's ability to withstand an applied stress
without failure. The applied stress may be tensile, compressive, or shear. A material's strength
is dependent on its microstructure. The engineering processes to which a material is subjected
can alter this microstructure. This book provides a variety of material strength research
including an extensive overview on the state of the art ceramic composite material BIOLOX
delta which, since 2001, has successfully implanted more than 500,000 artificial hip joints.
Due to the unique strength and toughness of this material, the risk of fracture has been
substantially reduced when compared to conventional ceramic materials. Several different
aspects of ionomer research from a physical property standpoint is discussed as well,
including the history and current trends in ionomer research and a discussion on the
immediate needs in this field. Furthermore, particle modeling (PM) as an innovative
particulate dynamics based modeling approach is examined as a robust tool for simulating
fracture problems of solids under extreme loading conditions, including situations of collapse,
impact, blasting or high strain rate tension/compression. This book includes research on the
ability of particle modeling to correctly predict dynamic fragmentation of materials with good
accuracy.
Ceramic-matrix composites (CMCs) have been considering as potential structural
materials for advanced energy-generation systems and propulsion systems. SiC fibers with
low oxygen content and high crystallinity, which derived from polycarbonsilane, are the
backbone as reinforcements in load-bearing CMCs. For high temperature application, the
most desired critical properties of SiC fibers are high strength and stiffness as well as the
reliable retention of these properties throughout the service life of application. Low fiber
strength and thermal stability could result in low fracture toughness and accelerate sub-critical
crack propagation in CMCs. Thus, the mechanical durability and microstructure stability of
SiC fibers are major concerns under severe environments. Furthermore, in practical service
environments, rarely is one degradation mechanism operative, but several mechanisms
operate simultaneously, leading to the environment-pertinent degradation mechanism is
complex for the SiC materials. In order to enhance the understanding of reliability and
durability of CMCs applied to high temperature and oxidative environments, the
investigations on the high temperature mechanical properties and microstructure for SiC-
based fibers subjected to severe environments were integrated into Chapter 1.
The field of ionomers is an often overlooked and under-utilized branch of polymer
research. Although ionomers can be broadly described as a class of polymers that contain any
viii Gustavo Mendes and Bruno Lago
number of ionic groups, from a structural property standpoint only a low percent of ionic
groups are necessary to impart significantly improved properties over the nonionic version of
the same polymer. Current trends in the field of ionomers are highly focused on the field of
fuel cell technology. There appears to be a significant hole remaining in the study of
imparting strength to materials using ionic groups. This hole is very significant from an
industrial point of view, and has a large commercial potential. There are very few
commercially available ionomers, which shows how little this field has been explored to date.
Chapter 2 will focus on several aspects of ionomer research from a physical property
standpoint: 1) A history of ionomer research, 2) Current trends in ionomer research - a) stand-
alone polymers, b) nanocomposites, c) blends; 3) A commentary on the immediate needs in the
field of ionomer research.
Layered composites subject to impacts can fail by delamination, i.e. by debonding
between laminae, if the stress waves cause damaging phenomena to take place mainly within
the resin-enriched interlaminar phases. To simulate delamination at the structural level,
processes dissipating energy are lumped onto fictitious zero-thickness interlaminar surfaces,
and softening interface constitutive laws are adopted to describe the progressive failure of the
interlaminar phases.
Since delamination occurs inside very narrow regions, results of experimental testing on
whole composites need to be accurately and reliably filtered to calibrate the interface
constitutive laws. To this aim, Chapter 3 proposes a sigma-point Kalman filter approach. The
performances of the proposed methodology, in terms of constitutive parameter estimations
and dynamic delamination tracking, are assessed through pseudoexperimental testings on a
two-layer composite, and real testings on multi-layer glass fiber reinforced plastic
composites.
An extensive overview about the state of the art of the ceramic composite material
BIOLOX®delta is given. The unique properties rely on a well defined alumina based fine
composite microstructure which is mainly achieved by high temperature solid body reaction
of the different ceramic phases during sintering. Zirconia comprises 17 % of the total volume.
The tetragonal phase of zirconia is stabilized chemically and mechanically.
The high strength and toughness of the material depend on transformation toughening of
the zirconia which is clearly shown by various experimental results. The excellent mechanical
properties are reproduced batch by batch with a very low scatter.
As presented in Chapter 4, the outstanding properties of the material BIOLOX®delta
support advantageous properties of the final product, e.g. ceramic hard-hard bearings for hip
arthroplasty. The burst load of the components is significantly increased. It is shown that the
design of the components is also very important for the reliability and the ultimate properties
of the system. Wear properties at severe conditions are significantly improved by using the
new composite material BIOLOX®delta in comparison to pure alumina.
Phase transformation of zirconia from the tetragonal to the monoclinic phase due to
hydrothermal aging is extensively discussed. Due to the particular distribution and
stabilization of the zirconia particles instable aging effects are not possible in this material.
After very long time of accelerated aging conditions an increase of monoclinic phase is found
– however, it is shown that dynamic and static properties of BIOLOX®delta are not
influenced by this effect.
Preface ix
Chapter 5 studies particle modeling (PM), which is an innovative particulate dynamics

based modeling approach. It has been demonstrated as a robust tool for simulating fracture
problems of solids with dynamic fragmentation under extreme loading conditions. These
loading conditions can include situations of collapse, impact, blasting or high strain rate
tension/compression, as well as thermally-induced breakage problems.
Initially, PM was developed for the purpose of mimicking the microscopic material
process at macroscopic level. This method can be conceptually illustrated by fully dynamic
particles (or “quasi-particles”) placed at the nodes of a lattice network without explicitly
considering their geometric size. The potential can be specified for particle-particle
interactions via axial springs. Theoretically, PM is an upscale of the molecular dynamics
(MD) model applicable to various length scale problems. This is possible if a proper
equivalent macroscopic potential is found, and, in case of lattice spacing decreasing to a few
Angstroms, a MD model at zero Kelvin with, say, Leonard-Jones potential is recovered. In its
current form, PM has been developed as a tool applicable to real engineering problems.
The advantages of PM over the existing discrete element based methods can be
summarized as follows: (1) Sample in theory. Four conservative/equivalent rules (mass,
potential energy, Young’s modulus and tensile/compression strength) are applied to preserve
the equivalent material properties. (2) Easy for implementation. Since the physical size of
each particle is ignored other than its equivalent mass, the algorithm of coding a PM
computation is fairly easy.
Current research work has exhibited that PM is able to correctly predict dynamic
fragmentation of materials with a good accuracy. In modeling an epoxy plate with randomly
distributed holes in tension, the PM result of the final crack pattern compared favorably with
the associate experiment; for the simulations of impact study of two polymeric materials
(nylon, 6-6 and vinyl ester) subject to a rigid falling indenter, the modeling results of resistant
force, energy, deflection and drop speed of indenter vs. time quantitatively agree fairly well
with the according empirical observations.
Electrical steels are the core materials for electrical motors or transformers. Those
materials for motors are played an energy conversion roll from electricity to motion.
However, energy losses are accompanied with this conversion. To minimize these losses is a
key technology to conserve our environment.
Numerous researches on the grain-oriented electrical steels reported. Those researches
especially for transformers are focused on the reducing the losses at supplying the electricity
from power plants. On the other hand, home or industrial appliances are the power consuming
devices, and the most effective point on the energy loss reduction. These home appliances are
used small motors using non-oriented electrical steels.
In Chapter 6, several researches on the non-oriented electrical steels are discussed and
focused on the metallurgical control of the steels to reduce the core loss for generating waste
heats and motor building innovation technologies for decreasing the building factor of the
core losses.
In the metallurgical part, some additive elements as phosphorus, aluminum and
manganese for improving magnetic properties reviewed. Moreover some contaminating
elements as vanadium, titanium and zirconium are discussed especially for precipitation
studies in the steels have been done. These precipitations are inhibited the grain growth at
final annealing or stress relief annealing. These inhibited small grains increase the core losses.
x Gustavo Mendes and Bruno Lago
For studying motor building technologies, compression stress effect, shearing stress
effect are discussed. Even though the best core materials are used for manufacturing motors,
those building deteriorations make worse for the motor efficiency. Therefore, those
technologies are also important for reducing the carbon dioxide emission.
The aim of Chapter 7 was to investigate regional root canal push-out bond strengths for a
fiber-reinforced post system varying the application method of the luting agent.
Recently extracted maxillary incisors (n=30) were sectioned transversally at the labial
cemento-enamel junction, and the roots treated endodontically. Following post space
preparations, fiber-reinforced posts (Endo Light-Post; RTD) were placed using adhesive
system and resin cement provided by the manufacturer. Three equal groups (n=10) were
assessed according to the technique used to place the luting agent into post space: using a
lentulo spiral, applying the cement onto the post surface, injecting the material with a specific
syringe. Each root was sliced into three discs (2 mm thick) representing the coronal, middle
and apical part of the bonded fiber post. Push-out tests were performed for each specimen to
measure regional bond strengths. Results were statistically analyzed using two-way ANOVA
and Tukey tests (α = 0.05). All fractured specimens were observed using a scanning electron
microscope to identify the types of failure.
The results indicated that bond strength values were significantly affected by the
application method of the resin cement (p < 0.05). The "syringe technique" and the "lentulo
technique" showed higher bond strength values compared with the "post technique". No
significant differences were recorded among the post space thirds. Microscopic analysis
revealed a prevalence of post/cement and mixed failures.
The best performance in terms of push-out bond strengths for the post system tested was
obtained when the luting agent was applied into the post space either with a specific syringe
or using a lentulo spiral. There were not differences in bond strength among root thirds.
In Chapter 8, the microstructures of a ceromer (Artglass®) reinforced by either glass
fibers (GlasSpan®) or polyethylene fibers (Connect®) were characterized and compared and
the influence of the fiber reinforcement on the flexural strength of the resulting products
evaluated. With this objective, seven bars of each material were produced. One bar of each
material was separated for microstructural analysis. The microstructural samples were
subjected to metallographic polishing and finishing, and then analyzed using optical
microscopy at different magnifications. The images obtained were treated using an image
processing computer program (Image Pro Plus) in order to quantify the microstructure by
calculating the mean diameter and mean volume fraction of fibers. The flexure tests were
made by three-point bending, using six samples of each material. After statistical analysis, the
results showed that the mean diameter of the glass fibers (4μm) was smaller than the
polyethylene ones (23.6 μm). The mean volume fraction of glass fibers (0.42) was larger than
that of the polyethylene fibers (0.28) and the mean center-to-center distance between fibers
was smaller in the glass fibers material (33 μm) than in the polyethylene fibers material (61
μm). The flexural strength of both glass and polyethylene fiber-reinforced materials was
statistically equal, despite the fiber volume fraction being statistically larger in the fiber glass
material.
The purpose of Chapter 9 is to describe the influence of anisotropy on the geomechanical
strength properties of two Spanish slates with different chemical and physical characteristics.
From laboratory testing results of slates under point load and uniaxial compression and the
use of indirect methods, as it is the measurement of P velocities, principal parameters were
Preface xi
calculated for this rock material. As it is well known under uniaxial compressive strength
slates are strong and also very strong rock when loading is parallel (90o) o perpendicular (0o)
to the main anisotropic planes. In contrast it is a weak rock with minimum strength values for
angles between 45 to 60o of inclination of anisotropy planes. The correlation equations were
calculated between different parameters. Despite weak correlation between different
geotechnical properties were found and when all lithologies are considered together
correlation of geomechanical properties is weak. However when each lithology is considered
separately the geomechanical properties can be coherently defined. Linear and polynomial
equations were found for the point load and uniaxial strength correlations with the inclination
of the anisotropy. Different strength fields were calculated when uniaxial strength and point
load test plot and its comparison include the inclination of the anisotropy planes on slates.
Uniaxial compressive strength and P wave velocity appears to be strongly influenced by
uniaxial strength and good polynomial correlations resulted. Plots of slates with other
sedimentary type of rocks from Cantabrian Zone, CZ, revealed the hardness and highest
strength of slates when loading is perpendicular to the main anisotropy planes.
In: Strength of Materials ISBN: 978-1-60741-500-8
Editors: G. Mendes and B. Lago, pp. 1-60 © 2009 Nova Science Publishers, Inc.
Chapter 1
HIGH TEMPERATURE MECHANICAL PROPERTIES

AND MICROSTRUCTURE OF SIC-BASED FIBERS
UNDER SEVERE ENVIRONMENTS
Jianjun Sha1,2*
1
Shool of Aeronautics and Astronautics, Dalian University of Technology
116024, Dalian, China
2
Ceramic Materials Engineering, University of Bayreuth,
D-95440 Bayreuth, Germany
Abstract
Ceramic-matrix composites (CMCs) have been considering as potential structural materials
for advanced energy-generation systems and propulsion systems. SiC fibers with low oxygen
content and high crystallinity, which derived from polycarbonsilane, are the backbone as
reinforcements in load-bearing CMCs. For high temperature application, the most desired
critical properties of SiC fibers are high strength and stiffness as well as the reliable retention
of these properties throughout the service life of application. Low fiber strength and thermal
stability could result in low fracture toughness and accelerate sub-critical crack propagation in
CMCs. Thus, the mechanical durability and microstructure stability of SiC fibers are major
concerns under severe environments. Furthermore, in practical service environments, rarely is
one degradation mechanism operative, but several mechanisms operate simultaneously,
leading to the environment-pertinent degradation mechanism is complex for the SiC materials.
In order to enhance the understanding of reliability and durability of CMCs applied to high
temperature and oxidative environments, the investigations on the high temperature
mechanical properties and microstructure for SiC-based fibers subjected to severe
environments were integrated into this review.
*
Corresponding address: School of Aeronautics and Astronautics, Dalian University of Technology, 116024,
Dalian, China
2 Jianjun Sha
1. Introduction
Non-oxide Ceramic-matrix Composites (CMCs), have been extensively studied during
the last two decades. Currently, CMCs have been proposed to use as the structural materials
in application of high temperature technologies, such as advanced nuclear energy systems [1-
3], various stationary gas turbine engines and aerospace propulsion systems [4-6].
As structural engineering materials, one of the important advantages of CMCs is the
significant improvement of toughness by using continuous ceramic fibers compared to their
monolithic ceramics. In continuous ceramic fiber reinforced CMCs, the improved toughness
is attributed to several energy dissipating mechanisms when the matrix crack is occurring
under applied stress, such as crack deflection, fiber bridging and fiber sliding [7]. The energy
dissipation procedure can be illustrated schematically in Figure 1. These energy dissipations
enhanced the fracture toughness and resulted in a non-catastrophic failure mode. If the
conditions for de-bonding are satisfied, fibers bridge the crack faces in the wake of the crack
tip, subsequent to matrix cracking. The stress bore by bridge fibers applies traction forces to
the crack faces that reduce the stress intensity at the crack tip. Under specific condition crack
propagation does not occur without additional applied stress. The performance of CMCs
could be improved through the optimization of fiber/interface/matrix, and now the fabrication
process is still in developing and progressing.
σ
σ Fiber bridging
Fiber fracture
Interface
Fibers
Crack tip
crack Crack
Crack deflection
Matrix
σ σ
Monolithic composite
Figure 1. Schematic of crack propagation and principle for improved toughness in CMCs.
Based on the simple theory of mixture for the ultimate tensile strength (UTS) calculation
of continuous fiber reinforced ceramic matrix composites [8], in the case of CMCs with I-D
reinforcement alignment, when the reinforced fiber has a smaller modulus and a similar or
even large strength, significant matrix cracks must be occurred before reaching the UTS of
High Temperature Mechanical Properties and Microstructure… 3
fiber. In other words, when the applied stress is beyond that of the crack initiation of matrix,
the residual stress is mainly carried by the fiber alone. Pull out of the fiber can significantly
improve the fracture toughness. The fibers are backbone in CMCs and play a very important
role on the mechanical properties of CMCs.
Among of CMCs, SiC fiber reinforced SiC matrix composite (SiC/SiC) has been
considered as one of the most potential candidate materials, because it possessed many
attractive properties for structural engineering applications under severe environments, such
as excellent mechanical and chemical stability. It is well know, the fracture behavior of
monolithic silicon carbide is brittle and fails catastrophically. However, if the SiC fibers with
an appropriate coating as reinforcement are incorporated with the silicon carbide matrix by
specific fabrication process to form the ceramic matrix composite, the fracture characteristics
of silicon carbide materials can be significantly improved. Figure 2 shows the typical polished
morphologies and the fracture surface of near-net shape SiC/SiC composite. It is apparent that
reinforcing fibers incorporated with dense matrix through an appropriate interphase. The
significant fibers pull out could improve the fracture toughness and result in a pseudo-ductile
fracture behavior. The tough ceramics as structural materials have the potential for being used
up to about 1500 °C which is much higher than the operation temperature of superalloy
(maximum 1100 °C close to the melting point). In different fields, such as advanced nuclear
energy system, gas turbines for power/steam co-generation, heat exchangers and so on, they
give different requirements for materials performance, but common features for high
temperature technologies are excellent mechanical performance and environmental durability.
Figure 2. Typical polished morphology of SiC/SiC composite and its fracture surface with long fiber
pull out.
The major advantages for SiC/SiC composite applied to engineering are: (i) high specific
strength, (ii) superior high temperature strength and creep resistance, (iii) low thermal
expansion coefficient and high thermal conductivity, (iv) low neutron irradiation-induced
radioactivity in nuclear environments.
4 Jianjun Sha
SiC fibers with low oxygen content and high crystallinity, which derived from
polycarbonsilane, are the backbone as reinforcements in load-bearing CMCs. For high
temperature application, the most desired critical properties of SiC fibers are high strength
and stiffness as well as the reliable retention of these properties throughout the service life of
application. Low fiber strength and thermal stability could result in low fracture toughness
and accelerate sub-critical crack propagation in CMCs. The key to the successful application
of high temperature ceramic matrix composites (CMC) is the judicious selection and
incorporation of ceramic fiber reinforcement with the proper chemical, physical and
mechanical properties. Thus, the mechanical durability and the microstructure stability of SiC
fibers are major concerns under severe environments. In practical service environments,
rarely is one degradation mechanism operative, but several mechanisms operate
simultaneously, which leading to the environment-pertinent degradation mechanism is
complex for the SiC materials.
For understanding the environmental durability and describing the response of reinforced
fibers to service environments and further evaluation of reliability of CMCs, the investigation
of thermal mechanical properties on SiC-based fibers in complex situation is essential. In this
chapter, the investigations on mechanical properties and microstructure of SiC-based fibers
are reviewed in terms of varied environments; some issues concerning the environment-
pertinent properties are discussed.
2. Materials System and Characterization Technique

2.1. Materials System
The first SiC-based Nicalon fiber was produced by Nippon Carbon which allowed non-
oxide ceramic matrix composites to be developed. It made possible to use the SiC fiber as the
reinforcement for high temperature structural materials in very severe environments. In order
to improve the flexibility and strength so that the preform can be woven in complex shape,
the fine diameter SiC fibers were developed. However the first SiC-based fibers are
inherently limited by oxidation at very high temperatures. As a result of this limitation a
renewal of interest has occurred in oxide resistance of SiC fibers by approaching to the near
stoichiometric composition or addition of small amount additives to improve thermal stability
at elevated temperatures. Efforts have been made to improve the high temperature properties
of fine diameter SiC fibers by making them with compositions increasingly approaching
stoichiometry. Based on the chemical composition and the fabrication process, the
development of SiC-based fibers could be categorized into three generations as illustrated in
Figure 3.
The first fine diameter SiC fiber (Nicalon NL200) was synthesized by Yajima in Japan in
1970 [9]. The Nicalon fiber could be viewed as the representative of the first generation of
SiC fibers (Figure 3). This fiber is thermodynamically unstable at high temperature, because
it consists of SiC-nanocrystals (average size: 1-2 nm) and free carbon embedded in an
amorphous SiCxOy matrix. The amorphous SiCxOy phase decomposes at temperature beyond
1300 ˚C [10-11], with a significant gaseous species evolution and SiC crystal growth [11-12].
In order to improve the high temperature resistance of the Si-C-O fiber, a new fabrication
process of the fiber, radiation curing method [13-15], has been developed. Irradiation curing
with an electron beam was applied to make the fiber infusible and cross linking.
β-SiC crystallite (~2 nm) Amorphous (Si-C-O) β-SiC crystallite (~5 nm) β-SiC crystallite (>20 nm)
Reduced oxygen Stoichiometric

Crystallized
Low thermal stability Improved thermal stability Increased elastic modulus

Low strength Modest oxidation resistance Enhanced creep resistance
Low stiffness Enhanced oxidation resistance
Amorphous Si-C-O Fibers Low-Oxygen Si-C Fibers Crystalline SiC Fibers

NicalonTM Hi-NicalonTM NicalonTM Type-S
TyrannoTM-LoxM, etc. TyrannoTM-ZMI, etc. TyrannoTM-SA
Figure 3. Illustration of R&D of SiC-based Fibers.
To avoid the thermal instability caused by the decomposition of oxycarbide phase

(SiCxOy), in 1990, a nearly oxygen-free SiC fiber, Hi-Nicalon (Nippon-Carbon) was
developed by melt spinning, electron beam curing and pyrolysis of a polycarbosilane
precursor (PCS) under anaerobic conditions [15-16]. This fiber had a much higher thermal
stability than the standard Nicalon fiber and was viewed as the representative of the second
generation. However, the Hi-Nicalon fiber consists of not only SiC nanocrystals (average
crystal size: 5 nm) but also excess of free carbon which affects its oxidation and creep
resistance.
To reduce the free carbon content and eventually improve the high temperature properties
of the fibers, extensive efforts have been devoted to develop near stoichiometric and high
crystallized SiC fibers. The precursor fiber can be sintered at high temperatures that excess
carbon and oxygen are lost as volatile species to yield polycrystalline and near-stoichiometric
SiC fiber. These fibers are advanced SiC fibers and generally called the third generation of
SiC fibers (Figure 3), including Hi-Nicalon type S fiber [17], Tyranno SA fibers [18] and
Sylramic SiC fiber [19]. The third generation of SiC fibers is oxygen-free and near-
stoichiometric (atomic ratio: C/Si=1.00–1.08). Furthermore, their grain size is relatively large
(20–200 nm) and their thermal stability is excellent.
For enhancing the environmental durability of CMCs, SiC-based fibers with high
crystallinity and near stoichiometry would be preferential. Based on this standing point, the
following SiC-based fibers were used for the work presented in this chapter (Table 1).
6 Jianjun Sha
Table 1. SiC-based fibers used for the work in this chapter and their properties
provided by manufacture
Oxygen Modulus Diameter

SiC fiber C/Si Strength (Gpa) Density (g/cm3)
(wt%) (Gpa) (μm)
HNL 1.39 0.5 2.8 270 2.74 14

HNLS 1.05 0.2 2.6 420 3.1 12
TySA 1.07 <0.5 2.6 400 3.0 7
Note*: HNL=Hi-NicalonTM fiber (500 fiber/yarn), HNLS=Hi-NicalonTM Type S (500 fiber/yarn) and
TySA=TyrannoTM SA fiber (1600 fiber/yarn).
Both HNL [20] and HNLS [21] fibers were fabricated by Nippon Carbon Co., Japan.
TySA [18] fibers were fabricated by Ube Industry Co. Ltd., Japan. It is clear that: the HNL
fiber (SiC1.39O0.014) consists of a mixture of SiC nano-crystals and free carbon; the HNLS
(SiC1.05O0.007) and TySA (SiAl0.02C1.07O0.03) fibers have near stoichiometry and high
crystallinity. Noting the TySA fiber contains small amount of alumina (less than 1 wt%) in
order to improve its thermal stability.
2.2. Methodology
Due to varied manufacturing approaches, SiC fibers are being produced with different
surface morphologies and internal microstructures, particularly regarding size and population
of defects, grains and grain boundary phases. The strength and creep resistance of SiC fibers,
which are dependent on intrinsic and extrinsic factors, such as material itself and service
environment, are first properties examined. The conventional evaluation methods developed
for engineering materials, could not be applied to the SiC fibers with fine diameter and brittle
nature. The methodology for the evaluation of these mechanical properties in precise way is
crucial, and it is described as follows in detail.
2.2.1. Single Fiber Tensile Test Technique
Single fiber tensile test technique is used to evaluate the tensile properties of SiC fibers.
Generally, in each condition, about 30 single fibers were selected at random from the fiber
yarn and cut into 50 mm lengths. The 50 mm length fiber is mounted on a paper cardboard
frame. Tensile tests were carried out at ambient conditions using an Instron test machine
equipped with a 2.5 Newton load cell (Figure 4). The tensile test generally followed ASTM-
recommended procedures [22]. The individual fiber was carefully separated and selected
randomly from the yarn of each fiber type. The fiber diameters were determined from one end
of projecting fragments. To do this, each 50 mm length fiber was mounted by centering and
fastening its ends with Aradi glue onto a paper cardboard frame with a 25.4 mm distance
between bonding points that defined the fiber gauge length (Figure 4). To prevent the fiber
pull-out from the bonding point of Aradi glue before reaching the failure load, the glue coated
at least a 5 mm length of the fiber ends and set for several days for complete drying. Before
tensile test, the paper cardboard was cut very carefully along the center line across the hole so
that the load was completely applied to the fiber. Load was applied at a constant displacement
rate of 0.3 mm/min (equivalent to a strain rate: 2.2×10-4 s-1). However, it is very difficult to
collect the fiber fragments after tensile test in normal way, because the sudden release of very
high stored energy at fracture will make the fragment breaking into many pieces. Therefore, a
procedure to decay the release of stored energy for capturing the broken fragments was
developed. The developed procedure was very effective to capture the fiber fragments. In the
case of fiber fractured at the edge of bonding point, the fracture of fiber might be caused by
bending moment due to poor alignment, which doesn’t reflect the true strength of a fiber.
Thus, this test was viewed as invalid.
Single filament Single filament

tensile test tensile specimen
19.0 mm
Load Cell: 2.5N
Cardboard Adhesive
63.5 mm
25.4 mm
Instron (Model 5581) Fiber
Strain Rate: 2 x 10-4 /s
Figure 4. Test method for measuring the tensile strength of single fiber and specimen geometry.
In the capturing of fracture fragments, a small rectangular plastic film (6.0×8.0 mm2) was
used and coated on one side with glycerin. This plastic film with glycerin on one side was
carefully bridged across the center hole of cardboard　(Figure 4), and the fiber to be tested
was completely wetted by glycerin. The glycerin effectively damped the shock wave in the
fracture of fiber and it usually fractured only at one location. Each fiber segment remained to
its half of the mounting frame, which is important for later SEM examination.
Because the fracture of ceramic materials generally originates from the critical flaws,
assuming those flaws in the fiber are distributed randomly in location, then the strength of the
fiber is determined by the strength at its weakest point (weakest link rule). Test on randomly
selected fibers will show a considerable dispersion in failure strengths because of the
presence of flaws. The strength of fibers can be shown generally to follow the classical two-
parameter Weibull distribution.
The two-parameter Weibull theory of statistical fracture was applied to characterize the
fracture behavior of brittle SiC fibers [23].
According to weibull’s statistical theory, the probability of failure Fi, of fiber subjected to
nominal tensile strength, σ, is given as
8 Jianjun Sha
σ −σ u m
Fi = 1 − exp[− L( ) ] (1)
σ0
where m is the Weibull modulus of the fiber, L is the gage length of the fiber, σu the stress
below which fiber is assumed to have zero failure probability and σ0 the Weibull scale
parameters. Both σ0 and m are constant for a given material and assuming σu=o. The Weibull
modulus, m, of the fiber can be determined by rearranging equation (1) into the form
1
ln ln( ) = m ln( σσ0 ) + c (2)
1 − Fi
where c is constant. Actually, m is the slope in a two parameter weibull plot, which can be
obtained by least squares fitting to the linear relationship of equation (2). In equation (2), the
probability of fiber failure Fi at the nth ranked sample from a total of N specimens is obtained
from the mean rank method as Fi=n/(N+1).
The Weibull average strength (σavg) was calculated from the relation σavg=σ0Г(1+1/m),
where Г(1+1/m) is a Gamma function [24].
∞
Γ(ζ ) = ∫ e − x X ζ −1dx. (3)
0
2.2.2. Bending Stress Relaxation Test
For evaluating the creep and rupture strength of individual fibers, the conventional tensile
creep test procedure is to subject an individual fiber specimen of length L and diameter D to a
constant tensile load P at a constant test temperature T and to measure fiber elongation ΔL
versus time t until the fiber finally fractures at rupture time tR. Creep strain is then determined
from
εc=ΔL(t, T, σ, G)/L (4)
where σ=4P/πD2 is the applied stress and G symbolizes effects from the environments.
Rupture time typically is also a function of temperature, stress, length and environment,
tR=tR(T, σ, L, G). (5)
Unfortunately, measuring the creep of SiC fibers under tensile loading is difficult. This is
especially true with fine diameter fibers which are often degraded by an air test environment
and can be easily fractured during grip and strain sensor attachment. Another problem is the
accuracy of the creep strain, because in many cases it is hard to define the gauge length
(including the cold grip and hot grip) during the tensile test. To avoid these problems, in this
study, a modified bend stress relaxation (BSR) method was utilized to evaluate the creep
resistance of SiC fibers, and attempts were made to relate the BSR with tensile creep for fine-
diameter fibers. An schematic illustration of the BSR test jig was shown in Figure 5. For
evaluating the environmental effect on the creep resistance of SiC-based fiber, a modification
was made on the conventional method [25] as shown in Figure 5. This improvement makes
the tested specimen to be sufficiently exposed to the test environment.
Fiber loop
Fiber loop R0
R0
Conventional Method Modified method

Figure 5. Comparison between conventional and modified bend stress relaxation test method.
In this method, the fiber with a length of 2-5 cm are wound around the rod at a constant
surface strain and held at desired temperature for given times in controlled environment. For
small diameter fibers in ambient conditions, the bending modes of different applied strain can
be achieved by tying the fiber into small loops with different radius, R0. The fiber loop is then
subjected to a specific time (t), temperature (T), and environmental treatment. After
treatment, the applied stain is then removed by release the fiber loop from the test jig or
broken the fiber loop at one point at room temperature. The stress relaxation-induced effects
are measured in terms of the residual radius of fiber loops, Ra. If the fiber remains completely
elastic during treatment, the broken loop will be straight with no curvature, i.e., Ra=∞. If the
creep-induced stress relaxation occurs, the Ra will be finite and typically will decrease with
increasing the treatment time and exposure temperature.
To quantify the stress relaxation occurred during thermal exposure, a parameter m, stress
relaxation parameter was defined, which is the ratio of final to initial stress at any local
position in the fiber as illustrated in Figure 6. That is
m = σ (t ,T , ε 0 ) / σ 0 (0,T , ε 0 ) = E • ε e (t ,T , ε 0 ) /( E • ε 0 ) = [ε 0 − ε c (t ,T , ε 0 )] / ε 0 (6)
where ε0, εe, and εc are the local initial strain, final elastic strain and total creep-induced strain,
and all of these strains vary within the fiber. For convenience, one can assume that (1) εc is
linearly proportional to the ε0 regardless of the stress direction and (2) it can be measured at
room temperature by relation εc=z/Ra. z is the distance from the neutral axis in the fiber loop
plane. The first assumption of linear strain dependence is generally valid for polycrystalline
materials which stress relax due to grain boundary sliding mechanisms that are either
elastically or diffusionally accommodated. That is, a stress power dependence of n ≈1 (εc ∝
σｎ ) is typically observed throughout both the primary and secondary creep stages. The second
assumption implies that at each local position within the fiber, stress relaxation not only is
proportional to ε0 but follows the same time-temperature dependence. This typically requires
a fiber with a uniform isotropic microstructure that creeps with an n ≅1 power dependence. If
these assumptions apply, the BSR m ration is independent of position and initial applied
10 Jianjun Sha
strain. It is then only a function of treatment time and temperature and can be determined by
the simple relation:
m (t, T)=1- R0/Ra (7)
Fiber
Loop
Ra
R0
Thermal exposure (t, T)
Initial loop at room temperature Broken loop after stress relaxation
z
m=1-εc/ε0=1-R0/Ra
εc: Creep strain (z/Ra)
ε0: Initial applies strain (z/R0)
d
R0: Initial curvature
d: Fiber diameter Ra: Residual curvature
Neutral axis
Figure 6. Schematic representation of the test principle of the bend stress relaxation originally
developed in Ref. [25].
In comparison to tensile creep test which conducted under a dead load with accessories
for strain measurement and a defined gauge length, the BSR offers many advantages
including the ability to simultaneously study many fibers of small diameter and short length
under same time, temperature, and controlled environmental conditions.
Here, it is obvious that stress relaxation parameter, m, can be determined based on the
extent of permanent deformation occurred during stress relaxation. An m value which
approaches 1 indicates that no permanent deformation occurred during the high temperature
exposure, while a m value of 0 indicates that the stress completely relaxed. Hence, fibers are
considered more thermally stable against creep as m values increase from 0 to 1 [25].
Practically, the BSR test also offers insight into the behavior of bent fibers in woven
performs as well as conditions for “creep-forming” fibers into complex shapes. And also, it
eliminates the need for furnace with long uniform hot zones, for mechanical grips, for remote
sensors and for multiple experimental runs that are often required to establish time,
temperature and stress dependencies and also to determine statistical variations. Second, for
polycrystalline fibers, which generally creep with stress power dependencies near unity, if the
BSR m-ratios are independent of applied strains, and thus equal to those stress relaxation
ratios that would be measured in a pure tensile test. Furthermore, by BSR test, it will be
beneficial to understand the basic mechanisms which controlled the creep behavior of SiC
fibers with fine diameters [25].
2.2.3. Microstructural Characterization
This section described some techniques that will help to clarify why the mechanical
properties were changed and how the microstructure influenced the mechanical properties.
Among these techniques, the facilities frequently used are optical microscopy (OM), scanning
electron microscopy (SEM) and X-ray diffractometer (XRD).
Optical microscopy with a video was used to examine the macrostructure of materials. It
is also useful in the determination of fiber loop diameter in the BSR test. In BSR test, a
photograph was taken of the loops before and after thermal treatment. The initial applied
curvature R0 or residual curvature Ra was measured by fitting a circle on the fiber in the
photograph, and then the curvature could be obtained by a graphic technique.
XRD is very useful in the identification of the crystal phase and the estimation of the
crystallite size. The X-ray diffraction (XRD) patterns were recorded by means of X-ray
diffractometer (Rigaku) with a rotating anode (Cu-Kα radiation).
X-ray scattering of the atom planes in the crystals gives a diffraction pattern
characteristic of the crystal structure. Diffraction peaks correspond to scattering of specific
planes which are defined by the structure factor and Bragg’s law,
n • λ = 2d • sin θ (8)
where n=1, 2, 3,…, λ is the X-ray wavelength, d is the planar spacing and θ is the diffraction
angle. The relative intensity of the diffraction pattern varied with the diffraction plane to aid
in structure identification. Comparing the experimental diffraction pattern to a known pattern
allows the crystal structure to be identified.
The specimen was prepared by attaching the powder sample on the glass slide with
double-side adhesive tape. The powder was obtained by pulverizing the fiber tow of about 0.1
g in a mortar. During pulverizing, in order to prevent the spray of fiber fragments from
pulverizing, the alcohol was mixed with powder to make viscous slurry. After careful milling
and drying, powder was put on the glass slide with double-side adhesive tape, and it was
pushed to be attached tightly.
During scanning, the XRD operated at 40 kV and 20 mA was used to identify the crystal
phase in the fibers. All of the scans were run at 2˚/min with a time interval of 0.05 s for
sampling. The range of 2θ was 10˚-90˚. The apparent crystallite size (D111) of the β-SiC
crystalline phase present in the samples was calculated from the half-value width of (111)
diffraction peak using Scherrer’s formula:
D = K • λ /( H w • cosθ ) (9)
where K is a constant (taken as 0.9), λ the CuKα wavelength (i.e., λ=0.154056), Hw the half-
value width of β-SiC (111) peak and θ the Bragg angle (θ=17.5˚ for β-SiC (111)).
FE-SEM (Field-Emission Scanning Electron Microscope, model; JEOL JEM-2010),
which provides narrow probing beams as well as high electron energy resulting in both
improved spatial resolution and minimized sample charging and damage, is a powerful
weapon in the characterization of dimension and microstructure such as examination of
surface morphologies and fractograph.
12 Jianjun Sha
The FE-SEM was employed to determine the typical diameter variation across a cross
section of fiber yarn and along the fiber length. The selected fiber was attached on the
specimen holder with the double-sided carbon tape. The fiber diameter was determined from
SEM image with high magnification (x5000). Special care was taken in the register of fiber
fragment so that the diameters represented the fibers that we want to investigate.
To examine the fracture surface, firstly a technique described in section 2.2.1 was
adopted to obtain the fracture fragments, and the fracture location was noted. In order to take
a high quality picture, the clean fracture surface is needed and it can be gotten by washing the
fragment in ultrasonic bath contained alcohol for about 30 s. Each fiber segment for the
successful tests was gripped with a narrow tip tweezers and broken off at the bonding point.
The clean segments were mounted on double-sided carbon tape applied to the circular
side surface of cylinder specimen holder (10 mm copper cylinder in diameter). Usually about
10 segments were mounted with each pair of matching fiber fracture surfaces, and keep the
fracture surface with a protruding length about 2 mm above the specimen holder surface. And
also, the fragments should be perpendicular to the horizontal surface of holder. Then, the
fracture surfaces of the aligned fragments could easily be located, identified and imaged by
SEM.
3. Basic Characteristics
3.1. Fiber Diameter Variation Analysis
Accurate determination of fiber diameter is necessary for the estimation of fiber’s

strength, because the use of a nominal/mean fiber diameter to determine individual fiber
strength is not precise. This is the likely situation for most polymer-derived SiC fibers that are
processed using a spinning method [26]. For instance, a 10% error in diameter would result in
about a 21% error in the strength calculation. Such errors cause additional scatter in the
Weibull strength distribution, which results in a low value of Weibull modulus. In such cases,
to properly determine the fiber diameter in the assessment of fiber strength is important.
The fiber diameter variation from fiber to fiber across a tow and along the single fiber
length was assessed by image analysis from SEM.
3.1.1. Fiber Diameter Variation within a Tow
To investigate the fiber diameter variation across a tow, a yarn of each fiber type was
scattered and mounted on the plane surface of copper specimen holder, and then carbon tape
was used to fix this fiber bundle. The picture was taken on these fibers one by one and the
number of selected fibers is as large as possible. Figure 7 showed the fiber diameter variation
across a fiber tow. Mean diameter and standard deviation were also calculated in Table. 2.
From this result, the HNLS fiber type showed smallest diameter variation across the tow,
which indicated this fiber type has more uniform diameter within a cross section of its tow.
The HNL fibers displayed a relatively wide fiber diameter variations within a tow (10.78-
16.60 um). Noting the average diameter for each fiber type is possibly different from batch to
batch. In these fiber types, the measured mean fiber diameter values given in Table 2 agree
fairly well with the manufacture’s value in Table 1. The HNL and HNLS fiber in diameter
size distribution have a standard deviation, 9.6% and 4.8%, respectively, and the TySA fiber
has a relatively high standard deviation, 13.2%.
Diameter of selected filament (um)


17 15 10
16
(a) HNL (b) HNLS (c) TySA
14 9
15
14 13 8
13 12 7
12
11 11 6
10 10 5
0 20 40 60 80 0 20 40 60 0 20 40 60 80 100
Selected filament within a tow Selected filament within a tow Selected filament within a tow
Figure 7. Fiber diameter variation within a tow: (a) HNL fiber, (b) HNLS fiber, (c) TySA fiber.
Table 2. Diameter variation within a tow determined by image analysis from SEM
Fiber Types Hi-NicalonTM Hi-NicalonTM Type S TyrannoTM SA
Number of fibers counted 80 56 112

Mean Diameter (um) 13.99 12.5 7.13
Standard Deviation (um) ±1.34 ±0.6 ±0.94
Minumum/Maximum 16.60/10.78 14.04/10.85 9.67/5.20
3.1.2. Fiber Diameter Variation along the Fiber Length
To investigate typical fiber diameter variation along a fiber length, three individual fibers
with a length of 30 cm were pulled out randomly from a tow and cut sequentially into 1 cm
segments. Before pulling out the individual fiber, soaking the fiber tow in acetone for 2 days
and followed by washing in boiling water for 1 minute. This step is quite necessary in aiding
the fiber separation, pulling, and reduction of the handling damage to the fiber. For viewing
by SEM, the 1 cm length segments were fastened sequentially on the flat specimen holder by
carbon tape. For reducing the charging effects, the segments should be connected well with
the specimen holder. The diameter was determined directly from image taken by SEM. It
should be noted that the ends of each segment were carefully retained in register so that the
diameters represented the variation of the diameter along the fiber length at 1 cm interval. The
fiber diameter variations at 1 cm interval along 30 cm length fiber for the randomly selected
fibers were shown in Figure 8. The typical diameter variation exhibited by the 30 cm fibers
was about ±1~3 um for HNL fiber. The HNL fibers exhibited a cyclic diameter variation with
a repeat distance of about 15 cm. The HNLS fiber did not exhibit strong cyclic diameter
variation, but a rather abrupt rate of change in diameters of about ±0.25 um/cm was observed.
The two of three TySA fibers exhibited a very similar variation tendency in diameter, and the
least variation ration of about <0.25 um/cm was also observed along the fiber length. But for
the third set of TySA fiber, it exhibited high diameter variation rate of >0.25 um/cm.
14 Jianjun Sha
D e v . F ro m A v g . D ia . (u m )
4 1 1.5
D ev. From A vg. D ia. (um )

3 (a) HNL (b) HNLS 1.25 (c) TySA
D e v . F ro m A v g . D ia . (u m )
0.75
2 1
0.5 0.75
1 0.25 0.5
0 0 0.25
0
-1 -0.25 -0.25
-2 -0.5 -0.5
-0.75 -0.75
-3 -1
-4 -1 -1.25
-1.25 -1.5
0 5 10 15 20 25 30 35
0 5 10 15 20 25 30 35 0 5 10 15 20 25
Position along filament length (cm) Position along filament length (cm) Position along filament length (cm)
Figure 8. Fiber diameter variation along the fiber length of 30 cm at a 1 cm interval: (a) HNL fiber, (b)
HNLS fiber, (c) TySA fiber.
3.2. XRD Patterns
Figure 9 shows XRD patterns for HNL, HNLS and TySA fiber. Obvious β-SiC peaks
were observed in these patterns, but HNLS and TySA fiber showed a relatively sharp peak.
This indicated the HNLS and TySA fiber have large crystallite size and high crystallization.
This is in agreement with the manufacture’s information that these fibers are near
stoichiometric and high crystallization. In the case of HNL fiber, it has been reported by
manufacturer that excess free carbon and amorphous phase existed at the grain boundaries.
(111)
TySA, DSiC(111) =22.7 nm

(220)
(222)
HNLS, DSiC(111) =11.4 nm
HNL, DSiC(111)=4.0 nm
60 80
ta
Figure 9. XRD patterns of as-received SiC-based fibers, (β-SiC).

Using Scherrer’s equation described in section 2.2.3, the apparent crystallite size of β-
SiC, was calculated from the half-value width of the (111) peak as shown on top right in
Figure 9.
3.3. Tensile Properties and Fracture Surface
In Figure 10, the tensile strength distribution and related tensile properties of three fiber
types were shown in a two-parameter Weibull plot.
3.0
HNL σ0=2.78 GPa
2.0 HNLS σAvg.=2.5 GPa
S.D.=0.18
1.0 TySA m=3.6
ln(ln(1/(1-Fi)))
0.0
σ0=3.3 GPa
-1.0 σAvg.=2.97 GPa
S.D.=0.42
-2.0 m=3.7
σ0=3.5 GPa
-3.0 σAvg.=3.25 Gpa
S.D.=0.17
-4.0 m=4.6
-5.0
-1 -0.5 0 0.5 1 1.5 2
ln (σi [GPa])
Figure 10. Two-parameter Weibull plot for three fiber types indicating the related tensile properties.
The m values listed in Figure 10 are slightly lower than that of in Refs. [27-28], but they
are very similar to the value obtained in Ref. [29]. As we know, the strength of ceramic fibers
is associated with the gauge length (weakest link rule) and fiber diameter. Long gauge length
and poor uniformity of fibers might be responsible for low Weibull modulus. Furthermore,
the performance of fibers also varied from batch to batch.
To examine the fracture surface of individual fiber segments, each fiber segment for the
successful test was gripped with narrow tip tweezers and broken off at the frame edge.
Generally, for the brittle materials, such as ceramics and glass, their fracture originated from
the critical flaw surrounded by the mirror zone, mist zone and hackle zone. Figure 11 is an
schematic illustration of crack initiation and propagation route of the SiC-based fiber showing
fracture mirror zone surrounding the critical flaw.
During microstructure observation, special care was taken on the features of critical flaw
size (rc), flaw type and mirror size (rm). The obvious fracture mirror zone was observed on the
fracture surface of most of HNL and HNLS fiber fragments except for TySA fiber.
16 Jianjun Sha
For the set of HNL fiber successfully tested, the critical flaws with different dimension
were observed on the surface or inner area of fibers, but the locations of most of critical flaws
are near the surface of fibers. In the case of HNLS fiber, the critical flaws were mainly
identified as the inner flaws (inner pore or inclusion).
Hackle
zone
Mist
zone
Mirror
zone
Critical flaw
rm
rc
Figure 11. Schematic illustration of fracture surface of fiber, showing fracture originated from critical
flaw surrounding by mirror and hackle regions.
Typical features of fracture surfaces for the SiC-based fibers are shown in the following
SEM micrographs.
In Figures 12 (a)-(b), a mating pair of fracture surfaces shows a surface critical flaw and
the surrounding mirror, mist and hackle regions. Both sides of the mating fracture surfaces
exhibited a well defined void. For this particular fiber, the diameter (d), the mirror radius (rm)
and the critical flaw radius (rc) were measured to be 15, 2.05 and 0.74 um, respectively. The
tensile strength (σf) is 2.59 GPa.
Figures 12 (c)-(d) are a typical pair of mating fracture surfaces showing a critical flaw of
inner pore type. It was very clear that the inner pore was observed in each surface. For this
sample, d=13.0 um, σf =4.22 GPa, rm=1.05 um and rc=0.34 um.
More attention was paid to HNLS fiber. As observed on the fracture surface of HNLS
fiber, most of critical flaws are inner flaws and identified as the second inclusions. Figure 12
(e)-(f) shows a pair of mating fracture surface of HNLS fiber exhibited a pore (Figure 12 (e))
and inclusion (Figure 12 (f)) on the opposite fracture surface. For this type of flaw, it is
possible that a second inclusion was pulled out freely from a pore. For this HNLS fiber with a
diameter of 14 um, it gave an strength (σf) of 3.34 GPa. The mirror radius and the critical
flaw radius are rm=2 um and rc=0.47 um, respectively.
1um 1um
aa bb
1um 1um
c
a d
b
1um 1um
ea bf
Figure 12. A typical pair of mating fracture surfaces: (a) and (b) for HNL fiber, showing surface critical
flaw is a void, for this sample, d=15 um, σf=2.59 GPa, rm=2.05 um and rc=0.74 um; (c) and (d) for HNL
fiber, showing an internal pore (critical flaw) in each surface, for this sample, d=13.0 um, σf=4.22 GPa,
rm=1.05 um and rc=0.34 um; the opposite fracture surfaces of HNLS fiber appeared a remaining pore (e)
and second phase inclusion (f) surrounded by the mirror and hackle regions, for this HNLS sample, d=
14 um, σf=3.34 GPa, rm=2 um and rc=0.47 um.
18 Jianjun Sha
In order to know the composition of this type of flaw, a higher magnification of a typical
inclusion is shown in Figure 13, and energy-dispersive spectroscopy (EDS) scan was
performed across this inclusion. The three line scans qualitatively show the variation of the C,
O and Si concentrations in the region of the critical flaw. The relative concentrations of C, O
and Si are almost unchanged until the inclusion is encountered by the scan, where the C
concentration abruptly increases and the Si concentration abruptly decreases at the region of
the inclusion. The O concentration within the region of inclusion appears to somewhat lower
than its base-line level. The pertinent data for this particular fiber are given in the caption of
Figure 13. Based on this information, the inclusion region (critical flaw) appears to be a
carbon-rich region. Inclusion with similar characteristics was also observed in previous study
of HNL fiber [27].
1um (a)
(b)
CKα
OKα (c)
SiKα (d)
Figure 13. An enlarged SEM view of: (a) an inclusion-type critical flaw, (b), (c) and (d) corresponding
to the C, O and Si EDS line scan across the inclusion.
Linking these micrographs to the production process, the defects such as voids or
inclusions from impurities or un-melted precursors may exist in polycarbonsilane-derived
fibers. These defects may generate local internal stress concentration during the process and
lead to the crack formation under tension. And also, under the same processing parameters, it
is likely that the stress concentration will vary with fiber diameter, since it is easier to relax
the stress concentration in a fiber with a smaller diameter. Wanger [30] suggested that the
spinneret hole has laminar flow properties which change with diameters during the
fabrication. This may also result in the flow density variation with varying fiber diameters.
The variation in flaw size is certainly a factor affecting the ceramic fiber’s strength and thus
needs to be incorporated into the fracture statistics.
In the case of TySA fiber with the small diameter (about 7.5 um), no critical flaw was
observed on the fracture surface (Figure 14). The fracture surfaces showed a trans-crystallite
fracture behavior. This fracture behavior could be partially related to the residual stresses
caused by the addition of alumina in this fiber. Existence of residual stresses in the grain
boundary of TySA fiber is quite possible because of significant mismatch in the coefficient of
thermal expansion between SiC and Alumina (SiC: 3.3×10-6/K; Alumina: 9.1-9.9×10-6/K) and
high sintering temperature (higher than 1700 ºC). In TySA fibers, the change in the extension
stability of micro-crack in the residual stress filed might improve the grain boundary strength.
The increase in grain boundary strength could explain the trans-crystalline fracture surface of
TySA fiber.
1μm
Figure 14. Typical fracture surface observation of TySA fiber showing trans-crystallite fracture.
3.4. Correlation between Tensile Strength and Fiber Diameter
The dependencies of the fiber tensile strengths on diameter for the HNL, HNLS and
TySA fibers are illustrated in Figure 15, respectively.
In Figure 15, the fiber tensile strengths exhibit significant scatter. Nevertheless, the
general tendency that fibers with larger diameters have lower strengths is consistent. The fiber
with large diameter will be easy to cause the stress concentration around the defects. The
similar phenomenon for Nicalon fibers was also observed in Ref. [31], and quantitative
explanation has been given from the fracture mechanics. The tensile strength σ vs. normalized
diameter (d/d0) data were fit with an empirical power law dependence of the form
σ=K0(d/d0)-n (10)
20 Jianjun Sha
where d0 is the average diameter of the data set, n is the power law exponent and K0 is the
average strength expected for a fiber with diameter d0. The values of K0 and n, determined
from linear least squares fitting of the fiber tensile strength data to Equation (10), are included
in the plot for each fiber type.
d0=14.5 d0=12.9 d0=7.4

K0=2.22 K0=2.13 K0=2.55
6 6 6 n=1.34
HNL n=0.62 HNLS n=1.98 TySA
Fiber strength (GPa)
5 5 5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
0.7 0.8 0.9 1 1.1 1.2 0.7 0.8 0.9 1 1.1 1.2 0.7 0.8 0.9 1 1.1 1.2
Normalized fiber Diameter (d/d0) Normalized fiber Diameter (d/d 0) Normalized fiber Diameter (d/d 0)
Figure 15. The dependence of tensile strength on the fiber diameter for: (a) HNL, (b) HNLS, (c) TySA
fiber types. The solid curve was obtained by fitting the data point in each plot with Equation (10).
Although the data is significant scatter, the general tendency that fibers with larger diameters have
lower strength is consistent.
3.5. Correlation between Tensile Strength and Mirror Size
Prior to understanding the correlation between the tensile strength and the mirror size, the
dependence of critical flaw size on its corresponding mirror size was examined as shown in
Figure 16.
2 2 (b) HNLS
Critical flaw radius (um)
(a) HNL
1.5 1.5
1 1
0.5 0.5
0 0
0 1 2 3 4 5 6 0 1 2 3 4 5
Mirror radius (um) Mirror radius (um)
Figure 16. The critical flaw radius vs. the mirror radius: (a) HNL fiber, the slop of fitting straight line is
0.38; (b) HNLS fiber, the slop of fitting straight line is 0.39.
Although the data is a little scattering in Figure 16, surprisingly, the slops of fitting line in
HNL and HNLS fibers have a very similar value. This gave us an information that critical
flaw size (rc) was linearly related to the mirror size (rm). A linear relationship between the
critical flaw size (rc) and the mirror size (rm) is commonly observed for brittle fracture of
ceramics and glass [26-27,32-33]. The ratio rc/rm in this study is about 0.39 which is
somewhat larger than the 0.33 for Hi-Nicalon fiber in previous study [27], and 0.19–0.22
range of values observed for ceramic grade (CG) Nicalon fiber [26]. Possibly, this difference
could be attributed to the definition of mirror size and the crystallization of PCS-derived
fibers. On the other hand, it could also be associated with the accuracy in estimating the
actual flaw sizes from the SEM micrographs.
In Figure 17, the individual fiber strengths are compared to their corresponding mirror
sizes for the HNL and HNLS fibers. The best fit straight line has an slope of -0.48 for HNL
fiber, and -0.49 for HNLS fiber (Figure 17). These values are very close to the -0.5 for brittle
ceramic materials according to the Griffith theory　[34-35]. The data scatter appeared in the
plots of tensile strength vs. mirror size for the HNL and HNLS (Figure 17) could be consist
with the trend that fiber with larger diameters are weaker as reported previously for the Hi-
Nicalon™ fiber [35] as well as for other polymer-derived SiC fibers [24]. In general, for the
ceramic fibers, the critical flaw density varied with varying fiber diameters, and the flaw
within the fiber with larger diameter is more easily to cause the stress concentration.
Furthermore, the error in the diameter measurement of fibers will also cause the data scatter
in the calculated fracture strength. We have observed that the diameter of fiber varied along
it’s length. And also, the strength of SiC fibers is sensitive to the surface critical flaw. For this
batch of HNL and HNLS fibers, the fraction of critical flaws occurred at the fiber surface is
relatively low, but mostly were distributed internally with only a slight preference for being
located nearer to the fiber surface than to the fiber center.
6 4
Tensile strength (GPa))
(a) HNL (b) HNLS

5 3
4
3 2
2
1
1
0 0
0 1 2 3 4 5 6 0 1 2 3 4 5
Mirror size (um) Mirror size (um)
Figure 17. Tensile strength vs. mirror size: (a) HNL fiber, (b) HNLS fiber. Fitting straight line
represented slops of approximate 0.5.
22 Jianjun Sha
3.6. Fracture Toughness and Critical Fracture Energy
Because the relation between the fracture strength and the critical flaw size observed the
Griffith theory according to the results in Figure 16 and Figure 17, thus, the fracture
mechanics principle could be applied to calculate the fracture toughness and the critical
fracture energy.
3.6.1. Fracture Toughness
Fracture mechanics predicts a relation between flaw radius, fracture strength (σf) and the
fracture toughness (K1c) for brittle materials [26,31-32], where K1c is the mode 1 fracture
toughness of the SiC fiber.
σf(rc)1/2=YK1c=constant (11)
In Equation (11), Y is a geometric factor which depends on the critical flaw shape and
location and its relative size compared to the fiber dimension. Y is 1.56 for a small, centrally
located penny-shaped flaw in a plane normal to the tensile axis given in the Ref. [31].
Additionally, it has been extensively demonstrated that the product of strength, σf, and
the square root of mirror size obeyed following formula [26,31-33,36]
σf=Am(rm)-0.5 (12)
where Am is the mirror constant.

Substituting σf in Equation (11) with Equation (12), the fracture toughness, K1c, could be
expressed as:
KIc=Am(rc/rm)0.5/Y (13)
6 4
Tensile strength (GPa)
5 (a) HNL (b) HNLS

3
4
3 2
2
1
1
0 0
0 0.5 1 1.5 0 0.5 1
(rm)-0.5 (r m)-0.5
Figure 18. The fiber tensile strength vs. the square root of the fracture mirror radius: (a) HNL fiber, the
slope yields the mirror constant Am=3.93 MPam1/2; (b) HNLS fiber, the slope yields the mirror constant
Am=4.33 MPam1/2.
In Figure 18, the tensile strength σf vs. (rm)-0.5 are plotted for HNL and HNLS fiber,
respectively. The data were fit to a linear relation by regression analysis. The mirror constant
Am, defined as the slope of the fit straight line, was determined to be 3.93 MPam1/2 for HNL
fiber, 4.33　MPam1/2 for HNLS fiber, respectively. From K1c=Am(rc/rm)0.5/Y, using rc≈0.39 rm
and Y=1.56, the calculated K1c is 1.56 MPam1/2 for HNL fiber, 1.74　MPam1/2　for HNLS
fiber. Since Am is an average value, the K1c-value determined for these fibers also is an
average value.
The K1c for polycrystalline SiC is ≈ 2 MPa⋅m1/2, while that for most amorphous ceramics
is ≈0.5 to 1 MPa⋅m1/2 [36].
3.6.2 Critical Fracture Energy
Attempts [32-33] have been made to relate the critical flaw radius to the critical fracture
energy, γc, which can be obtained from the following equations,
rc = Y 2 2 Eγ c / σ 2f (14)
1
σ f • rc 2 = Y 2 Eγ c (15)
where Y is a geometric factor, E the modulus of elasticity. By substituting the fracture

strength in Equation (15) with Equation (11), the critical fracture energy could be simplified
as:
K 12c
γc = (16)
2E
The critical fracture energy calculated with Equation (16) is 4.5 J/m2 for HNL fiber, 3.6
2
J/m for HNLS fiber. The low critical fracture energy for HNLS fiber could be attributed to
the low strain to failure (HNL: 1%, HNLS: 0.65%).
The Griffith theory presents a criterion for propagation of preexisting flaws that generally
determines the failure of brittle materials. After enough energy has been supplied to the crack,
it will propagate at velocity which increases as its length increases. Since the driving force
depends on crack length, crack velocity will increase until it approaches a terminal velocity.
As the crack approaches the terminal velocity, the sum of the potential energy resulting from
its increasing length and the kinetic energy resulting from its motion becomes greater than the
energy that can be used to increase the velocity of the crack. Small cracks are nucleated
around the tip of the main crack, forming mist, but there is insufficient energy to propagate
these secondary cracks very far. Limited velocity increases allow propagation of such
secondary cracks to form hackle. Finally, when enough energy is available, the crack can
branch macroscopically.
24 Jianjun Sha
4. Mechanical Properties and Microstructure Under Various

Environments
4.1. Heat Treatment at Elevated Temperatures
Because the CMCs may be fabricated above the fiber’s processing temperature [37-38],
in which case, the performance of fibers could be changed by high temperature treatment.
Thus, the identification of correlation between performance and heat treatment temperature is
essential for exploring the optimum condition for high performance CMCs fabrication and
application.
4.1.1. Correlation between Tensile Strength, Crystal Size and Heat

Treatment Temperatures
Figure 19 shows average room temperature tensile strengths and apparent crystallite sizes
for three types of fibers after heat treatment in Ar for 1 h at elevated temperatures. The
apparent crystallite size of β-SiC, DSiC, was calculated from the half-value width of the (111)
peak by using the Scherrer’s formula. From the dependence of crystallite size on the heat
treatment temperature as shown in Figure 19, following features were observed: (i) the grain
coarsening of HNL fiber started at 1400 °C; (ii) the crystallite size of β-SiC in HNLS and
TySA fiber remained almost constant as heat treatment temperature <1600 °C, while higher
temperature heat treatment caused an continuous coarsening in crystallite size of SiC in
HNLS. The crystallite size of β-SiC in TySA fibers appears to be little dependent on the heat
treatment temperature. The crystallite sizes for as-received HNL, HNLS and TySA fibers are
4.0 nm, 11.4 nm, 22.7 nm, while they are 15.5, 30.3, and 25.5 nm for fibers heat treated at
1900 °C for 1 h, respectively.
4.0 50.0
◆・◇ HNL
■・□ HNLS
Tensile strength, σ/GPa
▲・△ TySA 40.0

Crystallite size, nm
3.0
30.0
2.0
20.0
1.0
10.0
0.0 0.0
1000 1200 1400 1600 1800 2000
Initial
Annealing temperature, ºC
Reprinted by permission of Elsevier from [69].
Figure 19. Tensile strength and its relation to the crystallite size of SiC fibers heat treated at elevated
temperatures in Ar for 1 h.
The grain coarsening could be attributed to the coalescence of β-SiC nanocrystals due to
either decomposition of amorphous phase or diffusion of Si and C atoms at grain boundaries
during exposure at high temperatures. For the bulk materials with clean grain boundaries, the
grain growth proceeds through large grains incorporating the small one by grain boundary
diffusion. Especially, for grains with small size, the grain boundary diffusion operates much
more readily. As observed in two Nippon Carbon fibers, the grain coarsening is more obvious
in heat treated state than those of as-received state (TEM observations have revealed that
grain size is about 5 nm for HNL fiber [39], 20 nm for HNLS fiber [40], 200 nm for TySA
fiber [41]). On the other hand, the residual trace oxygen may play a role in the Si and C grain
boundary transport by accelerating diffusion [39], because the oxygen is not necessarily
eliminated from the fiber as reported in the literature [42], even for the HNLS fiber which
was fabricated at very high temperature.
Considering the starting temperature for grain coarsening in Figure 19, the grain size
might be related primarily to the maximum temperature at which the fibers were fabricated.
The fabrication temperatures have been presented for two Nippon Carbon fibers (HNL:1350
ºC, HNLS: 1600 ºC) and for TySA fiber (about 1800 ºC) in the literature [43]. From the
Figure 19, it can be seen that the crystallite size increased when heat treatment temperature is
above the fabrication temperature as expected. For HNL fiber, the grain coarsening occurred
at relatively low temperature is due to the decomposition of amorphous phase at about 1300
ºC. On the other hand, the thermally activated diffusion play an important role on the grain
coarsening of SiC materials at high temperatures.
If we make a further comparison in crystallite size between two Nippon Carbon fibers
again, we can see that a large difference in crystallite size was observed for two fibers het
treated at same temperature. As above mentioned, the HNL fiber has an small starting grain
size, which was expected to have a high diffusivity at grain boundaries and result in a large
grain size as heat treating at high temperatures. However, an unexpected phenomenon was
observed between two Nippon Carbon fibers. This can be attributed to the excess carbon in
HNL fiber. TEM observation revealed that heat treatment of the HNL fiber results in a
gradual organization of the free carbon phase in terms of the size of the carbon layer and the
number of stacked layers as increasing temperatures [39]. Takeda et al. [44] have investigated
the properties of polycarbosilen-derived silicon carbide fibers with various C/Si
compositions, and revealed that microstructure and mechanical properties are quite dependent
on the C/Si composition. Grain growth is suppressed with increase in excess carbon. In other
studies [45-46], the carbon suppressing growth and coalescence of the SiC microcrystals was
also observed. Sasaki [46] found that carbon disappeared above 1500 ºC heat treatment in SiC
fiber using Raman study. And then an abrupt increase of crystal size at 1500 ºC was observed.
The grain growth has a significant effect on the strength of SiC-based fibers. For the
HNL fiber, crystallization degraded its strength at all HTT. In both near stoichiometric fibers,
strength degradations occurred at the temperatures where crystallite size began to increase.
Fibers with larger grain size generally have relatively lower strengths, but it should be noted
that HNL fiber showed more rapid strength degradation than HNLS fibers above 1400 ºC heat
treatment as shown in Figure 19. HNL fiber has smaller crystal sizes comparing to that of
HNLS fiber. The growth of SiC crystals reduces the bonding forces at the grain boundaries.
Since the manufactures are always seeking the optimal fabrication temperature at which the
superior thermal stability and excellent mechanical strength can be obtained simultaneously,
thus, the upper fabrication temperatures are typically fixed by those temperature conditions
26 Jianjun Sha
above which performance degradation of the fibers occurred. The dependence of strength on
temperature in Figure 19 is in agreement with those of previous studies [18,39-40,46].
Ichikawa et al. [40] reported that HNLS was quite stable chemically after one hour exposure
in an argon gas at 1800 ºC, since no structural decomposition occurred and it exhibited a good
strength of 1.9 GPa. TEM observation shows that this heat treated HNLS fiber has a SiC
grain size of approximately 200 nm, which is about 10 times larger than that of the as-
received fiber.
As for TySA fiber, this is a sintering fiber, which is prepared by the reaction of a
polycarbosilane (PCS) with aluminium acetylacetonate, and subsequently converted into the
Tyranno SA fiber, by decomposition with an evolution of CO and SiO (1500 ºC <T<1700 ºC)
and sintering (about 1800 ºC). TySA fiber retained most of its initial strength, because no
significant grain coarsening was observed even heat treated at 1900 ºC. Excellent strength
retention has been observed in a former work [18].
Fiber strength is controlled not only by grain size, but also by critical flaw size and the
residual stresses etc, which were generated from fabrication process and gas evolution during
heat treatment at elevated temperatures. The mismatch in thermal expansion coefficient
between excess carbon and SiC grain, could cause strength loss, and the contribution of
residual stresses from the gas evolution to strength loss could increase with increasing the β-
SiC grain size.
On the other hand, the HNL fiber has smaller crystal size comparing to that of HNLS
fiber, but it showed more rapid strength degradation than HNLS fiber above 1400 ºC heat
treatment as shown in Figure 19; both HNLS and TySA fiber have near-stoichiometric
composition and high-crystallite structure, but they showed different strength retention. This
observed phenomenon implied that other mechanisms must be responsible for strength
degradation of SiC fibers besides the coarsening of crystallite size.
One of the most possible reasons for low strength retention of HNL fibers is residual
stresses which were generated from phase transformation and the mismatch in the coefficient
of thermal expansion between excess carbon and SiC grain. Sacks [47] produced a laboratory
fiber (UF fiber) with the similar composition as HNL fiber. There was no loss in strength with
heat treatments up to 1700 ºC and then the strength decreased rapidly with further heat
treatments up to 1900 ºC. He believed that strength is controlled by the residual tensile
stresses which developed as a result of the mismatch in the coefficient of thermal expansion
between SiC and C. This situation should be true. The coefficient of thermal expansion of
carbon/graphite (2.0-3.0×10-6/K) is less than SiC (3.9-4.0×10-6/K). When fibers were cooled
from high heat treatment temperature to room temperature, the SiC grains want to contract,
while carbon grain will resist their contraction. This action-reaction will put SiC in tension
and carbon in compression. This residual tension stresses could have a contribution to the
total stress loss. This case can be applied to each fiber type which contains excess carbon, but
here it should be more significant in HNL fiber because of high excess carbon (C/Si=1.38). In
both HNL fiber and TySA fiber, the size of carbon grain increased with increasing heat
treatment temperature [39,48-49]. For the TySA fiber, this fiber originally has a very large
crystallite size. In previous studies [42,48-49], a carbon-rich core was revealed in TySA fiber,
which results from the production process. Colomban et. al. estimated carbon grain size and
SiC grain size in TySA fiber from Raman spectroscopy [48-49]. The Carbon grains appear
approximately 2-3 times smaller on the fiber’s core (0.9-1.7 nm) than on its periphery (1.7-2.6
nm). The grain size of SiC in fiber core is much smaller than edge region. Likely, this is due
to that carbon suppressed the grain growth of β-SiC. Furthermore, this fiber contains the
small amount of alumina (less than 1 wt %) as sintering additive, which will also inhibit the
grain growth of SiC. As a result, the TySA fiber showed an excellent thermal stability in Ar
atmosphere. The growth of the carbon corresponds to a decrease of localized spin centers
[50]. The growth of the carbon grain might result in an increase of residual stress, however,
this evidence is insufficient because the magnitude of residual stresses is strongly dependent
on the volume fraction of carbon phase in a bulk material.
4.1.2. Microstructure
Figure 20 shows SEM morphologies of the fibers after heat treatment at high
temperatures in Ar for 1 h. The HNL fibers heat treated at temperatures below 1400 ºC had a
smooth surface, which is almost no difference from that of as received fibers. Heat treatment
at 1400 ºC caused slight coarsening of fiber surface. Obvious changes in appearance were
observed for the HNL fibers heat treated at 1780 ºC as shown in Figure 20 (a). The fibers
showed a porous microstructure and large grains deposition on the fiber surface (Figure 20
(a)). Such large grains are not observed within bulk of the fiber, due to the presence of free
carbon which inhibit the grain boundary or/and gaseous diffusion. For the HNLS fibers heat
treated below 1600 ºC, their microstructure did not vary compared to the as-received fibers.
After heat treated at 1600 °C, although the individual SiC grain grown on the fiber surface,
but fiber surface still remained smooth and it appeared no structure degradation. The fiber
heat treated at 1780 °C exhibited a rough surface with deposition of bulk SiC grains, but it
still remained a relatively dense structure as shown in Figure 20 (b). TySA fibers showed
outstanding thermal stability in microstructure comparing with other SiC fibers and didn’t
exhibit obvious structure damage even heat treated at 1900 ºC shown in Figure 20 (c).
(a):1780 °C (b): 1780 ° C (c): 1900° C
1 um 1 um
1um
1600 ºC 1600 ºC 1900 ºC
Figure 20. SEM photographs for SiC fibers heat treated in Ar for 1 h: (a) HNL fibers at 1780 ºC, (b)
HNLS fiber at 1780 ºC, (c) TySA fibers at at 1900 ºC; typical fracture surface observation for: (d) HNL
fiber at 1600 ºC, (e) HNLS fiber at 1600 ºC, (f) TySA fiber at 1900 ºC.
Obvious differences were found in subsequent observations of fracture surfaces.

The　fracture of as-received HNL and HNLS fibers mainly originated from the inner critical
flaw (inclusion-type or inner pore-type critical flaw). After heat treatment at 1600 ºC, most of
the examined HNL and HNLS fibers fractured at surface flaw as shown in Figure 20 (d)-(e),
and flaw size slightly increased with increasing heat treatment temperature. The critical flaw
size and mirror size were measured. The critical flaw sizes (rc) are: 0.90 μm for as received
HNL fiber, 1.07 μm for 1600 °C heat treated HNL fiber. In the case of HNLS fiber, the
critical flaw sizes (rc) are: 0.84 μm for as-received fibers, 0.90 μm for 1600 °C heat treated
28 Jianjun Sha
fibers. Figure 20 (f) showed that fracture surface of the TySA fibers after annealing at 1900
ºC did not reveal obvious difference in the fracture mode comparing with the as-received
fibers. The fracture origin and mirror zone are invisible on the fracture surface and fracture
surface showed a trans-crystallite fracture behavior. The trans-crystallite fracture behavior
could be partially related to a high compression residual stresses in SiC caused by addition of
alumina in this fiber. Existence of compression residual stresses in the grain boundary of
TySA fiber is quite possible because of significant mismatch in the coefficient of thermal
expansion between SiC and Alumina (SiC: 3.9-4.0×10-6/K; Alumina: 8.0-9.0×10-6/K) and
high sintering temperature (higher than 1800 ºC). In TySA fibers, the change in the extension
stability of micro-crack in the compression residual stress filed might improve the grain
boundary strength. The increase in grain boundary strength could explain the trans-crystalline
fracture behavior of TySA fiber.
Linking the tensile strength data in Figure 19 with the microstructure examination in
Figure 20 again, the decomposition of amorphous phase, grain coarsening and residual stress
at high temperatures in HNL fiber could be responsible for strength and microstructure
degradation. Above 1600 ºC, the outward growth of huge grains was observed as shown in
Figure 20 (a) and (b), and these huge grains might act as the critical flaw during the fracture
of fiber. Observation of surface morphologies (Figure 20 (a)-(c)) and fracture surface (Figure
20 (d)-(f)) provided some information for the strength degradation of SiC fibers. Due to near-
stoichiometric composition in HNLS fiber, it’s damage was limited on the surface of fiber.
The formation of porous structure in HNL fiber could be attributed to the rapid evolution
of gases at the earlier stage of high-temperature exposure according reaction (17).
SiCx Oy → SiC ( s) + C ( s) + SiO( g ) + CO( g ) (17)
It should be noted that large grains grown outward from the surface of fibers at 1780 ºC
appear to be β-SiC crystals, which were produced by following gas-phase reactions [51-52].
SiO ( g ) + 3CO ( g ) → SiC ( s ) + 2CO2 ( g ) (18)
SiO ( g ) + 2C ( s ) → SiC ( s ) + CO ( g ) (19)
Reaction (19) could occur because of presence of free carbon in surface and body of
HNL fibers [20]. According to above result, the reaction (17)-(19) are quite dependent on the
quantity of amorphous phase and content of carbon in SiC fibers. The use of graphite crucible
in this work might also cause the reaction:
CO2 ( g ) + 2C ( s ) → 2CO( g ) (20)
Combing the reaction (18) and (20) indicating the gas-phase reaction proceeded mainly
by reaction (19). Additionally, the CO-CO2 gas mixture might modify the microstructure of
SiC fibers at high temperatures [53].
Considering the surface degradation of HNLS fibers heat treated above 1600 ºC (Figure
20 (b)), as we know, the quantity of amorphous phase and oxygen content should be very
small in this fiber. Thus, thermal decomposition of the amorphous phase is almost negligible
in this fiber. As for the large grains deposited on the surface of HNLS fiber, it can be
explained by reactions (18)-(20), because the carbon layer on the surface of HNLS fiber (80
nm) is thicker than that of HNL fiber (20 nm) [42]. Concerning the origin of gas species,
other mechanism could be responsible for it. Active oxidation is quite possible. The transition
from passive oxidation to active oxidation occurs at oxygen partial pressure of 10-25 Pa and
1-2.5 Pa at 1500 ºC for HNL and HNLS fiber [52], respectively. At same oxygen partial
pressure level, the increase of heat treatment temperature will accelerate the transition from
passive-to-active oxidation [51].
For the TySA fiber, the excellent microstructure stability could be attributed to the high
processing temperature (over than 1800 °C) and addition of alumina [18]. The small amount
of alumina addition could inhibit the grain growth and enhanced the oxidation resistance. This
higher stability can also be linked to the silica protective layer formed on the surface of fiber
[48,54].
Combining the fracture properties with microstructure characterization of SiC-based
fibers, we can not deny the existence of other degradation mechanisms such as contaminants
during heat treatment and metallic impurities introduced during process [55-57]. The
existence of metallic impurities within the fibers is possible, because all these fibers are
polymer derived. The metallic impurities can easily enter the fibers during the various steps
of polymer handling and can cause rapid or abnormal grain growth in local areas. There are at
least two indirect observations supporting above mentioned mechanism: (i) Observation of
fracture surface in Figure 20 for the HNL and HNLS fiber showed that the strength-limiting
flaws after heat treatment are larger than the average grain size, indicating rapid defect
growth in selected areas of the fiber and thus suggesting the possible existence of metallic
impurites; (ii) the UF fiber showed high strength retention than HNL fiber [47]. This suggests
that the UF fiber during processing did not introduce the metallic impurities to the degree that
employed for the HNL fiber.
4.1.3. BSR Creep Resistance
Figure 21 shows dependence of 1-h BSR creep resistance m on HTT, which was tested at
1400 ºC. Heat treatments of the fibers above the processing temperature resulted in improved
creep resistance as shown in Figure 21. The creep resistance of heat treated HNL fiber above
1400 ºC was significantly improved. Likely, this could be attributed to the increased grain
sizes, high crystallization of β-SiC. Such microstructural changes are expected to inhibit
diffusion-controlled creep processes. For the 1600 ºC heat treated HNL fiber, the creep
resistance was better than those of as-received near stoichiometric fibers although the fact that
the grain sizes were much smaller than those of the latter fibers. This result indicated that the
improved creep resistance depended on not only the crystallization and grain growth, but also
the composition at grain boundaries. The excess carbon distributed at the grain boundary of
the HNL fiber inhibits the coalescence of β-SiC, which results in a stable grain boundary
structure. This implies that stability of Grain boundaries (GB) plays an important role on the
creep resistance of SiC fiber. This assumption was also demonstrated by TySA fiber. The
enhanced creep resistance of the TySA fiber was obtained prior to increase its crystallite size.
As a result of Al addition to TySA fiber, the complex oxide would be formed at GB by heat
treatment and they can stabilize the grain boundary to improve the creep resistance. The
stability of GB could be affected by GB composition.
30 Jianjun Sha
1.0 50.0
◆・◇ HNL
■・□ HNLS
Stress relaxation ratio, m

0.8 ▲・△ TySA 40.0
Cristallite size [nm]

0.6 30.0
0.4 20.0
0.2 10.0
0.0 0.0
Initial 1200 1400 1600 1800 2000
1000
Heat Treatment Temp. [oC]
Figure 21. 1-h stress relaxation ratio and crystallite size for heat treated HNL, HNLS and TySA fibers,
creep tests were performed at 1400 ºC in Ar.
4.1.4. Fracture Toughness and Critical Fracture Energy
Based on the principles described in section 3.6, the fracture toughness and critical
fracture energy was calculated for the heat treated fibers, the resultant value of fracture
toughness was shown in Figure 22. The fracture toughness decreased with increasing the heat
3 5
HNL
HNLS Fracture energy γc (J/m2)
K1c (MPa.m1/2)
2.5 4
2 3
1.5 2
1 1
0.5 0
As-received
950 1150 1350 1550 1750
Heat treatment Temperature (℃)
Figure 22. Dependence of fracture toughness and critical fracture energy on heat treatment
temperatures.
treatment temperature, but it did not show strong dependence on the heat treatment
temperature. For the as-received fibers, the carbon layer covered on the surface of fibers, can
blunt the critical flaw and reduce the stress concentration on the surface flaw. However, this
carbon layer can be removed by reaction with residual oxygen from fiber itself and
atmosphere. In this case, the propagation of preexisting surface flaws will become easy. In
addition, flaws produced by decomposition, active oxidation and large grain deposition can
exist on the fiber’s surface at high temperature resulting in the low fracture toughness and
critical fracture energy. At fairly high strain rate (0.3 mm/min) at which the strengths were
measured, these flaws would propagate gradually until they become critical because of stress
concentration around the flaws.
4.2. Annealing and Creep in Various Oxygen Partial Pressures
SiC fibers with high thermal stability are considering as the promising reinforcement in
CMCs. However, the mechanical and thermal stabilities of SiC fibers as reinforcements in
CMCs are very sensitive to their purity, crystallinity and service environments [58-60]
including thermal and loading history.
For high temperature applications, the CMCs are often subjected to oxidative
environments with different oxygen partial pressures. In such case, the SiC materials would
be oxidized in passive/active oxidation regimes [51]. As we know, the performance
degradation of SiC materials in oxidizing environments strongly depends on the oxidation
mechanism. Jacobson [61] has discussed the oxidation degradation mechanism of SiC
materials in varied environments, but it is still insufficient because of the complexity of
service environments. The key question concerns the oxidation kinetics: passive and active
oxidation. This topic has given rise to much controversy for SiC materials, because the
temperature boundaries for the oxidation kinetics are quite dependent not only on the
materials themselves (purity and crystallinity), but also on the specific service environment
(exposure temperature, oxygen partial pressure and mechanical state). Furthermore, rarely is
one mechanism operative in performance degradation of SiC materials. In practice, several
mechanisms operate simultaneously.
Therefore, for understanding the mechanical and thermal stabilities and failure
mechanism of SiC fibers over a wide range of temperatures and varied environments, the part
of this chapter reviewed the microstructure features and high temperature properties of SiC
fibers under annealing and creep in various oxygen partial pressures at elevated temperatures,
and attempted to clarify the correlation between the environment with mechanical and
thermal stabilities. Thus, SiC fibers were annealed and crept in air (O2: 20%, dew point: 3
ºC), high-purity Ar (HP-Ar, O2: 2 ppm, dew point: -5.5 °C) and ultra high-purity Ar (UHP-
Ar, O2: 0.1 ppb: dew point: -5.5 °C) under flowing atmosphere with a pressure of 105 Pa and
held for 1 hour at desired temperatures ranging from 1000 to 1500 °C. Furthermore, the
surface morphologies of fibers under annealing and creep were compared by the observation
of field-emission scanning electron microscope (FE-SEM).
32 Jianjun Sha
4.2.1. Morphologies of Fibers
(a) Under Annealing and Creep in Air
Figure 23 shows the morphologies of SiC fibers under annealing and creep at 1500 °C for
1 h. The SiO2 film formed uniformly on the surface of fibers during annealing at elevated
temperatures (Figures 23). From the cross-section of fibers under annealing at 1500 °C
(Figure 23 (c-1)-(c-3), we could see that the silica layer consisted of a concentric sheath and
the fiber’s surface was blanketed well by silica layer. And also, no further oxidation between
SiO2 film and SiC fiber surface was observed.
1500 ºC 1500 ºC 1500 ºC (c-1)

(a-1) (b-1)
1μm 1μm 1 um
1500 ºC (a-2) 1500 ºC (b-2) 1500 ºC (c-2)
1μm 1μm 1 um
1500 ºC (c-3)
1500 ºC (a-3) (b-3) 1500 ºC
1μm 1μm
Figure 23. Morphologies of SiC fibers annealed: ((a-1) and (c-1) for HNL; (a-2) and (c-2) for HNLS;
(a-3) and (c-3) for TySA fibers)) and crept ((b-1) for HNL; (b-2) for HNLS; (b-3) for TySA fibers) in
air at 1500 ºC.
For the HNL fibers under annealing (Figure 23 (a-1)) and creep (Figures 23 (b-1)) at
1500 °C, the cracks were found within silica film. Some patterns were also observed within
the silica film formed on the surface of HNL fibers (Figures 23 (b-1) under creep condition.
Because of the near stoichiometric composition, the HNLS (Figure 23 (a-2)-(c-2)) and
the TySA fiber (Figure 23 (a-3)-(c-3)) showed a relatively smooth surface compared with that
of HNL fiber. No significant cracks were found within silica layer even under crept condition.
A relatively rough surface of silica on the HNLS fiber crept at 1500 °C was observed (Figure
23(b-2)). In contrast, a smooth silica film coated on the surface of the TySA fiber (Figure
23(a-3)-(b-3)). Maeda et al. [62] have studied a sintered Al2O3-containing SiC for periods up
to 3000 h. The actual kinetics involved at least four different parabolic stages for this
material. They attributed these to the various microstructural changes in the scale:
crystallization of amorphous silica, transformation of those crystalline phase, and viscosity
changes in the oxide scale due to migration of the additives. This indicated the oxidation of
TySA fiber is much complex because of addition of alumina in this fiber.
(b) Under Annealing and Creep in HP-Ar
Figure 24 shows the morphologies of SiC fibers under annealing and creep at 1500 °C for
1 h.
1500 ºC (a-1) 1500 ºC (b-1)
Pit
1μm 1μm
1500 ºC (a-2) 1500 ºC (b-2)
1μm 1μm
1500 ºC (a-3) 1500 ºC (b-3)
1μm 1μm
Figure 24. Morphologies of SiC fibers annealed ((a-1) for HNL; (a-2) for HNLS; (a-3) for TySA fibers)
and crept ((b-1) for HNL; (b-2) for HNLS; (b-3) for TySA fibers) in HP-Ar at 1500 ºC.
34 Jianjun Sha
For the HNL fibers under annealing and creep at 1300 °C, fibers surface didn’t show
obvious oxygen attack. Under creep at 1400 °C, the bubbles were formed on the surface of
fiber, but it didn’t appear under annealing condition except for the slight grain growth. The
HNL fibers annealing at 1500 °C have coarse surface accompanying with the formation of
pits (Figure 24 (a-1)). In particular, for the fiber under creep at 1500 °C (Figure 24 (b-1)), it
gave a porous microstructure with many pits formation.
Below 1500 °C, annealed HNLS fiber displayed a fine-grained and pore-free structure,
namely, no obvious oxygen attack on surface was observed. At 1500 °C, few coarsening
grains were sitting on the annealed fiber surface (Figure 24 (a-2)), but the fiber surface is still
dense and smooth. Under creep condition, pits appeared on the surface of fibers (Figure 24
(b-2)). This result indicated that stress applied by BSR test could enhance the oxygen attack.
When TySA fibers were annealed in HP-Ar, a quite stable morphology was observed as
shown in Figure 24 (a-3). From the observation of cross section of TySA fibers annealed in
Ar at 1400 °C, a very thin silica film was formed on the surface. Under creep condition, the
grain of TySA fibers showed a little coarsening at 1500 °C. In a previous study, the TySA
fiber was oxidized in passive oxidation regime at 1500 °C even the oxygen partial pressure is
much low (1 Pa). The passive oxidation of TySA fiber at extremely low oxygen partial
pressure appears to be attributed to the addition of a minute of alumina. Alumina, as a
oxidation product of aluminum, reacts with the SiO2 film to form an alumino-silicates. At
1500 °C, the softening of the alumino-silicat film was observed because of low melting point.
The alumino-silicates have a high oxygen permeability, presumably enhancing the passive
oxidation of fibers [63-64]. On the other hand, SiO2 can also be formed when SiC was
annealed in low oxygen pressures by reaction of SiC with alumina [65]:
SiC+2/3Al2O3 = SiO2 + 4/3 Al + C (21)
This reaction is slightly, but such development of SiO2 film could suppress the active
oxidation of TySA fiber at extremely low oxygen partial pressure.
(c) Under Annealing and Creep in UHP-Ar
Figure 25 shows the morphologies of SiC fibers under annealing and creep at 1500 °C for 1 h.
No obvious oxidation was observed for HNL fibers under annealing at 1300 °C, but for
fibers under creep condition, bubbles and large grains were formed on the surface. At 1400
°C under annealing condition, many large pits was produced on the surface of fibers,
meanwhile, the tensile side of crept fiber showed the needle-like grains with a length of about
5 um [66]. In particular, for the Hi-Nicalon fibers under annealing and creep at 1500 °C, the
fibers were oxidized more severely and much coarse-grained surface was produced (Figure 25
(a-1) and Figure 25 (b-1)). It is clear these huge crystals grew outward from the fiber surface.
The HNLS fibers show a stable microstructure in annealed condition at temperature
below 1400 °C. The grain coarsening was observed under crept at 1400 °C. At 1500 °C, the
annealed fibers appeared the formation of pits (Figure 25 (a-2)), while in the crept fibers, the
fibers showed a porous structure (Figure 25 (b-2)).
1500 ºC (a-1) 1500 ºC (b-1)
1μm 1μm
1500 ºC (a-2) 1500 ºC (b-2)
1μm 1μm
1500 ºC (a-3) 1500 ºC (b-3)
1μm 1μm
Figure 25. Morphologies of SiC fibers annealed ((a-1) for HNL; (a-2) for HNLS; (a-3) for TySA fibers)
and crept ((b-1) for HNL; (b-2) for HNLS; (b-3) for TySA fibers) in UHP-Ar at 1500 ºC.
For TySA fiber under creep at 1500 °C (Figure 25 (b-3)), the surface was also damaged
slightly, but in annealed fibers no obvious change was observed (Figure 25 (a-3)).
As we know, SiC materials may be oxidized in passive and/or active oxidation regime,
depending on their microstructure, the oxygen partial pressure and exposure temperature [67-
70].
The thermochemical correlation and oxidation dynamics for active and passive oxidation
of silicon carbide have been investigated experimentally and theoretically in the literatures
[51,67-70]. These data were combined together and plotted into a new plot as shown in
Figure 26. The passive-to-active oxidation transition is strongly affected by factors such as
the type of silicon carbide, temperature and oxygen partial pressure. The oxygen partial
pressures are, 2.1×104 Pa in air, 0.2 Pa in HP-Ar and 1×10-5 Pa in UHP-Ar, respectively. The
36 Jianjun Sha
oxygen partial pressures and test temperatures for this work were shown in Figure 26 by three
lines. It is clear that proposed test conditions distributed in different regions.
Air
1.E+ 05
1.E+ 04
Oxygen partial pressure/ Pa

1.E+ 03 Nicalon
Passive oxidation
1.E+ 02
Hi-NIcalon
iC
DS
1.E+ 01 Hi-Nicalon S
CV Tyranno SA
e
1.E+ 00
in
rl
de
ei
1.E- 01
hn
re ive
Sc
HP-Ar
n
n ss
gio
1.E- 02
itio pa
Active oxidation
ns to
tra tive
1.E- 03 UHP-Ar
Ac
e
lin
1.E- 04
er
eid Passive oxidation
1.E- 05
hn
Active oxidation
Sc
1.E- 06
1 20 0 1 40 0 1 60 0 1 80 0 2 00 0
Temperature/K
Data from [51,67].
Figure 26. Oxygen partial pressure for the transition from passive to active oxidation at elevated
temperatures (at 1500 ºC, pO2 for the transition is: Niclaon: 100-250 Pa, HNL:10-25 Pa,
HNLS:1-2.5 Pa).
Based on the surface and cross section morphologies in Figure 23, it is thought that fibers
were mainly oxidized in passive oxidation regime characterized by the formation of silica
film when they exposed in air at high temperatures, but the passive oxidation was enhanced
under crept condition due to increased oxygen permeation. Cracks in the silica layer observed
in Figure 23 are mainly due to the difference in coefficient of thermal expansion (CTE)
between the fiber core and the silica layer. Because the CTE of SiC fiber is less than that of
Silica, on cooling, a tensile residual stress will be applied to silica layer. On the other hand,
beta SiO2 transforms into alpha SiO2 at 300-370 °C with an accompanying volume change
[71], can also generate the stress in silica layer resulting in the formation of cracks. The
passive oxidation formed SiO2 layer can refrain the further oxidation of SiC.
In low oxygen partial pressure atmosphere, the oxygen partial pressure for the transition
from passive to active oxidation is a key point in the microstructure change of SiC materials.
A previous study has found that the oxygen partial pressure for the transition from passive
oxidation to active oxidation was pO2 = 10-1∼ 2 Pa at 1300 ºC [51]. The pO2 value was about
0.2 Pa (HP-Ar) and 10-5 Pa (UHP-Ar). Hence, these oxygen partial pressures are well below
or fall into the range of 10-1∼ 2 Pa, indicating that the occurrence of active oxidation in the
proposed conditions is possible.
From the observation of surface morphologies, we can see the active oxidation initialized
at different temperatures in different atmospheres [66]. The temperatures for active oxidation
in crept fibers, however, were shift to low values, indicating the transition from passive
oxidation to active oxidation was enhanced under creep condition. The enhanced active
oxidation in creep condition might be caused by the rupture of silica scale on the surface of
fiber due to the stress applied by BSR test. Subsequently, a stress concentration would occur
around the flaws generated by active oxidation or gas evolution, and then oxygen attack on
the SiC fibers will be accelerated, leading to the formation of bubbles and pits. The formation
of bubbles in the silica scale may provide some indications of pressure buildup [72],
especially, when the passive and active oxidation proceeded concurrently.
Due to the near stoichiometric composition and high crystallinity of HNLS and TySA
fibers, their active oxidation was gradual in comparison with other SiC fibers [58-60,67,72].
It is obvious that HNLS fibers crept in HP-Ar and annealed and crept in UHP-Ar at 1500 °C
were slightly oxidized in active oxidation regime. Noteworthy is that the active oxidation is
infinite slow if the oxygen partial pressure is very low [73].
4.2.2. Tensile Properties
Figure 27 shows the dependence of mean strength on the testing environments. The
fiber’s strength decreased with decreasing the oxygen partial pressure. It should be noted
during the specimen preparation that fibers with low strength became very difficult to set
without breaking them. The mean strength we gave will consequently not take the weakest
fibers into account (no strength could be obtained). Due to this shortcoming, overestimation
of tensile strength is likely. As observed for HNL fibers after annealing in UHP-Ar at 1500
°C for 1 h, the fibers are too fragile to measure the strength. However, still an enough strength
for HNLS and TySA fibers was retained even after annealing in UHP-Ar as shown in Figure
27.
1.2 A r-rece ived

A ir
H P -A r
1
U H P -A r
0.8
0.6
0.4
0.2
0
H N1 L H 2N LS T3yS A
Figure 27. The tensile strength retention of SiC-based fibers annealed at 1500 ºC under varied
atmosphere.
38 Jianjun Sha
The strength degradation of fibers in different oxygen partial pressures could be

attributed to different microstructure and oxidative mechanism. The strength for the fiber
annealed in air decreased only slightly (Figure 27). This is due to the SiO2 layer, which acts as
environmental barrier coating, stopping the further oxygen attack.
The fracture origins on the fracture surface of fibers were shown in Figure 28. The
fracture mainly originated from surface flaws after annealing in different atmosphere. This
means the surface of fibers was damaged by oxidation. The relative low strength retention for
fiber annealed in UHP-Ar suggested changes in flaw population and flaw size. The low
oxygen partial pressure accelerates the transition from passive to active oxidation resulting in
a coarsening and pitting surface. At high temperature under low oxygen partial atmosphere,
SiC is relatively easy to be oxidized actively.
HNL (a-1) HNL (b-1) HNL (c-1)
Porous structure
Surface defect
Inner pore/ inclusion
Huge grain/defect
1um 1 m 1 m
HNLS (a-2) HNLS (b-2) HNLS (c-2)
1 um 1 um 1 um
TySA (a-3) TySA (b-3) TySA (c-3)
Surface defect
1 m 1μm 1μm
Figure 28. Fracture surface of fibers after annealing at 1500 ºC: (a-1)-(a-3) in air, (b-1)-(b-3) in HP-Ar,
(c-1)-(c-3) in UHP-Ar.
4.2.3. Creep Resistance
Figure 29 shows the dependence of 1-h BSR creep resistance m on temperatures under
various oxygen partial pressures for HNLS fibers. HNLS fiber exhibited excellent creep
resistance even exposed at 1400 °C. At temperatures above the 1300 °C, the BSR
examinations indicate that creep resistance of HNLS fibers under high oxygen partial pressure
is somewhat lower than that in low oxygen partial pressure. Namely, a weak dependence of
BSR creep resistance on oxygen partial pressure is observed.
Test temperature/K
1773 1673 1573 1473 1273
1
Stress relaxation parameter m
0 .9 HNLS
0 .8
0 .7
0 .6
0 .5
0 .4
0 .3 Air
0 .2 HP-Ar
0 .1 UHP-Ar
0
0 .5 0 .6 0 .7 0 .8
Reciprocal temperature/1000(K-1)
Figure 29. Effect of oxygen partial pressures on the BSR creep resistance of HNLS fiber.
As for the creep behavior of SiC fibers, generally, it can be explained by their oxygen
content, grain size and second phase in the grain boundaries. The creep resistance of SiC
fibers slightly decreased with increasing the oxygen partial pressure, which is likely
controlled by the oxidation and grain coarsening. Under creep condition, grain coarsening of
SiC has been observed for longer time creep experiments, which may contribute to a
decelerating creep rate [74-75]. Especially, grain coarsening is relative easy in low oxygen
partial pressure and at high temperature due to coalescence of β SiC grain [73]. The grain size
increased with decreasing oxygen partial pressure has been observed in other studies [59,67].
Furthermore, under the creep test, the grain coarsening could be accelerated by applied stress
[59]. The somewhat high creep resistance in low oxygen partial pressures might be attributed
partially to the concurrent grain coarsening during BSR test. On the other hand, the low creep
resistance in air could be partially explained by the resistance effect of silica layer. The
outmost SiO2 sheath formed in air after BSR test will counteract part of the initial applied
stress leading to a low stress relaxation parameter m, especially, for the fibers with fine
diameters.
40 Jianjun Sha
4.3. Thermal Exposure Under Loading
Most previous studies have been concerned with the degradation in mechanical properties of
SiC fibers after high temperature exposures [58,76-78]. There exist few studies on the
performance change of SiC fibers in low oxygen partial pressure atmosphere under loading [79].
Therefore, for understanding the degradation mechanism of CMCs under loading in low
oxygen partial pressure environments, the loading tests were performed on SiC fiber yarns by
applying different dead loads at elevated temperatures in Ar atmosphere. After each loading
test, the room temperature tensile properties and microstructure were characterized to clarify
the performance degradation mechanism.
4.3.1. Tensile Properties
Figure 30 showed the room temperature tensile strength distributions of SiC fibers with
different conditions. The details of tensile properties for each fiber type was indicated as
follows: (i) The effect of applied load on the strength degradation of HNL fiber was observed at
1250 °C; both the retained strength and Weibull modulus decreased with increasing applied
load in HNL fiber, but the effect of the applied load at 1250 °C was much smaller in HNLS
fiber (Figure 30 (a) and Figure 30 (e)). (ii) In the strength distributions of fibers after yarn-
loading test at 1250 °C and 1300 °C for 3 hours under an applied load of 201 g, the effect of
exposure temperature on the strength degradation is obvious at same applied load of 201 g as
shown in Figure 30 (b) and Figure 30 (e). (iii) A weak time effect on the tensile properties was
observed within the limited time difference employed here (Figure 30 (c) and Figure 30 (e)).
Likely, the time difference between 1h and 3h is too short. (iv) The strength retention and
Weibull modulus decreased in both HNL and HNLS fibers under combined effect of applied
load and exposure temperature as shown in Figure 30 (d) and Figure 30 (e). The combined
effect of applied load and exposure temperature is more obvious in properties degradation of
SiC fibers. Also, from the comparison of strengths between HNL and HNLS fibers after yarn-
loading tests, the observed strength decreases are greater in HNL fibers. The near-stoichiometric
and high-crystallite SiC fiber, HNLS, degraded gradually in tensile properties with increasing
load and temperature, as shown in Figure 30 (e). This phenomenon seems to be related to the
oxidation resistance and thermal stability of SiC fibers under loading.
2 2
As-received HNL
As-received HNL
1250C/90g/3h HNL (a) 1250C/201g/3h HNL (b)
1 1250C/201g/3h HNL
1 1300C/201g/3h HNL
As-received HNLS As-received HNLS
ln(ln(1/(1-Fi)
ln(ln(1/(1-Fi)
0 1250C/90g/3h HNLS 0 1250C/201g/3h HNLS

1250C/201g/3h HNLS 1300C/201g/3h HNLS
-1 -1
-2 -2
-3 -3
-4
-4
-2 -1 0 1 2
-2 -1 0 1 2
lnσi (GPa) lnσi (GPa)
Figure 30. Continued on next page.

2 2 As-received HNL
As-received HNL
1300C/201g/1h HNL (c) 1300C/annealed/1h HNL (d)
1 1300C/201g/3h HNL 1 1250C/90g/3h HNL
As-received HNLS 1300C/201g/3h HNL
ln(ln(1/(1-Fi)
ln(ln(1/(1-Fi) As-received HNLS
0 1300C/201g/1h HNLS 0
1300C/201g/3h HNLS 1300C/annealed/1h HNLS
1250C/90g/3h HNLS
-1 -1 1300C/201g/3h HNLS
-2 -2
-3 -3
-4 -4
-2 -1 0 1 2 -2 -1 0 1 2
lnσi (GPa) lnσi (GPa)
5
4.5
HNL: Mean strength (GPa)
HNLS: Mean strength (GPa) (e)
4 HNL: Weibull modulus
HNLS: Weibull modulus
3.5
3
2.5
2
1.5
1
0.5
0
As-received
1 1300C/annealed/1h
2 1250C/90g/3h
3 1250C/201g/3h
4 1300C/201g/1h
5 1300C/201g/3h
6
Figure 30. Room temperature tensile strength distributions for HNL and HNLS fibers after loading test
with different conditions: (a) effect of loading, (b) effect of exposure temperature, (c) effect of exposure
time, (d) combined effect of exposure temperature and loading, (e) tensile properties of SiC fibers with
different conditions.
4.3.2. Morphology
Figure 31 showed the microstructure features of as-received and annealed HNL fibers
at 1300 °C for 1 h. No obvious differences in the surface morphologies were observed
between the as-received and annealed fibers (Figures 31 (a) and (c)). For the as-received
HNL fiber (Figure 31 (b)), the inner pore acted as the critical flaw during the tensile test (In
the examined HNL fibers, about 74% of critical flaws were identified as inner flaws with
different dimensions). But for the annealed HNL fibers (Figure 31 (d)), their fractures
mainly originated from a surface flaw (about 78% of total fractures originated from the
surface).
42 Jianjun Sha
(a) (c)
5 um
5 um
(b) (d)
Surface flaw
Inner pore
2 um 2 um
Figure 31. SEM images of HNL fibers: (a) Surface morphology and (b) fracture surface of as-received
fibers; (c) surface morphology and (d) fracture surface of annealed fibers at 1300 °C for 1 h.
Figure 32 showed the surface morphologies and fractographs of fibers tested in 1300
ºC/201 g/3h. A quite different morphology is apparent for these fibers. A rough surface with
extensive grain growth and micro pores was observed in HNL fibers as shown in Fig 32 (a).
According to the observation of the fracture surface (Figure 32 (b)), the HNL fibers fractured
at an irregular groove, which extended from surface to interior and was significantly different
from that of annealed fibers in shape and size (Figure 31 (d)). There are only a few individual
large grains grown on the surface of HNLS fibers (Figure 32 (c)), and the fracture of the
HNLS fiber also originated from the irregular surface flaw (Figure 32 (d)), but the flaw size is
smaller than that of HNL fiber.
In both HNL and HNLS fiber, the critical flaw was surrounded by a distinctive mirror
zone and hackle zone.
In Figure 30, the tensile properties of SiC fibers decreased with increasing load and
temperature, which suggested changes in flaw population and flaw size, as observed from
surface morphologies and fractographs (Figure 31~Figure 32).
(a) (c) Large grain
pore
Extensive grain growth
5 um 5 um
(b) (d)
Critical flaw Critical flaw
1 um 1 um
Figure 32. SEM images of SiC fibers after loading test in 1300 °C/201 g/3 h: (a) Surface morphology
and (b) fracture surface of HNL fiber; (c) Surface morphology and (d) fracture surface of HNLS fiber.
Fracture of HNL and HNLS fibers under loading mainly originated from an irregular
surface flaw, and features are associated with brittle failure (Figure 32 (b) and Figure 32 (d)).
By comparing the microstructure of the loaded fibers with that of the as-received and the
annealed fibers (Figure 31), it is obvious that the fiber’s surface was attacked by enhanced
oxidation under loading (Figure 32). The irregular groove could be a stress corrosion crack
(SCC) caused by combination of the oxidation and the applied load.
Generally, the oxidation behavior of silicon carbide at a high temperature depends on
ambient oxygen partial pressure. Passive oxidation occurs at high oxygen partial pressures
and results from the formation of SiO2 that grows on the surface of exposed fibers. Passive
oxidation could protect the materials from further oxidation attack. For active oxidation, pits
and cracks occurred at low concentrations of O2 where the SiO2 formation rate is too low to
seal the surface of materials. The passive-to-active oxidation transition is strongly affected by
factors such as the type of silicon carbide, total gas pressure and gas flow rate, as well as
temperature and oxygen partial pressure [51,70].
Thus, based on the surface morphologies and fractographs in Figure 31, it was thought
that annealed fibers were mainly oxidized in passive oxidation mode characterized by
formation of silica. The passive oxidation formed a thin silica film on the surface of annealed
44 Jianjun Sha
fibers and this thin silica film could protect the fiber from further oxidation attack resulting in
no critical structural change. In this case, the thermal decomposition of amorphous phase
which proceeded by reaction (17) should also be suppressed [67,80-81], because the gas
species cannot be removed fast enough through the silica film by diffusion or migration
[67,70].
Under loading conditions, the rupture of silica scale on the surface of fiber would be
easier and a stress concentration occurs around the flaws. At that time, the oxidation was
enhanced by loading, leading to the stress corrosion crack, as observed in Figure 32 (b) and
Figure 32 (d). In this case, the active oxidation for SiC material could occur by following
reactions [51]:
SiC(s)+O2(g) ⇋ SiO(g)+CO(g) (22)
SiC+3/2O2(g) ⇋ SiO(g)+CO2(g) (23)
The oxygen partial pressure for the transition from passive oxidation to active oxidation
was pO2 = 10-1∼ 2 Pa at 1300 ºC and total pressure of 100-800 Pa [51]. The total pressure in
this study was 105 Pa; hence, the pO2 value in furnace chamber was about 0.2 Pa (oxygen
concentration in Ar: 2 ppm). This oxygen partial pressure falls into the range of 10-1∼ 2 Pa,
indicating that the occurrence of active oxidation is quite possible. A high total pressure will
increase the limits for the transition from passive-to- active oxidation [51].
In addition to the HNL fibers, the oxidation of free carbon and the decomposition of
amorphous intergranular phase by reaction (17) also yield a damaged structure [67,80], which
accelerated the active oxidation because of the fiber’s high permeability to oxygen. On the
other hand, due to the acceleration of C/Si diffusivity at grain boundaries and reaction
between SiO and free carbon, the grain growth of HNL fibers was enhanced by applied stress
as shown in Figure 32 (a). The grain growth also has a contribution to the strength
degradation [69]. Because of the near stoichiometric composition and high crystallinity of
HNLS fibers, its active oxidation was gradual in comparing to that of HNL fiber.
In previous studies [26,36,58-59,76-78], many researchers have observed that the
strength of ceramic fibers was associated with critical flaws. As observed in Figure 32, a clear
mirror zone around the critical flaw on the fracture surface corresponds to the smooth
propagation of a crack. In the present work, because most of the critical flaws are irregular in
shape, it is very difficult to define the flaw size.
Thus, for clarifying the degradation mechanism, the fracture mirror size on the fracture
surface was measured. There is a tendency that mirror size increased with increasing load and
temperature. For fibers tested in 1250 ºC/ 90 g/3 h and 1300 ºC /201 g/3 h, the measured
mean mirror sizes on their fracture surfaces were about 3.37 and 3.67 μm for HNL fiber, 2.97
and 3.34 μm for HNLS fiber, respectively.
The increased mirror size could be related to the increased critical flaw size. Critical flaw
size could be associated with the oxidation and creep resistance of SiC fibers under loading.
The creep resistance evaluated by bending stress relaxation method has indicated that applied
stress of HNL fiber by bending at 1300 ºC was nearly complete relaxation [69]. In contrast,
the HNLS fibers showed excellent creep resistance at 1300 ºC. Having a low oxidation and
creep resistance could easily cause the new flaw nucleation and growth and slow crack
propagation. Meanwhile, the excess carbon and amorphous phase in HNL fiber might also
have a contribution to the flaw population and flaw size. Finally, the strength of fibers was
limited by the degradation of fiber’s microstructure, which could be attributed to the stress
corrosion caused by oxidation and loading. For easily understanding the oxidation corrosion,
following mechanism was proposed. For annealed fibers, a thin silica film formed on the
surface of fibers and prevented the gas species passing through the fiber surface. This would
be possible because of the use of the alumina furnace wall which resulting in the real oxygen
partial pressure might be higher than the equilibrium pressure. In this case, the oxidation and
the thermal decomposition of amorphous intergranular phase [67,82] were suppressed. Under
loading condition, the grain growth of β-SiC and the decomposition of amorphous
intergranular phase were enhanced by applied load. The SiO and CO gases were transported
through the fiber surface. SiC crystals grew on the fiber surface due to the reaction between
SiO and free carbon. Meanwhile, the rupture of silica scale on the surface of fiber would be
easier during loading test [70,83]. Once the flaw was produced by oxidation and
decomposition of amorphous phase, the stress concentration occurred around these flaws and
preexisted crack tip. At that time, the active oxidation was enhanced by loading because of
the rupture or removal of the protective silica film on the fiber surface, leading to the stress
corrosion cracks, as schematically illustrated in Figure 33. The stress corrosion cracks acted
as the critical flaws in the fracture of fibers. The active oxidation is very detrimental to
properties of materials and it proceeded mainly by reactions proposed in literatures [51,58-
59,67].
Figure 33. Schematic of strength degradation of SiC fibers under loading.
5. Tensile Creep Prediction by Long Time BSR Test

The understanding of creep behavior and creep mechanism of SiC materials is essential
for real application. However, there are significant difficulties in the experimental
measurement of tensile creep of advanced small diameter (about 7-14 µm) SiC fibers. In
order to evaluate the creep resistance of these advanced SiC fibers, and also to clarify the
creep mechanism to support continuing optimization of the properties, the apparent activation
46 Jianjun Sha
energy of creep was calculated by applying a cross-cut method to the results of the long-term
BSR tests [84].
Sixta et al. [85] studied the flexural creep of sintered SiC between 1100 ºC and 1500 ºC
for stress 200-100 MPa. The flexural creep rate exhibits linear stress dependence with an
apparent activation energy of 230 ± 80 kJ/mol. The activation energy for grain boundary
diffusion has been reported for high-purity SiC to be 563 kJ/mol [86] and 611 kJ/mol [87] for
carbon and silicon in β-SiC, respectively. Given n=1 (linear stress dependence) and a
relatively low activation energy, it is concluded that grain boundary sliding (GBS),
accommodated by grain boundary transport of SiC, is the controlling creep mechanism in
their test.
Besson et al. [88] investigated the compressive creep behavior of a Si3N4/SiC
nanocomposite in the 1150 ºC -1350 ºC temperature range under stresses from 45 to 180
MPa. The stress exponent equals 1 and the apparent activation energy is 580 kJ/mol. Honda et
al. [89] performed compression tests on Al-doped β-SiC at 2123-2223 K, the stress exponents
were from 1.1 to 1.4 in the temperature range of 2123-2223 K. the apparent activation energy
is 760 kJ/mol. In both above-mentioned studies, these authors considered GBS
accommodated by diffusion as the critical creep mechanism.
Lane et al. [90] investigated a sintered polycrystalline α-SiC containing minor amounts of
impurities. The creep was performed within a range of temperatures and stresses of 1547 ºC-
1747 ºC and 138-148 MPa. The stress exponent increased from 1.44 to 1.71 with temperature.
The activation energies were between 338-434 kJ/mol and 802-914 kJ/mol for temperatures
below and above 1650 ºC, respectively. These authors concluded: (1) the creep mechanism at
low temperature is GBS accommodated by grain boundary diffusion, and at high
temperatures the controlling mechanism becomes GBS accommodated by lattice diffusion;
(2) the parallel mechanism of dislocation glide contributes increasingly to the total strain as
the volume of precipitates declines as a result of progressive coalescence with increasing
temperature.
As we know, creep of CMCs involves stress transfer between the matrix and fibers
caused by their different creep rates, that may lead to fiber failures or matrix cracking [91-
93]. When the matrix is elastic and creep resistant, fiber creep induces stress transfer from the
fibers onto the matrix that may cause matrix cracking. Thus, the understanding of creep-
related properties of fibers is essential in the identification of time-dependent failure
mechanism of CMCs.
In literatures [25,94-97], the creep behavior of SiC-based fibers has been investigated.
The apparent activation energies for creep obtained in these literatures are consistent with
activation energy of self-diffusion for carbon [86,98] and silicon [99-100] in α-and β-SiC,
which are 713-840 kJ/mol and 695-910 kJ/mol, respectively. Based on the results of these
studies, creep is a thermally activated process and controlling creep mechanism is grain
boundary sliding. Particularly, in the tensile creep behavior of SCS CVD SiC fibers, some
authors observed that the fibers exhibited only primary creep, which was characterized by a
continuously decreasing creep rate for progressively longer times, and that the creep rate was
proportional to an exponential power of time [94-96].
The thermally activated creep can be described using traditional Bailey-type relationship
[101].
ε c = At pσ n exp(−Q / RT ) (24)
where t is the creep time and A and p are creep parameters, σ elastic stress, n stress exponent,
Q apparent activation energy, R universal gas constant, T absolute temperature. However,
there are significant difficulties in the experimental measurement of tensile creep of advanced
SiC-based fibers because of their fine diameters (about 7-14 μm). Fortunately, the bend stress
relaxation (BSR) [25] test has been demonstrated to be an effective method for comparing the
relative creep resistance of a wide variety of ceramic fibers. And also, some researchers have
been attempted to relate BSR data with that of tensile creep [95-96]. Their results showed
good correlation between BSR and tensile creep data.
5.1. Bend Stress Relaxation and its Relation to the Tensile Creep
In BSR test, the stress relaxation parameter, m, is defined as the ratio of final to initial
stress at any local position in the fiber. That is:
m = σ (t , T , ε 0 ) / σ (0, T , ε 0 ) (25)
For convenience, two assumptions were made [25]; one is assume that creep strain (εc) is
linearly proportional to the initial strain (ε0) regardless of stress direction, and the other one is
that εc can be measured at room temperature by the relation εc =z/Ra (z is the distance from
fiber neutral axis, see Figure 6). The first assumption of linear strain dependence is generally
valid for polycrystalline materials which relax stress by the grain boundary sliding (GBS)
mechanism, because the stress power dependence of n ≅ 1 is typically observed throughout
creep stage [85,88,94-96,102]. The second assumption implies that at each local position
within the fiber, stress relaxation is not only proportional to ε0 but follows the same time-
temperature dependence. This typically requires a fiber with a uniform microstructure that
creeps with n ≅ 1 power dependence. This assumption is suitable to the advanced SiC-based
fibers with a near-stoichiometric composition and high-crystalline structure. The stress
exponent of n=1 is indicative of a diffusional creep mechanism [103-105].
If these assumptions apply, the BSR ratio m, obtained by a method illustrated in Figure 1,
is independent of position and initial applied strain. On the other hand, in a BSR test, the
stress, σ(t), in a material in response to a constant bend strain, ε0, as a function of time, t, can
be described by the relaxation modulus in bending, Mb(t), where
σ (t ) = M b (t )ε 0 (26)
Substituting the σ (t) in Equation (25) with Equation (26),　and then the Equation (25)
could be expressed as:
R0
m = M b (t ) ⋅ ε 0 / Eε 0 = M b (t ) / E = 1 − (27)
Ra
48 Jianjun Sha
where R0 and Ra are the initial and residual curvature of fiber loops as indicated in Figure 6. It
was also indeed that the negligible dependence of m ration on the initial applied strain was
observed [66]. This result further supported that the above assumptions were appropriate. On
the other hand, the negligible initial strain dependence of m also evidenced the grain
boundary sliding (GBS) as the principal creep mechanism during the stress relaxation of SiC-
based fibers. The rest of BSR tests in this study were performed at a constant initial strain for
times ranging from 1-100 h at elevated temperatures in air. Furthermore, the apparent
activation energy of thermally controlled creep was calculated by a cross-cut method from
time-temperature dependence of stress relaxation parameter.
Since the individual fiber type with uniform microstructure displayed a strain-
independent m ratio, the predictions of tensile creep of fibers from BSR data should be
possible according to the previous results [25,94-96]. Because the εc is linearly stress-
dependent, it is reasonable to relate tensile creep with the relationship between the tensile
relaxation modulus, Mt(t), and tensile creep compliance, Jt(t) [96]:
1 ε (t ) + ε e
J t (t ) = = c (28)
M t (t ) σ
If Mb(t) = Mt(t) and substituting Mt(t) in Equation (28) with Equation (27), so a normalized
creep strain (NCS) could be defined as [96]:
ε c (t ) 1
NCS = = −1 (29)
εe m(t )
If assuming the stress exponent, n, is equal to 1 in Equation (24) and combining the Equation
(24) with Equation (29), the stress relaxation parameter could be correlated with the tensile
creep described by the form:
ε c (t ) 1
NCS = = − 1 = A0 • E • t p (30)
εe m(t )
where A0=A•exp(-Q/RT) is creep parameter (constant for the specific temperature). The time
exponent, p, determined from a plot of log NCS vs log t, is constant.
5.2. BSR Tests at Elevated Temperatures
Figure 34 showed the stress relaxation parameter (m) as a function of temperature and
time for three SiC fiber types. The duration of the stress relaxation tests was 1, 10, 25, 50 and
100 h. The stress relaxation follows an S-shaped curve. If we take m = 0.5 as an arbitrary
value for which we can compare test results for these fibers, two trends are evident. First, the
relaxation temperature for m = 0.5 in HNL fibers increased with an increase of the heat
treatment temperature. Second, at the level of m = 0.5, the 1h BSR tests show a higher
relaxation temperature than that of longer time BSR test. These results suggest that thermally
activated process plays an important role in the creep behavior of SiC fibers.
Test temperature [ºC] Test temperature [ºC] Test temperature [ºC]

1400 1000 1
1400 1000 1500 1200
1 1
Stress relaxation parameter, m 0.9 HNL: As received-1 h
0.9 HNL: 1400 HTT-1 h 0.9 HNL: 1600 HTT-1 h
0.8 HNL: 1400 HTT-10 h 0.8
0.8 HNL: As received-10 h HNL: 1600 HTT-10 h
HNL: 1400 HTT-25 h
0.7 0.7 0.7 HNL: 1600 HTT-25 h
0.6 0.6 0.6 HNL: 1600 HTT-50 h
0.5 0.5 0.5

0.4 ?(1/T) 0.4 0.4
0.3 0.3 0.3
0.2 0.2 0.2
0.1
(a) 0.1 (b) 0.1 (c)
0 0 0
0.5 0.6 0.7 0.8 0.5 0.6 0.7 0.8 0.5 0.6 0.7
Reciprocal temperature, [10-3 K-1] Reciprocal temperature, [10-3 K-1] Reciprocal temperature, [10-3 K-1 ]
Test temperature [ºC] Test temperature [ºC]

1500 1000 1500 1200
1 1
Stress relaxation parameter, m
0.9 0.9 TySA: As received-1 h

0.6 0.6
0.5 0.5
HNLS: As received-1 h
0.4 0.4
HNLS: As received-10 h
0.3 HNLS: As received-50 h
0.3
0.2 0.2
0.1 (d) 0.1 (e)
0 0
0.5 0.6 0.7 0.8 0.5 0.6 0.7
-3 -1
Reciprocal temperature, [10-3 K-1] Reciprocal temperature, [10 K ]
Figure 34. Temperature dependence of m value of SiC fibers for (a) HNL fibers; (b) 1400 °C heat
treated HNL fibers; (c) 1600 °C heat treated HNL fibers; (d) HNLS fibers; (e) TySA fibers.
Since the stress relaxation mechanism is thermally activated process for ceramic
materials, the rate-controlling activation energy, Q, can then be determined from the Δ(1/T)
spacing between the curves at a constant m value (cross-cut method) in Figure 34. The Δ(1/T)
spacing corresponds to the apparent activation energy for stress relaxation in a cross-cut
method.
The Q value for a given m value, can be calculated from the relationship:
t
ln( 2 )
Q = R.
t1 (31)
1 1
−
T1 T2
That is, for 1 order of magnitude change in time, Q=2.3 R/Δ(1/T), where R is the gas
constant (8.314 J/mol. K).
In present work, the activation energies were found (from m =0.3, 0.5, 0.7, which
represent the high temperature, moderate temperature and low temperature regions.) to be 563
kJ/mol, 598 kJ/mol and 445 kJ/mol for the as-received HNL fibers (Figure 34 (a)); 707
50 Jianjun Sha
kJ/mol, 692 kJ/mol, 500 kJ/mol for the as-received HNLS fibers (Figure 34 (d)); 774 kJ/mol,
707 kJ/mol and 524 kJ/mol for the as-received TySA fibers (Figure 34 (e)). From these
results, we can see the Q value increases with increasing the test temperature. The same
change in apparent activation energy was also found in earlier studies [90,96-97]. The large
activation energy at high temperature regions could be related to the concurrent
microstructure change during BSR test, such as grain growth, the crystallization and loss of
oxygen due to the decomposition of amorphous phase (SiCxOy) at grain boundary. As we
know, the creep behavior of ceramics can be explained by oxygen content and grain size. The
crystallite sizes of as received fibers are 4.0 nm for HNL, 11.4 nm for HNLS, 22.7 nm for
TySA, while the crystallite sizes are strongly dependent on the heat treatment temperature
[69]. Furthermore, the grain growth could also be enhanced by applied stress. Thus, the effect
of grain growth must be taken into account at high temperature test. All the grain boundary
sliding mechanisms have a negative grain size exponent, which means that smaller grains will
result in a faster creep rate. Further comparison in Q value for heat treated HNL fiber
supports above result. The Q value for HNL fibers increased with increasing the heat
treatment temperature (Figures 34 (b)-(c)), for instance, at the level of m=0.5, Q is 622 kJ/mol
and 929 kJ/mol for HNL fibers heat treated at 1400 ºC (crystallite size: 5.3 nm) and 1600 ºC
(crystallite size: 8.0 nm), respectively.
The activation energies obtained in our work are in acceptable agreement with the
activation energies of carbon self-diffusion (713-840 kJ/mol) [86,98] and silicon self-
diffusion (695-910 kJ/mol) [99-100] in SiC.
5.3. Prediction of Tensile Creep from BSR Data
Figure 35 shows the normalized creep strain (NCS) versus time in log-log form for a
specific test condition. The NCS was calculated from BSR data with Equation (30). The creep
parameters listed in Table 3 are the best fit in Figure 35. The parameter, A0, is dependent on
the temperature, while time exponent, p, is somewhat high in high temperature region. The
fibers exhibit very similar creep behavior, which suggests that they all creep mainly via a
similar mechanism and the difference in the individual parameter given in Table 3 is due to
compositional or microstructural difference among these fibers. For each fiber type, the
tendency of NCS behaved similarly at different temperatures.
Furthermore, since the bend stress relaxation tests were performed in air, the oxidation of
surface of SiC fibers would enhance the surface stress relaxation [76]. Especially, for the long
time BSR test of fine-diameter SiC fibers, the Silica layer carrying a part of initial applied
stress by stress sharing mechanism might be possible, which results in an overestimation of
NCS. This was reflected by somewhat high time exponent in high temperature region.
According to the oxidation kinetics, the silica thickness obeys a parabolic law for T ≤1400 ºC,
which has been well investigated in literatures [106-107]. SEM observation on cross section
of crept fibers at 1400 ºC for 25 h, has revealed that fibers have a silica layer of about 1.0 um.
The result in present work is in agreement well with previous work [96], although the
fiber types are different. Further comparison was made in BSR test between present and
previous work [96]. It can be seen that HNL fiber has an similar BSR creep resistance with
that of SCS-6 CVD SiC fiber (higher than 5% excess carbon); the near-stoichiometric HNLS
and TySA fiber behaved similarly in BSR creep resistance with that of 2 mil CVD SiC fibers
(nearly pure SiC, excess carbon less than 1%). Both CVD SiC fibers appeared that the stress
exponent, n, is approximate 1.
10 10 100
HNL-1600HTT/1200 (a)
(a)
HNLS/1400 (b) TySA/1200
(c)
HNLS/1300 TySA/1400
Normalized creep strain
HNL-1600HTT/1000 HNLS/1200 TySA/1300

HNL-1400HTT/1000
1 10
0.1 1
0.01 0.1 0.1

1 10 100 1 10 100 1 10 100
Time (h) Time (h) Time (h)
Figure 35. Normalized creep strain (NCS) predicted from BSR data versus time for HNL fiber, (b)
HNLS fiber, (c) TySA fiber. The parameters presented in Table 3 are the best fit of these curves.
Figure 36 showed the tensile creep strain, which was predicted by BSR data. Generally,
this prediction showed a similar time and temperature dependence with that of the primary
creep stage. As pointed out by other researchers [96], for silicon carbide fibers which have
uniform microstructures, the BSR predictions usually are very near the magnitude of tensile
creep strain.
2 2.5 4
HNL-1400HTT/1000 C/1000 MPa
(a)
HNLS/1400 C/670 MPa
(b) TySA/1200 C/320 MPa
(c)
Tensile creep strain (%)
1.8 HNL-1600HTT/1000 C/1000 GPa

HNLS/1300 C/670 MPa 3.5 TySA/1300 C/320 MPa
1.6 HNL-1600HTT/1200 C/1000 MPa 2 HNLS/1200 C/670 MPa
3 TySA/1400 C/320 MPa
1.4
1.5 2.5
1.2
1 2
0.8 1 1.5
0.6 1
0.4 0.5
0.5
0.2
0 0 0
0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 7000
Time (minute) Time (minute) Time (minute)

Figure 36. Showed the tensile creep strain, which was predicted by BSR data parameters using
Equation (30): (a) HNL fiber; (b) HNLS fiber; (c) TySA fiber.
In addition, the actual tensile creep strains from Refs. [41] were also listed in Table 3, and
compared with that of BSR predictions. It was observed that tensile creep strains from BSR
predictions showed same order of magnitude with the actual tensile creep strains. Noting in many
tensile creep tests of ceramic fibers, the data scattering in creep strain is significant. On the other
hand, because of the lack of tensile creep data on advanced SiC-based fibers, this comparison was
made among the studies with different fiber batch. The properties of SiC fibers are different from
batch to batch and they are still in developing. Further demonstration from tensile creep test with
same fiber batch would be necessary. Nevertheless, in present work, the BSR data predicted the
same time and temperature dependence of tensile creep for advanced SiC-based fibers.
Table 3. Tensile creep parameters from the best fit of BSR data with Equation (30)
A0 (GPA) εc (%) Actual (from tensile

Fiber Type Experimental NCS (50h) p Best Fit εc (%) Best Fit
Bestfit creep test)
HNLS 670 MPa 670 MPa
1200 ºC 0.45 0.000435 0.24 0.19 (50h)
1300 ºC 1.38 0.000955 0.32 0.83 (50h) 0.36 (50h)#1
1400 ºC 4.56 0.002721 0.28 1.72 (50h)
TySA 320 Mpa 320 Mpa
1200 ºC 0.54 0.000269 0.4 0.21 (50h)
1300 ºC 2.85 0.000833 0.4 0.66 (50h) 1.72 (20h)#1
1400 ºC 11.5 0.000346 0.4 2.72 (50h)
HNL 1000 Mpa 1000 Mpa
As received 0.3 (1200 ºC, 3 5h)#2
HNL 1400 C HTT/1000 ºC 0.23 0.000366 0.22 0.22 (50h)
HNL 1600 C HTT/1000 ºC 0.16 0.000274 0.20 0.14 (50h)
HNL 1600 C HTT/1200 ºC 0.66 0.000793 0.28 0.75 (50h)
.24 (0.83h) 0.13 (0.39h), 0.2 (0.83h)#2
Note: NCS=1/m-1=A0·E·tp, εc=σ·A0·tp, σ=E•εe, t in minutes, σ in Gpa; HTT: Heat Treatment Temperature; #1: [41]; #2: [107].
Furthermore, Combining the analysis of activation energies with the predicted result
(creep rate is linearly dependent on the applied stress) in our work, suggests that the main
mechanism responsible for creep of SiC fibers is controlled by GBS accommodated by
diffusion (prevalent at low temperatures and small grain size) [103-104]. Since the silicon
self-diffusion coefficient is one order of magnitude slower than the carbon self-diffusion
coefficient [99-100], so the silicon is the controlling species during the diffusion process. The
strain rate for a pure diffusion lattice mechanism is given by Nabarro-Herring creep model
[104].
The time exponent p and its slight variation at high temperature indicate that additional
mechanism with similar activation energy might be operating besides GBS accommodated by
diffusion. Lane et al. [90] found an increase of activation energy with temperature due to the
transition from GBS accommodated by grain boundary diffusion and low dislocation activity,
to GBS accommodated by bulk diffusion and high dislocation activity. The dislocation
activity increased with the coalescence of the precipitate phase, because the interaction of
dislocation with the precipitates makes it difficult to gild. This mechanism would be possible
in present test, because the SiC fiber used is not so pure.
6. Conlusion
The review of the recent studies on the mechanical durability and microstructure stability
of SiC-based fibers presented in this chapter has shown that they are promising reinforcement
for CMCs, which have been proposed as potential structural materials for advanced energy
systems and propulsion systems. Especially, the development of advanced SiC fibers with
near stoichiometry and high crystallinity has improved their mechanical and thermal
stabilities significantly. Mechanical and microstructure characterization have led to a deeper
knowledge of the relationship between the microstructure and the mechanical behaviour. The
main completion can be summarized as follows:
(i) Most of the initial strength for SiC fibers with near stoichiometric composition is
retained up to very high temperature in inert atmosphere. The strength of fiber is
sensitive to the critical flaws caused during fabrication process or by exposure to
service environments. The heat treatment above the processing temperature could
improve the creep resistance due to the crystallization, the grain coarsening and
composition changes at grain boundaries (GB). The GB composition could affect the
stability of GB boundaries.
(ii) The environment-pertinent degradation mechanism is complex for SiC materials
subjected to realistic application. When the SiC-based fibers were exposed to the
inert atmosphere with different oxygen partial pressures, the SiC-based fibers can be
oxidized in passive/active oxidation regimes. The strengths of SiC-based fibers are
strongly dependent on the oxygen partial pressures, being decrease with decreasing
the oxygen partial pressures when they are oxidized in active oxidation regime.
Strength degradation was caused by different oxidation mechanism in different
atmospheres. In contrast, no obvious dependence of creep resistance on oxygen
partial pressures was observed. On the other hand, the microstructure observation
revealed that the oxidation and applied stress/loading can result in the nucleation and
54 Jianjun Sha
growth of new flaw leading to the stress corrosion. The stress corrosion is
detrimental to the integrated performance of materials.
(iii) The apparent activation energies were calculated by cross-cut method from the long
time bend stress relaxation (BSR) tests. The activation energies calculated from BSR
data are in agreement with the activation energies of grain boundary diffusion of
carbon and silicon in β-SiC, suggesting the main mechanisms responsible for creep
of SiC-based fiber is controlled by GBS accommodated by diffusion. The activation
energies are different at low and high temperature regions. This phenomenon could
be related to the concurrent microstructure change during BSR test at high
temperatures. The tensile creep of SiC-based fibers was predicted from the BSR data
by a defined normalized creep strain (NCS). Tensile creep predicted from BSR data
could reflect the tendency of primary creep of the SiC-based fibers with similar time
and temperature dependence.
Finally, the appropriate understanding of the environmental durability of SiC fibers is

essential for the reliable evaluation of CMCs. Although progress in some aspects is achieved,
but a number of key issues still remain open. Currently there is much interest in SiC fiber
reinforced CMCs. Stress corrosion of CMCs is another interesting point because the stress
and corrosion (oxidative and corrosive environments) often act synergistically. The
synergistic linkage of mechanical stability and corrosion under stress is an important area of
research towards applying SiC fiber reinforced CMCs as structural components.
Acknowledgements:
This work was partially supported by the National Nature Science Foundation of China
(grant No. 50871092). The author acknowledges the support of the NSFC.
References
[1] Naslain, R. Design, preparation and properties of non-oxide CMCs for application in
engines and nuclear reactors: An overview. Composite Science and Technology, 2004,
64, 155-170.
[2] Kohyama, A. Advanced ceramic composite materials for fusion/fission reactor
systems. Ceramics (in Japanese), 2004, 39, 838-842.
[3] Katoh, Y; Kohyama, A; and Hinoki, T. Progress in SiC-based ceramic composites for
fusion applications. Fusion Science and Technology, 2003, 44, 155-162.
[4] Spriet, P; and Habarou, G. Applications of CMCs to turbojet engines: overview of the
SEP experience. Key Engineering Materials, 1997, 127–131, 1267–1276.
[5] Ohnabe, H; Masaki, S; Onozuka, M; Miyahara K; and Sasa, T. Potential application of
ceramic matrix composites to aero-engine components. Composites: Part A, 1999, 30,
489–496.
[6] Staehler, JM; and Zawada, LP. Performace of four ceramic-matrix composites
divergent flap inserts following ground testing on an F110 turbofan engine. J. Am.
Ceram. Soc., 2000, 83, 1727-38.
[7] Evans, AG; and Zok, FW. The physics and mechanics of fiber-reinforced brittle-matrix
composites. J. Mater. Sci., 1994, 29, 3857-3896.
[8] Donald, LW; and Mcmillan, PW. Review ceramic-matrix composites. J. Mater. Sci.,
1976, 11, 949-972.
[9] Yajima, S; Hasegawa, X; Hayashi, J; and Iiuma, M. Synthesis of continuous SiC fibers
with high tensile strength and modulus. J. Mater. Sci., 1978, 13, 2569–2576.
[10] Mazdiyasni, KS; West, R; and David, LD. Characterization of organosilicon-infiltrated
porous reaction-sintered Si3N4. J. Am. Ceram. Soc., 1978, 61, 504-508.
[11] Mah, T; Hecht, NL; McCullum, DE; Hoenigman, JR; Kim, HM; Katz, P; and Lipsitt,
HA. Thermal stability of SiC fibers (Nicalon). J. Mat. Sci., 1984, 19, 1191-1201.
[12] Johnson, SM; Brittain, RD; Lamoreaus RH; and Rowcliffe, DJ. Degradation
mechanisms of Silicon-carbide fibers. J. Amer. Ceram. Soc., 1988, 71, 132-135.
[13] Okamura, K; Sato, M; Seguchi, T; and Kawanishi, S. Application of electron beam
irradiation for preparation of SiC fiber Hi-Nicalon for ultra-high temperature.
Composite Science and Technology, 1994, 51, 135-144.
[14] Okamura, K; and Seguchi, T. Application of radiation curing in the preparation of
polycarbosilane-derived SiC fibers. J. Inorganic and Organometallic Polymers, 1992,
2, 171-179.
[15] Takeda, M; Imai, Y; Ichikawa, H; Ishikawa, T; Segushi, T; and Okamura, K. Properties
of the low oxygen conent SiC fibers on high temperature treatment. Ceram. Eng. and
Sci. Proc., 1991, 12, 1007-1018.
[16] Takeda, M; Imai, Y; Ichikawa, H; Ishikawa, T; Kasai, N; Segushi T; and Okamura, K.
Ceram. Eng. and Sci. Proc., 1992, 13, 209-217.
[17] Ichikawa, H. Recent advances in Nicalon ceramic fibers including Hi-Nicalon Type S.
Annales De Chimie-Science Des Materiaux, 2000, 52, 523-528.
[18] [18] Ishikawa, T; Kohtoka, Y; Kumagawa, K; Yamamura T; and Nagasawa, T. High-
strength alkali-resistant sintered SiC fiber stable to 2200 ºC. Nature, 1998, 391, 773-
774.
[19] [19] Yun, HM; and Dicarlo, JA. Comparison of the tensile, creep and rupture strength
properties of stoichiometric SiC fibers. Ceram. Eng. and Sci. Proc., 1999, 20, 259-272.
[20] [20] Takeda, M; Sakamoto, J; Imai, Y; Ichikawa, H. Thermal stability of the low-
oxygen-content silicon carbide fiber, Hi-NicalonTM. Composites Science and
Technology, 1999, 59, 913-919.
[21] Takeda, M; Urano, A; Sakamoto, J; and Imai, Y. Microstructure and oxidative
degradation behaviors of silicon carbide fiber Hi-Nicalon type S. J. Nucl. Mater., 1998,
258-263, 1594-1599.
[22] ASTM D3379-75 (reapproved 1989), Standard test method for tensile strength and
young’s Modulus for High-Modulus Single-Fiber materials.
[23] Weibull, W. A statistical distribution function of wide applicability. Journal of Applied
Mechanics-Transcations of the ASME, 1951, 18, 293-297.
[24] Patankar, SN. Weibull distribution as applied to ceramic fibers. J. Mater. Sci. Lett.,
1991, 10, 1176-1181.
[25] Morscher, GN; and Dicarlo, JA. A simple test for thermomechanical evaluation of
ceramic fibers. J. Am. Ceram. Soc., 1992, 75, 136-140.
56 Jianjun Sha
[26] Taylor, ST; Zhu, YT; Blumenthal, WR; Stout, MG; Butt, DP; and Lowe, TC.
Characterization of Nicalon fibers with varying diameters: Part I strength and fracture
studies. J. Mater. Sci., 1998, 33, 1465-1473.
[27] Youngblood, GE; Lewinsohn, C; Jones RH; and Kohyama, A. Tensile strength and
fracture surface characterization of Hi-Nicalon™ SiC fibers. J. Nucl. Mater., 2001, 289,
1-9.
[28] Serizawa, H; Lewinsohn, CA; Youngblood, GE; Jones, RH; Johnston, DE; and
Kohyama, A. High temperature properties of newly developed advanced SiC fibers.
Key Engineering Materials, 1999, 164-165, 287-290.
[29] Eckel, AJ; and Bradt, RC. Strength distribution of reinforced fibers in a Nicalon
fiber/chemically vapor infiltrated silicon carbide matrix composite. J. Am. Ceram. Soc.,
1989, 72, 455-458.
[30] Wagner, HD. Stochastic concepts in the study of size effects in the mechanical strength
of highly oriented polymeric materials. Journal of Polymer Science Part B-Polymer
Physics, 1989, 27, 115-149.
[31] Zhu, YT; Blumenthal, WR; Taylor, ST; and Lowe, TC. Analysis of size dependence of
ceramic fiber and whisker strength. J. Am. Ceram. Soc., 1997, 80, 1447-1452.
[32] Sawyer, LC; Jamieson, M; Brikowski, D; Haider, MI; and Chen, RT. Strength,
structure, and fracture properties of ceramic fibers produced from polymeric precursors.
Journal of the American Ceramic Society, 1987, 70, 798-810.
[33] Mecholsky, JJ; Freiman, SW; and Rice, RW. Effect of flow shape and size on strength
and fracture in glass. Report of NRL Progress, 1975, 11, 7-10.
[34] Griffith, AA. The phenomenon of rupture and flow in solids. Phil. Trans. Roy. Soc.
London A, 1920, 221, 163-198.
[35] Jones, RH; Lewinsohn, CA; and Youngblood, GE. Properties of advanced fibers for
SiC/SiC composite applications in fusion energy systems. High temperature ceramic
matrix composites III, 1999, 164-1, 405-408.
[36] Thouless, MD; Sbaizero, O; Sigl, LS; and Evans, AG. Effet of interface mechanical-
properties on pullout in a SiC-fiber-reinforced lithium aluminum silicate glass-ceramic.
Journal of the American Ceramic Society, 1989, 72, 525-532.
[37] Dong, SM; Katoh, Y; and Kohyama, A. Preparation of SiC/SiC composites by hot
pressing, using Tyranno-SA fiber as reinforcement. J. Am. Ceram. Soc., 2003, 86, 26-
32.
[38] Lee, SP; Katoh, Y; Park, JS; Dong, SM; Kohyama, A; Suyama, S; and Yoon, HK.
Microstructural and mechanical characteristics of SiC/SiC composites with modified-
RS process. J. Nucl. Mater., 2001, 289, 30-36.
[39] Chollon, G; Pailler, R; Naslain, R; Laanami, F; Monthioux M; and Olry, P. Thermal
stability of a PCS-derived SiC fibre with a low oxygen content (Hi-Nicalon). J. Mater.
Sci., 1997, 32, 327-347.
[40] Ichikawa, H. Advances in Nicalon ceramic fibers including Hi-Nicalon type S. Annales
De Chimie-Science Des Materiaux, 2000, 25, 523-528.
[41] Bunsell, AR; and Berger, MH. Fine diameter ceramic fibres. J. Euro. Ceram. Soc.,
2000, 20, 2249-2260.
[42] Dong, SM; Chollon, G; Labrugere, C; Lahaye, M; Guette, A; Bruneel, JL; Couzi, M;
Naslain, R; and Jiang, DL. Characterization of nearly stoichiometric SiC ceramic fibers.
J. Mater. Sci., 2001, 36, 2371-2381.
[43] Ichikawa, H; and Ishikawa, T: Silicon carbide fibers (organometallic Pyrolysis). In:
Kelly, A; Zweben, C; Chou, T. Comprehensive composite Materials. Oxford, England:
Elsevier Science Ltd.; 2000; 107-145.
[44] Takeda, M; Saeki, A; Sakamoto, J; Imai, Y; and Ichikawa, H. Properties of
polycarbosilane-derived silicon carbide fibers with various C/SiC compositions.
Composite Science and Technology, 1999, 59, 787-792.
[45] Yajima, S; Okamura, K; Matsuzawa, T; Hasegawa, Y; and Shishido, T; Anomalous
characteristics of the microscrystalline state of SiC fibers. Nature, 1979, 279, 706-707.
[46] Sasaki, Y; Nishina, Y; Sato, M; and Okamura, K. Raman-study of SiC fibers mae from
polycarbosilane. J. Mater. Sci., 1987, 22, 443-448.
[47] Sacks, MD. Effect of composition and heat treatment conditions on the tensile strength
and creep resistance of SiC-based fibers. Journal of the European ceramic society,
1999, 19, 2305-2315.
[48] Havel, M; and Colomban, Ph. Skin/bulk nanostructure and corrosion of SiC-based
fibers: a surface Rayleigh and Raman Study. Journal of Raman Spectroscopy, 2003, 34,
786-794.
[49] Havel, M; and Colomban, Ph. Rayleigh and Raman images of the bulk/surface
nanostructure of SiC based fibers. Compoistes: Part B-Engineering, 2004, 35, 139-147.
[50] Delhaes, P; and Carmona, F. Physical-properties of noncrystalline carbons. Chemistry
and Physics of carbon, 1981, 17, 89-174.
[51] Schneider, B; Guette, A; Naslain, R; Cataldi, M; and Costecalde, A. A theoretical and
experimental approach to the active-to-passive transition in the oxidation of silicon
carbide: experiments at high temperatures and low total pressures. J. Mater. Sci., 1998,
33, 535-547.
[52] Shimoo, T.; Morisada Y.; and Okamura K. Oxidation behavior of Si-M-C-O fibers
under wide range of oxygen partial pressures. J. Mater. Sci., 2002, 37, 4361-4368.
[53] Shimoo, T; Okamura, K; and Morita, T. High temperatue exposure of low-oxygen SiC
fibers (Hi-Nicalon) in CO-CO2 gas mixture. J. Mater. Sci., 2004, 39, 7031-7039.
[54] Kumagawa, K; Yamaoka, H; Shibuya, M; and Yamakura, T. Fabrication and
mechanical properties of new improved Si-M-C-(O) tyranno fiber. Ceram. Eng. Sci.
Proc. 1998, 19, 65-70.
[55] Toreki, W; Batich, CD; Sacks, MD; Saleem, M; Choi, GJ; and Morrone, AA. Polymer-
derived silicon-carbide fibers with low-oxygen content and improved
thermomechanical stability. Composites Science and Technology, 1994, 51, 145-159.
[56] Ramberg, CE; and Worrell, WL. Oxygen transport in silica at high temperatures:
implications of oxidation kinetics. J. Am. Ceram. Soc., 2001, 84, 2607-2616.
[57] Mazdiyasni, KS; and Aangvil, A. Effect of impurities on SiC whisker morphology. J.
Am. Ceram. Soc., 1985, 68, C142-C144.
[58] Shimoo, T; Takeuchi, H; and Okamura, K. Thermal stability of polycarbosilane-
derived silicon carbide fibers under reduced pressures. J. Am. Ceram. Soc., 2001, 84,
566-570.
[59] He, G; Shibayama, T; T and akahashi, H; Microstructural evolution of Hi-NicalonTM
SiC fibers annealed and crept in various oxygen partial pressure atmospheres. J. Mater.
Sci., 2000, 35, 1153-1164.
[60] Shimoo, T; Okamura, K; Ito, M; and Takeda, M; High-temperature stability of low-
oxygen silicon carbide fiber heat-treated under different atmosphere. J. Mater. Sci.,
58 Jianjun Sha
2000, 35, 3733-3739. [61] Jacobson, NS. Corrosion of Silicon-based ceramics in

combustion environments. J. Am. Ceram. Soc., 1993, 76, 3-28.
[61] Maeda, M; Nakamura, K; Ohkubo, T; Ito, J; and Ishii, E. Oxidation behavior of silicon
nitride under static and cyclic conditions. Ceramic International, 1989, 15, 247-53.
[62] Singhal, SC. Oxidtion-kinetics of hot-pressed silicon-carbide. J. Mater. Sci., 1976, 11,
1246-1253.
[63] Singhal, SC. Oxidation of silicon-based structural ceramics. Journal of the
Electrochemical Society, 1976, 123, 253-253.
[64] Gallardo-Lopez, A; Munoz, A; Martinez-Fernandez, J; and Dominguez-Rodriguez, A;
High-temperature compressive creep of liquid phase sintered silicon carbide. Acta
mate., 1999, 47, 2185-2195.
[65] Sha, JJ; Hinoki, T; and Kohyama, A. Mechanical and thermal stabilities of Hi-Nicalon
fiber under annealing and creep in various oxygen partial pressure. Corrosion Science,
in press, 2008.
[66] Shimoo, T; Okamura, K; and Morisada, Y Active-to-passive oxidation transition for
polycarbosilane-derived silicon carbide fibers heated in Ar-O2 gas mixtures. J. Mater.
Sci., 2002, 37, 1793-1800.
[67] Shimoo, T; Morisada, Y; and Okamura, K; Oxidation behavior of Si-M-C-O fibers
under wide range of oxygen partial pressures. J. Mater. Sci., 2002, 37, 4361-4368.
[68] Sha, JJ; Nozawa, T; Park, JS; Katoh, Y; and Kohyama, A. Effect of heat treatment on
the tensile strength and creep resistance of advanced SiC fibers. J. Nucl. Mater., 2004,
329-333, 592-596.
[69] Nickel, KG. Corrosion of Non-oxide ceramics. Ceramics international, 1997, 23, 127-
133.
[70] Shimoo, T; Toyoda, F; and Okamura, K. Thermal stability of low-oxygen silicon
carbide fiber (Hi-Nicalon) subjected to selected oxidation treatment. J. Am.Ceram. Soc.,
2000, 83, 1450-1456.
[71] Luthra, KL. Some new perspectives on oxidation of silicon carbide & silicon Nitride. J.
Am. Ceram. Soc., 1991, 74, 1095-1103.
[72] Jaskowiak, MH; and Dicarlo, JA. Pressure effects on the thermal-stability of silicon-
carbide fibers. J. Am. Ceram. Soc., 1989, 72, 192-197.
[73] Giannuzzi, LA; Lewinsohn, CL; Bakis, CE; T and ressler RE. Microstructural
development of SCS-6 SiC fibers during high temperature creep. Journal of Materials
Research, 1998, 13, 1853-1860.
[74] Dicarlo, JA; Yun, HM; and Hurst, JB. Fracture mechanisms for SiC fibers and SiC/SiC
composites under stress-rupture conditions at high temperatures. Applied mathematics
and computation, 2004, 152, 473-481.
[75] Sha, JJ; Park, JS; Hinoki, T; Kohyama, A; and Yu, J. Tensile properties and creep
behavior of SiC-based fibers under various oxygen partial pressures. Materials Science
forum, 2005, 475-479, 1333-1336.
[76] Pysher, DJ; Goretta, KC; Hodder, RS; and Tressler, RE. Strength of ceramic fibers at
elevated-temperatures. J. Am. Ceram. Soc., 1989, 72, 284-288.
[77] Sha, JJ; Park, JS; Hinoki, T; and Kohyama, A. Hot corrosion, oxidation and their
effects on the tensile strength of SiC fiber in alkaline melts. Materials Transactions,
2005, 46, 1032-1035.
[78] Sha, JJ; Park, JS; Hinoki, T; and Kohyama A. Tensile behavior and microstructural
characterization of SiC fibers under loading. Materials Science and Engineering A,
2007, 456, 72-77.
[79] Shimoo, T; Tsukata, I; Seguchi, T; and Okamura, K. Effect of firing temperature on the
thermal stability of low-oxygen silicon carbide fibers. J. Am. Ceram. Soc., 1998, 81,
2109-2115.
[80] Bodet, R; Bourrat, X; Lamon, J; and Naslain, R. Tensile creep-behavior of a silicon
carbide-based fiber with al low-oxygen content. J. Mater. Sci., 1995, 30, 661-677.
[81] Shimoo, T; Tsukata, I; Seguchi, T; Okamura, K. Effect of firing temperature on the
thermal stability of low-oxygen silicon carbide fibers. J. Am. Ceram. Soc., 1998, 81,
2109-2115.
[82] Gogotsi, YG; and Grathwohl, G. Stress-enhanced oxidation of silicon nitride ceramics.
J. Am. Ceram. Soc., 1993, 76, 3093-3104.
[83] Sha, JJ; Park, JS; Hinoki, T; and Kohyama, A. Bend stress relaxation of advanced
SiGbased fibers and its prediction to tensile creep. Mechanics of Materials, 2007, 39,
175-182.
[84] Sixta, ME; Zhang, XF; and Jonghe, LC. Flexural creep of an in situ-toughened silicon
carbide. J. Am. Ceram. Soc., 2001, 84, 2002-2028.
[85] Hon, MH; and Davis, RF. Self-diffusion of C-14 in polycrystalline beta-SiC. J. Mater.
Sci., 1979, 14, 2411-2421.
[86] Sargent, PM; and Ashby, MF. Deformation-mechanism maps for silicon-caride. Scri.
Metall., 1983, 17, 951-57.
[87] Besson, JL; Doucey, B; Lucas, S; Bahloul-Hourlier, D; and Goursat, P. SiCN
nanocomposite: creep behaviour. J. Euro. Ceram. Soc., 2001, 21, 959-968.
[88] Honda, S; Nagano, T; Kaneko, K; and Kodama, H. Compressive deformation behavior
of Al-doped β-SiC at elevated temperatures. J. Euro. Ceram. Soc., 2002, 22, 979-985.
[89] Lane, JE; Carter CH; and Davis, RF. Kinetics and mechanisms of high-temperature
creep in silicon-carbide: III, sintered alpha-silicon carbide. J. Am. Ceram. Soc., 1988,
71, 281-295.
[90] Holmes, JW; Park, YH; and Jones, JW. Tensile creep and creep recovery behaviour of
a SiC-fiber-Si3N4-matrix composite. J. Am. Ceram. Soc., 1993, 76, 1281-1293.
[91] Lamouroux, F; Steen, M; and Valles, JL; Uniaxial tensile and creep-behavior of an
alumina fiber-reinforced ceramic-matirx composite: expermental-study. J. Euro.
Ceram. Soc., 1994, 14, 529-537.
[92] Evans, AG; and Weber, C; Creep damage in SiC/SiC composites. Mater. Sci. Eng. A,
1996, 208, 1-6.
[93] Dicarlo, JA. Creep of chemically vapor-deposited SiC fibers. J. Mater. Sci., 1986, 21,
217-24.
[94] Lewinsohn, CA; Giannuzzi, LA; Bakis, CE; and Tressler, RE. High-temperature creep
and microstructural evolution of chemically vapor-deposited silicon carbide fibers. J.
Am. Ceram. Soc., 1999, 82, 407-13.
[95] Morscher, GN; Lewinsohn, CA; Bakis, CE; and Tressler, RE. Comparison of bend
stress relaxation and tensile creep of SVD SiC fibers. J. Am. Ceram. Soc., 1995, 78,
3244-3252.
[96] Shoji, S; Katase, Y; Okamura, K; Creep behavior of SiC fibers at high temperatures
using the BSR method. Key Eng. Mater., 2003, 247, 187-190.
60 Jianjun Sha
[97] Hong, JD; and Davis, RF. Self-diffusion of C-14 in high-purity and N-doped alpha-Sic
single-crystals. J. Am. Ceram. Soc., 1980, 63, 546-552.
[98] Hong, JD; Davis, RF; and Newbury, DE. Self-diffusion of Si-30 in alpha-SiC single-
crystals. J. Mater. Sci., 1981, 16, 2485-2494.
[99] Hong, JD; Hon, MH; and Davis, RF; Si-30 diffusion in alpha-SiC and beta-SiC.
American Ceramic Society Bulletin, 1979, 58, 348-348.
[100] Shames, H; and Cozzarelli, FA. Elastic and Inelastic Stress Analysis. Prentice Hall,
Englewood Cliffs, NJ; 1992.
[101] Ogasawara, T; Ishikawa, T; Ohsawa, Y; Ochi, Y; and Zhu, S; Tensile creep behavior
and thermal stability of orthogonal three-dimensional woven Tyranno ™ ZMI
fiber/silicon-titanium-carbon-oxygen matrix composites. J. Am. Ceram. Soc., 2002, 85,
393-400.
[102] Coble, RL. A model for boundary diffusion controlled creep in polycrystalline
materials. J. Appl. Phys., 1963, 34, 1679-82.
[103] Herring, C. Diffusional viscosity of a polycrystalline solid. J. Appl. Phys., 1950, 21,
437-445.
[104] Cannon, WR; Langdon, TG. Creep of ceramics: mechanical characteristics. J. Mater.
Sci., 1983, 18, 1-50.
[105] Shimoo, T; Toyoda, F; and Okamura, K. Oxidation kinetics of low-oxygen silicon
carbide fiber. J. Mater. Sci., 2000, 35, 3301-3306.
[106] Chollon, G; Railler, R; Naslain, R; and Olry, P. Correlation between microstructure and
mechanical behaviour at high temperatures of a SiC fibre with a low oxygen content
(Hi-Nicalon). J. Mater. Sci., 1997, 32, 1133-47.
Chapter 2
IONOMERS AS CANDIDATES FOR STRUCTURAL

MATERIALS
Daniel J. Klein
Ashland Performance Materials,
5200 Blazer Parkway, Dublin, Ohio, USA, 43017
Abstract
The field of ionomers is an often overlooked and under-utilized branch of polymer research.
Although ionomers can be broadly described as a class of polymers that contain any number
of ionic groups, from a structural property standpoint only a low percent of ionic groups are
necessary to impart significantly improved properties over the nonionic version of the same
polymer. Current trends in the field of ionomers are highly focused on the field of fuel cell
technology. There appears to be a significant hole remaining in the study of imparting strength
to materials using ionic groups. This hole is very significant from an industrial point of view,
and has a large commercial potential. There are very few commercially available ionomers,
which shows how little this field has been explored to date.
This chapter will focus on several aspects of ionomer research from a physical property
standpoint:
1) A history of ionomer research
2) Current trends in ionomer research
a) stand-alone polymers
b) nanocomposites
c) blends
3) A commentary on the immediate needs in the field of ionomer research
Introduction
(1)The field of ionomers is a very broad area of research. There are vast numbers of
published works, including books, journal articles, and patents, which account for the wide
range of ionomers. Due to the large number, and numerous applications of ionomers, a
complete review of the field will not be discussed here. However, there are several good
62 Daniel J. Klein
review articles, chapters, and books [1-8] that will give the reader a good understanding of the
overall subject matter. However, one of the goals of this chapter is to hit on some key areas of
ionomer research as they could possibly relate to potential structural materials. The focus of
this chapter will be based on commercially viable options for ionomers.
(2)The word ionomer has many different meanings depending upon the researcher. In this
chapter the word ionomer will be defined according to Eisenberg [1]. This is defined as
“polymers in which the bulk properties are governed by ionic interactions in discrete regions
of the material.” From a structural polymer standpoint it is most beneficial to keep the
ionomer ion content relatively low. This is due to the fact that polymers containing high ion
contents tend to be brittle.
(3)From a microstructural standpoint there are different structures that form depending
upon the ion content. It is generally accepted that the ions coordinate into either multiplets or
clusters [3]. Both represent sites where the mobility of the polymer chain segments in the
proximity of the ions is reduced relative to the regions that are not restricted. Multiplets are
usually only small amounts of ionic coordination to tie the chains together (figure 1) [4].
When the regions are of sufficient enough size, and are in close proximity with one another,
clusters form. Both act as non-covalent crosslink sites. In the region of restricted mobility the
polymer has its own set of properties separate from the rest of the polymer. Obtaining the
desired properties of the ionomer is highly dependent upon proper selection of ion
concentration and counterion (type, size, charge, etc.). These, along with the polymer
structure and polarity, help determine whether the ionomers form multiplets or clusters.
Figure 1. Example of a multiplet based on poly(styrene-co-sodium neutralized methacrylic acid).
(4)An interesting aspect of ionomers is that they can be used in several different
applications. Examples include: use as a stand-alone polymer, be used to incorporate various
Ionomers as Candidates for Structural Materials 63
fillers in nanocomposite applications, and can act as a compatibilizer between two dissimilar
polymers as part of a blend. Each of these subject areas will be discussed in this chapter.
(5)Although there has been a lot of research in the field of ionomers, and many research
groups continue to explore this field to date, the field still remains a largely untapped resource
from a commercial polymer standpoint. An extensively thorough review of each topic will not
be covered, but rather several potentially important systems will be discussed. The goal of
this chapter is to introduce the reader to the field of ionomers as structural materials, and to
urge the reader to consider the possibilities that do exist in this field.
Advantages/Disadvantages of Ions in Polymers

(6)There are many ways in which ionic character can be incorporated into polymers.
From an industrial point of view, most of these techniques are undesirable since they require
an added step in order to convert the polymer into its ionic form. Additional steps add
additional cost to the final product. However, the benefits associated with conversion to the
ionomeric state, such as increased glass transition temperatures (Tg), tensile properties,
fatigue properties, scratch resistance, and optical transparency may be worth the extra effort
and cost. And since there are many commercially available ionic or potentially ionic
monomers that can be incorporated into polymers during the polymerization reaction, the
overall cost can be minimized. This is more cost effective than developing new monomers to
achieve the same benefits as can be obtained through ionic forces.
(7)One major benefit of having ionic segments in polymers is that they can form a type of
crosslink, either as a multiplet or cluster, in the system. As opposed to most covalent
crosslinks, the ionic crosslinks can be broken and reformed. This allows thermoplastic
polymers to gain mechanical strength as a result of the ionic crosslinks, but still be able to
maintain the thermoplastic nature of the overall polymer. Over time the thermoformed
polymer will again reform the ionic crosslinks, and regain the prior properties. Hence, the
ionomer has properties of both a thermoset and thermoplastic polymer.
(8)Due to the number of counterions available, the properties can readily be tailored
through proper selection of these counterions. Additionally, multiple counterions can be
combined to vary the properties relative to the ionomer containing only one type of
counterion [9-13]. The polarity and size of the counterion can alter the properties of the
polymer. This wide range of counterions adds an easy benefit to changing the properties of
the polymer with little effort.
(9)Since ionic segments can cause aggregation, there are segments in the polymer
microstructure that have limited mobility. In semi-crystalline polymers this can cause a
reduction in the percent crystallinity and crystallization rate due to the limited mobility of
segments of the polymer chains. This reduction in crystallinity and crystal size can lead to
polymers that have a greater transparency than their non-ionic analogs. This can be beneficial
in applications in which optical clarity is an important quality.
(10)The introduction of ionic groups into polymers has a large influence on the melt
rheological behavior of the system. This can potentially be a big disadvantage in that the ionic
forces can create such a high viscosity that the polymer cannot be processed using traditional
thermal methods. Achieving a balance between ionic content and processability helps achieve
64 Daniel J. Klein
desirable properties. Due to the potential for the ionic groups to impart ionic crosslinking, the
polymer could behave similar to a thermoset at room temperature yet remain a thermoplastic.
(11)A potential disadvantage of ion-containing polymers is that the ionic content
typically causes the polymer to absorb more water than if there were no ionic portions present
[5-7]. The water absorption is tied to the counterion used in the ionomer. This water
absorption can break-up the ionic crosslinks in the polymer, and weaken the polymer. Also,
water absorption is very undesirable in environments in which moisture is avoided, such as in
electronic applications. Proper selection of the counterion can minimize water absorption, yet
still take advantage of the ionic character of the polymer.
(12)Another disadvantage of ionomers is a tendency to exhibit stress relaxation over
time, especially in elastomeric ionomers [8-10]. Under a load ionomers tend to display a high
level of permanent set. This is believed to be due to an ion-hopping mechanism in which the
ions migrate to new sites, which thereby forms new ionic crosslinks. This then prevents the
ionomers from regaining their original dimensions, leaving a high degree of permanent set.
Roles of Ions in Properties of Polymers

(13)There are a wide range of polymers that can be used as the base polymer for ion-
containing polymers. The incorporation of ionic or potentially ionic units is very amenable to
the chemistry used to make them. Structurally the location of the ionic segments in polymers
can be varied based upon the chemistry used to make the polymer. The locations of the ionic
groups includes: mono or telechelic, block, random, and combinations or variations of such.
The main polymer chains are typically linear, star, hyperbranched, or dendrimeric. However,
from a commercial standpoint linear polymers are preferred.
(14)There has been significant progress in the past 40-50 years as to determine what is
happening in ionomers that causes such a dramatic change in properties with slight changes in
the concentration of ionic units. Much of these changes in properties centers on the formation
of multiplets and/or clusters. These two are much more complicated than just serving as
crosslinks. However, multiplets and clusters are not always possible in polymeric systems.
These are highly dependent on the polarity of the polymer structure.
(15)The generally accepted model for ion multiplets and clusters is based upon the
Eisenberg-Hird-Moore (EHM) model [4], which is a core-shell structure (Figure 1). In this
model the ions coordinate in the center of the shell, and the polymer segments in the
immediate vicinity have restricted mobility. As the number of segments away from the center
increases the mobility increases. Hence, what are formed are portions of the polymer with
distinct properties from one another. As the number and location of these multiplets increases,
regions of greater restricted mobility increases. This leads to cluster formation in which
properties such as a well-defined cluster Tg can result. These regions essentially act as
separate phases within the polymer system. Such cluster formations leads to stronger virtual
crosslinks than what is achieved with multiplets. And, like multiplets, cluster formations are
able to be broken and reformed.
(16)When the ion content becomes sufficiently high, which is dependent on the polymer
structure and counterion, the ionomer can become brittle. The restriction in mobility does not
allow the polymer chains to disentangle or release pressure, which leads to a structure that has
little toughness. Therefore, controlling the upper limit of ionic content is necessary in order to
reap the maximum benefits of its incorporation into the polymer.
Research in Ionic Polymers

(17)As mentioned previously among the many uses of ionomers is that they can act as
structural materials by themselves, be used as a matrix to incorporate fillers, or benefit by
incorporation as a compatibilizing agent between two incompatible polymers as part of a
blend. By far most of the research has been pursued in studying the ionomers by themselves.
There has been a lot of research in the past few years in incorporation of fillers, such as clays,
to create new nanocomposites. Also, the incorporation of ionomers into two highly
incompatible polymers as a polymer blend has garnered much interest of late.
(18)Care must be taken when comparing ionomers vs. non ion-containing polymers. For
example, when comparing the properties of ion-containing sulfonated polystyrene it is best to
compare it with sulfonated polystyrene rather than comparing it to polystyrene. There is the
possibility that simple functionalization of the polymer, rather than the fact that it contains
ions, is the reason for the change in properties. Although the ions in these polymers impart
changes in the properties of the overall polymer, it is also possible that the non-ionic portions
also cause similar changes. This is important due to the fact that an additional step can be
avoided should the conversion to an ionomer not be necessary. Simple functionalization may
give the desired properties by themselves.
(19)A good example of comparing the acid neutralized vs. ion-functionlized polymer is
poly(ethylene-co-methacrylic acid). This polymer is commercialized under the name Nucrel®
by DuPont. When Nucrel® is neutralized to various degrees using cations, such as Li+, Na+,
K+, or Zn2+, it falls under the name Surlyn®. There has been a tendency to compare Surlyn®
directly to polyethylene (either low density or high density). However, when one compares
the properties of Nucrel® and Surlyn® there are many similarities between the two. The
methacrylic acid segments cause the overall crystallinity of the polymer to be reduced relative
to polyethylene. In fact, there is evidence that the methacrylic acid segments interact with one
another in a similar fashion as the ionic segments of Surlyn® to form multiplets [14-16].
Therefore, Surlyn® should be compared to other ionomeric polyethylene copolymers, or to
the respective acid-neutralized analogs, rather than to polyethylene itself.
(20)A wide range of polymers can be considered as structural materials. Perhaps the most
well-known ionomeric structural polymer is Surlyn®. Due to the number of research groups
that have studied Surlyn® over the years there has been a wealth of information gathered that
can be translated to other ionomeric systems. Selection of the valency of the metal salt
counterion plays an important role in the final properties of this polymer. The fact that
multiplets and clusters form in this polymer [17-22] shows that there are regions of restricted
mobility in the polymer. This leads to a separate phase in the polymer that displays its own
separate properties from the bulk polymer.
(21)Surlyn® has applications as golf-ball covers due to the impact/scratch resistance,
clarity, and water-resistance of the polymer. The success of Surlyn® shows that ion-
containing polymers can be commercialized as long as there is a need for polymers that have
those specific properties. This work transitions into other research that has taken place which
has potential for commercialization.
66 Daniel J. Klein
Ionomers as Stand-Alone Polymers

(22) Several research groups have synthesized and studied ionomeric polyurethanes.
Polyurethanes are a common class of commercial polymers that can have their properties
altered through proper selection of the properties of the soft and hard segments, such as
molecular weight and chemical structure. Typically the potentially ionic segments are
incorporated into polyurethanes through a modified chain-extending unit. Polyurethanes have
an advantage in that these polymers can easily be modified to have either cationic or anionic
counterions. Selection of the molecular weights of the hard and soft segments allows for an
adjustment of regularly spaced ionic sites along the polyurethane backbone. This allows for a
control of the concentration of ionic groups in the polymer.
(23)”Anionomeric” polyurethanes were prepared from polypropylene glycol,
dimethylolpropionic acid, and isophorone diisocyanate where polymers were neutralized
using triethylamine (TEA) or N-methyldiethanolamine (MDEA) [23]. The tensile properties
were determined on these two ionomers, and then subsequently compared to the non-
neutralized analog (Figure 2). The researchers claimed that there were no significant
differences between the polymers. It was found that the ionomer using TEA as the counterion
had slightly higher tensile strength and elongation than when MDEA was the counterion. This
may have been due to the differences in strengths of the ionic interactions of the two systems.
Using DMTA it was found that the ionomers both had greater phase separation than the non-
neutralized analog.
Figure 2. The effect of counterion on the tensile properties of polyurethanes.
(24)In similar work, polyurethanes “cationomers” were synthesized in a similar approach

to the previous example. The chain extender for these polyurethanes was methylene
diisocyanate, with poly(caprolactone) glycol being the comonomer [24,25]. The resulting
polymer was converted to the ion-containing polymer through addition of MDEA. It was
stated that the ionization causes a reduction in the order of the hard segments. As a result, it
was found that the tensile strengths and tensile moduli increased upon ionization (table 1). It
was concluded that ionization increases the hard domain cohesion. Also, it was shown that by
increasing the molecular weight between the urethane and the tertiary nitrogen groups that the
tensile strength, tensile modulus, and Shore A hardness decreased, and the ultimate
elongation increased. This is a reflection of the decrease in the ionic crosslinking upon the
decrease in the density of ionic sites. This work also showed that increased ionic content led
to increased water uptake. This resulted in a decrease in the tensile properties, but could be
regained through removal of the water.
Table 1. The effect of the degree of quaternization on the physical properties

of polyurethanes
Tensile Tensile
Tensile modulus at modulus at
Elongation Shore A
DQ1 strength 100% 300%
(%) hardness
(MPa) elongation elongation
(MPa) (MPa)
10 4.64 0.77 1.55 1279 48
20 7.47 1.81 2.46 1082 57
30 19.28 2.21 3.86 998 65
40 20.56 2.64 4.64 836 69
60 21.87 5.04 5.71 788 73
80 24.96 5.38 8.07 770 78
100 29.31 6.49 9.68 714 82
1
degree of quaternization
Table 2. Molecular weight and molecular weight distribution data of polyurethanes
Chain extender MW (g/mol) MWD

ED 27 x 104 1.7
MDA 8 x 104 1.2
BDDS-0.21 21 x 104 1.5
BDDS-0.81 9 x 104 1.3
BDDS-1.41 7 x 104 1.3
BDDS-2.01 4 x 104 1.3
BDDS-1.4-ArgMe2 7.4 x 104 1.3
BDDS-1.4-Asp2 7.5 x 104 1.3
BDDS-1.4-Gly2 7.6 x 104 1.3
BDDS-1.4-Lys2 6.1 x 104 1.3
1
sulfonyl content was 0.2, 0.8, 1.4, and 2.0 wt%
2
carboxyl content was 1.4 wt%
(25) In another study [26], segmented ion-containing polyurethanes were prepared using
different hard segment lengths. The chain-extending unit was the disodium salt of 4,4΄-
diamino-2,2΄-biphenyldisulfonic acid (BDDS). The non-ionic chain extenders that were used
were methylenedianiline (MDA) and ethylenediamine (ED). Table 2 lists the molecular
68 Daniel J. Klein
weights and molecular weight distributions of the polymers in this study. It was found that the
tensile strength could be enhanced to about double of the polymers containing the non-ionic
chain extenders, with a 2-30 fold increase in tensile modulus (table 3). Increasing ionic
content led to decreasing ultimate elongations, which was due to increasing levels of ionic
crosslinking. It was speculated that the decrease in tensile strength from BDDS-1.4 to BDDS-
2.0 was due to molecular weight differences.
Table 3. Mechanical properties of polyurethanes
Tensile modulus at Ultimate

Tensile Strength
Chain extender 100% elongation elongation (%), ±
(MPa), ± 10%
(MPa), ± 15% 12%
ED 12.8 3 1340
MDA 10.1 34 365
BDDS-0.21 ND3 6 >3900
BDDS-0.81 6.0 8 750
BDDS-1.41 28 60 380
BDDS-2.01 11 100 240
BDDS-1.4-ArgMe2 26 18 450
BDDS-1.4-Asp2 36 140 390
BDDS-1.4-Gly2 34 100 400
BDDS-1.4-Lys2 37 32 380
1
sulfonyl content was 0.2, 0.8, 1.4, and 2.0 wt%
2
carboxyl content was 1.4 wt%
3
not determined
(26)In the same study the ionic groups in the ionic polyurethanes were also varied to
determine the effect of sulfonic and carboxylic groups in many polymeric systems. The
greater polarity of the sulfonyl group vs. the carboxyl groups should show several differences
in the properties of the resultant polymers. Four different amino acids were incorporated into
the polyurethanes, each of which contains carboxyl groups. It was found that at equivalent ion
contents that the ones containing carboxyl groups had higher tensile moduli for the Asp and
Gly, but not for the ArgMe and Lys, when compared to the ones containing sulfonyl groups
(BDDS-1.4) at equivalent ion contents. The researcherrs concluded that the long hydrocarbon
chains of the ArgMe and Lys groups hinder close packing of ion aggregates, thereby lowering
the tensile moduli. However, the results of this work show that the tensile properties of the
polyurethanes can be significantly improved through incorporation of a small percentage
(1.4%) of ionic groups.
(27)To generalize that the same carboxyl vs. sulfonyl ionic trends holds true for all
polymers would be incorrect. On the one hand it was found that there are dramatic differences
in the properties of carboxylic acid vs. sulfonic acid-containing polystyrene [27]. However,
this dramatic effect was not seen in the properties of ionomeric poly(styrene-ethylene-
butadiene) [28]. Therefore, it appears that the trends should be determined for each class of
ionomers.
(28)Ionomeric polystyrenes have been studied for well over 50 years, with the first patent
reported in 1954 [29]. Although these polymers are rarely studied for their stand-alone
properties, much insight can be gained through a study of these systems. Sulfonyl groups can
be incorporated either through a sulfonation reaction of polystyrene, or through
copolymerization with styrene sulfonate. Pure poly(styrene sulfonate) is a brittle polymer.
With decreasing levels of ionic content this polymer gains structural strength. In essence, at
less than 100% sulfonation of polystyrene the polymer can be referred to as a polystyrene
copolymer (a copolymer of styrene and styrene sulfonate).
(29)The addition of small amounts of ionic content into polystyrene has a very significant
effect on the physical properties of the polymer, such as increased toughness, fatigue
resistance, and improved tensile properties. As mentioned previously the level of crystallinity
tends to decrease as ions are incorporated into the polymer. The same holds true for ion-
containing polystyrene. Incorporation of sulfonate groups into polystyrene does restrict chain
mobility, which causes a reduction in crystallinity [30]. However, the neutralization of the
sulfonic acid groups further restricts chain mobility, which causes an even further reduction in
the crystallization rate. Also, the crystal morphologies can vary from those of pure
polystyrene. This causes a decrease in the melting point of the ion-containing polystyrene
relative to the pure polystyrene. The decrease in number and type of crystals in ionomeric
polystyrene means that the optical clarity should increase relative to polystyrene
(30)In terms of physical properties of sulfonated polystyrenes, it was found that the
tensile strengths of thin films increased up to approximately 7 mol% Na+ ionic content before
its properties started to decrease (Figure 3) [31]. The authors also measured the toughness of
the polymer with increasing ion content. The maximum was reached at the same level as the
maximum tensile strength (Figure 4). This percentage is approximately the same as found in
other studies [32-33]. This level appears to be the critical ion level in which cluster formation
starts. The toughness of the 7% ion-containing polystyrene neutralized using Na+ was
approximately two times that of polystyrene. This increase in properties is very attractive in
applications using polystyrene in which increased mechanical properties are desirable.
Figure 3. The effect of ion content on the tensile strength of polystyrene ionomers.
70 Daniel J. Klein
Figure 4. The effect of ion content on the toughness of polystyrene ionomers.
Figure 5. The effect of ion content on the cycles to initiate damage (Ni) and cycles to fracture (Nf) of
Na+ neutralized sulfonated polystyrene ionomers.
(31)Fatigue resistance is an important property in polymers. It was believed that

incorporation of ionic groups into polystyrene should increase fatigue resistance. This stems
from the ionic groups forming a type of crosslinking that would stabilize crack formation
when it occurs. From a fatigue standpoint it was found that the fatigue resistance increases
with increasing ion content (Figure 5) [34]. At lower ion contents, in which multiplets are
dominant, the fatigue resistance improvement was lower than when clusters were able to
form. The clusters were able to form at slightly higher concentrations than when multiplets
are formed, which fell at a concentration similar to that mentioned in the previous paragraph.
Also, a study of divalent cations revealed that, at equivalent concentrations as the monovalent
cations, the properties were significantly improved. DMTA analyses revealed that ionomers
based upon Ca2+ had a larger rubber plateau modulus, indicative of greater ionic crosslinking,
than those from K+ and Cs+ (Figure 6) [35]. Thus, it was found that the ionomers using Ca2+
as the counterion had three times the fatigue lifetime of polystyrene. Hence, proper selection
of ion concentration in polystyrene maximizes fatigue resistance, thereby improving
polystyrene as a structural material.
Figure 6. The effect of counterion on the storage modulus of polystyrene ionomers.
Figure 7. Stress-strain curves for K+, Mg2+, Zn2+, and Zr4+ neutralized sulfonated polyisoprene
ionomers.
72 Daniel J. Klein
(32)A study of telechelic ionomers based upon polyisoprene was conducted in which a
comparison was made between carboxylate and sulfonate groups [36]. In this work it was
found that the polymers based upon sulfonate salts were several orders of magnitude greater
than the respective carboxylate salts in terms of tensile stress (Figure 7 and 8). The
counterions used in this study were K+, Mg2+, Zn2+, and Zr4+. Polymers based upon the K+
counterion led to the greatest tensile strengths in this study. Although the Mg2+ counterion is
smaller and divalent, whereas the K+ is larger and monovalent, polymers based upon the Mg2+
counterion had lower tensile strengths at equivalent stresses. It was speculated that the Mg2+
counterion would lead to smaller, more stable ionic domains. This was reflected in the longer
time creep and stress relaxation experiments. Also, it was found that the sulfonated ionomers
exhibited stronger aggregation than the carboxylated analog.
Figure 8. Stress-strain curves of K+, Mg2+, and Zn2+ neutralized carboxylated polyisoprene ionomers.
(33)In a study of carboxy-telechelic ionomers based upon polyisoprene, the carboxy

groups were neutralized with various divalent cations [37]. Among these were two alkaline
earth cations (Ca2+ and Sr2+) and three rare earth cations (Ni2+, Zn2+, and Cd2+). The ionomers
neutralized using the alkaline earth cations displayed the highest tensile moduli and tensile
strengths, and the lowest percent elongations in the series (Figure 9a and 9b). Through SAXS
studies it was concluded that the alkaline earth cations lead to polymers with larger
aggregates than the rare earth cations. Researchers speculated that the polymer chains form
loops when the two aggregates merge. Because the alkaline earth cations can accommodate
larger numbers of cations due to the larger core radii of both cations compared to the rare
earth cations, this leads to larger numbers of entanglements [38]. This translates into the
higher tensile moduli for the ionomers from the alkaline earth cations compared to the rare
earth cations. EXAFS confirmed that the alkaline earth cations formed more cohesive ionic
microdomains. This caused strain hardening of these two polymers, whereas the polymers
using the rare earth cations could relax through ion-hopping. Hence, the authors showed a
significant difference in tensile properties through proper selection of the counterion.
Figure 9. Stress-strain curves of carboxy-telechelic polyisoprenes; (a) neutralized using Ca2+ and Sr2+;
(b) neutralized using Ni2+, Zn2+, and Cd2+
(34)Polyesters are a commercially important class of polymers that have also been
studied as ionomers. Research on polyester ionomers has focused on the random and
telechelic incorporation of ionic groups. One study involved the incorporation of ionic
segments into poly(ethylene terephthalate) (PET). The use of 5-(sodiosulfo)isophthalate (5-
SSI) led to direct incorporation of ionic groups without further modification of the polymer
[39]. In this study the ion content was varied from 0 to 9 mol% 5-SSI groups. The Tgs of the
ionic and the acid-neutralized analogs were identical, and the increasing concentration of the
5-SSI groups had little effect on the Tgs. Variation in the Tgs ranged only ± 4.5 ºC, but did
show a minimum at 4 mol% 5-SSI groups (Figure 10), after which point the Tg increased.
Incorporation of the 5-SSI groups did decrease the crystallinity and crystallization rate
relative to unmodified PET, but increased the crystallization temperature. Also, the melting
points decreased with increasing 5-SSI content. Although no physical properties were
reported, one would expect beneficial properties typical of ionomers to be extended to this
system, which could bring valuable benefits to PET, such as improved tensile properties and
optical clarity.
(35)Ionomeric polyester liquid crystalline polymers have also been synthesized in order
to increase the tensile and compressive properties relative to the nonionic analog. In this
system the polymer was based upon 4-hydroxybenzoic acid, 6-hydroxy-2-naphthoic acid, and
5-SSI. The divalent cations Mg2+, Ca2+, Ba2+, and Zn2+ were used as the counterions [40]. The
tensile properties of the resulting ionomers and the acid-neutralized analog are listed in table
4. It was speculated that the poor properties of the ionomers based on Zn2+ and Mg2+ were
due to the low molecular weight, as evidenced by the low intrinsic viscosity values. At
equivalent concentrations the Ca2+ ion is more effective than the Na+ ion in increasing the
physical properties. Such behavior has been documented by several researchers [32,41]. The
increase in the physical properties was explained through SEM analysis, which showed that
the non-ionic polymer failed at localized locations, while the ionic analogs showed no
localized failure. These marked improvements in the properties of the polymer relative to the
74 Daniel J. Klein
acid-neutralized analog are good incentives to benefit from a change to the ionomeric form of
this polymer system.
Figure 10. The effect of ion content on the melting point of Na+ neutralized sulfonated PET.
Table 4. Mechanical properties and intrinsic viscosities of ionomeric liquid

crystalline polymers
Tensile Strength Tensile Modulus Ultimate Intrinsic viscosity

Cation
(MPa) (GPa) elongation (%) (dL/g)
None 142 13.3 1.27 5.12
Na+ 162 14.9 1.22 5.66
Mg2+ 106 16.4 0.61 2.50
Ca2+ 351 23.4 1.59 4.83
Ba2+ ND1 ND1 ND1 2.24
Zn2+ 86 10.0 0.85 3.88
1
ND = not determined because films were brittle
Ionomers in Nanocomposites
(36)The field of nanocomposites has gathered much interest in the use of ionomers as
compatibilizing agents. Typically non ion-containing polymers have been used in dispersing
the fillers such as layered silicates. These layered silicates are typically of the monmorillonite
(MMT) variety. The key to obtaining a good dispersion is to maximize exfoliation of the filler
particles. Although techniques do exist in which this can be attained without using any
addition of ionomeric compatibilizing agents, these techniques tend to be complex. Such
other techniques include solution intercalation and in situ polymerization techniques, along
with traditional melt intercalation.
(37)One approach that can be used to get good exfoliation of the filler is to use ionomers.
The use of such polymers should take advantage of the charged surfaces of many different
types of fillers. The interactions between the ionomers and the charged filler surfaces are
reasons why ionomers are good candidates for exfoliation of these fillers. Along these lines,
the use of crystalline polymers is also beneficial to take advantage of the thermal stability of
such structures.
(37’)One such ionomer is Na+ neutralized sulfonated poly(butylene terephthalate) (PBT).
Studies on this ionomer has shown that the ionic groups aggregate to form ionic domains
[42]. The incorporation of sulfonate groups was easily obtained through incorporation of the
dimethyl-5-sodioisophthalate monomer [43]. Significant increases in the tensile moduli were
observed when the MMT contained alkylammonium ions, and the polymer contained various
levels of sodium sulfonate groups (table 5). This modulus increased with increasing sodium
sulfonate concentration. However, this increase in moduli was not seen when the clay
contained sodium ions. It was determined that the alkylammonium-modified clays have
smaller particle sizes than the sodium-modified clays. This leads to greater interactions with
the polymer, which translates into a higher amount of energy required to deform the
nanocomposite. The high degree of exfoliation was determined through TEM analysis.
Table 5. Young’s Moduli of non-ionic and ionomeric PBT
Clay PBT (psi) PBT-3% SO3-Na+ (psi) PBT-5% SO3-Na+ (psi)

None 169000 ± 7500 165000 ± 7400 157000 ± 5800
Na+ MMT 182000 ± 50 178000 ± 5100 178000 ± 2000
R4N+ MMT 196000 ± 3800 210000 ± 7300 215000 ± 7600
(37a)Ionomeric sulfonated PET/MMT nanocomposites were also recently prepared to

take advantage of the benefits of incorporation of MMT [44]. The ionic segments were
incorporated through addition of dimethyl-5-sodiosulfoisophthalate during the synthesis of
the polymer. The levels of sulfonation were controlled from 0-8 mol%. Nanocomposites were
prepared using 5 wt% MMT. Increasing levels of sulfonate groups led to better dispersion of
MMT. Interestingly it was found that the ionomer alone showed no evidence of
crystallization, but with the addition of MMT the ability to crystallize returned. The
researchers speculated that the MMT could be acting as a nucleation site in the system, or due
to a combination of the MMT with the polymer to form the nucleation site. Also, it was found
that addition of MMT increased the thermal stability of the polymer relative to the unfilled
analog (Table 6).
Table 6. Decomposition temperatures of ionomeric PET and ionomeric

PET/MMT nanocomposites
Sample Td, 5% (°C) Td, 10% (°C) Wt. at 600 °C

SPET2a 299 327 0.0
SPET6b 258 288 1.4
SPET2M5c 324 344 6.9
SPET6M5d 343 357 11.8
a
sulfonated PET, 2 mol%
b
sulfonated PET, 6 mol%
c
sulfonated PET, 2 mol%; 5% MMT
d
sulfonated PET, 6 mol%; 5% MMT
76 Daniel J. Klein
(38)A recent study of polyurethane ionomers based on poly(tetramethylene oxide)/4,4´-

diphenylmethylene diisocyanate which contained 3 mol% quaternary ammonium groups was
used to incorporate MMT [45]. Simple blending of the non-ionomeric polyurethane (PU) did
not lead to exfoliated nanocomposites, as evidenced by WAXS. Incoroporation of the
polyurethane ionomer (PUC) led to apparent exfoliation of the filler, as determined using
WAXS and TEM. This led to nanocomposites that displayed an increase in moduli with
increasing MMT content. As is typical with most nanocomposites of this nature, there was a
subsequent decrease in the ultimate elongation (table 7). The authors concluded that the
changes in the tensile properties of the nanocomposites upon use of PUC was due to its strong
interaction with the exfoliated MMT layers, which resulted in a less phase-separated
morphology.
Table 7. Mechanical properties of ionomeric polyurethanes/MMT nanocomposites
Maximum stress Elongation at Young’s modulus

Sample
(kgf/mm2) break (%) (kgf/mm2)
PU 4.88 806 4.11
PUC 4.90 824 4.15
1 wt% MMT/PUC 4.93 760 5.60
3 wt% MMT/PUC 3.86 568 6.26
5 wt% MMT/PUC 3.66 495 7.20
7 wt% MMT/PUC 3.45 487 9.43

Figure 11. The effect of MMT(HT2) concentration on the mechanical properties of Surlyn®/MMT
nanocomposites. (BUR = ratio of the diameter of the film to the diameter of the die. MD = machine
direction. TD = transverse direction.)
78 Daniel J. Klein
Table 8. Mechanical properties of LDPE, Nucrel®, and Surlyn®/MMT nanocomposites
Tensile Tensile
Organoclay MMT Relative Elongation
Polymer modulus strength
type (wt%) modulus1 (%)
(MPa) (MPa)
LDPE None 0.0 114 1.00 13.6 108
HT1 2.5 155 1.36 14.4 87
5.0 172 1.51 14.4 80
7.5 194 1.70 14.1 73
10.0 218 1.91 14.0 67
HT2 2.5 178 1.56 14.2 83
5.0 227 1.99 14.3 77
7.5 280 2.46 14.3 70
10.0 375 3.29 14.2 62
Nucrel®2 None 0.0 118 1.00 13.9 136
HT1 2.5 151 1.31 14.3 120
5.0 180 1.52 14.5 108
7.5 220 1.86 14.9 99
10.0 260 2.20 15.2 90
HT2 2.5 189 1.60 14.7 111
5.0 259 2.20 16.5 99
7.5 328 2.78 17.5 91
10.0 425 3.60 18.0 82
Nucrel®3 None 0.0 73 1.00 15.4 185
HT1 2.5 112 1.53 16.1 176
5.0 133 1.82 16.7 165
7.5 178 2.44 17.4 148
10.0 220 3.01 18.5 133
HT2 2.5 147 2.01 17.8 156
5.0 203 2.78 19.1 143
7.5 254 3.48 20.6 134
10.0 353 4.83 22.2 120
Surlyn®4 None 0.0 262 1.00 21.3 194
HT1 2.5 349 1.33 21.2 117
5.0 410 1.56 21.0 111
7.5 465 1.77 22.1 119
10.0 563 2.15 23.2 116
HT2 2.5 403 1.54 22.3 127
5.0 560 2.14 23.8 111
7.5 732 2.79 26.6 72
10.0 919 3.51 29.4 65
1
Relative modulus = modulus with organoclay/modulus without organoclay
2
Contains 3.9 wt% methacrylic acid
3
4
(39)Surlyn® has also been used as the matrix polymer for dispersing clays. In this specific
study the authors tested the properties of various blow-molded Surlyn®/MMT
nanocomposites at different clay concentrations [46]. The clays that were used in this study
were both single (HT1) and dual tail hydrogenated tallow oil (HT2). It was found that the
tensile moduli increase as the MMT concentration increases. The tensile strength and
elongation do not change much in the concentration range tested in this study (Figure 11).
The authors compare the Surlyn®/MMT nanocomposites vs. LDPE/MMT nanocomposites.
Due to the structural differences between these polymers these systems should not be directly
compared to one another. The high crystalline content of LDPE is broken up by the
incorporation of methacrylic acid monomeric units. A more direct comparison should be
made between Surlyn®/MMT nanocomposites vs. Nucrel®/MMT nanocomposites. However,
favorable results showing increased tensile moduli, tensile strengths, and improved impact
resistance bode well as candidates for structural materials.
(40)Subsequent work by the same researchers did compare two grades of Nucrel® vs.
Surlyn® and LDPE as matrices for nanocomposites [47]. The Nucrels® used in this study
contained 3.9 and 8.9 wt% methacrylic acid groups, while Surlyn® contained 15.2%
methacrylic acid groups. The clays that were used in this study were both single and dual tail
hydrogenated tallow oil. It was found that exfoliation increased when a dual tail HT was used
vs. single tail HT clay. The tensile strengths and tensile moduli increased, and elongation
decreased, with increasing levels of clay (table 8). In the two Nucrels® that were studied the
tensile modulus increased with increasing methacrylic acid content. It was suggested that
Surlyn® exfoliated the clays the most due to the largest particle aspect ratios in the series.
Conversion of this Surlyn® to the acid-neutralized analog, and subsequent comparison to
Surlyn®, would have shown whether the ions were responsible for the greater exfoliation.
However, the lesson learned was that the tensile strengths and moduli could markedly be
increased through proper selection of polymer matrix, and selection of methacrylic acid
content.
Figure 12. The effect of MMT on the total fracture energy per unit area.
(41)Nanocomposites using Surlyn® 8945, a Na+ neutralized version of Surlyn®, as the

matrix were also studied to determine the improvement in fracture toughness relative to the
unfilled analog [48]. In this work the fracture toughness was maximized through proper
selection of the clay loading. What was found was that initially the fracture toughness
80 Daniel J. Klein
increases quickly with the addition of clay, but then decreases slowly as the concentration
continues to increase (Figure 12). At MMT concentrations of greater of 2.5% the fracture
energy starts to decrease. At MMT concentrations of 5% and greater the nanocomposites
display brittle failure. In applications where an increase in stiffness is required, these
nanocomposites, with low MMT concentrations, would be good candidates. The negative
associated with this system is that the addition of the clay causes a reduction in the overall
ultimate elongation and increase in brittleness relative to the unfilled system.
(42)In terms of tensile properties of Surlyn® nanocomposites, a series of three different
Surlyns® neutralized using three different cations were studied [49]. All three of the Surlyns®
contained similar methacrylic acid contents. The three cations that were studied were Na+,
Li+, and Zn2+. Analysis of the Surlyn®/clay nanocomposites revealed that the Surlyn® using
Li+ as the counterion had the worst exfoliation in the series. In each series the tensile moduli
and tensile strengths increased with increasing clay content, but the elongation decreased in
the same series (table 9). Because the Li+ neutralized Surlyn® did not exfoliate the clay
particles completely, the tensile properties reflected numbers closer to the matrix resin than
the other two polymers in the series. The Surlyn® neutralized using the Zn2+ counterion
yielded the nanocomposite with the highest relative modulus. Although no clear conclusion
was reached as to why this is the case, the authors do provide some possible reasons as to
why this occurs. It was speculated that the Zn2+ neutralized ionomer causes anhydride
formation, but not those using Na+ and Li+ cations.
Table 9. Mechanical properties of Surlyn®/MMT nanocomposites
Tensile Tensile
MMT Relative Elongation
Cation modulus strength
(wt%) modulusa (%)
(MPa) (MPa)
Zn2+ 0.0 176 1.00 172 19.3
Zn2+ 2.5 314 1.78 116 22.2
Zn2+ 5.0 447 2.50 86 24.2
Zn2+ 10.0 795 4.51 59 29.2
Na+ 0.0 260 1.00 194 21.0
Na+ 2.5 412 1.58 130 22.6
Na+ 5.0 568 2.18 119 25.9
Na+ 10.0 908 3.49 66 28.8
Li+ 0.0 292 1.00 136 21.0
Li+ 2.5 407 1.38 116 24.2
Li+ 5.0 491 1.68 104 24.8
Li+ 10.0 676 2.32 98 27.2
a
Relative modulus = (modulus at MMT wt% > 0%)/(modulus without MMT)
(43)Rubbers have also been used as matrices in nanocomposites. One such polymeric
system is based on poly(isobutylene-co-isoprene) (BIIR). In this work the BIIR matrix was
used to incorporate ion-exchanged MMT (NR+-MMT) and precipitated silica [50]. The BIIR
polymers were modified to make the triphenylphosphonium bromide salts (IIR-PPh3Br). The
researchers hypothesized that the quaternary phosphonium cations could displace the
ammonium ions of the modified clays. This could then lead to an interaction with the
particles, which would lead to exfoliation of the NR+-MMT. Incorporation of NR+-MMT into
the non ion-containing polymer that was crosslinked using ZnO did not lead to any significant
reinforcement of the system. Use of the ionomeric analog led to an increase in both the tensile
strengths and tensile moduli of the system (table 10). As noted by these researchers, and
which is common throughout ion-containing polymeric systems, this system exhibited stress
relaxation. It was also noted that this stress relaxation was not improved through
incorporation of NR+-MMT.
(44)As a continuation of this work the ionomer was used to incorporate precipitated
silica. The researchers indicated that there were no signs of silica agglomeration at any of the
tested concentrations. Based upon the tensile properties it appeared that the ion-containing
analog had a greater degree of exfoliation than the non ion-containing polymer. Strain
relaxation analysis, which involves the determination of the storage modulus as a function of
amplitude, revealed that the non ion-containing polymer containing silica had agglomeration
of the silica particles, while that ion-containing polymer did not.
Table 10. Mechanical properties of cured BIIR/NR+-MMT and

IIR-PPh3Br/NR+-MMT nanocomposites
Tensile Tensile
Filler level
Polymer/filler strength Modulus Elongation (%)
(wt%)
(MPa) (MPa)
IIR-PPh3Br/NR+- 0 5.0 0.55 320
MMT 3 4.4 0.64 280
5 8.1 0.79 235
15 5.4 2.15 78
BIIR-ZnO/ NR+- 0 3.1 0.33 375
MMT 5 1.8 0.43 285
15 1.2 0.65 285
IIR-PPh3Br/silica 0 5.0 0.55 320
15 3.6 1.18 270
30 3.9 3.60 60
BIIR-ZnO/silica 0 3.1 0.33 330
15 2.3 0.64 280
30 3.7 2.90 200
(45)An ion-containing poly(ethylene-graft-maleic anhydride) was used to incorporate

silica and MMT [51]. TEM analysis clearly showed the exfoliation of the particles. This is
very unlike standard HDPE, which led to silica aggregates rather than platelets. In both
versions of the nanocomposites the tensile strengths and tensile moduli increased with
increasing filler levels (table 11). A reduction in the ultimate elongation was seen with
increasing levels of filler, which was expected. A comparison of the nanocomposites using
MMT vs. silica fillers showed that those from silica displayed greater ductility than those
from MMT. The authors concluded that the nanocomposites using silica provided better
overall melt viscosity, stiffness, and mode of failure than those using MMT.
(46)Wood flour, a common filler used in the wood plastics industry, has been used as a
filler in a HDPE matrix to form nanocomposites. Four different grades of Surlyn® were used
82 Daniel J. Klein
as the compatibilizing agents [52]. Two of the Surlyns® were Zn2+ neutralized, and two were
Na+ neutralized. Use of the Na+ neutralized ionomers led to an improvement in the tensile
properties at concentrations above 8%. At this point it was found that the mode of failure
occurred along the shear plane. This is in contrast to lower concentrations of this ionomers in
which the mode of failure was similar to that of HDPE/wood flour blends in which the
polymers cracked. The ionomers neutralized using Zn2+ displayed significant improvement in
the mechanical properties at concentrations of 2%. Whereas the modulus of elasticity
decreased with increasing ionomer content, the modulus of rupture increased. This implies
that there is an increase in toughness of the composite due to increased compatibility. In terms
of toughness it was found that the Na+ neutralized ionomers displayed a better improvement
in toughness compared to the Zn2+ neutralized ionomers. In the Zn2+ neutralized series there
was no evidence of an improvement or deterioration of the toughness with a change in the
concentration.
Table 11. Mechanical properties of ion-containing poly(ethylene-graft-maleic

anhydride)/MMR-NR4+ and SiO2 nanocomposites
Filler level Tensile modulus Yield Stress Elongation

Filler
(wt%) (MPa) (MPa) (%)
MMT-NR4+ 0 290 ± 20 26 ± 1 2190 ± 340
MMT-NR4+ 1 320 ± 10 25 ± 1 1570 ± 230
MMT-NR4+ 5 320 ± 60 27 ± 1 690 ± 310
MMT-NR4+ 9 410 ± 30 30 ± 2 740 ± 300
MMT-NR4+ 13 420 ± 60 27 ± 3 12 ± 1
MMT-NR4+ 29 470 ± 125 29 ± 7 11 ± 5
SiO2 0 30 ± 10 26 ± 1 2190 ± 340
SiO2 1 330 ± 40 27 ± 1 1150 ± 380
SiO2 5 340 ± 20 28 ± 1 1250 ± 290
SiO2 9 360 ± 50 30 ± 1 2010 ± 360
SiO2 13 400 ± 50 31 ± 2 1220 ± 360
(47)Polypropylene (PP)/Vectra B composites were achieved through addition of Zn2+

neutralized Surlyn® [53]. Blending Vectra B with ionomer led to an immiscible blend, as
evidenced by two Tgs using DMTA. Blending PP with ionomer led to a miscible blend, as
determined by the single Tg using DMTA. Tensile testing revealed that the binary blends had
yielding, while the compatibilized ternary blends did not. The higher modulus values of the
ternary blends it indicated that Vectra B was a large contributor to the mechanical response.
Ionomers as Blend Compatibilizers

(48)The field of polymer blends is one in which the work is limited by the compatibility
of the polymers. The low entropy of mixing causes many polymer blends to be
thermodynamically unstable, which leads to low interfacial adhesion [54]. In order to
compatibilize blends, the use of additives that reduce the interfacial tensions of the polymers
are used [55]. In recent years there has been an interest in using various ionomers as the
compatibilizing agents. Due to the commercial availability of several ionomers, this portion
of the field of polymer blends can be pursued relatively quickly. As such, there has been a
significant amount of work published in the past decade using ionomers in this application.
(49)From an industrial standpoint polymer blends are of high interest due to the
availability of the starting polymers. Through proper selection of the polymers used in the
blends, specific properties can be targeted that benefit from the advantages of the polymers
used in the blend. Also, rather than investing in research and development of new polymers,
blending of several commercially available polymers may accomplish the same goals. Hence,
blending of polymers offers an attractive alternative to the synthesis of new polymers.
(50)Polyurethane (PU) was blended with a Zn2+ neutralized poly(ethylene-co-methacrylic
acid-co-isobutyl acrylate) (EMI-Zn) ionomer [56]. Most of the tensile properties of the blends
were lower than that of pure PU (table 12). In one case the ionomer was neutralized with H+
to determine the effect of the ions on the tensile properties. What was found was that that
there was a significant difference between the propeties of the Zn2+ neutralized ionomer vs.
the H+ neutralized analog. This implies that the blend had better compatibility when using an
ionomer vs. the non-ionomeric version. The thermal properties of the blends were not much
different than that of PU, but there were differences in crystallinity. In the case of the 90/10,
70/30, and 50/50 blends the crystallinity increased relative to PU.
Table 12. Mechanical and crystallinity data of PU/EMI-Zn blends
Tensile Tensile
PU/EMI-Zn strength
Elongation at
modulus χc, PU χc, EMI-
break (%) (%) Zn (%)
(MPa) (J/cm3)
100/0 57 ± 10 731 ± 57 220 ± 20 33.8 -----
90/10 42 ± 7 639 ± 84 142 ± 37 39.3 6.7
70/30 28 ± 3 624 ± 75 98 ± 23 41.1 7.4
70/30a 24 ± 3 518 ± 69 74 ± 15 NDc ND
50/50 32 ± 4 657 ± 66 128 ± 21 39.2 12.2
30/70 27 ± 2 743 ± 39 117 ± 13 18.7 11.7
10/90 25 ± 3 747 ± 55 106 ± 18 ----- 11.1
0/100 31 ± 3 562 ± 48 109 ± 15 ----- 8.8
a
The Zn2+ was replaced with H+
b
χ = percent crystallinity
c
c
ND = not determined
(51)PU was also blended with HDPE using EMI-Zn as the ionomeric compatibilizing
agent [57]. Optimization of the properties was achieved by melt mixing all of the components
for 15 minutes at the same time, except for the 28.3/56.7/15 blend. Longer mixing times led
to a decrease in the properties, which may be due to ionomer agglomeration [58]. The
properties of the blends increased as the PU content increased (table 13). Use of Na+
neutralized EMI led to blends with inferior properties to that containing EMI-Zn. This was
thought to be due to the greater ionic character, which would lead to a decreased amount of
dispersion [59].
84 Daniel J. Klein
Table 13. Mechanical properties of PU/HDPE/EMI-Zn blends
Tensile Tensile
PU/HDPE/EMI- Mixing Elongation at
strength modulus
Zn time (min) break (%)
(MPa) (J/cm3)
66.7/33.3/0 15 9±1 28 ± 6 2.0 ± 0.1
56.7/28.3/15 15 14 ± 1 280 ± 6 46 ± 1
151 13 ± 1 52 ± 16 7±3
20 14 ± 4 209 ± 41 38 ± 7
50/50/0 15 11 ± 2 9±1 1.0 ± 0.1
47.5/47.5/5 15 13 ± 2 93 ± 20 7±3
45/45/10 15 14 ± 1 95 ± 30 11 ± 1
42.5/42.5/15 10 14 ± 1 103 ± 20 13 ± 4
15 14 ± 1 152 ± 30 21 ± 3
20 12 ± 1 144 ± 30 20 ± 4
33.3/66.7/0 15 9±1 8±2 0.4 ± 0.1
28.3/56.7/15 15 15 ± 1 19 ± 2 3±0
20 14 ± 2 185 ± 35 23 ± 7
25 11 ± 1 11 ± 4 2.5 ± 1
1
HDPE added to premixed PU/EMI-Zn
Table 14. Mechanical properties of HDPE/EVOH/Surlyn® blends
Tensile Tensile Tensile

HDPE/EVOH/ Elongation at
strength at strength at modulus
Ionomera break (%)
yield (MPa) break (MPa) (J/cm3)
66.7/33.3/0 ----- 22 ± 1 7±0 1
60/30/10 ----- 17 ± 1 7±2 1
53.3/26.7/20 20 ± 1 17 ± 1 78 ± 21 13
50/50/0 ----- 29± 3 7±0 1
45/45/10 29 ± 2 24 ± 2 37 ± 13 8
40/40/20 25 ± 3 26 ± 4 346 ± 36 78
a
Polymers were conditioned at 60% relative humidity, and quenched at 0 ºC
(52)Several blends using poly(ethylene-co-vinyl alcohol) (EVOH) have been pursued

with several polymers. EVOH offers beneficial gas barrier properties, processability, optical
clarity, and oil resistance [60, 61]. A largely detrimental property of EVOH is its absorption
of water, which causes a reduction in its gas barrier properties. One such blend system that
was attempted to overcome the water absorption issue was EVOH/HDPE [62]. It was
believed that the HDPE portion of the blend would lead to a decrease in the water absorption
ability compared to pure EVOH. In this study the ionomer compatibilizer was Zn2+
neutralized Surlyn®. The ratio of EVOH to HDPE was varied in order to determine the
concentration of ionomer necessary to compatibilize the blend. It was found that at a 2/1 ratio
of HDPE to EVOH that a larger concentration of ionomer was required to compatibilize the
blend than when the ratio was 1/1. However, even at the 1/1 ratio the 20% ionomer had a
dramatic effect on the tensile moduli and elongation values (table 14). Blending EVOH with
ionomer was enhanced through treatment in a humid environment. It is believed that the
water bridges the components to aid in the compatibilization of the two polymers, which is
why the ternary blends were also treated in a humid environment.
(54)EVOH/polypropylene (PP) blends were compatibilized using Surlyn® neutralized
using Na+ [63]. The reasoning behind this polymer blend was for the same reasons listed for
the EVOH/HDPE blends. The ratios of the PP to EVOH of 90/10 and 80/20 yielded the best
ionomer compatilization, which was proven in previous work [64,65]. It was determined that
increasing levels of ionomer did not lead to better tensile properties (table 15). In fact,
increasing levels of ionomer had a detrimental effect, which may be due to the tensile strength
and modulus of Surlyn®. PP/EVOH blends with ionomer displayed a better fracture parameter
than those without ionomer. The EVOH/PP blends, both with and without ionomer, did
increase the ductile-type of fracture, as opposed to PP that is not stable to crack propagation.
Table 15. Mechanical properties of PP/EVOH/Surlyn® blends
Polymer Young’s Yield Deformatio Break Deformatio

(PP/EVOH/ modulus strength n at yield strength n at break
Surlyn®) (GPa) (MPa) (%) (MPa) (%)
24.73 ± 21.74 ± 619.20 ±
100/0/0 1.39 ± 0.11 10.99± 0.46
0.31 0.77 4.70
27.07 ± 13.80 ± 113.80 ±
90/10/0 1.60 ± 0.15 8.34 ± 0.40
0.55 1.83 5.07
27.19 ± 10.63 ± 60.25 ±
90/10/2 1.84 ± 0.29 8.01 ± 0.45
0.53 0.58 13.90
25.75 ± 12.55 ± 92.58 ±
90/10/5 1.62 ± 0.10 8.29 ± 0.52
0.82 1.99 19.36
25.71 ± 282.00 ±
90/10/10 1.46 ± 0.11 8.26 ± 0.18 9.62 ± 0.48
0.28 33.77
27.13 ± 29.26 ±
80/20/0 1.85 ± 0.07 5.48 ± 0.61 8.50 ± 0.40
0.55 0.68
27.10 ± 29.17 ±
80/20/2 1.82 ± 0.09 5.98 ± 0.19 8.82 ± 0.17
0.37 0.62
27.52 ± 11.88 ± 34.25 ±
80/20/5 2.08 ± 0.37 6.54 ± 0.28
0.50 1.87 5.92
27.88 ± 12.92 ± 27.42 ±
80/20/10 1.58 ± 0.19 6.68 ± 0.59
0.29 1.49 5.63
67.80 ± 28.26 ± 20.69 ±
0/100/0 5.03 ± 0.83 5.09 ± 0.14
0.77 2.02 2.45
(30.33 ± 11.66 ± 154.70 ± 11.50 ± 174.40 ±
0/0/100
0.08) x 10-3 0.11 0.86 1.11 5.95
(53)EVOH was blended with poly(ethylene-co-cyclohexane-1,4-dimethanol) (PETG)

using two different versions of Surlyn® [66]. One of version of Surlyn® was neutralized using
Zn2+ and the other using Na+. As in the previous work a certain percent ionomer was required
based upon the ratio of EVOH to PETG. It was found that the ionomer neutralized using Na+
was able to compatibilize the EVOH/PETG blends at a lower concentration (5%) than the
86 Daniel J. Klein
analogous ionomer neutralized using Zn2+ (15%). These findings were supported using tensile
testing (table 16) and DMTA. It was also found that the ionomer neutralized using Na+ had
better barrier properties than that neutralized using Zn2+. The combination of the benefits of
EVOH (oxygen and carbon dioxide gas barrier properties and processing ease) and PETG
(toughness and clarity), along with the inherent properties of the ionomer lead to this high
strength material that targets a specific application.
Table 16. Mechanical properties of PETG/EVOH/Surlyn® blends
Tensile Tensile Tensile

PETG/EVOH/ Elongation at
strength at strength at modulus
Surlyn® break (%)
yield (MPa) break (MPa) (J/cm3)
66.7/33.3/0 ----- 34 ± 4 7±2 1±0
63.3/31.7/5 37 ± 3 36 ± 3 305 ± 60 89 ± 7
60/30/10 28 ± 1 32 ± 3 324 ± 45 78 ± 9
56.7/28.3/15 34 ± 4 33 ± 3 205 ± 45 59 ± 5
53.3/26.7/20 27 ± 1 30 ± 2 265 ± 21 62 ± 4
67/33/0 ----- 35 ± 3 8±2 2±1
63.3/31.7/5 34 ± 2 33 ± 3 311 ± 39 75 ± 12
60/30/10 32 ± 1 31 ± 4 331 ± 39 80 ± 11
56.7/28.3/15 27 ± 1 26 ± 4 287 ± 38 59 ± 10
53.3/26.7/20 26 ± 2 25 ± 4 289 ± 58 62 ± 11
Table 17. Mechanical properties of PET/nylon-6/Surlyn® blends
PET/nylon- Tensile stress Elongation Tensile modulus

6/Surlyn® (MPa) (%) (J/cm3)
100/0/0 46 ± 2 587 ± 41 156 ± 11
0/100/0 36 ± 2 423 ± 31 135 ± 4
50/50/0 42 ± 4 396 ± 43 128 ± 14
48.5/48.5/3 26 ± 2 279 ± 27 65 ± 6
47.5/47.5/5 32 ± 4 357 ± 60 88 ± 3
45/45/10 33 ± 3 333 ± 48 86 ± 6
42.5/42.5/15 33 ± 3 420 ± 38 101 ± 5
37.5/37.5/25 34 ± 2 452 ± 22 99 ± 11
33.3/33.3/33.3 35 ± 5 481 ± 54 117 ± 9
66.7/33.3/0 37 ± 3 413 ± 31 108 ± 6
56.7/28.3/15 38 ± 3 486 ± 65 127 ± 9
33.3/66.7/0 47 ± 5 505 ± 53 149 ± 19
28.3/56.7/15 30 ± 2 347 ± 46 72 ± 7
(55)Two common and well known commercial thermoplastics are PET and nylon-6.
Surlyn® neutralized using Zn2+ was used as the compatibilizing agent when blending these
two polymers [67]. Table 17 shows the mechanical properties of the blends from this work.
Aging studies were conducted on blends of the PET/nylon-6/ionomer in ratios of 50/50/5,
47.5/47.5/5, and 42.5/42.5/15. It was found that at the 5% ionomer level that the tensile
properties decreased over a month (tables 18-20). However, when 15% ionomer was used the
properties remained fairly constant. Also, the thermal properties and percent crystallinity did
not change over a month. These results show that the ionomer can be used as a
compatibilizing agent, and also as a stabilizer of the physical properties of the blends.
Table 18. Tensile stress aging studies of PET/nylon-6/Surlyn® blends
PET/nylon-6/Surlyn® 1 day 1 week 2 weeks 1 month

50/50/5 35 ± 1 36 ± 9 41 ± 1 40 ± 3
47.5/47.5/5 31 ± 3 22 ± 5 24 ± 2 20 ± 4
42.5/42.5/15 33 ± 3 31 ± 7 32 ± 2 33 ± 2
Table 19. % Elongation aging studies of PET/nylon-6/Surlyn® blends

50/50/5 396 ± 43 373 ± 34 327 ± 28 17 ± 3
47.5/47.5/5 323 ± 56 25 ± 9 25 ± 14 15 ± 7
42.5/42.5/15 420 ± 38 373 ± 31 402 ± 47 420 ± 30
Table 20. Tensile modulus aging studies of PET/nylon-6/Surlyn® blends

50/50/5 128 ± 14 103 ± 10 94 ± 8 3±1
47.5/47.5/5 88 ± 3 2±1 1±0 1±0
42.5/42.5/15 101 ± 5 82 ± 6 95 ± 7 118 ± 6
(56)Nylon-6 was blended with LDPE using Surlyn® neutralized using Na+ [68]. Based on
the successes of blending maleic anhydride-grafted polyethylene with nylon-6, this work
sought to use Surlyn® instead of a modified polyethylene. Although no tensile data was
conducted on these blends, the blends using the ionomer as the compatibilizing agent had a
very significant effect using only a small concentration. For example, the size of the dispersed
phase was decreased five-fold through addition of only 0.5 phr ionomer. Also, it was found
that the Na+ neutralized ionomer was more effective than the Zn2+ neutralized ionomer as a
compatibilizer. This was based on the lower dispersed phase size and higher thermal stability
of the resulting blend. The major drawback is that the Na+ neutralized Surlyn® absorbs more
water than the Zn2+ neutralized analog. There was no reported change in the crystallization
rate of the ternary blend relative to pure nylon-6.
(57)Nylon-6/poly(ethylene-co-vinyl acetate) (EVAc) blends were compatibilized using
Na+ neutralized Surlyn® [69]. Marked improvements in the tensile (table 21) and impact
(figure 13) properties were observed through the compatibilization of the blends using
ionomer. At the ionomer levels tested, all exhibited about a 3 times improvement in the
notched impact strength relative to the blend not containing ionomer. Analysis of the impact-
fracture area indicated that the energy was dissipated effectively, which can be attributed to
the good interfacial adhesion between the blend components.
88 Daniel J. Klein
Table 21. Mechanical properties of Nylon-6/EVAc/Surlyn® blends
Nylon- Tensile stress Elongation Tensile modulus

6/EVAc/Surlyn® (MPa) (%) (MPa)
80/20/0 38 264 820
80/19.6/0.4 42 350 862
80/19.2/0.8 40 365 848
80/18.8/1.2 42 382 830
80/18.4/1.6 40 325 825
Figure 13. The effect of concentration of Surlyn® on the impact strength of blends; N20=20% nylon-6;
S=Surlyn® concentration.
(58)Binary blends of poly(ethylene-co-acrylic acid) (PEA) and Zn2+ neutralized

carboxylated nitrile rubber (XNBR) were prepared in an attempt to incorporate rubber and
plastic in a single compatible blend (table 22) [70]. It was believed that the ionic segments
would help compatibilize the typically difficult blending of these two polymers with high
interfacial tension. Conversion of XNBR to the ionomeric form was accomplished through
melt mixing with ZnO and stearic acid. Physical properties of the ionomeric blends were
compared to that of the non-ionomeric blend (table 23). The optimum properties appeared to
be in blends 2 and 3. X-ray analysis determined that the percent crystallinity of these two
blends (5% and 11%, respectively) were less than of the starting polymers XNBR and PEA
(13% and 61%, respectively). The conclusion was that the improvement in properties was due
to the compatible blend rather than an increase in crystallinity.
Table 22. XNBR/PEA blend compositions
Component 1 2 3 4 5 6 7 8
XNBR 100 90 80 70 60 50 0 80
PEA 0 10 20 30 40 50 100 20
ZnO 20 20 20 20 20 20 20 0
Stearic acid 1 1 1 1 1 1 1 1
Table 23. Physical properties of XNBR/PEA blends
Property 1 2 3 4 5 6 7 8
Modulus at 50% Elongation (MPa) 1.6 5.3 6.4 8.8 10.3 ----- ----- 1.1
Modulus at 200% elongation (MPa) 2.9 10.9 10.8 11.3 ----- ----- ----- 1.7
Tensile strength (MPa) 13.8 35.7 26.5 13.3 12.4 12.5 16.2 1.8
Elongation at break (%) 765 650 555 260 76 45 24 670
Tear strength (KN/m) 50.0 88.8 101.6 88.0 70.0 45.6 65.4 24.5
Shore A hardness 57 70 75 85 87 90 90 44
Tension at 100% elongation (%) 10 10 20 45 ----- ----- ----- 50
Figure 14. The effect of ionomer concentration on impact strength of POM/Surlyn® blends.
Figure 15. The effect of ionomer concentration on tensile strength of POM/Surlyn® blends.
90 Daniel J. Klein
Figure 16. The effect of ionomer concentration on the elongation of POM/Surlyn® blends.
Figure 17. The effect of ionomer concentration on the impact strength of POM/MBS/Surlyn® blends
(POM/MBS = 80/20).
(59)An example of tailoring a blend for an application involved the blending of

poly(oxymethylene) (POM) with poly(methyl methacrylate-styrene-butadiene) (MBS) using
Na+ and Zn2+ neutralized Surlyn® [71]. The goals of this work were to reduce the crystallinity
of the POM while maintaining the high creep resistance, fatigue resistance, heat resistance,
and solvent resistance. In the binary blend of POM and Surlyn® it was found that the Zn2+
neutralized Surlyn® was the better compatibilizing agent, leading to better toughening (Figure
14-16). In the ternary blends the Zn2+ neutralized Surlyn® also led to better compatibility
between the components than without Surlyn® (Figure 17-19). From a tensile strength
standpoint the maximum was reached at 15% ionomer. The incorporation of MBS and
ionomer also led to a decrease in the crystallinity of the blend relative to pure POM.
Figure 18. The effect of ionomer concentration on the tensile strength of POM/MBS/Surlyn® blends
(POM/MBS = 80/20).
Figure 19. The effect of ionomer concentration on the elongation of POM/MBS/Surlyn® blends
(POM/MBS = 80/20).
Commentary and Current and Future Directions of Ionomer

Research in the Field of Structural Materials
Much of the current focus on ionomer research has been towards obtaining proton
exchange membranes (PEM) for fuel cell applications. On the one hand this is good for the
field of ionomer research. This potentially will lead to advancements in learning more about
the fundamentals of structure-property relationships in ionomeric structures. Also, new
92 Daniel J. Klein
ionomers may be formed which are not applicable for PEMs, but can be used for other
applications. However, when these new ionomers are formed one must be careful to not give
up on them due to their lack of success in the area of PEMs. Rather, there is a potential that it
can be used in other applications.
There continues to be a lack of commercially available ionomers. Many ionomers exist
that can be used in membrane applications, but cannot be used in dry environments due to
their brittle nature when dry. There continues to be a large interest in ionomers in the field of
dentistry. However, from a high volume industrial point of view there has not been a
significant contribution in the field of ionomers in a very long time.
The use of ionomers in nanocomposites has shown a large degree of success. This can
open the door to future ionomer research into making new ionomers rather than relying on the
standard few commercially available ionomers. Earlier in this chapter there was mention of a
few of polymers that were modified to make them ionic in order to incorporate MMT. In
these cases the polymer was easily amenable to functionalization in order to achieve this
result. Similar techniques can be used to alter other polymeric structures in order to achieve a
high degree of exfoliation of filler particles.
There appears to be a lot of room for advancement in the field of nanocomposites using
ionomers, including the incorporation of nanotubes. There has been significant progress over
the past few years, but it appears that there is a lot of room to grow and advance. The large
volume of research in basic ionomer structures should be helpful in the advancement of the
use of ionomers as matrices for nanocomposites.
Blending two or more incompatible polymers to make them compatible has been an art
form over the years. However, the use of ionomers to compatibilize these incompatible
polymers appears to have large commercial importance. From an economical viewpoint the
field of blending is very attractive compared to the development of new polymers.
Commercially available polymers have established themselves, and their long term use is
well-known. Taking advantage of these known properties and availability would allow for
making several new potentially important polymer blends.
A potential exists in combining several aspects of the field of ionomers. For example, the
use of ionomers as compatibilizing agents in blends may also be used in conjunction with
incorporation of fillers. Hence, there could be an opportunity in making polymer blend
nanocomposites. Blending targets specific applications based upon taking advantage of the
benefits of the properties of the parent polymers in the blend. However, addition of fillers to
these to make nanocomposites may be able to make stronger blends without losing any of the
beneficial properties of the blend.
There is a large volume of ionomer research that is constantly being published and
patented for the reader to consider. This chapter only touched on a few potentially
commercially important systems from a structural polymer standpoint. Creativity and
imagination are important in the use of ionomers as structural materials. Simple
functionalization of existing polymers may lead to the next great breakthrough in the field.
The reader is urged to look beyond what has been published and patented, and consider new
possibilities. There is a large potential market for the use of ionomers in structural materials,
and the field of ionomers still appears to be more of a niche field of study rather than a major
focus.
References
[1] Eisenberg, A. & King, M. Ion Containing Polymers: Physical Properties and Structure,
vol. 2. New York: Halsted-Wiley; 1977.
[2] Tant, M.R., Mauritz, K.A., & Wilkes, G.L. Ionomers: Synthesis, Structure, Properties
and Applications. New York: Van Nostrand Reinhold Company Inc.; 1997.
[3] Eisenberg, A. & Kim J-S Introduction to Ionomers. New York: John Wiley & Sons,
Inc.; 1998.
[4] MacKnight, W.J. & Earnest, T.R., Jr. (1981) J. Polym.. Sci. Macromol. Rev.,16, 41-122.
[5] Eisenberg, A. & Bailey, F, Eds. (1986) Coulombic Interactions in Macromolecular
Systems. ACS Symposium Series No. 302. American Chemical Society, Washington, DC.
[6] Tant, M.R. & Wilkes, G.L. (1988) An Overview of the Viscous and Viscoelastic Behavior
of Ionomers in Bulk and Solution. J. Macromol. Sci. Part C: Polym. Rev., 28, 1-63.
[7] Lantman, C.W., MacKnight, W.J., & Lundberg, R.D. (1989) Structural Properties of
Ionomers. Annu. Rev. Mater. Sci., 19, 295-317.
[8] Fitzgerald, J.J. & Weiss, R.A. (1988) Synthesis, Properties, and Structure of Sulfonate
Ionomers. J. Macromol. Sci. Part C: Polym. Rev., 28, 99-185.
[9] Tachino, H., Hara, H., Hirasawa, E., Kutsumizu, S., & Yano, S. (1994) Structure and
properties of ethylene ionomers neutralized with binary metal cations. Macromolecules,
27, 372-378.
[10] Kutsumizu, S., Hara, H., Tachino, H., Shimabayashi, K., & Yano, S. (1999) Infrared
spectroscopic study of the binary blends of sodium and zinc salt ionomers produced
from poly(ethylene-co-methacrylic acid). Macromolecules, 32, 6340-6347.
[11] Jeon, H.S. & Kim, J-S (2003) Dynamic mechanical properties and morphology of
sulfonated polystyrene ionomers neutralized with mixtures of various cations. Polym.
Bull., 49, 457-464.
[12] Phillips, A.K. & Moore, R.B. (2006) Mechanical and transport property modifications
of perfluorosulfonate ionomer membranes prepared with mixed organic and inorganic
counterions. J. Polym. Sci. Part B: Polym. Phys., 44, 2267-2277.
[13] Nishio, M., Nichioka, T.T., & Koyama, K. (2005) Rheological properties of ethylene
ionomer neutralized with binary metal cation. Polymer, 46, 261-266.
[14] Tadano, K., Hirasawa, E., Yamamoto, H., & Yano, S. (1989) Order-disorder transition
of ionic clusters in ionomers. Macromolecules, 22, 226-233.
[15] Hirasawa, E., Yamamoto, Y., Tadano, K., & Yano, S. (1989) Formation of ionic
crystallites and its effect on the modulus of ethylene ionomers. Macromolecules, 22,
2776-2780.
[16] Hirasawa, E., Yamamoto, Y., Tadano, K., Yano, S. (1991) Effect of metal cation type
on the structure and properties of ethylene ionomers. J. Appl. Polym. Sci., 42, 351-362.
[17] MacKnight, W.J., McKenna, L.W., & Read, B.E. (1967) Properties of ethylene-
methacrylic acid copolymers and their sodium salts: Mechanical relaxations. J. Appl.
Phys., 38, 4208-4212.
[18] McKenna, L.W., Kajiyama, T., & MacKnight, W.J. (1969) The γ relaxation of a
poly(ethylene-co-methacrylic acid) polymer and its salts. Macromolecules, 2, 58-61.
[19] Register, R.A. & Cooper, S.L. (1990) Anomalous small-angle x-ray scattering from nickel-
neutralized ionomers. 2. Semicrystalline polymer matrices. Macromolecules, 23, 318-323.
94 Daniel J. Klein
[20] Quiram, D.J., Register, R.A., & Ryan, A.J. (1998) Crystallization and ionic associations
in semicrystalline ionomers. Macromolecules, 31, 1432-1435.
[21] Sauer, B.B. & McLean, R.S. (2000) AFM and x-ray studies of crystal and ionic domain
morphology in poly(ethylene-co-methacrylic acid) ionomers. Macromolecules, 33,
7939-7949.
[22] Walters, R.M., Sohn, K.E., Winey, K.I., & Composto, R.J. (2002) Local acid
environment in poly(ethylene-ran-methacrylic acid) ionomers. J. Polym. Sci. Part B:
Polym. Phys., 41, 2833-2841.
[23] Lee, K.H. & Kim, B.K. (1996) Structure-property relationships of polyurethane
anionomer acrylates. Polymer, 37, 2251-2257.
[24] Al-Salah, H.A., Xiao, H.X., McLean, J.A., Jr., & Frisch, K.C. (1988) Polyurethane
cationomers. I. Structure-properties relationships. J. Polym. Sci., Part A: Polym. Chem.,
26, 1609-1620.
[25] Hsu, S.L., Xiao, H.X., Szmant, H.H., & Frisch, K.C. (1984) Polyurethane ionomers. I.
Structure-properties relationships of polyurethane ionomers. J. Appl. Polym. Sci., 29,
2467-2479.
[26] Santerre, J.P. & Brasch, J.L. (1997) Physical properties of nonionomeric and ionomeric
segmented polyurethanes: Effect of sulfonated, carboxylate, and quaternary ammonium
ions in the hard segment Ind. Eng. Chem. Res., 36, 1352-1359.
[27] Lundberg, RD; Makowski, HS Ions in Polymers, ACS Adv. Chem. Ser. 187; Eisenberg,
A Ed; American Chemical Society: Washington, DC, 1980, p. 21.
[28] Tierney, N.K., Trzaska, S.T., & Register, R.A. (2004) Matched random ionomers:
Carboxylate vs sulfonate. Macromolecules, 37, 10205-10207.
[29] Roth, H.H. Water-dispersible molding compositions and articles prepared therefrom.
U.S. Patent 2,691,639, Oct. 12, 1954.
[30] Orler, E.B. & Moore, R.B. (1994) Influence of ionic interactions on the crystallization
of lightly sulfonated syndiotactic polystyrene ionomers. Macromolecules, 27, 4774-
4780.
[31] Bellinger, M., Sauer, J.A., & Hara, M. (1994) Tensile fracture properties of sulfonated
polystyrene ionomers. 1. Effect of ion content. Macromolecules, 27, 1407-1412.
[32] Eisenberg, A; King, M. Ion Containing Polymers, New York, Halsted-Wiley, 1975.
[33] Hird, ,B. & Eisenberg, A. (1990) Phase dominance, multiple glass transitions, and
apparent activation energies in some poly(styrene-co-alkali methacrylate) ionomers. J.
Polym. Sci., Polym. Sci. Polym. Phys. Ed., 28, 1665-1675.
[34] Hara, M., Jar, P-Y, & Sauer, J.A. (1988) Fatigue behavior of ionomers. 1. Ion content
effect on sulfonated polystyrene ionomers. Macromolecules, 21, 3183-3186.
[35] Hara, M., Jar, P., & Sauer, J.A. (1990) Fatigue behavior of ionomers. 2. Effect of
counterion on sulfonated polystyrene ionomers. Macromolecules, 23, 4465-4469.
[36] Venkateshwaran, L.N., Tant, M.R., Wilkes, G.L., Charlier, P., & Jérôme, R. (1992)
Structure-property comparison of sulfonated and carboxylated telechelic ionomers based
on polyisoprene. Macromolecules, 25, 3996-4001.
[37] Tant, M.R., Song, J.H., Wilkes, G.L., Horrion, J., & Jérôme, R. (1986) Mechanical
properties of carboxylato-telechelic polyisoprene. Polymer, 27, 1815-1818.
[38] Register, R.A., Foucart, M., Jérôme, R., Ding, S., & Cooper, S.L. (1988) Structure-
property relationships in elastomeric carboxy-telechelic polyisoprene ionomers
neutralized with divalent cations. Macromolecules, 21, 1009-1015.
[39] Guo, X-Y, Gu, L-X, & Feng, X-X (2002) The glass transition, crystallization, and
melting characteristics of a class of polyester ionomers. J. Appl. Polym. Sci.,, 86, 3660-
3666.
[40] Xue, Y., Hara, M., & Yoon, H.N. (1998) Ionic naphthalene thermotropic copolyesters:
Divalent salts and tensile mechanical properties. Macromolecules, 31, 7806-7813.
[41] Eisenberg, A; King, M. Ion Containing Polymers, New York, Halsted-Wiley, 1975.
[42] Holliday, L. Ionic Polymers. London: Applied Science; 1975.
[43] Ma, X; Sauer, JA; Hara, M. (1995) Poly(methyl methacrylate) based ionomers. 1.
Dynamic mechanical properties and morphology. Macromolecules, 28, 3953-3962.
[44] Chisholm, B.J., Moore, R.B., Barber, G., Khouri, F., Hempstead, A., Larsen, M., Olson,
E., Kelley, J., Balch, G., & Caraher, J. (2002) Nanocomposites derived from sulfonated
poly(butylene terephthalate). Macromolecules, 35, 5508-5516.
[45] Gorda, KR; Peiffer, D.G. (1992) Properties of sulfonated poly(butylene terephthalate). J.
Polym. Sci., Polym. Phys. Ed., 30, 281-292.
[46] Li, Y., Ma, J., Wang, Y., & Liang, B. (2005) Synthesis and characterization of
sulfonated poly(ethylene terephthalate)/montmorillonite nanocomposites. J. Appl.
Polym. Sci., 98, 1150-1156.
[47] Jeong, E.H., Yang, J., Hong, J.H., Kim, T.G., Kim, J.H., & Youk, J.H. (2007) Effective
preparation of montmorillonite/polyurethane nanocomposites by introducing cationic
groups into the polyurethane main chain. Eur. Polym. J, 43, 2286-2291.
[48] Shah, R.K., Krishnaswamy, R.K., Takahasi, S., & Paul, D.R. (2006) Blown films of
nanocomposites prepared from low density polyethylene and a sodium ionomer of
poly(ethylene-co-methacrylic acid). Polymer, 47, 6187-6201.
[49] Shah, R.K., Kim, D.H., & Paul, D.R. (2007) Morphology and properties of
nanocomposites formed from ethylene.methacrylic acid copolymers and organoclays.
Polymer, 48, 1047-1057.
[50] Yoo, Y., Shah, R.K., & Paul, D.R. (2007) Fracture behavior of nanocomposites based
on poly(ethylene-co-methacrylic acid) ionomers. Polymer, 48, 4867-4873.
[51] Shah, R.K. & Paul, D.R. (2006) Comparison of nanocomposites prepared from sodium,
zinc, and lithium ionomers of ethylene/methacrylic acid copolymers. Macromolecules,
39, 3327-3336.
[52] Parent, J.S., Liskova, A., & Resendes, R. (2004) Isobutylene-based ionomer composites:
siliceous filler reinforcement. Polymer, 45, 8091-8096.
[53] Lee, J.A., Kontopoulou, M., & Parent, J.S. (2005) Synthesis and characterization of
polyethylene-based ionomer nanocomposites. Polymer, 46, 5040-5049.
[54] Li, T. & Yan, N. (2007) Mechanical properties of wood flour/HDPE/ionomer
composites. Composites Part A, 38, 1-12.
[55] Vallejo, F.J., Eguiazábal, J.I., & Nazábal, J. (2000) Compatibilization of PP/Vectra B
“in situ” composites by means of an ionomer. Polymer, 41, 6311-6321.
[56] Paul, D.R. & Newman, S. Polymer Blends, Vol. 1. New York: Academic Press; 1979.
[57] Utracki, L.A. & Favis, B.D. Polymer alloys and Blends. In: Cheremisinoff, N.P., editor
Encyclopedia of Engineering Materials. New Jersey: Marcel Dekker; 1988.
[58] Papadopoulou, C.P. & Kalfoglou, N.K. (1998) Polyurethane/olefinic ionomer blends. 1:
Compatibility characterization. Polymer, 39, 7015-7021.
[59] Papadopoulou, C.P. & Kalfoglou, N.K. (1999) Polyurethane/HDPE blends: 2.
Compatibilization with an olefinic ionomer. Polymer, 40, 905-911.
96 Daniel J. Klein
[60] Willis, J.M. & Favis, B.D. (1988) Processing-morphology relationships of

compatibilized polyolefin/polyamide blends. Part I: The effect of an ionomer
compatibilizer on blend morphology. Polym. Eng. Sci., 28, 1416-1426.
[61] Bazuin, C.G., Rancourt, L., Villeneuve, S., & Soldera, A. (1993) Miscibility of blends
of poly(2,6-dimethyl 1,4-phenylene oxide) with polystyrene-based ionomers and
copolymers. J. Polym. Sci., Part B: Polym. Phys., 31, 1431-1440.
[62] Kit, K.M., Schultz, J.M., & Gohil, R.M. (1995) Morphology and barrier properties of
oriented blends of poly(ethylene terephthalate) and poly(ethylene 2,6-naphthalate) with
poly(ethylene-co-vinyl alcohol). Polym. Eng. Sci., 35, 680-692.
[63] Ahn, T.O., Kim, C.K., Kim, B.K., Jeong, H.M., & Huh, J.D. (1990) Binary blends of
nylons with ethylene vinyl alcohol copolymers: Morphological, thermal, reheological,
and mechanical behavior. Polym. Eng. Sci., 35, 341-349.
[64] Samios, C.K. & Kalfoglou, N.K. (1998) Compatibilization of poly(ethylene-co-vinyl
alcohol) (EVOH) and EVOH/HDPE blends with ionomers. Structure and properties.
Polymer, 39, 3863-3870.
[65] Montoya, M., Abad, M.J., Barral, L., & Bernal, C. (2006) Mechanical and fracture
behavior of polypropylene/poly(ethylene-co-vinyl alcohol) blends compatibilized with
ionomer Na+. Eur. Polym. J., 42, 265-273.
[66] Abad, M.J., Ares, A., Barral, L., Cano, J., Díez, F.J., García-Garabal, S., López, J., &
Ramírez, C. (2004) Use of a sodium ionomer as a compatibilizer in polypropylene/high-
barrier ethylene-vinyl alcohol copolymer blends: The processability of the blends and
their physical properties. J. Appl. Polym. Sci., 94, 1763-1770.
[67] Abad, M.J., Barral, L., & Eguiazabal, J.I. (2005) Influence of the ethylene-(methacrylic
acid)-Zn2+ ionomer on the thermal and mechanical properties of blends of poly(propylene)
(PP)/ethylene-(vinyl alcohol) copolymer (EVOH). Polym. Int., 54, 673-678.
[68] Samios, C.K. & Kalfoglou, N.K. (2001) Acrylic-modified polyolefin ionomers as
compatibilizers for poly(ethylene-co-vinyl alcohol)/aromatic copolyester blends.
Polymer, 42, 3687-3696.
[69] Samios, C. & Kalfoglou, N.K. (1999) Compatibilization of poly(ethylene
terephthalate)/polyamide-6 alloys: Mechanical, thermal, and morphological
characterization. Polymer, 40, 4811-4819
[70] Lahor, A., Nithitanakul, M., & Grady, B.P. (2004) Blends of low-density polyethylene
with nylon compatibilized with a sodium-neutralized caboxylate ionomer. Eur. Polym.
J., 40, 2409-2420.
[71] Bhattacharyya, A.R., Ghosh, A.K., & Misra, A. (2003) Ionomer compatibilized
PA6/EVA blends: mechanical properties and morphological characterization. Polymer,
44, 1725-1733.
[72] Antony, P., Bandyopadhyay, S., & De, S.K. (2000) Synergism is properties of
ionomeric polyblends based on zinc salts of carboxylated nitrile rubber and
poly(ethylene-co-acrylic acid). Polymer, 41, 787-793.
[73] Wang, X. & Cui, X. (2005) Effect of ionomers on mechanical properties, morphology,
and rheology of polyoxymethylene and its blends with methyl methacrylate-styrene-
butadiene copolymer. Eur. Polym. J., 41, 871-880.
In: Strength of Materials ISBN 978-1-60741-500-8
Editors: G. Mendes and B. Lago, pp. 97-131
c 2009 Nova Science Publishers, Inc.
Chapter 3
FAILURE OF L AYERED C OMPOSITES S UBJECT

TO I MPACTS : C ONSTITUTIVE M ODELING
AND PARAMETER I DENTIFICATION I SSUES
Stefano Mariani∗
Politecnico di Milano, Dipartimento di Ingegneria Strutturale
Piazza L. da Vinci 32, 20133 - Milano (Italy)
Abstract
Layered composites subject to impacts can fail by delamination, i.e. by debonding
between laminae, if the stress waves cause damaging phenomena to take place mainly
within the resin-enriched interlaminar phases. To simulate delamination at the struc-
tural level, processes dissipating energy are lumped onto fictitious zero-thickness in-
terlaminar surfaces, and softening interface constitutive laws are adopted to describe
the progressive failure of the interlaminar phases.
Since delamination occurs inside very narrow regions, results of experimental test-
ing on whole composites need to be accurately and reliably filtered to calibrate the
interface constitutive laws. To this aim, here we propose a sigma-point Kalman filter
approach. The performances of the proposed methodology, in terms of constitutive pa-
rameter estimations and dynamic delamination tracking, are assessed through pseudo-
experimental testings on a two-layer composite, and real testings on multi-layer glass
fiber reinforced plastic composites.
Keywords: composites, delamination, impact loading, interface constitutive modeling, pa-

rameter identification, sigma-point Kalman filter.
1. Introduction
Foreign objects striking the outer surface of composite structures may cause permanent
damage, or even sudden failure [1]. On the basis of the velocity of the striker, impacts
can be roughly distinguished into two main classes: low-velocity impacts, characterized
by small or negligible effects of inertial forces on the damage/failure mode; high-velocity
∗
E-mail address: [email protected]
98 Stefano Mariani
impacts, whereby inertial forces strongly affects the damage/failure event. In case of a high-
velocity impact, the eventual failure of layered composites can be due to the propagation
of interlaminar cracks only (delamination), or to a local perforation accompanied also by
intralaminar damage [2].
In this work we focus on impact-induced delamination and provide a review of a clas-
sical approach to debonding in layered bodies [3–8]. Within this approach it is assumed
that laminae always behave elastically (i.e. intralaminar damaging phenomena are disre-
garded, or thought to be negligible), and interlaminar resin-enriched phases are lumped
onto zero-thickness surfaces, along which debonding can occur because of the impinging
shock waves. To model delamination, softening interface constitutive laws are adopted
to link the tractions acting upon each interlaminar surface with the displacement jumps
occurring across it. If these laws are able to phenomenologically describe the microme-
chanical processes leading to debonding, the above approach furnishes accurate results at
the structural level, see e.g. [4]. However, calibration of the interface laws is still an open
issue: since damaging and cracking phenomena linked to delamination take place inside
very narrow regions, direct testing on the interlaminar phases can not be devised. Instead,
inverse analysis procedures can be adopted to efficiently manage experimental data in order
to estimate uncertain constitutive parameters [6, 9–11]; needles to say, the aforementioned
experimental data have to be informative, i.e. they have to carry information on the current
response of interphases to the impact loading.
As for model calibration purposes, standard filtering procedures have been proven ac-
curate enough in case of static loadings [12, 13]; in case of impacts the inverse prob-
lem becomes stiff since composite failure, once incepted, usually occurs almost instan-
taneously [2, 14]. To deal with this issue, we recently adopted the extended Kalman filter
(EKF) [11, 15]; when compared to alternative approaches (like, e.g., neural networks and
least squares), the EKF has the great advantages of being able to work in real-time and of
being explicitly linked to the physics of the ongoing delamination processes. Moreover, the
EKF exploits the evolution in time of the measured fields in order to continuously enhance
model calibration.
As pointed out by several authors (see, e.g., [16, 17]), the EKF looses stability when
nonlinearities become dominant in the equations governing the inverse problem. This is due
to the fact that the EKF replaces nonlinear functions with their relevant tangent surfaces,
leading to possibly biased or even divergent parameter estimates. An alternative approach
to treat nonlinearities in a stochastic framework recently led to the proposal of the so-called
sigma-point Kalman filter (SPKF), also termed unscented Kalman filter [16]. Instead of
linearizing the governing equations, thereby introducing approximations, the SPKF samples
the statistics of the current state of the system and of model parameters to draw a set of
sigma-points. These sigma-points are then let to evolve according to the actual, nonlinear
dynamics of the problem. The filter estimates are eventually computed by averaging the
information conveyed by all the evolved sigma-points through an ad-hoc numerical scheme
[16, 18]. When compared to the EKF, the SPKF can achieve a much higher accuracy in
the model calibration task [19], often preventing the aforementioned bias and divergence
occurrences.
If used to deal with the nonlinearities accompanying delamination in layered compos-
ites, the SPKF has to furnish accurate estimates of uncertain interface model parameters
Failure of Layered Composites Subject to Impacts 99
m
W1
n
s11
W 2 s21
G1
x3
G2
W3
x1 x2
Figure 1. geometry of a three-dimensional layered continuum (here nΓ = 2), and notation.
while tracking the state of the laminate. In this Chapter we analyze this framework in
details, trying to point out the main strengths of the SPKF. To this purpose, in Section
2. we present the equations governing the dynamics of a layered, possibly delaminating
body. Constitutive models for laminae and resin-enriched interphases are then discussed
so as to recognize the basic parameters, in need of an accurate estimation. An explicit
time integration scheme for the equations of motion of the composite, and an explanation
on how to handle model calibration via the SPKF are eventually offered. Section 3. deals
with sigma-point Kalman filtering: physical arguments are used to propose a slightly new
scheme for the drawing of the sigma-points. In Section 4. the performances of the SPKF
are assessed: pseudo-experimental, i.e. fictitious impact testings on a two-layer composite
are first considered; hence, real experiments on multi-layer glass fiber reinforced plastic
(GRP) composites are used to testify the robustness of the SPKF in promptly detecting
delamination.
Throughout the whole Chapter, a matrix notation is adopted, with uppercase and low-
ercase bold symbols respectively denoting matrices and vectors; a superscript T stands for
transpose, while a superposed dot represents time rates.
2. Dynamics of Layered Composites

2.1. Governing Relations
Let Ω be a three-dimensional, layered body; its smooth outer boundary, with unit out-
ward normal m, be constituted by the two disjoint parts Γu and Γτ , where displacement
and traction fields are respectively assigned. Ω is assumed to be crossed by nΓ non-
intersecting surfaces, or interfaces Γj , j = 1, ..., nΓ , see Figure 1. Each resulting portion
Ωj , j = 1, ..., nΓ + 1, of the bulk will be here referred to as layer, or lamina. Interfaces
in laminates are usually flat and parallel to each other, with a common orientation defined
by the unit vector n; opening (mode I) and sliding (mode II and mode III) displacement
discontinuities along each Γj take place along the n direction and in the sj1 − sj2 plane,
100 Stefano Mariani
respectively.
The equilibrium of Ω at time t is governed by the following equations:
C T σ + b̄ = ̺ü in Ω\Γ (1)
Mσ = τ̄ on Γτ (2)
(
N σ = −τ on Γ+
j
j = 1, ..., nΓ (3)
Nσ = τ on Γ−
j
where: Ω = ∪nj=1 Ω and Γ = ∪nj=1

Γ +1 j Γ
Γj ; Γ+ −
j and Γj are the sides of Γj respectively
belonging to layers Ωj and Ωj+1 , according to the notation of Figure 1. The two sides
can not be distinguished in the initial configuration (at t = 0) but, as soon as delamination
is incepted, they can experience a relative movement. In the equations above, adopting a
standard Voigt notation: σ is the stress vector, which gathers the independent components of
the stress tensor; b̄ and τ̄ are the assigned loads in the bulk Ω\Γ and along Γτ , respectively;
̺ is the bulk mass density; ü is the acceleration field in the bulk Ω\Γ; C is the differential
compatibility operator; M and N are the matrices respectively collecting the components
of the unit vectors m and n. To locally ensure equilibrium along each surface, sides Γ+ j
and Γ− j are respectively acted upon by the traction vectors −τ and τ .
Since we aim at modeling phenomena occurring within very short time intervals after
the impact, linearized kinematics proves sufficient. Compatibility thus reads:
ε = Cu in Ω\Γ (4)
u = ū on Γu (5)
where: ε is the strain vector, which gathers the independent components of the strain tensor;
u is the displacement field in the bulk Ω\Γ; ū is the assigned displacement along Γu .
Across each interface, the displacement discontinuity [u] is defined as:

[u] = u − u j = 1, ..., nΓ (6)
Γ+
j Γ−
j
The body is assumed to be initially at rest, in an undeformed and unstressed state, such
that: (
u0 = 0
in Ω (7)
u̇0 = 0
2.2. Constitutive Modeling

Energy dissipation in layered composites subject to impacts arises from: the spreading
of damage and the subsequent propagation of cracks within the interlaminar, resin-enriched
phases; the spreading of damage inside laminae. This latter dissipation mechanism can be
disregarded, as done here, if the impact energy is not high enough to cause a penetration
of the impactor inside the laminate. In such a case, laminae can be assumed to behave
elastically, according to:
σ = EΩε in Ω\Γ (8)
where E Ω is the elasticity matrix of the bulk. Each lamina is usually an orthotropic body;
though not explicitly shown in (8), E Ω can change from lamina to lamina because of a
different orientation of the axes of elastic symmetry.
To simulate strength reduction in the interlaminar phases, eventually leading to
debonding, softening interface constitutive laws prove efficient [3–8, 20, 21]. A detailed,
micromechanics-based representation of the damaging processes caused by the constrained
deformation field inside each interphase is not looked for; instead, a phenomenological re-
−
lationship is adopted to link the tractions τ acting upon the interface sides Γ+
j and Γj to
the displacement jump [u] occurring across Γj . This constitutive law might be conceived
as the macroscopic (homogenized through the interphase thickness) behavior of the inter-
phase. Since the ratio between the thickness of each interlaminar phase and the thickness of
the whole laminate is usually quite small, this approach can furnish accurate results at the
laminate length scale; on the contrary, an accurate representation of the micromechanical
phenomena preceding delamination is in need of a multi-scale approach (see, e.g., [22–24]).
Along each Γj , both opening/closing (along n) and sliding (in the sj1 − sj2 plane) dis-
continuities take place under general loading conditions. By way of a simplified scheme
adopted in [25–28], the local kinematics of an interface is governed by the effective dis-
placement discontinuity [u], defined as:
p
[u] = [u]2n + κ2 [u]2s (9)
where: [u]n = [u]T n and [u]s = |[u] − [u]n n| are, respectively, the opening and sliding
displacement discontinuities; κ is a coupling coefficient, which accounts for the interaction
between stretching and shearing deformation modes inside the interphase. Through an
incremental work equivalence the effective traction τ , work-conjugate to [u], turns out to
be (see [25, 27]):
r
τ2
τ = τn2 + s2 (10)
κ
where τn and τs are defined like [u]n and [u]s . The mechanical behavior of the interface
can now be furnished in terms of an effective τ − [u] relationship.
For quasi-brittle materials, like interphase resins, the tensile strength is typically much
smaller than the compressive one. Therefore, when dynamic delamination occurs without
intralaminar damage, the response of the interface under compressive stress states along n
can be assumed linear elastic, according to:
τ = K[u] if [u] < 0 (11)
where K is the stiffness of the interface. On the other hand, the response of the interface
under tensile stress states along n is characterized by strength reduction, leading to soften-
ing, beyond the attainment of a peak traction. A simple way to model the transition from the
102 Stefano Mariani
(a)
(b)
(c)
Figure 2. Effective traction-displacement discontinuity relationships under tensile loading.

(a) piecewise linear law (12); (b) linear-exponential law (13); (c) exponential law (14).
initial elastic regime to the subsequent softening one, is through a piecewise linear (PWL)
law (see Figure 2(a)):

τ = K[u]
 if [u] ≤ [u]e
τ = τM + Q ([u] − [u]e ) if [u]e < [u] ≤ [u]U (12)

τ =0 if [u] > [u]U

where: [u]e is the effective displacement discontinuity corresponding to the peak traction
τM = K[u]e ; Q is the (negative) slope of the softening branch; [u]U = 1 − K Q [u]e is
the effective displacement discontinuity at which the interaction between the crack faces
ceases.
The linear softening in (12) can be a too crude approximation of the post-peak regime
for some materials, whose response features an initial steep descent followed by a long,
much less steep tail (see, e.g., [29]). In such a case, the softening regime can be modeled
through an exponential law (hereafter referred to as linear-exponential, L-E law because of
the pre-peak linear elastic phase) according to (see Figure 2(b)):
(
τ = K[u] if [u] ≤ [u]e
(13)
τ = τM exp (−ς([u] − [u]e )) if [u] > [u]e
where ς is a model parameter that allows to match the slope of the softening branch just
beyond the attainment of the peak traction τM .
Sometimes, a smooth transition from the elastic regime to the softening one turns out to
be more representative of the actual interphase response. The nonlinear binding model,
originally proposed in [30, 31] for metals and bimetallic compounds and later adopted
also in nonlinear fracture mechanics [20, 21, 32], allows to describe such smooth transi-
tion through the following exponential (EXP) law (see Figure 2(c)):

[u]
τ = K[u] exp − (14)
[u]e
Besides the effective stiffness K and strength τM , a full characterization of the nonlin-
ear behavior of the interface has to match the fracture energy, or work of separation G. In
terms of effective quantities, G is defined as the amount of energy required to annihilate the
interaction between the opening/sliding crack faces, i.e.:
Z∞
G= τ d[u] (15)
0
From a model calibration perspective, parameters Q and ς in (12) and (13) can be tuned
to accurately match the actual G, since they do affect only the softening branch of the
interface law. On the other hand, after having assigned K in (14), only [u]e can be adjusted:
therefore, both τM and G can not be accurately matched. In the exponential law, in fact, the
following constraint holds: s
KG
2 =e
τM
(16)
104 Stefano Mariani
e being the Nepero number. To avoid problems related to this fictitious constraint, a modi-
fied exponential law is here formulated as follows (see Figure 3):
[u] q

τ = K[u] exp − (17)
[u]e
where q shows up as an additional constitutive parameter. In law (17): K still represents

the initial elastic stiffness; [u]e is a reference effective displacement discontinuity, while
peak traction τM is attained at [ū] ≡ (1/q)1/q [u]e . The effective peak traction and fracture
energy are affected by q, according to:
1/q
1 1
τM =K[u]e exp −
q q
(18)
K 2
G = [u]2e γf
q q
γf being the gamma function. The dependence of τM and G on the parameter q is depicted
in Figure 4: it can be seen that G is a monotonically decreasing function of q, whereas τM is
lower-bounded by the value corresponding to q = 1. Having tuned K, this law thus allows
the calibration of both τM and G.
All the above laws but the piecewise linear one, assume that the interaction between
the opening interface sides continues up to [u] → ∞, which seems not physical at the
macroscale. To simulate delamination growth a breakdown threshold therefore needs to be
introduced [33, 34]: as soon as the current traction τ reduces to a small fraction (say 5-10
%) of the peak value τM , the interaction is suddenly assumed to vanish.
When unloading from the tensile envelope occurs, i.e. when [u̇] < 0, the above interface
models can be viewed as either reversible, if τ always belongs to the envelope (leading to
interface healing if softening has already started), or irreversible, if τ decreases following a
radial path to the origin of the τ − [u] plane. These two alternative constitutive assumptions
lead to different entries in the interface tangent stiffness matrix E Γ , linking rates of τ and
[u] in the local n − sj1 − sj2 reference frame according to:
τ̇ = E Γ [u̇] (19)
For additional details, readers are referred to [27, 28].
2.3. Finite Element Formulation

The weak form of the equilibrium equations (1)-(3) reads:
Z Z Z nΓ Z
X
T T T
εv σ dΩ = v (b̄ − ̺ü) dΩ + v τ̄ dΓτ − [v]T τ dΓj ∀v ∈ U0 (20)
Γτ j=1 Γ
Ω\Γ Ω\Γ j
where: v is the test function; εv = Cv; U is the trial solution space, collecting displace-
ment fields u continuous in Ω\Γ, possibly discontinuous along each Γj and fulfilling the
boundary condition (5) on Γu ; U0 is the relevant variation space, with zero prescribed
Figure 3. effect of q on the modified exponential traction-displacement discontinuity law

(17).
(a) (b)
Figure 4. modified exponential law (17). Effects of q on (a) the effective peak traction τM
and on (b) the effective fracture energy G.
displacements on Γu . In (20), in view of the assumed linearized kinematics, the relation

−
Γj ≡ Γ+ j ≡ Γj for each interface has been exploited.
Allowing for the elastic bulk constitutive law (8), the following variational statement is
arrived at:
Z Z nΓ Z
X
T T
find u ∈ U : ̺ v ü dΩ + εv E Ω ε dΩ + [v]T τ dΓj
j=1 Γ
Ω\Γ Ω\Γ j
Z Z (21)
T T
= v b̄ dΩ + v τ̄ dΓτ ∀v ∈ U0
Ω\Γ Γτ
106 Stefano Mariani
Now, let the finite element approximation of the displacement and deformation fields in
Ω\Γ be (see [35] for the notation):
u∼
= Φuh (22)
ε∼
= CΦuh = B Ω uh (23)
where matrix Φ gathers the nodal shape functions, and vector uh collects the nodal dis-
placements.
If delamination is allowed to occur only along element boundaries, the discrete dis-
placement jump field can be written:

∼
= B Γj uh

[u] j = 1, ..., nΓ (24)
Γj
Owing to the discrete interpolation fields defined above, the semi-discretized equations
of motion of the composite turn out to be:
nΓ
X
M üh + K Ω uh + Rj = F (25)
j=1
where the mass matrix M , the bulk stiffness matrix K Ω , the internal force vectors Rj and
the external load vector F are, respectively:
Z
M= ̺ ΦT Φ dΩ
Ω\Γ
Z
KΩ = B TΩ E Ω B Ω dΩ
Ω\Γ
Z (26)
Rj = B TΓj τ dΓj
Γj
Z Z
T
F = Φ b̄ dΩ + ΦT τ̄ dΓτ
Ω\Γ Γτ
Smarter finite element formulations, like the extended or generalized ones [36,37], have
been recently formulated to simulate mixed-mode crack growth in homogeneous solids,
see e.g. [27, 38, 39]. These methodologies allow cracks to propagate not only along inter-
element edges, but also inside the elements; possible constraints imposed by the mesh lay-
out on crack trajectories, evidenced e.g. in [40], can be therefore alleviated. When deal-
ing with delamination in layered continua, where debonding occurs only along the a-priori
known interlaminar surfaces, crack description looks simple and the aforementioned feature
of the extended finite element method looses much of its advantages.
2.4. Time Integration

In our previous works [11, 15] it was shown that the time integration scheme can
strongly affect the stability of the filtering procedure and, therefore, the accuracy of model
parameter estimates.
In case of impact loadings, which cause the propagation of shock waves inside the com-
posite, the time marching algorithm has to dump the spurious high frequency oscillations
linked to space discretization. Otherwise, the numerically computed displacement and ve-
locity fields do not prove reliable enough to be compared to the experimental data.
We adopt here the explicit α−method [41, 42] to advance in time the solution of the
equations of motion (25), see also [19]. After having partitioned the time interval of interest
Nt −1
according to [t0 tN ] = ∪i=0 [ti ti+1 ], at the end of the generic time step [ti ti+1 ]
the solution to (25) is obtained according to the following predictor-integrator-corrector
splitting:
• predictor:
1
ũi+1 =ui + ∆t u̇i + ∆t2 ( − β)üi (27)
2
˜ i+1 =u̇i + ∆t(1 − γ)üi
u̇ (28)
where ∆t = ti+1 − ti ;
• explicit integrator:
    
j
Rji 
X X
üi+1 = M −1 F i+1+α − (1 + α) K ũi+1 + R̃i+1  + α Kui +
j j
(29)
j
Rji B TΓj τ i dΓj
R
where: F i+1+α = F (ti + (1 + α)∆t); = and R̃i+1 =
Γj
R T
B Γj τ̃ i+1 dΓj ;
Γj
• corrector:
ui+1 =ũi+1 + ∆t2 β üi+1 (30)

˜ i+1 + ∆tγ üi+1
u̇i+1 =u̇ (31)
In the above equations α, β and γ are algorithmic coefficients. To get a non-oscillatory

velocity field, α = −0.3 has been adopted in all the forthcoming simulations; furthermore,
β and γ have been finely tuned around the values allowing second-order accuracy in linear
elasto-dynamics.
To ensure accuracy of the filtering procedure, the time step size ∆t has been always set
so as to fulfill the Courant condition in the bulk of the composite. Moreover, to speed up
the explicit integrator phase (29), the mass matrix M has been diagonalized by means of a
standard row-sum lumping procedure [42].
108 Stefano Mariani
Account taken of the explicit format of the integrator stage, the space-time discretized
equations of motion of the laminate (state equations) can be formally written:
 
ui+1 
z i+1 = u̇i+1 = f zi (z i ) (32)
üi+1
 
where z is the structural state vector, and mapping f z turns out to be nonlinear because of
the softening interface behavior.
3. Constrained Sigma-point Kalman Filtering

3.1. Parameter Identification via Joint Kalman Filtering
According to a standard methodology [43], the calibration of constitutive laws can be
pursued by Kalman filtering if a state vector x is obtained by joining the structural state
vector z (see Eq. 32) with a vector ϑ gathering all the model parameters to be tuned. At
time ti this can be written:

zi
xi = (33)
ϑi
While the current structural state z is always at least partially observed, model parame-
ters to be identified can not be directly measured; by joining z and ϑ, state tracking can
consistently improve model calibration.
In case of irreversible constitutive laws, internal state variables must be gathered by x
too, see e.g. [12, 19].
Allowing for model and measurement errors, the state-space model describing the evo-
lution within the time interval [ti ti+1 ] of the joint state vector and its link with observa-
tions turns out to be: (
xi+1 = f i (xi ) + v i
(34)
yi = Hxi + wi
where: y is the observation vector, which collects the measured components of the state
vector; v is the process noises; w is the measurement noise. v, w are assumed to be
additive, uncorrelated white and Gaussian processes, with zero mean and covariances V
and W [44, 45]. Since z is defined according to Eq. (32), the observation equation in (34)
shows up as a linear relation between y and x. On the contrary, the interface behavior
renders the evolution equation f nonlinear.
By way of the EKF [12, 46], within the time step the nonlinear mapping f is expanded
in Taylor series, up to the first order, around the current estimates of the state vector and
of model parameters. Bounds on the required accuracy of the initialization of x, and on
the statistics of noises v and w to assure filter stability were provided for linear systems
in [47] and, more recently, for nonlinear systems in [48]. Even in the absence of filter
instabilities, the softening response of the interlaminar surfaces does not always guarantee
the achievement of an accurate model calibration, see [15, 19].
Table 1. Sigma-point Kalman filter.
• Initialization at t0 :
x̂0 =E[x0 ]
P 0 =E[(x0 − x̂0 ) (x0 − x̂0 )T ]
• At ti , for i = 0, ..., N
1. Predictor phase:
χ̂i,j =x̂i + ∆χi,j j = 0, ..., Nχ

χ̂−
i+1,j =f i (χ̂i,j )
Nχ
X
x̂−
i+1 = ωj χ̂−
i+1,j
j=0
P−
i+1 =R−i+1 +V
where
Nχ T
X
R−
i+1 = ωj χ̂− −
i+1,j − x̂i+1 χ̂− −
i+1,j − x̂i+1
j=0
2. Corrector phase:
x̂i =x̂− U −

i + Gi y i − H x̂i
P i =P − U −
i − Gi HRi
where
−
−1
GU
i =Ri H
T
HR− T
i H +W
To improve the results when nonlinearities become dominant, the SPKF has been re-
cently proposed [16, 49–51]. At the beginning of the time step, the probability distribution
of x is deterministically sampled through a set of sigma-points χ̂j , j = 0, ..., Nχ . These
sigma-points are then allowed to evolve according to the nonlinear mapping f . The statis-
tics of x at the end of the time step are finally obtained through a proper weighted averaging
scheme [18]. This filtering procedure is detailed in Table 1, where E[2] represents the ex-
pected value of 2.
The number of sigma-points and their location in the state vector space are accurately
chosen, so as to achieve high accuracy in the estimated probability distribution of x at the
end of each time step; when compared to the EKF, a better performance of the SPKF, also
in terms of model calibration, is therefore expected [16]. The enhanced accuracy of the
110 Stefano Mariani
SPKF is discussed next; even though these results have been already presented elsewhere,
they are here collected to show how possible constraints on parameter estimates, not dealt
with by the standard SPKF, can be managed.
3.2. Accuracy of a Constrained Sigma-point Transformation

In this Section we focus on the time interval [ti ti+1 ], but we avoid using indexes i
and i + 1 to simplify the notation.
Let x be a random vector, featuring at the beginning of the time step mean x̂ and
covariance P . We conceive x as the sum of the mean x̂ and a zero-mean disturbance
∆x = x − x̂. If x undergoes a nonlinear transformation, governed by a mapping f
analytic everywhere so that it can be expanded in Taylor series about x̂, at the end of the
time step we get:
∞
−
X 1 n
x = f (x) = f (x̂ + ∆x) =f (x̂) + D (35)
n=1
n! ∆x
where, with a slight abuse in notation, the n−th order term D∆x
n in the series expansion is:
Nx !n
X ∂f
n
D∆x ≡ ∆xℓ (36)
∂xℓ x=x̂
ℓ=1
Nx being the number of components of the state vector x. Since the derivatives of f in (36)
are evaluated at x = x̂, they are not random variables. The expected value of x− therefore
reads:
"∞ #
X 1
x̂− = E[x− ] =E [f (x̂)] + E Dn
n! ∆x
n=1
∞
(37)
X 1 n
=f (x̂) + E D∆x
n!
n=1
The relevant error covariance matrix is:

h T i
P − =E x− − x̂− x− − x̂−


∞
! ∞
!T 
X 1 X 1 m
=E  n
D∆x − E D∆x n m
D∆x − E D∆x 
n! m! (38)
n=1 m=1
∞ X
∞
X 1 1 h n m T
i n h m T i
= E D∆xD∆x − E D∆x E D∆x
n! m!
n=1 m=1
Now, let us suppose to sample the probability distribution of x through a set of sigma-
points χ̂j , j = 0, ..., Nχ , chosen around the current mean x̂ according to:
χ̂j = x̂ + ∆χj , j = 0, ..., Nχ (39)

where the terms ∆χj need to be determined.

Similarly to x, within the time step each sigma-point undergoes the transformation:
∞
X 1 n
χ̂− = f (χ̂j ) = f (x̂) + D (40)
j
n=1
n! ∆χj
where:
Nx !n
X ∂f
Dn
∆χj ≡ ∆χ (41)
∂xℓ x=x̂ j ℓ
ℓ=1
At the end of the time step, the information in the evolved sigma-points are collected
via a weighted averaging procedure to obtain:
Nχ
X
x̂−
SPT = ωj χ̂−
j
j=0
    (42)
Nχ ∞ Nχ
X X 1 X
= ωj  f (x̂) +  ωj D n
∆χj

n!
j=0 n=1 j=0
where ωj are the weights of the sigma-point transformation relevant to the mean of x. The
corresponding error covariance matrix is given by:
Nχ T
X
P−
SPT = ωj⋆ χ̂−
j − x̂ −
SPT χ̂−
j − x̂−
SPT
j=0
  T
Nχ ∞ X ∞ Nχ Nχ
X X 1 1  n X X
= ωj⋆ D ∆χj − ωr D n
∆χr ∆χj −
 D m ωs D m
∆χs

n=1 m=1
n! m!
j=0 r=0 s=0
(43)
where ωj⋆ are the weights of the sigma-point transformation relevant to the covariance of x.
If x is a Gaussian random vector, its probability distribution is symmetric with respect
to the mean x̂; therefore, all the odd central moments D∆x n , n = 1, 3, 5, ..., are zero. To be
compliant with this condition, couples of sigma-points are symmetrically placed around x̂,
according to [16]:

∆χ0
 =0
 √ Nχ
∆χk = +ψ P 1k k = 1, ..., 2 (44)
 √
∆χ Nχ

+k
= −ψ P 1k
2
√
Here: P represents the square root of matrix P , computed e.g. through a Cholesky
factorization; ψ is a scaling parameter; 1k is a unit vector aligned with component k in the
112 Stefano Mariani
state vector space. The series expansions (37) and (42) agree up to third order if:
P
Nχ
j=0 ωj = 1




 N

 P χ 1
j=0 ωj D ∆χj = 0

PNχ (45)
2
2
j=0 ωj D ∆χj = E D∆x





PNχ ωj D 3 = 0


j=0 ∆χj
To simplify the matter, let us assume ωj = ω for j = 1, ..., Nχ ; relations involving the first
and third order terms in (45) are then automatically satisfied. Relations involving the zeroth
and second order terms in (45) then furnish:
(
ω0 + Nχ ω = 1
(46)
2ψ 2 ω = 1
A further condition to set ω0 , ω and ψ can be furnished by matching the diagonal entries of
the fourth order terms (kurtoses) in (37) and (42). This leads to [19]:
Nχ 1 √
ω0 = 1 − , ω= , ψ= 3 (47)
6 6
Here we propose an alternative condition to determine ω0 , ω and ψ, partially exploiting
the features of the so-called scaled unscented transformation [52]. Let us assume that model
parameters have to satisfy the constraints:
ϑm ≤ ϑ ≤ ϑM (48)
where ϑm and ϑM respectively gather the minimum and maximum (if any) allowed values
of model parameters. This requirement must be fulfilled by each sigma-point χ̂− j , j =
0, ..., Nχ . For j = 0 the conditions (48) are automatically satisfied, since x̂ (and, therefore,
ϑ̂) is computed at the end of the previous time step by averaging sigma-points all fulfilling
the constraints. Further, if ϑ = B ϑ x, B ϑ being a Boolean matrix, conditions (48) are
satisfied by all the sigma-points if:
( )
ϑ̂ℓ − ϑm
ℓ ϑ M − ϑ̂
ℓ ℓ Nχ
ψ ≤ min ℓ
, ℓ
, k = 1, ..., ; ℓ = 1, ..., Nϑ (49)
ak ak 2
√
where ak = B ϑ P 1k , and Nϑ is the√number of model parameters in ϑ. In the forthcoming
examples, we initially assume ψ = 3 (according to what reported in 47) and reduce its
value if necessary, according to relation (49).
As for the error covariance matrix P − ⋆
SPT , by letting ωj = ωj = ω for j = 1, ..., Nχ , we
get:
 
∞ X ∞ Nχ Nχ Nχ
X 1 1 X X X
P−SPT =
 ∆χj D ∆χj + (ω0 + Nχ ω − 2)
ωj D n m T ⋆
ωj D n
∆χj ωk D m∆χk
T
n=1 m=1
n! m!
j=0 j=0 k=0
(50)
In case of Gaussian random variables, independently of the value of ω0⋆ , (38) and (50) agree
up to the third order. Weight ω0⋆ can be set by matching part of the fourth order terms
T
(specifically those involving D 2∆χj D 2∆χk in 50), thereby obtaining (see also [52]):
Nχ
ω0⋆ = 4 − − ψ2 (51)
ψ2
In what precedes we have assumed the mapping f to be analytic everywhere in the x

space. If f is not differentiable, low order terms of the Taylor series expansions of mean
and covariance get affected. It is difficult to quantify the discrepancies with respect to
the analytic case, because they depend on whether the sigma-points sample the loci of non-
differentiability. However, it can be generally said that the order of accuracy is detrimentally
affected.
4. Results
To assess the performances of the proposed filtering approach in calibrating the in-
terlaminar constitutive law while detecting impact-induced delamination, we first study a
simple problem consisting of a two-layer composite stricken by a homogeneous impactor.
Hence, two different impact tests on GRP composites [14, 53] are considered to mainly
show the accuracy in detecting delamination in real-time.
In all the cases, it is assumed that the contact between specimen and impactor is per-
fect (i.e. distributed all over their approaching surfaces) and that failure of the laminate
occurs because of the propagation of dilatational plane waves in the through-the-thickness
direction: the interlaminar surfaces are therefore subject to pure mode I loading.
4.1. Pseudo-experimental Testings

As a starting benchmark, a pseudo-experimental testing condition is conceived. The
pseudo-experimental response of the laminate to the impact loading has been computed by
adding a white noise (of assigned variance) to sampled outcomes of finite element analyses.
Even though this approach is sometimes criticized, being not clear whether one is test-
ing with it the filtering approach or its implementation, it helps in getting insights into the
performance of the filter in terms of stability and convergence rate. Indeed, calibration of
interlaminar constitutive models may become difficult if delamination occurs almost instan-
taneously: the filter has to be highly sensitive to model parameters to promptly react to the
information conveyed by measurements. This requires a careful setting of filter parameters,
like P 0 and V . It is worth mentioning that, to further complicate the problem, the effects
of the shape of the tensile envelope of interlaminar laws (at assigned strength and tough-
ness) on the overall response of a laminate subject to impacts have not been thoroughly
understood yet: depending on the loading and boundary conditions, on the composite ge-
ometry and on the stacking sequence, in some circumstances the shape affects the response,
whereas in others it does not [54].
The capabilities of the SPKF are therefore first assessed through a simple test: a two-
layer composite is stricken by a homogeneous impactor. The laminae and the impactor are
114 Stefano Mariani
Figure 5. impact on a two-layer composite. Space-time diagram (the vertical dashed line
here represents the possibly debonding surface when a brittle, homogeneous material is
subject to the same impact).
assumed to be isotropic and elastic, featuring Young’s modulus E = 10 GPa, Poisson’s ratio
ν = 0.35 and mass density ̺ = 1500 kg/m3 (see also [11]). Each lamina and the impactor
are 0.75 mm in thickness. Target mechanical properties of the interlaminar surfaces are
assumed:
2K =277.09 (N/mm3 )
τM =75 (MPa) (52)
G =0.15 (N/mm)
According to the space-time diagram of Figure 5, failure can occur only along the interlam-
inar surface because of the interaction of the two release waves propagating inwards from
the free surfaces of impactor and specimen.
Two different values of the velocity v of the impactor are considered. In a first case
v = 10.19 m/s leads to the propagation in the through-the-thickness direction of a com-
pressive/tensile wave of amplitude τ̄ = 50 MPa, which does not cause interface failure. In
a second case v = 20.38 m/s causes laminate failure, i.e. whole delamination, because of
the propagation of a compressive/tensile wave of amplitude τ̄ = 100 MPa. Outcomes of
the two tests are respectively reported in Figures 6 and 7, in terms of time evolution of the
free surface velocity velocity u̇r at the rear laminate surface, of the opening displacement
discontinuity [u] and of the normal traction τ (here and in what follows the subscript n has
been dropped to simplify the notation). Results are shown for all the constitutive models
described in Section 2.2., having assumed q = 1 for the modified exponential interface law
(17). For comparison purposes, the response of an interface-free specimen is reported too;
in such a case, the purely elastic behavior of the material leads to the propagation of sharp
fronts of a shock wave.
If delamination is not incepted (τ̄ = 50 MPa), the response is almost independent of
the shape of the interface law. Only in the presence of an interface that behaves according
to the exponential model, the pre-peak nonlinearity of the constitutive law (see Figure 2(a))
(a)
(b) (c)
Figure 6. impact on a two-layer composite, τ̄ = 50 MPa. Effects of the interlaminar laws on

the time evolution of (a) the free surface velocity u̇r , of (b) the displacement discontinuity
[u] and (c) relevant traction τ along the interface.
leads to a larger opening [u] of the interlaminar surface when subject to a tensile stress. For
any constitutive model, the signature of the interface is shown in Figure 6(a) by the delay in
the sudden changes of u̇r with respect to the reference, interface-free solution. This delay,
which grows in time, is caused by the compliance of the interface, that is additional to the
bulk one.
In case of failure (τ̄ = 100 MPa), the free surface velocity u̇r is affected by the interface
model only when the waves, traveling across the interlaminar surface while softening is
taking place, reach the rear surface; this occurs in the present case around 1µs after the
impact, see Figure 7(a). After failure, the rear lamina detaches from the front one and
freely flies off, as testified by the diverging [u] history in Figure 7(b). Part of the shock
waves then get confined inside the back lamina: this explains the subsequent doubling of
the drops in the u̇r evolution. It is worth noting that the time elapsed between softening
116 Stefano Mariani
(a)
(b) (c)
Figure 7. impact on a two-layer composite, τ̄ = 100 MPa. Effects of the interlaminar laws
on the time evolution of (a) the free surface velocity u̇r , of (b) the displacement discontinu-
ity [u] and (c) relevant traction τ along the interface.
inception (t ∼ = 0.72 µs) and whole failure of the interface (t ∼

= 0.9 µs), see Figure 7(c), is
very short; only the information on this failure event, constituting the so-called pull-back
signal (PBS), conveyed by the shock waves to the free laminate surface, can be used by the
filter to calibrate the interface law.
The effects on the PBS of the shape of the tensile envelope and of the strength τM and
toughness G values need to be assessed. In the absence of any dissipation mechanisms, in
this test the free surface velocity u̇r would drop to zero at t = 0.917 µs (see Figure 6(a));
in case of delamination, the minimum attained velocity after the arrival of the unloading
tensile wave and the shape of the PBS do furnish information on the interface response. To
understand the roles of interface law, strength and toughness, the results of a parametric
analysis are shown in Figure 8: for any interface model, τM and G are varied by 20%
at most with respect to the target values (52). The piecewise linear and linear-exponential
(a) (b)
(c) (d)
(e) (f)
Figure 8. impact on a two-layer composite, τ̄ = 100 MPa. Effects of the interface strength
τM (left column) and toughness G (right column) on the pull-back signal. (a-b) piecewise
linear law (12); (c-d) linear-exponential law (13); (e-f) exponential law (14).
118 Stefano Mariani
laws, having a common initial elastic phase in tension, lead to a common descending branch
in the PBS. At variance, the local tangent stiffness of the exponential law in the hardening
phase is affected by τM and G: the slope of the descending branch of the PBS is therefore
affected by τM and G too. Independently of the interface law, τM turns out to affect the
starting stage of debonding, whereas G affects its tail and the time needed to complete it;
this is clearly evidenced by the PBS, since τM modifies the minimum attained velocity,
whereas G influences only the ascending branch with no effects on the pull-back velocity.
From a model calibration perspective, it is clear that the SPKF has chances to improve
parameter estimates only in the time interval 0.9 ≤ t ≤ 1.2 µs. It is therefore hard to
figure out from the previous plots whether the effects of tensile envelope, τM and G can
be actually interpreted by the filter to improve model calibration. The performances of
the SPKF are hence tested here not only looking at parameter estimates, but also checking
its capability to detect whether a laminate is failing and, in case of delamination, where it
actually takes place.
Typical results of the filtering procedure are depicted in Figures 9 and 10; in this case
the pseudo-experimental data, which consist in the free surface velocity alone, have been
supposed very accurate, featuring a standard deviation δ = 0.33 m/s (the measurement
error covariance matrix W becomes scalar-valued, with entry W = δ 2 ). As far as the
process covariance matrix V is concerned, in case of pseudo-experimental testing it can
be assumed to be vanishing, since the filter employs the same structural model adopted to
get the pseudo-experimental data. Components of P 0 instead need to be finely tuned to
enhance filter convergence [11, 15].
Figure 9 shows the obtained estimates of τM and G as a function of their initial guess in
x̂0 (here respectively denoted by τM,0 and G0 ), for all the interface models. These estimates
evolve from τM,0 and G0 once the PBS is processed by the filter; after that, they become
stationary. The tracked state of the laminate is shown in Figure 10 in terms of predicted
interface opening [u] and free surface velocity u̇r . Knowing the target response of the
composite to the impact, here denoted by the dashed lines, allows to certify stability and
convergence of the SPKF, no matter if displacement is diverging in a part of the system and
what kind of interface constitutive law has been adopted. It can be seen that estimates get
enhanced as soon as the filter senses the PBS: in fact, the sudden changes in the estimate of
[u] show up only while processing the PBS, starting from t ∼ = 0.9 µs.
These outcomes testify that the SPKF is very efficient in tracking the state of the lam-
inate, i.e. in understanding whether the structure is failing or not. Model calibration is
instead less accurately accomplished: independently of the interface law, τM is quite pre-
cisely estimated, provided the initial guess τM,0 is not too far from the target value, whereas
G can be hardly inferred. These conclusions are in agreement with the results of the para-
metric analysis: while τM affects the whole PBS, G affects only its ascending branch.
Therefore, filtering out from u̇r the effects of G alone turns out to be extremely difficult.
In case of a much higher scattering of pseudo-experimental data (δ = 3.3 m/s), see
Figure 11, results loose accuracy as for the calibration task. Contrariwise, state tracking
maintain accuracy: even though measurements contain poor information, the SPKF is again
able to provide the evolution of the free surface velocity in the PBS.
(a) (b)
(c) (d)
(e) (f)
Figure 9. impact on a two-layer composite, τ̄ = 100 MPa (W=10−1 m2 /s2 ). Effects of the
initialization values τM,0 and G0 on the converged estimates of τM (left column) and G
(right column). (a-b) piecewise linear law; (c-d) linear-exponential law; (e-f) exponential
law.
120 Stefano Mariani
(a) (b)
(c) (d)
(e) (f)
Figure 10. impact on a two-layer composite, τ̄ = 100 MPa (W=10−1 m2 /s2 ). Evolution in
time of interface opening [u] (left column) and free surface velocity u̇r (right column), and
comparison among tracked state (orange squares), actual state (dashed lines) and pseudo-
experimental data (blue circles). (a-b) piecewise linear law; (c-d) linear-exponential law;
(e-f) exponential law.
(a) (b)
(c) (d)
(e) (f)
Figure 11. impact on a two-layer composite, τ̄ = 100 MPa (W=10 m2 /s2 ). Evolution in
time of interface opening [u] (left column) and free surface velocity u̇r (right column), and
comparison among tracked state (orange squares), actual state (dashed lines) and pseudo-
experimental data (blue circles). (a-b) piecewise linear law; (c-d) linear-exponential law;
(e-f) exponential law.
122 Stefano Mariani
(a)
(b)
Figure 12. impact on a 7-layer composite [53]. (a) space-time diagram (the vertical dashed
line here represents the possibly debonding surface when a brittle, homogeneous material
is subject to the same impact), and (b) experimentally measured free surface velocity.
4.2. Actual Experimental Testings

To finally check the performances of the SPKF when dealing with multi-layered com-
posites, we consider two of the experiments reported in [53] and [14].
In the first experiment (experiment FY06001 in [53]), the specimen is a 7-layer com-
posite plate; each lamina is 1.37 mm in thickness, and is made of a balanced 5-harness satin
weave E-glass and LY564 epoxy. The wave speed in the through-the-thickness direction is
3.34 km/s, while the mass density is ̺ = 1885 kg/m3 [53]. This laminate was subject to a
plane impact, stricken by an aluminum impactor (12.5 mm thick) flying at velocity v = 71
m/s. The relevant space-time diagram is shown in Figure 12, along with the free surface
velocity profile measured via a velocity interferometer for any reflector (VISAR).
Account taken of the high accuracy of the experimentally measured u̇r , the identifica-
tion procedure have furnished the results reported in Figure 13 in terms of time evolution
of estimates of τM and G as a function of their initialization values within the domain:
Cϑ = {50 ≤ τM,0 ≤ 250 (MPa), 0.5 ≤ G0 ≤ 2.5 (N/mm)} (53)

(a) (b)
(c) (d)
(e) (f)
Figure 13. impact on a 7-layer composite [53]. Evolution in time of the estimated values of
τM (left column) and G (right column). (a-b) piecewise linear law; (c-d) linear-exponential
law; (e-f) exponential law.
124 Stefano Mariani
(a)
(b) (c)
(d) (e)
(f) (g)
Figure 14. impact on a 7-layer composite [53]. Evolution in time of (a) free surface velocity
u̇r and (b-g) estimated interface openings [u]1 − [u]6 .
(a)
(b)
Figure 15. impact on a 11+11-layer composite [14]. (a) space-time diagram (the vertical
dashed line here represents the possibly debonding surface when a brittle, homogeneous
material is subject to the same impact), and (b) experimentally measured free surface ve-
locity.
Converged estimates of τM are in good agreement with the spall strength of 119.5 MPa
reported in [53]; on the other hand, final estimates of G are well representative for this kind
of composites. Figure 14 reports the estimated state of the specimen: the capability to track
the measured free surface velocity and to foresee delamination along the third interlaminar
surface away from the impact plane, is evidenced. This latter result, allowing also for wave
dispersion caused by interlaminar surfaces and inner inhomogeneities of the composite,
well agrees with the state-space diagram of Figure 12(a).
In the second experiment (experiment 1 in [14]), a GRP specimen, 7.02 mm thick, is
backed by another GRP plate, 6.91 mm thick; both laminates are made of 11 plies. The
wave speed in the through-the-thickness direction now amounts to 3.19 km/s, and the mass
density to ̺ = 1867 kg/m3 [14]. The specimen is stricken by a 5-layer GRP flyer, 2.96 mm
in thickness, flying at velocity v = 85 m/s. The corresponding space-time diagram, and the
free surface velocity profile measured via a VISAR are reported in Figure 15. Because of
the test set-up, the release waves interact causing delamination inside the back plate.
126 Stefano Mariani
(a) (b) (c)
(d) (e) (f)
(g) (h) (i)
(j) (k)
Figure 16. Impact on a 11+11-layer composite [14]. Evolution in time of (a) free surface
velocity u̇r and (b-k) estimated interface openings [u]1 − [u]10 in the back plate.
Results of the filtering process are reported in Figure 16 is terms of tracked free surface
velocity and estimated displacement jumps along all the interfaces inside the back plate
(sequence starts at the specimen-back plate contact surface). These estimations turn out
once again to be independent of the interface law and of the initialization values of τM and
G inside the domain:
Cϑ = {25 ≤ τM,0 ≤ 100 (MPa), 0.1 ≤ G0 ≤ 0.6 (N/mm)} (54)
While the free surface velocity is accurately tracked, delamination is foreseen to take place
along the 7-th interlaminar surfaces, in agreement with the results of [11, 14]. As far as
model calibration is concerned, outcomes turn out to be qualitatively in agreement with
those already reported for the previous tests.
5. Conclusion
In this Chapter we have addressed some issues related to constitutive modeling and pa-
rameter identification in finite element simulations of layered composites subject to impacts.
Assuming the impact energy to be high enough to cause damage spreading inside the inter-
laminar resin-enriched phases, but not high enough to result in penetration of the impactor
accompanied by intralaminar damage, a numerical scheme for structural-level analyses has
been revised. Within this scheme the laminae are assumed to behave elastically, whereas
dissipation mechanisms are lumped onto zero-thickness interlaminar surfaces. Along these
interlaminar surfaces strength reduction, eventually leading to delamination, is governed by
softening interface constitutive laws linking tractions to displacement jumps.
Interface laws are known to be difficult to calibrate, since direct testing on a single
interlaminar phase can not be devised. Here we have offered a sigma-point Kalman filtering
approach to estimate uncertain model parameters of the aforementioned interface laws. This
technique outperforms most of the customarily adopted ones, since it efficiently deals with
nonlinearities, which are a result of interlaminar strength degradation in the case under
study.
The performances of the filtering procedure have been assessed through pseudo-
experimental testings on a two-layer composite, and through real testings on multi-layer
glass fiber reinforced plastic composites. It has been shown that the state of the composite,
including delamination inception and growth, is always tracked with a noteworthy level of
accuracy. Due to the fast failure events, model calibration is instead less accurately per-
formed and sometimes requires initialization values of uncertain model parameters to be
appropriately chosen to avoid biased estimates.
References
[1] S. Abrate. Impact on composite structures. Cambridge University Press, 1998.
[2] H.D. Espinosa, S. Dwivedi, P.D. Zavattieri, and G. Yuan. A numerical investigation of
penetration in multilayered material/structure systems. International Journal of Solids
and Structures, 35:2975–3001, 1998.
[3] O. Allix and P. Ladevéze. Interlaminar interface modelling for the prediction of de-
lamination. Composite Structures, 22:235–242, 1992.
128 Stefano Mariani
[4] A. Corigliano. Formulation, identification and use of interface elements in the numer-
ical analysis of composite delamination. International Journal of Solids and Struc-
tures, 30:2779–2811, 1993.
[5] A. Corigliano and O. Allix. Some aspects of interlaminar degradation in composites.

Computer Methods in Applied Mechanics and Engineering, 185:203–224, 2000.
[6] A. Corigliano and S. Mariani. Simulation of damage in composites by means of inter-

face models: parameter identification. Composites Science and Technology, 61:2299–
2315, 2001.
[7] G. Alfano and M.A. Crisfield. Finite element interface models for the delamination
analysis of laminated composites: mechanical and computational issues. International
Journal for Numerical Methods in Engineering, 50:1701–1736, 2001.
[8] A. Corigliano. Damage and fracture mechanics techniques for composite structures. In
I. Milne, R.O. Ritchie, and B. Karihaloo, editors, Comprehensive Structural Integrity,
volume 3, chapter 9, pages 459–539. Elsevier Science, 2003.
[9] A. Corigliano and S. Mariani. Parameter identification of a time-dependent elastic-

damage interface model for the simulation of debonding in composites. Composites
Science and Technology, 61:191–203, 2001.
[10] A. Corigliano and S. Mariani. Identification of a constitutive model for the simula-
tion of time-dependent interlaminar debonding processes in composites. Computer
Methods in Applied Mechanics and Engineering, 191:1861–1894, 2002.
[11] S. Mariani and A. Corigliano. Impact induced composite delamination: state and pa-
rameter identification via joint and dual extended Kalman filters. Computer Methods
in Applied Mechanics and Engineering, 194:5242–5272, 2005.
[12] S. Bittanti, G. Maier, and A. Nappi. Inverse problems in structural elastoplasticity:

a Kalman filter approach. In A. Sawczuk and G. Bianchi, editors, Plasticity today.
Modelling, Methods and Applications, pages 311–329. Elsevier Applied Science Pub-
lishers, 1984.
[13] A. Corigliano, S. Mariani, and B. Orsatti. Identification of Gurson-Tvergaard material

model parameters via Kalman filtering technique - I. Theory. International Journal of
Fracture, 104:349–373, 2000.
[14] H.D. Espinosa, S. Dwivedi, and H.-C. Lu. Modeling impact induced delamination of
woven fiber reinforced composites with contact/cohesive laws. Computer Methods in
Applied Mechanics and Engineering, 183:259–290, 2000.
[15] A. Corigliano and S. Mariani. Parameter identification in explicit structural dynamics:

performance of the extended Kalman filter. Computer Methods in Applied Mechanics
and Engineering, 193:3807–3835, 2004.
[16] S.J. Julier, J. Uhlmann, and H.F. Durrant-Whyte. A new method for the nonlinear
transformation of means and covariances in filters and estimators. IEEE Transactions
on Automatic Control, 45:477–482, 2000.
[17] T. Lefebvre, H. Bruyninckx, and J. De Schutter. Kalman filters for nonlinear systems:
a comparison of performance. International Journal of Control, 77:639–653, 2004.
[18] K. Ito and K. Xiong. Gaussian filters for nonlinear filtering problems. IEEE Transac-
tions on Automatic Control, 45:910–927, 2000.
[19] S. Mariani and A. Ghisi. Unscented Kalman filtering for nonlinear structural dynam-
ics. Nonlinear Dynamics, 49:131–150, 2007.
[20] A. Corigliano, S. Mariani, and A. Pandolfi. Numerical modeling of rate-dependent

debonding processes in composites. Composite Structures, 61:39–50, 2003.
[21] A. Corigliano, S. Mariani, and A. Pandolfi. Numerical analysis of rate-dependent

dynamic composite delamination. Composites Science and Technology, 66:766–775,
2006.
[22] R. Lund and J. Fish. Space-time multiscale laminated theory. International Journal
of Multiscale Computational Engineering, 2:26–47, 2004.
[23] S. Mariani, A. Pandolfi, and R. Pavani. Coupled space-time multiscale simulations of

dynamic delamination tests. Materials Science, 23:509–519, 2005.
[24] R.R. Talreja Y.W. Kwon, D.H. Allen, editor. Multiscale Modeling and Simulation of
Composite Materials and Structures. Springer, 2007.
[25] G.T. Camacho and M. Ortiz. Computational modelling of impact damage in brittle
materials. International Journal of Solids and Structures, 33:2899–2938, 1996.
[26] Y. Arun Roy and R.H. Dodds. Simulation of ductile crack growth in thin allu-
minum panels using 3-D surface cohesive elements. International Journal of Fracture,
110:21–45, 2001.
[27] S. Mariani and U. Perego. Extended finite element method for quasi-brittle fracture.
International Journal for Numerical Methods in Engineering, 58:103–126, 2003.
[28] C. Comi and S. Mariani. Extended finite element simulation of quasi-brittle fracture
in functionally graded materials. Computer Methods in Applied Mechanics and Engi-
neering, 196:4013–4026, 2007.
[29] S. Li, M.D. Thouless, A.M. Waas, J.A. Schroeder, and P.D. Zavattieri. Mixed-mode
cohesive-zone models for fracture of an adhesively bonded polymer matrix composite.
Engineering Fracture Mechanics, 73:64–78, 2006.
[30] J.H. Rose, J. Ferrante, and J.R. Smith. Universal binding energy curves for metals and
bimetallic interfaces. Physical Review Letters, 47:675–678, 1981.
130 Stefano Mariani
[31] J.H. Rose, J.R. Smith, and J. Ferrante. Universal features of bonding in metals. Phys-
ical Review B, 28:1835–1845, 1983.
[32] X.-P. Xu and A. Needleman. Numerical simulation of fast crack growth in brittle
solids. Journal of the Mechanics and Physics of Solids, 42:1397–1434, 1994.
[33] A. Needleman and A.J. Rosakis. The effect of bond strength and loading rate on
the conditions governing the attainment of intersonic crack growth along interfaces.
Journal of the Mechanics and Physics of Solids, 47:2411–2449, 1999.
[34] C.M. Landis, T. Pardoen, and J.W. Hutchinson. Crack velocity dependent toughness
in rate dependent materials. Mechanics of Materials, 32:663–678, 2000.
[35] O. C. Zienkiewicz and R. L . Taylor. The finite element method, volume II. McGraw-
Hill, London, 1989.
[36] N. Moës, J. Dolbow, and T. Belytschko. A finite element method for crack growth
without remeshing. International Journal for Numerical Methods in Engineering,
46:131–150, 1999.
[37] T. Strouboulis, K. Copps, and I. Babuška. The generalized finite element method.
Computer Methods in Applied Mechanics and Engineering, 190:4081–4193, 2001.
[38] G.N. Wells and L.J. Sluys. A new method for modelling cohesive cracks using finite
elements. International Journal for Numerical Methods in Engineering, 50:2667–
2682, 2001.
[39] C. Comi, S. Mariani, and U. Perego. An extended FE strategy for transition from
continuum damage to mode I cohesive crack propagation. International Journal for
Numerical and Analytical Methods in Geomechanics, 31:213–238, 2007.
[40] M. Jirasek and T. Zimmerman. Embedded crack model: II. Combination with smeared
cracks. International Journal for Numerical Methods in Engineering, 50:1291–1305,
2001.
[41] H.M. Hilber, T.J.R. Hughes, and R.L. Taylor. Improved numerical dissipation for time
integration algorithms in structural dynamics. Earthquake Engineering and Structural
Dynamics, 5:283–292, 1977.
[42] T.J.R. Hughes. The finite element method. Linear static and dynamic finite element
analysis. Dover Publications, 2000.
[43] L. Ljung. System identification. Theory for the user. Prentice Hall, second edition,
1999.
[44] A. Gelb. Applied optimal estimation. The M.I.T. Press, 1974.
[45] E.A. Wan and A.T. Nelson. Dual extended Kalman filter methods. In S. Haykin,
editor, Kalman filtering and neural networks, pages 123–173. John Wiley & Sons,
Inc., 2001.
[46] R. E. Kalman and R. S. Bucy. New results in linear filtering and prediction theory.
ASME Journal of Basic Engineering, 83D:95–108, 1961.
[47] L. Ljung. Asymptotic behavior of the extended Kalman filter as a parameter estimator
for linear systems. IEEE Transactions on Automatic Control, AC-24:36–50, 1979.
[48] K. Reif, S. Günther, E. Yaz, and R. Unbehauen. Stochastic stability of the discrete-
time extended Kalman filter. IEEE Transactions on Automatic Control, 44:714–728,
1999.
[49] S.J. Julier and J.K. Uhlmann. A general method for approximating nonlinear trans-
formations of probability distributions. Technical report, RRG, Department of Engi-
neering Science, University of Oxford, 1996.
[50] S.J. Julier, J. Uhlmann, and H.F. Durrant-Whyte. A new approach for filtering nonlin-
ear systems. In Proceedings of the American Control Conference, pages 1628–1632,
Seattle, June 1995.
[51] E.A. Wan and R. van der Merwe. The unscented Kalman filter. In S. Haykin, editor,
Kalman filtering and neural networks, pages 221–280. John Wiley & Sons, Inc., 2001.
[52] S.J. Julier. The scaled unscented transformation. In Proceedings of the American
Control Conference, volume 6, pages 4555–4559, Anchorage, May 2002.
[53] F. Yuan, L. Tsai, V. Prakash, A.M. Rajendran, and D.P. Dandekar. Spall strength
of glass fiber reinforced polymer composites. International Journal of Solids and
Structures, 44:7731–7747, 2007.
[54] G. Alfano. On the influence of the shape of the interface law on the application of
cohesive-zone models. Composites Science and Technology, 66:723–730, 2006.
Chapter 4
CURRENT STATE OF THE ART OF THE CERAMIC

COMPOSITE MATERIAL BIOLOX®DELTA
Meinhard Kuntz1, Bernard Masson2 and Thomas Pandorf1

1
CeramTec AG, Plochington, Germany
2
CeramTec AG, Pechabou, France
Abstract
An extensive overview about the state of the art of the ceramic composite material
BIOLOX®delta is given. The unique properties rely on a well defined alumina based fine
composite microstructure which is mainly achieved by high temperature solid body reaction
of the different ceramic phases during sintering. Zirconia comprises 17 % of the total volume.
The tetragonal phase of zirconia is stabilized chemically and mechanically.
The high strength and toughness of the material depend on transformation toughening of
the zirconia which is clearly shown by various experimental results. The excellent mechanical
properties are reproduced batch by batch with a very low scatter.
The outstanding properties of the material BIOLOX®delta support advantageous
properties of the final product, e.g. ceramic hard-hard bearings for hip arthroplasty. The burst
load of the components is significantly increased. It is shown that the design of the
components is also very important for the reliability and the ultimate properties of the system.
Wear properties at severe conditions are significantly improved by using the new composite
material BIOLOX®delta in comparison to pure alumina.
Phase transformation of zirconia from the tetragonal to the monoclinic phase due to
hydrothermal aging is extensively discussed. Due to the particular distribution and
stabilization of the zirconia particles instable aging effects are not possible in this material.
After very long time of accelerated aging conditions an increase of monoclinic phase is found
– however, it is shown that dynamic and static properties of BIOLOX®delta are not
influenced by this effect.
1. Introduction
Since 2001 more than 500.000 artificial hip joints with components of the new high
performance ceramic composite BIOLOX®delta have been successfully implanted on a
134 Meinhard Kuntz, Bernard Masson and Thomas Pandorf
global basis. Due to the unique strength and toughness of this material the risk of fracture has
been substantially reduced when compared to conventional ceramic materials.
The outstanding properties of BIOLOX®delta rely on complex reinforcing mechanisms.
Therefore, it is necessary to assess if reinforcement is maintained throughout the life-time of
the artificial joint which is anticipated to exceed more than 20 years. Furthermore, it is shown
that the challenging production of BIOLOX®delta is reproduced at a high quality from batch
to batch.
Within the scope of this technical contribution the composite ceramic BIOLOX®delta is
extensively described and analyzed. The composition and the material properties are
presented based on data of regular production lots. It is shown that the advantageous
properties of this material are based on the reinforcing mechanisms which are activated due to
the unique composition of this material.
The particular effect of monoclinic phase transformation and hydrothermal aging is
described in detail based on general mechanisms and specific analysis of phase
transformation in BIOLOX®delta. Furthermore, experimental data are provided which
describe the long term properties of the material, in particular with respect to hydrothermal
phase transformation of zirconia in combination with wear and cyclic load conditions.
2. International Material Standards

BIOLOX®delta is a modern ceramic composite material for biomedical applications. The
main components of the composite are alumina and zirconia. There are ISO standards
available for bioceramics of high purity alumina (ISO 6474 - 1) and high purity zirconia (ISO
13356). However, these standards are not directly applicable for the composition of
BIOLOX®delta. The ISO organisation is already on the way to prepare a new standard which
is applicable for BIOLOX®delta and other similar composite materials (ISO 6474 - 2). The
new standard will be released presumably in 2010.
Meanwhile, it is helpful to refer and compare the properties of BIOLOX®delta to the
draft of the new ISO 6474-2 and the other international material standards which are
applicable for related high purity bioceramics.
ISO 6474 – 1 Implants for Surgery – Ceramic Materials - Part 1: Ceramic

Materials Based on High Purity Alumina
The current version of this standard was released in 1994. The material properties which
are defined here reflect typical properties of high quality pure alumina. Strength and
toughness which are required according to this standard are significantly lower than those
which are available with the composite material BIOLOX®delta.
Today, the experts of the ISO working group agree that some details of the current ISO
6474 do not represent the state of the art. Thus, a new version of ISO 6474 - 1 is being
preparedwhich is already published as a Draft International Standard (ISO/DIS). Most
technical details in this version are already finally implemented. Some details (fracture
Current State of the Art of the Ceramic Composite Material BIOLOX®delta 135
toughness, microstructure) are still under discussion. Presumably, the final release of the new
standard will be in 2009.
Comment: There is a similar ASTM standard F 603 for the same application and material
type. The required material properties are comparable.
ISO 6474 – 2. Implants for Surgery – Ceramic Materials - Part 2: Composite

Materials Based on a High Purity Alumina Matrix with
Zirconia Reinforcement
A standard which is applicable for alumina zirconia composite materials is under

preparation. Such composite materials are distinguished in those where the main phase is
alumina (ZTA = zirconia toughened alumina) and those where zirconia is the dominating
phase (ATZ= alumina toughened zirconia). Both material types are available for biomedical
applications.
The basic physical properties (e.g. hardness, thermal conductivity) of a composite are
primarily derived from the main phase. It is thus useful to describe alumina based zirconia
toughened materials parallel to pure alumina materials. Consequently, the new standard has
been proposed as part 2 of the established standard ISO 6474 for pure alumina. This concept
has been discussed at the ISO TC 150 meetings and was approved by the international
experts.
The new standard ISO 6474 - 2 is accepted as a Working Draft. It covers all material
properties (except of Young´s modulus) which are defined in ISO 6474 - 1. Additionally, the
specific subjects of hydrothermal aging and radioactivity, which are relevant for zirconia
toughened materials is also included.
ISO 6474 – 2 is defined such that a broad range of inorganic compositions are included.
A composition of ≥ 60 wt. % alumina and ≥ 10 wt. % zirconia is required. Other ingredients
are allowed. The new standard is not exclusively designated for BIOLOX®delta.
So far there is no ASTM standard for alumina zirconia composite biomaterials available.
ISO 13 356. Implants for Surgery –Ceramic Materials Based on Yttria-

Stabilized Tetragonal Zirconia (Y-TZP)
This standard was revised and published in 2008 as an International Standard (ISO).
In contrast to pure alumina, zirconia as a ceramic material can not be produced without a
significant amount of other substances for phase stabilization. Several elements are known
which are applicable. ISO 13356 is only focussed on Yttria as the stabilizing element. A
specific range of Y content is predetermined. It should be noted that the typical range of Y in
pure zirconia materials can be different to the required amount of Y in alumina zirconia
composites. This issue is thoroughly discussed in chapter 4.
Pure zirconia bioceramics can be applied either for biomedical bearings (e.g. hip or knee)
or for dental applications. ISO 13 356 does not distinguish between these different
applications.
As a specific issue of zirconia a test for accelerated hydrothermal aging is required.
Table 1 gives an overview of the required properties of the 3 standards. The material
properties according to the latest revised versions are chosen
Table 1. Required material properties according to the ISO standards 6474 – 1 (pure
alumina), ISO 6474 – 2 (alumina – zirconia composite) and ISO 13 356 (zirconia).
ISO Standard ISO 6474 - 1 ISO 13 356 ISO 6474 – 2

Material Unit Pure Alumina Zirconia Alumina Zirconia
Average Bulk ≥ 3,94 g/cm3 ≥ 6,00 g/cm3 ≥ 98,7 %
density
Chemical wt% Al2O3 ≥ 99,7 ZrO2+HfO2+Y2O3 ≥ Al2O3 60 - 90
Composition MgO ≤ 0,2 99,0 ZrO2+HfO2 10 - 30
Impurities ≤ 0,1 Y2O3 4,5 – 6,0 Additives ≤ 10
HfO2 ≤ 5 Impurities ≤ 0,2
Al2O3 ≤ 0,5
Others ≤ 0,5
Grain Size µm MV ≤ 2,5 MV ≤ 0,4 Al2O3 MV ≤ 1,5
SD ≤ 40 % ZrO2 MV ≤ 0,6
SD ≤ 40 %
Strength MPa ≥ 500 ≥ 800 ≥ 1000
Weibull Modulus ≥8 ≥ 10
(4 pt bending)
Young´s modulus GPa ≥ 380
Fracture MPa ≥ 2,5 ≥ 4,0
Toughness m
Hardness HV1 GPa ≥ 18 ≥ 17
Wear Resistance Info Info
Cyclic fatigue No failure No failure No failure
limit at 200 MPa at 320 MPa at 400 MPa
Amount of % ≤ 20
monoclinic phase
Accelerated ≤ 25 % monocl. phase Accomplish
Aging strength decrease not requirements
more than 20% described above
Radioactivity Bq / kg ≤ 200 ≤ 100
Note: The values of ISO 6474 – 1 & 2 are not finally fixed at the date of this publication.
3. Description of BIOLOX®delta
BIOLOX®delta is an alumina based composite ceramic. Approximately 80 vol.-% of the

matrix consist of fine grained high purity alumina which is very similar to the well known
material BIOLOX®forte (ISO 6474). As it is the case in any other composite material, the
basic physical properties like stiffness, hardness, thermal conductivity etc. are mainly
predetermined from the dominating phase. It was the basic idea for the development of the
new material to preserve all the desirable properties of BIOLOX®forte - as an excellent
bioceramic with more than 30 years clinical experience - but to increase its strength and
toughness.
These properties are substantially improved by implementation of reinforcing elements.
Figure 1 shows the microstructure of BIOLOX®delta.
Alumina matrix
Zirconia
Platelets
Figure 1. Microstructure of BIOLOX®delta.
Two reinforcing components are integrated into BIOLOX®delta. 17 vol.-% of the matrix
consist of tetragonal zirconia particles. The average grain size of the zirconia is around
0.27 µm. As a further reinforcing element, approximately 3 vol.-% of the matrix are built by
platelet shaped crystals of the ceramic composition strontium aluminate. The platelets stretch
to a maximum length of approximately 5 µm with an aspect ratio of 5 – 10. The reinforcing
ability of these ingredients is explained below.
Additionally to the reinforcing components, there are also stabilizing elements doped to
the material. Chromium is added which is soluble in the alumina matrix and increases the
hardness of the composite. The minor amount of chromium is the reason for the pink color of
the material, see Figure 2. Furthermore, some yttrium is added to the composite which is
solved in the zirconia and supports the stabilization of the tetragonal phase. In Table 2, the
composition is given:
Table 2. Raw material specification for BIOLOX®delta
Ingredient Formula Weight percent

Yttrium oxide Y2O3
Chromium oxide Cr2O3 1,4 – 2,0
Strontium oxide SrO
Zirconium oxide ZrO2 24,0 – 25,5
TiO2, MgO, SiO2, CaO, Fe2O3,
Other oxides < 0,22
Na2O, K2O
Alumina Al2O3 balance
During thermal treatment of the material the ingredients are transformed to the particular
composition with the 3 components. The basic transformation equations are known as
follows:
2 YCrO3 → Y2O3 + Cr2O3 (1)
SrZrO3 → SrO + ZrO2 (2)
Al2O3 + Cr2O3 → Al2O3 : Cr (3)
SrO + 6 Al2O3 : Cr → SrAl12− x CrxO19 (4)
ZrO2 + Y2O3 → ZrO2 : Y (5)
In Table 3 the volume fractions of the final products according to equations

(3) – (5) are given.
Table 3. Components of the final composite BIOLOX®delta
Component of the composite Formula Volume percent

Alumina, doped with Chromia Al2O3:Cr 80 %
Zirconia with Y-stabilization ZrO2:Y 17%
Strontiumaluminate (minor Cr-content) SrAl12-xCrxO19 3%
Figure 2. Ball heads and inserts of BIOLOX®delta.

4. Reinforcing Mechanism on BIOLOX®Delta

Benefit of Phase Transformation
The reinforcing elements, in particular zirconia, substantially increase fracture toughness

and strength of the material. Fracture toughness (KIC) is a measure for the ability of the
material to withstand crack extension. Strength (σc) is defined as the maximum stress within a
structure at failure of the component.
The correlation of strength and toughness is given in the fundamental equation of fracture
mechanics:
K IC = σ c ac Y (6)
where ac is the size of a typical critical defect in the material and Y the shape factor.
Consequently, when the fracture toughness of the alumina is increased also the strength is
directly improved. This basic principle is the concept of the development of BIOLOX®delta.
The microstructure is designed in order to provide an optimum of resistance against crack
extension.
The benefit in crack resistance which is obtained from incorporating zirconia into an
alumina matrix (as shown in Figure 3) is well known in the science of high performance
ceramics.
1 µm
Figure 3. Reinforcing mechanism in BIOLOX®delta at crack initiation and propagation. Yellow

particles represent tetragonal zirconia. Color change to red indicates monoclinic phase transformation.
Arrows show the region of compressive stresses due to phase transformation.
The figure represents a realistic part of the microstructure. The gray particles refer to the
alumina matrix, yellow to tetragonal zirconia. The phase transformation of zirconia is
indicated by the change to red color. In the case of severe overloading crack initiation and
crack extension will occur. High tensile stresses in the vicinity of the crack tip trigger the
tetragonal to monoclinic phase transformation of the zirconia particles. The accompanied
volume expansion leads to the formation of compressive stresses which are efficient for
blocking the crack extension.
The model as represented in Figure 3 has also been verified experimentally. Pezzotti et.al.
[PezTBP] analyzed the monoclinic phase transformation in the vicinity of an artificial crack
tip as shown in Figure 4.
As it is demonstrated in Figure 3 this reinforcing mechanism is fully activated within a

region of a few micrometers. For the macroscopic performance of the material it is very
important that immediately at the beginning of crack initiation also the reinforcing
mechanisms are activated. Regarding Figure 3 one should keep in mind that the average
distance between the reinforcing zirconia particles is approx. 0,3µm, i.e. similar to the grain
size. Thus, the reinforcement is activated immediately when any microcrack is initiated.
The reinforcing ability of zirconia particles is a consequence of the phase transformation,
i.e. the spontaneous change from the tetragonal to the monoclinic phase [Han00]. The phase
transformation is accompanied by a volume change of 4 % of the zirconia particle.
Spontaneous phase transformation is a well known principle in material science. For example,
the properties of high performance steels also rely on phase transformation from austenite to
martensite.
Grey intensity
proportional to
monoclinic phase
content
Raman Spectroscopy, G. Pezzotti.
Figure 4. Monoclinic phase transformation in the vicinity of an artificial crack tip.
It should be emphasized that the ability of phase transformation is the precondition for
any benefit of the zirconia within the material. The composite is designed such that phase
transformation occurs when it is needed, i.e. to prevent microcrack initiation and propagation
at a high mechanical stress level.
Experiment: What Happens when Phase Transformation Is Suppressed?
It has been shown experimentally that the ability of zirconia phase transformation in
BIOLOX®delta is necessary for the excellent mechanical properties. The experiment has been
designed such that the experimental material was identically produced to BIOLOX®delta but
with a significant higher amount of Y2O3. Yttria is known for stabilizing the tetragonal phase
of zirconia. Consequently, in the case of a too high amount of yttria, the ability of phase
transformation is suppressed. This has been shown in the experiment.
The experimental material W3530 has been produced equivalently to the production of
BIOLOX®delta. In Figure 5 it is shown that the microstructure is identical. In Table 4 the
properties of the two materials are compared.
®
BIOLOX delta Experimental material W3530
Figure 5. Microstructure of regular BIOLOX®delta and experimental material W 3530
Table 4. Comparison of regular BIOLOX®delta and high stabilized

experimental material
Material properties BIOLOX®delta W3530

Ratio Y2O3 / ZrO2 [mol %] 1,3 3,0
Final density [g/cm3] 4,37 4,38
Grain size [µm] 0,54 0,55
Strength [MPa] 1392 777
Hardness [HV1] 1757 1747
Monoclinic phase content [%] 5 ≤1
Fracture toughness 6,5 5,1
The basic properties of regular BIOLOX®delta and the experimental material W3530 are
identical, i.e. microstructure, density, grain size and hardness. The fundamental difference is
the ratio of Y2O3/ZrO2. In BIOLOX®delta the amount of yttria is significantly lower. As can
be seen from the data, the higher amount of yttria in the experimental material leads - as
expected - to a lower content of monoclinic phase, because phase transformation is
suppressed. As a consequence the fracture toughness and the strength of the experimental
material are much lower than the properties of BIOLOX®delta. In particular, the strength of
the experimental material W3530 is only 55% of the normal strength of BIOLOX®delta.
From this result it is immediately clear that the ability of phase transformation is necessary to
obtain a high performance composite ceramic. The phase transformation can be easily
suppressed by chemical stabilization (using yttria). However, suppressing phase
transformation means loosing the excellent mechanical properties of the material.
Stabilization of the Zirconia Tetragonal Phase
In BIOLOX®delta the content of yttria has been optimized during the material
development. It should be noted that the Y2O3/ZrO2-ratio is lower than in normal pure
zirconia materials (3Y-TZP), because the stabilization of the tetragonal phase in
BIOLOX®delta is also influenced by “mechanical stabilization”, i.e. the embedding of
zirconia in the stiff alumina matrix.
cub
2000
temperature [°C]
1500
tet + cub
tet
1000
500
mon mon + cub
2 4 6 8
Y2O3 [mol %]
Figure 6. Background of stabilizing effects: Y-doping and embedding of zirconia particles in a matrix.
As can be seen in Figure 6 doping of Y2O3 into ZrO2 shifts the temperature of tetragonal
– to - monoclinic phase transformation towards lower temperatures. Thus, doping with Y
means “chemical stabilization”. At Y2O3 content lower 10% the stable phase at room
temperature is monoclinic. So additionally the embedding of the zirconia particles in a matrix
as well as surface stresses in the small particles also act as stabilizing mechanisms. The
surrounding material will oppose the transformation and it is the strain energy that is involved
in this constraint that allows the tetragonal phase to be retained at room temperature [Gre89].
This effect is referred to as “mechanical stabilization”
5. Material Production and Properties

BIOLOX®delta is a comparatively complex composite material where 4 different
ingredients are mixed during powder preparation and undergo solid phase transformation at
high temperature treatment as explained in chapter 2. The ability of reproducing such a
material in high quantities with excellent quality batch by batch is the key qualification of
CeramTec as manufacturer.
It is the purpose of this section both to summarize the important material properties and
to elucidate the reproducibility of the production.
The important production and analytical steps are as follows:
Figure 7. Schematic description of processing and material data generation of a single batch.
As shown in Figure 7 the material data are obtained for every powder batch. In Table 5
the material data as obtained for all powder batches in 2007 are summarized.
Table 5. Material properties of BIOLOX®delta batches in 2007
4-pt.
Weibull Hardness Monoclini Grain Size Fracture
Density Bending
Modulus HV 10 c content Alumina Toughness
Strength
g/cm3 MPa GPa % µm MPa m
Average 4,37 1411 14,9 17,2 4,4 0,54 6,4
Std. Dev. 0,007 50 3,1 0,09 1,1 0,027 0,20
The physical background of these parameters should be discussed. There are parameters
which are almost invariable due to the physical nature of the property, in particular density
and hardness. In the normal case of regular production there is only very little scatter with
these data. However, it is important to analyze these parameters for every batch because any
deviation of the expected results would indicate an insufficient production lot.
On the other hand, it is worth highlighting the low alumina grain size and the very low
scatter of this value. During sintering and final densification of any ceramic the particles build
a dense matrix but simultaneously grain growth also occurs. It is the goal of adequate
sintering to achieve full density but to suppress grain growth, as a fine microstructure is
necessary for good mechanical properties. Obviously, the sintering of BIOLOX®delta is very
well reproduced batch by batch. The grain size is low in comparison to pure alumina because
the dispersed zirconia particles prevent grain growth of the alumina matrix.
The fracture toughness is a measure for the reinforcing mechanisms in the material. As
described in chapter 3 the high fracture toughness depends on the transformation mechanism
of the zirconia particles. Obviously also this value shows very low scatter. The average
fracture toughness is 6,4 MPa√m. In contrast, the “overstabilized” material W3530 described
in chapter 3 has a significant lower fracture toughness K1c 5,1 MPa√m. Thus, it can be
derived from the evaluation of fracture toughness from batch to batch that the desired
transformation toughening is working properly.
As explained in chapter 3 the fracture toughness should be discussed in context with the
monoclinic phase content which is determined on a polished flat surface of a specimen. The
monoclinic phase content [in %] as obtained from the regular X-ray diffraction is relative to
the total zirconia content, not to the total volume of the material. Thus, in any case the
monoclinic phase content of the total material can be determined by simply referring to the
zirconia fraction of 17 vol.%. Example: 10% monoclinic phase content is equivalent to 1,7%
relative to the total volume of the material.
According to the materials specification the monoclinic phase content after polishing
(intrinsic monoclinic phase content) is ≤ 10% of ZrO2 which is regularly determined by X-ray
diffraction. The sensitivity of this technique is around 1% monoclinic phase content. As it is
evident from the data the average intrinsic monoclinic content of ZrO2 of 4,4% is above the
sensitivity limit. This indicates again that the sound material is in disposition of phase
transformation. Note that the monoclinic content of the experimental material W3530 is
below the sensitivity limit.
In contrast to the other parameters discussed above the strength of the material shows
significantly higher scatter. It is important to understand that failure in ceramics is always
triggered by imperfections of the microstructure. High performance ceramics can only be

achieved when the natural defects in the material are very small. Typical relevant
imperfections in BIOLOX®delta are within a range of 5 – 50µm. Accordingly, the scatter of
natural defect size directly matches the scatter of strength which is described by the
Weibull´s modulus. A high modulus indicates low scatter. For the high performance material
BIOLOX®delta a Weibull’s modulus of ≥ 7 is tolerated in the specification. As can be seen
from the data the normal scatter of the strength is much lower (i.e. higher modulus).
6. Correlation of Material and Component Properties

In chapter 5 the extensive efforts of analyzing the material properties of BIOLOX®delta
batch by batch have been discussed. These material properties are determined according to
ISO 6474 which is applicable for pure alumina and currently being extended for alumina-
composite materials such as BIOLOX®delta (ISO 6474-2). This type of data is very familiar
for evaluation of the performance of ceramics. In this chapter it is intended to discuss shortly
how these material data correlate to component properties, e.g. the strength of ball heads and
inserts.
In general, the performance of any system depends on the intrinsic material properties,
the design and manufacturing quality of the components and the system, the external load and
the particular environment, and finally the quality of mounting and installation. The use of
high performance materials inevitably promotes the performance of a system - however, the
other factors may be even more decisive for the success of a system. These complex
correlations must be necessarily evaluated by design analysis, modeling, simulations, risk
analysis and many other tools. In order to eliminate any influences of design features most of
the material testing has been performed using 4-point bending bars.
In Figure 9 the setup of the regular 4-point bending test as recommended in ISO 6474-2
and the burst test according to ISO 7206 are shown. The bending test reveals the intrinsic
strength of the material whereas the burst test is designed in order to simulate the in-vivo load
of ceramic ball heads.
4-Point Bending Test Burst Test
Figure 8. Schematic set-up of 4 point bending test and ball head burst test.
The strength measured with the bending test is – to a first approach - an intrinsic material
parameter1, whereas the burst load as obtained from the burst test depends on the materials
strength and the design of the ball head and the taper. This is directly shown by comparison of
strength and burst load of different ball heads made of BIOLOX®forte and BIOLOX®delta,
see Table 6.
Table 6. Comparison of strength and burst load of BIOLOX®forte and BIOLOX®delta
ratio
Parameter Test / Design Unit BIOLOX®forte BIOLOX®delta
delta / forte
4-point
Strength MPa 620 1400 2,3
bending
Burst load 28-12/14 L kN 54 85 1,6
Burst load 36-12/14 M kN 110 131 1,2
All burst loads are far above the required value of 46kN. The maximum in vivo load at
worst case conditions is approximately 10 kN.
The data of the burst tests given in Table 6 are obtained from ball heads with identical
geometry, Ti test taper and the same test setup. Thus, the advantage of BIOLOX®delta ball
heads in the burst load only comes from the higher strength of the material in comparison to
the pure alumina BIOLOX®forte. The strength of BIOLOX®delta is more than twice the
strength of BIOLOX®forte, whereas the ratio of the burst strength values is lower. This is
explained by the ductile deformation of the Ti taper during the burst test which steadily
increases the contact area of the conical bore of the ceramic ball head and the metal taper.
However, the burst load of identical ball heads is always higher when a high strength material
is used.
It is also seen that the burst strength strongly depends on the ball head size. A larger ball
head shows a higher burst load. It is concluded that the benefit of using a high performance
material is higher when applied to a challenging design, e.g. a ball head with lower wall
thickness. Nevertheless, the use of the high strength material always increases the safety
margin of the component.
7. Wear Performance of BIOLOX®Delta

At normal wear conditions (e.g. standard wear simulator) the wear of a hard-hard couple
of BIOLOX®delta is identical to the excellent performance of the well proven pure alumina
BIOLOX®forte. There is only a minor difference in hardness of these both materials which
does not compromise the normal wear behavior.
However, a significant advantage of BIOLOX®delta is identified in the case of worst
case wear simulation as shown in Figure 9. In this experiment microseparation of ceramic
ball head and insert during each load cycle has been simulated which leads to highly localized
1
Due to the statistical nature the strength also slightly depends on the specimen size and the stress distribution. Size
effects can be mathematically balanced. In order to obtain results which can be directly compared to each
other the standardized set-up of the bending test should be used.
forces at the contact area. This experimental setup was supposed to simulate e.g. low tension
of the soft tissue after surgery as it is discussed frequently by orthopedic experts. It was
concluded from various retrievals that in some cases a well defined stripe-shaped area shows
a more intense worn surface than the normal surface of the ball head. This phenomenon is
known as “stripe wear”.
In the experiment, heavy wear conditions were simulated due to the highly localized
contact area. It was found after 5 mio microseparation load cycles that the wear volume of
BIOLOX®delta couples (both ceramic ball head and insert were made of BIOLOX®delta)
was 7 times lower than that with the coupling made of pure alumina BIOLOX®forte [Cla05].
The wear rate of the mixed couplings (either ball head or insert made of BIOLOX®delta, the
remaining made of pure alumina BIOLOX®forte) ranged between those values. It is important
to mention that due to the small differential hardness between BIOLOX®forte and
BIOLOX®delta can be combined in a ceramic-ceramic coupling without running the risk of
excessive wear or other adverse effects.
®
Figure 9. Wear performance of BIOLOX delta and BIOLOX®forte at simulated micro separation.
Obviously BIOLOX®delta shows an excellent “stripe wear tolerance”. The analysis as

given in [Cla05] shows that in the stripe wear region the monoclinic phase content is strongly
increased. This indicates the mechanism behind the excellent stripe wear tolerance of
BIOLOX®delta. As a first approach it is assumed that at these special test conditions a very
high localized stress acts in the contact area which may be able to introduce damage in the
surface. In this case the reinforcing mechanism as described in Figure 3 is activated in
BIOLOX®delta which supports maintaining the surface quality under these extreme
conditions. This assumption is supported by the Raman analysis of the worn surface. In the
stripe wear region a high monoclinic phase content was found which indicates that phase
transformation for reinforcement took place.
8. Discussion of Hydrothermal Aging

Mechanism of Hydrothermal Aging
Aging is a relevant issue for all zirconia containing ceramics. The transformation from
the tetragonal to the monoclinic phase can be triggered in a hydrothermal environment.
“Hydrothermal” means that this particular aging effect only takes place in aqueous
environment at elevated temperatures. A critical temperature range for hydrothermal aging is
around 120 – 200°C. However, a very slow aging effect is potentially possible even at human
body environment. The kinetics and the threshold of aging activation strongly depend on the
grain size, volume fraction, amount and type of stabilizing elements (e.g. Y, Ce, Mg, Ca) and
mechanical stabilization of the zirconia [Che99].
1st step adsorption 2nd step

H2O
surface
Zr O Zr Zr OH OH Zr
stress
O O O O
vacancy
Y Zr Y Zr
3rd step 4th step

migration
Zr OH OH Zr Zr OH OH Zr
O migration O O O
lattice
Y OH Zr expansion Y OH Zr
Figure 10. Model of hydrothermal aging.
Albeit the hydrothermal aging is extensively studied in the literature, the exact
mechanism behind this effect is still not perfectly described. Most of the available models
refer to interaction of the hydroxide ions and oxygen vacancies in the zirconia lattice. The
oxygen vacancies inevitably are introduced in the lattice at solution of Y2O3 in ZrO2 due to
different valence of the cations. Figure 10 represents a model of the interaction of hydroxide
ions with Y-doped zirconia with 4 steps:
1. Adsorption of water molecules at the surface

2. Dissolution of the oxygen bond
3. Migration of hydroxide ions to oxygen vacancy

4. Reorganisation of the atomic lattice
The critical step within this model is the dissolution of the oxygen bond which is very
strong in zirconia. Thus, a thermal activation at elevated temperatures is required.
Hydrothermal Aging in BIOLOX®delta
Phase transformation caused by hydrothermal aging is an undesired effect. However, as it

has been shown in the previous chapters, monoclinic phase transformation is necessary for the
reinforcement and the high strength and toughness of the material BIOLOX®delta. Thus, a
certain amount of phase transformation in hydrothermal environment is not a matter of
concern. It depends on the specific composition of the material if there is a critical level of
phase transformation where the material can be damaged. In the years 2001 – 2002 some
batches of the pure zirconia material Prozyr® of the company Desmarquest showed
catastrophic failure in-vivo due to monoclinic phase transformation at the surface [Che06]. It
has to be emphasized that such a damage is impossible in BIOLOX®delta due to the fact that
only 17vol% consist of zirconia. Consequently, the ultimate upper limit of monoclinic
zirconia is only a total of 17% whereas 83% of the material remain not affected by phase
transformation and hydrothermal aging.
As mentioned above, hydrothermal aging is accelerated at elevated temperatures. For
pure zirconia in biomedical applications a standardized accelerated aging test is
recommended in ISO 13 356. The aging conditions are autoclaving at 134°C and 2 bar water
steam for 5h. In the new standard ISO 6474-2 being prepared for alumina zirconia composites
an aging time of 10h is recommended. As a rough estimate it is proposed that 1h autoclaving
is equivalent to 2 – 4 years in human body environment. However, this transfer of aging
kinetics depends on individual properties of a material.
The aging behavior of BIOLOX®delta has been extensively analyzed using accelerated
hydrothermal aging by autoclaving. Moreover, the aging has been combined with severe
mechanical static and dynamic testing in order to understand if any material degradation
occurs during aging.
In fact, the monoclinic phase content in BIOLOX®delta is increased after long term
aging. In Figure 11 quantitative analysis of monoclinic phase transformation using Raman
spectroscopy are shown. The aging conditions (121°C, 1 bar) for this analysis were slightly
different from the recommendation in the ISO standard. Equal to Figure 4 the gray intensity
indicates monoclinic phase distribution. As can be seen the monoclinic content is increased
after extreme long exposure time. After 300 h the monoclinic phase content reaches a relative
value of 66,7%, i.e. 11,3% of the total volume of the composite are monoclinic zirconia.
Figure 11. Increase of monoclinic phase content after very long term hydrothermal aging in an
autoclave. Autoclaving conditions 121°C, 1 bar. Pictures adopted from [Pez08]. Quantitative
monoclinic phase content obtained by direct communication to the author.
The autoclaving time used in this study is extremely long, much longer than
recommended in the ISO standard. The monoclinic phase content seems to reach a saturation
since the increase from 200h to 300 h is only marginal.
Aging Kinetics of BIOLOX®delta
As described above the aging effect of zirconia containing materials depends on the
individual composition. It is of particular interest to estimate the aging effect of a bioceramic
under in-vivo-conditions, i.e. 37°C in serum. As it was the case in the Prozyr® disaster a high
amount of phase transformation occurred after a relatively short time in-vivo. On the other
hand, it is well known that even pure zirconia is usually much more stable against
hydrothermal aging when produced appropriately.
For a given material, the long term aging (in years) at low temperatures can be predicted
based on the assumption that nucleation and growth of phase transformation are thermally
activated. The thermal activation can be assessed by a systematic variation of temperature and
duration time in hydrothermal environment and measurement of monoclinic phase
transformation before and after the test. The most common method for such a prediction is
based on the model of Mehl-Avrami-Johnson (MAJ-theory), which was first applied to
zirconia based materials by Prof. Chevalier (Lyon, France) [Che99]. For example, it was
found that for a conventional zirconia material with 3 mol% Yttria content (3Y-TZP) one
hour in autoclaving conditions (134°C, 2 bar water steam) is equivalent to 3-4 years in-vivo.
For other zirconia materials, a correlation of only 2 years was found. It is important to
understand that this relation is not a universal law for zirconia containing materials but
instead the individual outcome of the particular ceramic composition.
The aging kinetics of BIOLOX®delta was analyzed in close cooperation with Prof.
Chevalier. The complete study will be accomplished and published in 2009. In the following,
the preliminary results which are available at the current state are summarized.
The accelerated aging tests were performed in water steam at 142°C, 134°C, 121°C, and
105°C. Additionally one test was performed at 90°C in water. It is assumed that steam and
liquid hydrothermal environment are comparable by means of the effect on hydrothermal
aging. More tests have also been launched at 70°C and 50°C in water. However, due to the
long duration time which is necessary to detect any effect at these temperatures results are
expected within the next 2 years. The monoclinic phase content was determined using X-ray
diffraction. In Figure 12 the results of the tests are shown on a logarithmic time scale.
80
90°C
70 105°C
121°C
monoclinic fraction (%)
60
134°C
50 142°C
Pezzotti
40
30
20
10
0
0,1 1 10 100 1000 10000
time (h)
Figure 12.:Monoclinic phase transformation of BIOLOX®delta in hydrothermal environment. For

comparison, also data from [Pez08] is included.
For evaluation of these data, the following modified MAJ equation was applied
[ (
Vm = Vm0 + (Vmax − Vm0 ) 1 − exp − (b t )
n
)] (7)
0
where Vm is the monoclinic phase content, Vm is the initial monoclinic phase content prior to
the test, Vmax is an apparent upper bound of the monoclinic phase, t is the time and n the time
exponent. b is the factor which describes the temperature dependence of the aging effect. It is
derived from fitting the data according to the following Arrhenius type equation:
⎛ Q ⎞
b = b0 exp⎜⎜ − ⎟⎟ (8)
⎝ RT ⎠
where b0 is a material constant, Q is the activation energy, R the universal gas constant and T
the temperature.
Evaluation of the data given in Figure 12 reveals an activation energy Q of 108 kJ/mol.
This is a particular high value in comparison to a similar material (alumina zirconia
composite) which was discussed in [Pez08], where an activation energy of 78kJ/mol is
referred. The reason for this discrepancy is, as discussed above, the low amount of oxygen
vacancies in BIOLOX®delta due to the unique chemical composition. High activation energy
means strong influence of temperature on aging. Consequently at low temperatures the aging
rate of in BIOLOX®delta is significantly lower.
From the modified MAJ equations the hydrothermal aging effect under in-vivo
conditions can be estimated. The following equivalence to the accelerated hydrothermal aging
was found:
1 h at 134°C is equivalent to 3,9 years in-vivo
According to the proposal of the new ISO standard 6474-2 hydrothermal aging is
simulated under autoclaving conditions at 134°C for 10 hours. These conditions are thus
equivalent to 39 years in-vivo for BIOLOX®delta which seems to be a realistic upper bound
of the expected live time of an artificial joint.
Several accelerated aging tests according to these conditions (134°C, 2 bar water steam,
10 h) have been performed with BIOLOX®delta. The initial monoclinic phase content prior to
aging depends on the surface finishing. On a ground surface the initial monoclinic phase
content is higher than on a polished surface due to the more intensive interaction of the
ceramic surface to the diamond tool. There is no effect on strength of the higher initial
monoclinic phase content - the test specimens for the regular evaluation of strength are
ground, not polished. Table 7 gives the data of monoclinic phase increase after 10 h
autoclaving
Table 7. Analysis of zirconia phase transformation at accelerated aging
Initial monoclinic phase Monoclinic phase content

Surface finishing
content after 10 h autoclaving
5% 8%
21 % 33 %
There is obviously a large difference of initial monoclinic phase content between

polishing and grinding. However, the relative increase of monoclinic phase transformation
does not depend on the surface condition. As discussed in chapter 5 the monoclinic phase
content is given relative to the total amount of zirconia. When the effect of phase
transformation shall be assessed, it is useful to refer the monoclinic content to the total
volume of the material, i.e. taking into account the zirconia volume fraction of 17 vol %.
Thus, after grinding and autoclaving the absolute monoclinic phase volume content is 5,6%
(33% × 0,17).
The accelerated aging according to Figure 11 was performed at 121°C 1 bar, i.e. a lower
temperature and steam pressure than recommended in the ISO standards. It is also possible to
evaluate the equivalence of this treatment to in-vivo conditions.
1 h at 121°C is equivalent to 1,3 years in-vivo.
It should be noted that there is a strong discrepancy of this equivalence factor to the data
which was used and discussed in [Pez08], because in this publication the activation energy of
a different material was used. It is evident that the aging kinetics must be carefully analyzed
and discussed for the individual material.
Effect of Hydrothermal Aging on Strength of BIOLOX®delta
It is necessary to examine if the phase transformation reveals any damage or loss in

strength of the material. For this purpose experiments were designed in order to combine
accelerated aging and mechanical loading at a very high stress level. The specimens were
prepared according to the 4-point bending configuration as it is shown in Figure 9 (left).
These bend bars were exposed to accelerated aging and then to loading-unloading cycles at
high stresses. After aging and cyclic loading the residual strength was determined.
Two stress levels (300 MPa and 600 MPa) were chosen for the cyclic loading tests. The
lower stress level was applied for 20 Mio cycles, the higher stress level for 5 Mio cycles. All
tests were performed in Ringer’s solution. The accelerated aging was simulated by 5 h and
100 h treatment in standard autoclaving conditions (134°C, 2 bar water steam). It should be
noted that again an aging time much longer than required from the standard was used. For
each step of the experiments (as received, after aging, after aging and cycling) the monoclinic
phase content was measured using X-ray diffraction.
As the most amazing result the yield of specimens surviving all the tests was 100 % in all
cases. Even most severe conditions (i.e. 100 h autoclaving, 600 MPa cyclic load) did not
reveal any premature failure. It should be recalled that this stress level represents 4 times the
highest load level at worst case conditions in-vivo. We can thus conclude that the reliability
of BIOLOX®delta exceeds by far the necessary requirements for reliable surgical
components.
Table 8. Residual strength and monoclinic phase content after diverse treatments.
Monoclinic phase is given relative to the total volume of the material
Autoclaving no cyclic 300MPa, 600MPa,

duration load 20*106 cycles 5*106 cycles
Strength [MPa] 1346 1433 1284
0h
Monoclinic phase content [%] 18 33 43
Strength [MPa] 1332 1248 1361
5h
Strength [MPa] 1234 1308 1300
100 h
Table 8 shows the results of the post – test analysis including residual strength and
monoclinic phase content. There is –as expected - a marginal natural scatter in residual
strength. However, statistical analysis using Student’s t-test did not reveal any significant
deviation of all strength results.
50
Monoclinic Content [%]

40
30
100 h ageing
20
5 h ageing
10 No ageing
0
no 300 MPa 600 MPa
cycling 20 Mio cycles 5 Mio cycles
Figure 13. Increase of monoclinic phase content at cyclic loading.
In contrast, there is a clear tendency of an increase in monoclinic phase content both,

after autoclaving and after cyclic loading, which is illustrated in Figure 13. For example, the
test series without autoclaving shows an increase of monoclinic phase content from 18 % in
the initial state to 47 % after 5 Mio cycles at 600 MPa. It must be concluded that the cyclic
mechanical loading at a high stress level (600 MPa) which represents already the strength of
pure alumina activated the reinforcing ability of the material. As discussed under Figure 3, a
high mechanical stress triggers localized phase transformation which prevents any further
crack propagation. Obviously the increased amount of monoclinic phase content does not
deteriorate the strength of the material. This important conclusion is independent from the
source of the phase transformation. In other words, when the phase transformation is
activated either by accelerated aging, cyclic fatigue or a combination of both, the residual
strength remains at the initial level.
References
[Che99] Chevalier J., Cales B., and Drouin J.M., Low temperature aging of Y-TZP
ceramics, J.Am.Cer.Soc. 82 (1999) 2150-54
[Che06] Chevalier J., What future for zirconia as a biomaterial, Biomaterials 27 (2006)
535-543
[Cla05] Clarke I.C., Pezzotti G., Green D.D., Shirasu H., and Donaldson T., Severe
Simulation Test for run-in ear of all-alumina compared to alumina composite THR,
Proceedings 10th BIOLOX Symposium, 11-20 (2005)
[DeA02] De Aza, A.H., Chevalier J., Fantozzi G., Schehl M., and Torrecillas R., Crack
growth resistance of alumina, zirconia and zirconia toughened alumina ceramics
for joint prostheses, Biomaterials 23, 937-945 (2002)
[Gre89] Green D.J., Hannink R.H.J., and Swain M.V., Tranformation Toughening of
Ceramics, CRC Press Inc., Boca Raton, Florida (1989) ISBN 0-8493-6594-5
[Gre04] Gremillard L., Chevalier J., Epicier T., Deville S., and Fantozzi G., Modeling the
ageing kinetics of zirconia ceramics, J.Eur.Cer.Soc., 24, 3483-3489 (2004)
[Han00] Hannink R.H.J., Kelly P.M., and Muddle B.C., Transformation Toughening in
Zirconia-Containing Ceramics, J.Am.Cer.Soc. 83 [3] 461-87 (2000)
[Ohm99] Ohmichi N., Kamioka K., Ueda K., Matsui K. and Ohgai M., Phase
Transformation of Zirconia Ceramics by Annealing in Hot Water, J. Cer. Soc.
Jap., 107 [2] 128-133 (1999)
[Pez06] Pezzotti G. Environmental Phase Stability of Next Generation Ceramic composite
for Hip Prostheses, Key Engineering Materials Vols. 309-311, 1223-1226 (2006)
[Pez08] Pezzotti G. Yamada K., Sakakura S., and Pitto R.P., Raman Spectroscopic
Analysis of Advanced Ceramic Composite for Hip Prosthesis, J.Am.Ceram.Soc. 91
[4] 1199-1206 (2008)
[PezTBP] Pezzotti G. Yamada K., Porporati A., Kuntz M., and Yamamoto K., Raman
Spectroscopic Analysis of Advanced Ceramic Composite for Hip Prosthesis: Part
II Fracture Behavior (to be published)
Chapter 5
PARTICLE MODELING AND ITS CURRENT SUCCESS

IN THE SIMULATIONS OF DYNAMICS
FRAGMENTATION OF SOLIDS
G. Wang1*, A. Al-Ostaz1, A.H.D. Cheng1 and P. Radziszewski2

1
Department of Civil Engineering, University of Mississippi,
Oxford, MS, 38677-1848, USA
2
Mechanical Engineering Department, McGill University,
Montreal, QC, H3A 2T5, Canada
Abstract
Particle modeling (PM) is an innovative particulate dynamics based modeling approach. It has
been demonstrated as a robust tool for simulating fracture problems of solids with dynamic
fragmentation under extreme loading conditions. These loading conditions can include
situations of collapse, impact, blasting or high strain rate tension/compression, as well as
thermally-induced breakage problems.
Initially, PM was developed for the purpose of mimicking the microscopic material
process at macroscopic level. This method can be conceptually illustrated by fully dynamic
particles (or “quasi-particles”) placed at the nodes of a lattice network without explicitly
considering their geometric size. The potential can be specified for particle-particle
interactions via axial springs. Theoretically, PM is an upscale of the molecular dynamics
(MD) model applicable to various length scale problems. This is possible if a proper
equivalent macroscopic potential is found, and, in case of lattice spacing decreasing to a few
Angstroms, a MD model at zero Kelvin with, say, Leonard-Jones potential is recovered. In its
current form, PM has been developed as a tool applicable to real engineering problems.
The advantages of PM over the existing discrete element based methods can be
summarized as follows:
(1) Sample in theory. Four conservative/equivalent rules (mass, potential energy,
Young’s modulus and tensile/compression strength) are applied to preserve the equivalent
material properties.
(2) Easy for implementation. Since the physical size of each particle is ignored other than
its equivalent mass, the algorithm of coding a PM computation is fairly easy.
*
E-mail address: [email protected]. Tel: + 662 915 5369, Fax: + 662 915 5523. Corresponding author: Ge Wang
158 G. Wang, A. Al-Ostaz, A.H.D. Cheng et al.
Current research work has exhibited that PM is able to correctly predict dynamic
fragmentation of materials with a good accuracy. In modeling an epoxy plate with randomly
distributed holes in tension, the PM result of the final crack pattern compared favorably with
the associate experiment; for the simulations of impact study of two polymeric materials
(nylon, 6-6 and vinyl ester) subject to a rigid falling indenter, the modeling results of resistant
force, energy, deflection and drop speed of indenter vs. time quantitatively agree fairly well
with the according empirical observations.
1. Introduction
The prompt development of modern computer technology and computational methods
enables scientists, more easily than ever before, to numerically examine the nature of
materials, to verify or modify the existing theories, and even to discover new phenomena.
Dynamic fracture modeling also greatly benefits from this advancement.
The dynamic fracture process itself is an exceedingly complex, multi-scale physical
phenomenon. Material failure exhibits non-linearly, localized presence and a dependence on
dynamic loading and loading conditions, etc. Hence, an ideal solver of dynamic fracture
mechanics is required to be able to handle all the above-mentioned factors, especially to deal
with the discontinuity of material that dominantly occurs in a dynamic fracture process which
becomes an overwhelming difficulty to numerical approaches.
At the microscopic level, fracturing is a process that material becomes separated due to
the successive failures of atomic bonds. Since the intrinsic strength properties at atomic
structure level are available, molecular dynamics (MD) analysis has been used in modern
scientific research at nano-scale. However, although MD simulation has benefited from the
rapid development of modern computer technology and is becoming increasingly popular, the
present state of computational power is still far from being able to support simulation at the
macroscopic level. For example, we currently still cannot simulate a 1 × 1 × 1 cm3 cubic
copper body at atomic level because the body consists of 10 24 copper atoms, a number so
large that no computer in the world can handle it. The second difficulty is its inability to reach
the laboratory time scales. For instance, the laboratory fracture experiments generally last in
−6 −9
microseconds ( 10 second), while the MD model time steps are typically in the nano ( 10 )
−12
or pico ( 10 ) second range. As such, MD is limited narrowly to solving nano- to
micrometer scale problems. Hence modeling fracture problems directly at macroscopic scale
is a prevalent pursuit.
An all-round search to the existing optional modeling techniques reveals that the modern
numerical methods for dynamic fracture simulations at macroscopic level can be generally
classified into two categories of approaches. One is continuum mechanics based and the other
is discrete element based. Examining the state-of-the-art of the research, we conclude that the
continuum mechanics based approaches, such as the finite element method (FEM), have
difficulty in solving dynamic fracture problems, particularly in dealing with the simulation of
the collapse of structure and its fragmentation under extreme loading conditions. For this
reason, another domain of alternative modeling approach —the discrete element based
models, has become more and more popular.
Discrete element models share a common concept of “discrete material”, which can still
be further classified into two sub-categories. The first one is classified as a discrete element
Particle Modeling and Its Curent Success in the Simulations... 159
method, in which the physical size and shape of each element is considered in the
computation, such as the applied element method (AEM) [Meguro & Tagel-Din, 2000], and
particle finite element method (PFEM) [Oñate, et al., 2004] - another particle-related model.
In practice, this type of models is complex and difficult to be implemented. For instance,
AEM is laborious in keeping track of the instantaneous contact positions and the evolving
geometry of all elements for a proper updating of the dynamic bonds; PFEM is expensive in
remeshing and redefining boundaries at each time step.
The second branch is named particle (or particulate) dynamics method, in which the
physical size and shape of each discrete cell is not explicitly considered in the computation.
Consequently, each cell is treated as a particle, with the equivalent mass lumped at its center.
Smoothed particle hydrodynamics (SPH) [Monaghan, 2005], and particle modeling (PM) are
currently the mainly two popular particulate based models. SPH adopts a kernel probability
density function to define for each particle a reaction domain at each time step, in terms of a
particle number density; it is therefore highly expensive in computation. In contrast, PM is
rather simple in theory and easy for implementation; hence, it is attracting increasing interest,
despite a good number of SPH codes.
Particle modeling (PM) method was originally proposed by Greenspan [1981, 1997] for
the purpose of mimicing the microscopic material process at macroscopic level. PM is fully
dynamic with particles (or “quasi-particles”) placed at the nodes of a periodic (equilateral
triangular) lattice. Conventionally, only the nearest neighboring particles are accounted for in
the interaction via axial spring connections. Similar to MD, potential can be specified for
particle-particle interactions, and, in the case of lattice spacing decreasing to a few
Angstroms, we recover a molecular dynamics (MD) model at zero Kelvin with, say, a
Leonard-Jones potential. This implies that PM is an upscaled MD and can be applied to
various length scale problems if a proper equivalent macroscopic potential is used. Obviously,
by making use of the similar processes as atomic interactions, it is possible to solve fracture
problems at macro-scale level. However, as no direct link with the material properties was
attempted by Greenspan, Greenspan’s PM method remains a conceptual model without a
demonstrated success in fracture modeling.
Wang and Ostoja-Starzewski [2005a] developed a newly modified PM with a
conventional linkage pattern as shown in Figure 1. In this new PM, equivalent material
properties are formulated using four physical conditions to determine continuum-level
Young’s modulus and tensile strength, while maintaining the conservation of mass and
energy of the particle system, and satisfying the interaction laws among all the particles. The
theoretical foundations of the current PM can be outlined in the following four aspects:
(a) Derivation is based on equivalence/equality of mass, energy, elastic moduli and

tensile strengths between both atomic and quasi-particle systems.
(b) Derivation is carried out in the setting of a 3-D face-centered cubic (f.c.c) lattice
network (Figure 2) for both atomic and quasi-particle structures.
(c) Either Lennard-Jones, polynomial or quadratic potentials can be employed for
nonlinear or linear axial linkage.
(d) In principle, it can be applied for various length scale problems of solids.
Wang and Ostoja-Starzewski, 2006.
Figure 1. Lattice structure in PM.
Wang and Ostoja-Starzewski, 2005a.
Figure 2. Meshing system for a 3D material body.
In this paper, current PM modeling development is to be addressed. In the coming

sections, first we will introduce the theory of PM; then report some significant success and
also defects experienced through the applications. Meanwhile, the potential solutions to the
modeling deficiencies are proposed; and finally, in the conclusion part, an outlook of future-
going PM development is addressed.
2. Methodology of PM
In particle modeling (PM), the nonlinear interaction force is considered between nearest-
neighbor (quasi-) particles and assumed to be of the same form as in MD:
G H
F =− +
r p rq (1)
Here G, H, p and q are positive constants, and q > p ≥ 1 to obtain the repulsive effect
that is necessarily (much) stronger than the attractive one, r being the distance between two
particles.
Ashby & Jones [1980] presented a simple method to evaluate continuum-type Young’s
modulus E and tensile stress σ (r ) of the material from F (r ) , namely
S0
E=
r0 (2)
and
σ ( r ) = NF (r ) (3)
where S 0 = ( dF / dr ) r = r0 , and r0 is the equilibrium spacing between contiguous particles. N

is the number of bonds/unit area, equal to 1 / r02 . Tensile strength, σ TS , results when
dF (rd ) / dr = 0 , that yields,
σ TS = NF (rd ) (4)
Just as in MD, the non-linear dynamical equation of motion for each particle Pi of the
PM system is given by
G
d 2 ri ⎡⎛ G H ⎞ rGji ⎤
mi 2 = ∑ ⎢⎜ − pi + qi ⎟ ⎥ i ≠ j
⎜ rij ⎟⎠ rij ⎦⎥
dt ⎣⎢⎝ rij , (5)
G
where mi and r ji are mass of Pi and the vector from Pj to Pi . Note that if an equilateral
triangular lattice structure is adopted in 2-D, the resulting Poisson ratio equals 1/4 (or 1/3)
when a 3-D (respectively, plane) elasticity formulation is adopted [Ostoja-Starzewski, 2002,
2007].
The derivation of four parameters in Equation (1) from MD structures is conducted on a
cubic body with volume V (= A × B × C ), in Figure 2. A face-centered cubic (f.c.c) lattice for
both atomic and quasi-particle structures is chosen. If p, q and r0 are given, then, by
conditions of mass and energy conservation, G and H can be derived. Consequently, Young’s
modulus is evaluated by Equation (2) and tensile strength by Equation (4). To represent an
expected material property, we would have to do many sets of testing until a unique ( p, q ) is
found to match both Young’s modulus and tensile strength of the material. The complete
derivation process is described below.
First, for the atomic structure (MD model), we have:
⎛ Ga r 1− pa H a r 1−qa ⎞
interaction potential energy (ergs): φ a = ⎜⎜ + ⎟ × 10 −8
⎟ (6)
⎝ 1 − pa 1 − qa ⎠
Young’s modulus (GPa) is obtained from Equation (2) and tensile strength (MPa) from
Equation (4).
Total number of atoms in A × B × C cubic material body:
⎛ A × 10 8 ⎞ ⎛ B × 10 8 ⎞ ⎛ C × 10 8 ⎞
N * = ⎜⎜ + 1⎟⎟ × ⎜⎜ + 1⎟ × ⎜ + 1⎟
sin 60 0 ⎟ ⎜ r 6 /3 ⎟
⎝ r a ⎠ ⎝ a r ⎠ ⎝ a ⎠ (7)
In Eqs. (6) and (7), ra is equilibrium position of the simulated material in atomic
structure, and p a , q a are the exponential parameters in atomic structure. Note that, for a
Lennard-Jones interaction case, pa = 7 and qa = 13 .
Next, for the quasi-particle structure (PM model), we have interaction force (dynes) as in
Equation (1).
Interaction potential energy (ergs):
G r 1− p H r 1− q H r 1− q
φ = + , for p > 1; φ = G ln r + , for p =1 (8)
1− p 1− q 1− q
total number of quasi-particles in PM system: N = imax × j max × k max (9)
We now postulate the equivalence of MD and PM models. From the mass conservation,
we calculate the mass of each quasi-particle m based on atomic mass m a :
m = N * × ma / N (10)
From the energy conservation, we have:
(N × φ )r =r0
(
= N * × φa )
r = ra
(11)
under the requirement: F (r0 ) = 0 (12)
From equations (11), (12), we now derive Young’s modulus E:

for p = 1 :
1− q
(N *
× φa )
r = ra
(1 − q )
G = Hr , H = 1− q 1− q
, E = −Gr0−3 + qHr0− q − 2 (13)
N (1 − q )r ln r0 − r
o
0 0
for p > 1 :
(N ×φ )
*
a r =ra (1− p)(1− q)
G = Hr 1− q
, H= r0q−1 , E = − pGr0− p − 2 + qHr0− q − 2 (14)
N( p − q)
o
Similarly, tensile strength can be obtained under dF ( rd ) / dr = 0 . Evidently, the four

parameters ( p, q ) , r0 and V affect E and σ TS .
We have established the equations for G, H, p and q, and carried out a parametric study to
find the differing effects on p, q, V and r0 [Wang and Ostoja-Starzewski, 2005a]. Herein,
we summarize the obtained rules as follows:
(i) The larger the values of ( p, q ) are adopted, the larger is E generated. This is
typically associated with the material becoming more brittle than ductile, albeit there
is a range of toughness to choose from. Also, with E going up, there is a
fragmentation into a larger number of pieces.
(ii) In the case of p = 1 , the larger r0 spacing is adopted, the higher is Young’s modulus
of the PM material. On the contrary, in the special case of p ≠ 1 , there is an opposite
trend. In any case, this increase or decrease does not change very much.
(iii) In the case of p ≠ 1 , while keeping the volume fixed, an increase of r0 produces a
decrease of Young’s modulus. The situation is again opposite in the case of p = 1.
(iv) A uniform augmentation of volume V by dilation in all three coordinate directions
(xyz), at any ( p, q ) combination, results in Young’s modulus increasing first
strongly and then leveling off.
For brittle materials, a general format of linear dynamical equation is often employed
[Wang, et al., 2008a],
− S (r − r0 ) for rc ≤ r ≤ rt
F ={ 0 (15)
0 otherwise
with r being the distance between two particles, the stiffness S 0 = E • r0 by Equation (2), E
the Young modulus and rmax the failure displacement of material, r0 the equilibrium
spacing between the contiguous particles.
In Equation (15), rc and rt are the fracture positions applied for compression and
tension, respectively, which need to be empirically determined.
The leapfrog method, with second-order accuracy, is employed in all PM simulations.
The leapfrog formulas relating position, velocity and acceleration for particles Pi
( i = 1,2,..., N ) [Greenspan, 1997] are
G G (Δt ) G
Vi ,1 / 2 = Vi , 0 + ai , 0 (starter formula) (16)
2
G G G
Vi ,k +1 / 2 = Vi ,k −1 / 2 + (Δt )ai ,k , k = 1,2,3,... (17)
G G G
ri ,k +1 = ri ,k + (Δt )Vi ,k +1 / 2 k = 0,1,2,... (18)
G G G
where Vi ,k , ai ,k and ri ,k are the velocity, acceleration and position vectors of particle i at
G
time t k = kΔt , Δt is the time step. Vi ,k +1 / 2 stands for the velocity of particle i at time
t k = (k + 1 / 2)Δt , and so on. Notably, the leapfrog method is of second-order
2
accuracy: O ((Δt ) ) .
The safe time step is after the derivation result by Hockney & Eastwood [1999]:
1 dF
ΩΔt << 2 , Ω=( )1 / 2 (19)
m dr max
To readily describe the breakage effect on material, we define a concept of fracture

density [Wang, et al., 2008b]. By this definition, the local fracture density of particle i , f iden . ,
is equal to the ratio of its current number of broken bonds, N bi to its original number of
bonds, N oi , i.e.,
N bi
fiden . = (20)
N oi
It is clearly seen that a big f iden . value indicates a severe failure locally occurring at i.
3. PM Applications: Success and Deficiencies

3.1. Success of PM Applications
Since the establishment of the newly modified PM [Wang and Ostoja-Starzewski, 2005a]
up to now, particle modeling (PM) technique has been constantly improved and has found
good success in a number of applications including impact induced dynamic crack
propagation and fragmentation [Wang et al., 2008a-f; 2007a-d; 2006; 2005a-d; Ostoja-
Starzewski and Wang, 2006, 2005, 2004].
3.1.1. Validation Work
The first successful application of PM has been achieved for simulation of dynamic
fragmentation in an (elastic-brittle) epoxy plate (8.25 cm x 33.02 cm), containing non-
uniformly distributed circular holes in tension [Ostoja-Starzewski & Wang, 2006]. As
demonstrated in Figure 3(a, b), PM forecasting of the crack pattern agrees well with the
associated empirical observation [Al-Ostaz & Jasiuk, 1997].
Ostoja-Starzewski and Wang, 2006.
Figure 3. Experimental and modeling results of epoxy in tension, (a) experiment [Al-Ostaz and Jasiuk,
1997], (b) PM simulation.
Other two validations of the PM are conducted by simulating the impact of a rigid
indenter on polymeric materials (nylon-6, 6 and vinyl ester), the associated experiments
conducted at the University of Mississippi. Figure 4(a, b) shows the comparison of the
experimental result of fracture pattern of nylon-6,6 due to the impact of a rigid indenter while
Figure 5(a, b) shows the comparison of the similar impact study of vinyl ester.
From Figure 4(a, b), it found that (1) PM modeling crack pattern agrees with the
associated observation, and (2) it is seen that measured load peak happens around t ≈ 1.70
ms, and the measured deflection at the load peak is 3.03 mm, with the total impact energy
equal to 1.2 J. The corresponding PM simulated result shows that the load peak happens
around t ≈ 1.66 ms, and the deflection at load peak is 3.0 mm, and the total impact energy
calculated is 1.2 J. Although the simulated load profile is not exactly the same as the
experiment, we observe similar characteristics, including the fluctuating profile with roughly
the same period. The simulated peak load is also reasonable close to the experimental value.
Hence we conclude that the PM simulation compares favorably with the experimental
measurements.
Wang, et al., 2008a.
Figure 4. The study of the failure of nylon-6, 6 due to the impact of a rigid indenter (a) experimental
results; (b) PM results. Maximum drop velocity of indenter is 1.87 m/s.
From Figure 5(a, b), it is seen that measured load peak happens around t ≈ 1.38 ms, and
the measured deflection at the load peak is 2.60 mm, with the total impact energy equal to 0.7
J. The corresponding PM simulated result shows that the load peak happens around t ≈ 1.42
ms, and the deflection at load peak is 2.66 mm, and the total impact energy calculated is 0.76
J. Similarly to Figure 4(a, b), although the simulated load profile is not exactly the same as
the experiment, we observe similar characteristics, including the fluctuating profile with
roughly the same period. The simulated peak load is also reasonable close to the experimental
value. Hence we conclude that the PM simulation compares favorably with the experimental
measurements.
After the confidence with the model is gained, a number of typical extreme loading
problems can be explored with PM technique. Next, some selected PM applications will be
presented.
Experiment conducted at the University of Mississippi.
Figure 5. Comparison of experimental and modeling results of vinyl ester due to the impact of a rigid
indenter: load, energy, deflection and indenter speed curves versus time. Maximum drop velocity of
indenter is -1.91 m/s.
3.1.2. Miscellaneous Applications
(A) High Speed Collision and High Strain Rate Tension/Compression

of Material Blocks
Simulation of high speed collision of material blocks is quite a challenge to continuum

based models as dynamic discontinuity of material geometry is impossibly predefined and
stress singularity must be elaborately considered in the vicinity of cracks. However, the basic
treatment of PM, in which fracture is created when a bond (spring) is broken by translational
force, provides PM a unique power to be able to quite easily overcome these problems. In
PM, for each particle, the fracture density of bond illustrated in Equation (20) can plot exactly
its connection with the remaining assembly of material. Thus, the discontinuity of material
can be easily traced. Else, since the consideration of PM is not based on the stress intensity
factor, energy release rate, or the plastic process zone near or at the crack tip, therefore, in
reality PM does not create a stress singularity. It is worth to point out that, although the near
fracture tip behavior cannot be modeled as the classical continuum fracture mechanics,
following Equation (1) PM can model the nonlinear constitutive behavior near the crack, as
well as everywhere else.

Figure 6. PM simulations: (a) high speed collision of two materials blocks, (b) high strain rate
compression, and (c) high strain rate tension.
Figure 6(a-c) displays PM simulations of high speed collision, tension and compression
of material blocks with heterogeneity. Different material properties plotted in different colors
in Figure 6 are assigned with different (p, q) into Equation (1).
(B) Blasting Simulations
Simulation of blasting problems is another important application of PM. Figure 7(a, b)

displays the PM blasting simulations of a wall with and without a retrofitting treatment. From
Figure 7(a) it is found that for the case without the retrofitting enforcement attached to
infrastructure, debris flow of the material will fly into the area behind the wall after the
structure is fractured, and consequently can cause severe damage to the creatures wherein. In
contrast, for the case with the retrofitting consideration, due to the resistance from the
Figure 7. Examples of PM simulations of blasting (a) homogeneous material, (b) retrofitting material
structure.
enforcement of the layered material, there are no material fragments intruding into the area
behind the wall, as shown in Figure 7(b); hence, people and facilities in the area will be well
protected.
(C) Crack Formation and Propagation in Different Materials
Crack formation and propagation can result in a sudden catastrophic failure of material. It
is obvious that many lives may survive if the failure propagation speed can be postponed to
have more time for the victims to escape or be rescued from the disaster once it happens.
Thus, investigations of crack formation and propagation within different materials are
necessary. PM technique can be easily applied for the study of crack formation and
propagation in different materials, say ductile or brittle. Equation (1) with different (p, q) can
result in different materials. Table 1 illustrates the physical outcomes by using (p, q) = (3, 5),
(5, 10) and (7, 14), under equilibrium lattice spacing r0 = 0.2 cm. Figure 8 displays the
according interaction force profile vs. the three above-employed (p, q) values.
Figure 8. Interaction force of PM under r0 = 0.2 cm, with respect to (p, q) = (3, 5), (5, 10) and (7, 14).
Table 1. Physical outcomes with (p, q) = (3, 5), (5, 10) and (7, 14) under equilibrium
lattice spacing r0 = 0.2 cm
(p, q) (3, 5) (5, 10) (7, 14)

6
G 2.473 ×10 7
1.781×10 1.102 ×105
H 9.892 ×105 5.698 ×102 1.411
E (GPa ) 15.457 69.557 150.706
86.205
σ TS ( MN / m2 ) 263.570 441.534
Necking position 1.29 • r0 1.15 • r0 1.11 • r0

Table 1 illustrates that with the increase of (p, q), (i) the Young’s modulus and tensile
strength values of the resultant material also increase. So does the necking position of the
interaction force profile shown in Figure 8 as well as in Table 1. This indicates that the
material with big (p, q) tends to be brittle and vice versa.
Figure 9(a-c) displays, respectively, the initial crack formation, propagation and final
crack pattern under (p, q) = (5, 10); Figure 10(a-c) shows, respectively, the initial crack
formation, propagation and final crack pattern under (p, q) = (7, 14).
Figure 9. Time-dependant fracture of 2D plate with initial crack-tip under r0 = 0.2 cm, (p, q) = (5,
10). Streching rate = 40 cm/s. (a) T = 3.66 ms , (b) T = 3.72 ms , and (c) final crack pattern.
Figure 10. Time-dependant fracture of 2D plate with initial crack-tip under r0 = 0.2 cm, (p, q) = (7,
14). Streching rate = 40 cm/s. (a) T = 2.58 ms , (b) T = 2.64 ms , and (c) final crack pattern.
Comparing Figures 9 and 10, we obtain some very interesting results:
(a) Crack develops sooner and propagates faster in brittle materials than in ductile
materials (compare Figs. 9(a, b) with 10(a, b)).
(b) Crack propagation tends to behave in an unstable manner in brittle materials while in
a steady fashion in ductile materials (compare Figs. 9(c) with 10(c)).
The benefit from this research is that the above-obtained results may help fabricate a
high-resistance retrofitting layered structure optimally comprised of different materials. This
fabricated enforcement structure is then coated to the infrastructure to postpone the failure
propagation speed, as shown in Figure 11. Prolonging the failure process with this treatment,
more lives can escape from the disaster once failure occurs.
Figure 11. A fibrication of a retrofitting layered structure with different material properties attached to
the infrastructure to postpone the failure propagation speed.
(D) Thermally-Induced Breakage of Ores
It has been found that most minerals on the earth or on the other planets in space (e.g.
Moon, Mars, etc.) are composed of two types of materials: little or no heat generated, such as
calcite, etc., and heat generated, such as pyrite, etc. Therefore, a rapid heating of ore minerals
in a microwave-transparent matrix can generate thermal stress of sufficient magnitude to
create micro-cracks along grain boundaries, and this type of microcracking might have the
potential to improve ore grindability and increase liberation of individual mineral phases.
Wang, et al, 2008b.
Figure 12. PM simulation of fracturing efficiency results of rock using different microwave power
density. (a) Pd = 1.0 × 109 , T = 1.37× 10−2 s ; (b) Pd = 1.0 × 1011 , T = 7.5× 10−3 s .
Wang, et al, 2008b.
Figure 13. Mechanical energy consumption of breakage by a constant uniaxial compression of

20.0 cm / s applied after different time length of microwave radiation on the heterogeneous material
(a) T = 1.0× 10 −3 s , (b) T = 2.0× 10 −3 s , (c) T = 3.0× 10 −3 s , (d) T = 4.0× 10 −3 s .

Microwave power density Pd =1.0 × 1011 W / m3 .
If microwave energy can indeed induce microcracking around phase grain boundaries of ores,
the reductions in required comminution energy and enhanced liberation of valuable mineral
would occur. In practice, experimental approaches cannot easily reach a precise insight into
the entire thermal fracture process of ores because most minerals are brittle materials with a
complex three dimensional structure. In this situation, numerical investigations are powerful.
To meet this demand, a thermal-based PM has been established and been employed to explore
this problem [Wang, et al, 2008b].
Figure 12 (a, b) shows the global averaged fracture density results with respect to the two
different microwave input powers. It shows that heating first causes thermal expansion of
pyrite (interior circular part). Consequently the calcite part (exterior rectangular part) is
fractured by the pyrite expansion from the inside. In detail, we see that breakage originates
from the region of interface between the two material phases and propagates into the calcite
phase. It demonstrates (i) that heating pyrite in a microwave-transparent matrix, calcite,
indeed creates micro-cracks along grain boundaries, and this type of microcracking can
improve ore grindability and increase liberation of individual mineral phases; and (ii) that the
bigger power density of microwave is applied, the sooner the fracture is generated and the
higher breakage efficiency is reached.
Figure 13 illustrates the modeling results of a study in which, first we use microwave to
heat the pyrite/calcite sample in different durations, and then a mechanical breakage approach
is applied. We find that the longer we heat up the material, the more mechanical energy is
saved from the breakage cost. This reveals that it is an ideal option to combine microwave
breakage technique with mechanical methods to aid the mechanical breakage of ores and
mineral assemblages. This new technique not only can help enhance mining efficiency on the
earth, but also can be applied to aid human space exploration activities.
3.2. Deficiencies of PM and Potential Solutions
Since PM employs the simple, nearest-neighbor and axial linkage mechanism, and the
according computations are conducted on a regular triangular lattice network, three following
technical deficiencies are unavoidable:
1) A bias in the fracture propagation direction inherent to the geometry of the lattice
network employed, say, along the 60° direction for the equilateral triangular, nearest
neighborhood network. This deficiency of the regular lattice model has been
demonstrated by Jirasek, et. al. [1995a, b] and Schlangen [1995]. A mesh bias
example is illustrated in Figure 14.
2) A fixed Poisson ratio. Lattice model (LM) theory has addressed that the equivalent
Poisson’s ratio on a 2-D equilateral triangular lattice is fixed to the value of 1/3
whereas it is ¼ for a 3-D structure [Ostoja-Starzewski, 2002].
3) Isotropic solver. An equilateral triangular lattice network is limited to a valid
description of isotropic materials [Ostoja-Starzewski, 2007].
Wang, 2005b.
Figure 14. Example of mesh effect by lattice model: compression simulation (red line indicates crack path).
The above three situations can be remedied by the use of the more advanced lattice
models. For example, Ostoja-Starzewski [2002, 2007] has manipulated several types of spring
systems, including central ( α ), angular ( β ) and the mixed ( α - β ) interactions, coupled
with different lattice networks (triangular, rectangular, etc.) as well as multiple neighborhoods
(nearest, second neighboring particle interaction, etc.), that allows the modeling of various
Poisson ratio materials, elimination of mesh bias problem, as well as the potential for
applying to anisotropic materials.
Ostoja-Starzewski [2002] addressed combined axial-angular ( α - β ) scheme that
theoretically works for various Poisson ratio and anisotropic materials as in Figure 15. For
simplicity, assigning both of six axial and six angular springs equal, an isotropic LM solver is
obtained that works for Poisson ratios ranging from (-1, 1/3]. For the cases with Poisson’s
ratio ranging from 1/3 up to 1, a ‘triple honeycomb lattice’ can meet this demand [Ostoja-
Starzewski, 2002]. As is shown in Figure 16, this isotropic technique considers nearest
neighbors but sets up three axial spring constants α1 , α 2 and α 3 in each triangular unit cell,
respectively. Synder et al [1992] derived the Young’s modulus and the Poisson’s ratio from
this technique as follows.
2 3(α1 + α 2 + α 3 )
E= (21)
3{1 + 2(α1 + α 2 + α 3 ) / 9[(1/ α1 ) + (1/ α 2 ) + (1/ α 3 )]}
2
ν = 1− (22)
{1 + 2(α1 + α 2 + α 3 ) / 9[(1/ α1 ) + (1/ α 2 ) + (1/ α 3 )]}
We have proposed a newly developed hybrid lattice particle modeling (HLPM) technique
that combines the strengths of the LM and the PM [Wang, et al., 2008c,d]. In Table 2, the
strengths and weaknesses of the traditional LM and PM are summarized and compared. It is
clearly seen that HLPM contains the strengths of both LM and PM.
Ostoja-Starzewski, 2002.
Figure 15. Schematic of axial-angular ( α - β ) scheme.

Figure 16. A triangular lattice with hexagonal unit cell for α1 − α 2 − α 3 model, considering nearest
neighbor particle interactions; after Snyder et al [1992].
Table 2. Comparison of the lattice model (LM), the particle model (PM), and the hybrid
lattice particle model (HLPM)
Hybrid Lattice
Lattice Model (LM) Particle Model (PM) Particle Model
(HLPM)
Particle interaction spring (axial/angular), Lennard-Jones spring (axial/angular)
beam, etc. potential (axial only) mimicking the
Lennard-Jones
potential
Interaction not limited to nearest nearest neighbor only not limited to nearest
neighborhood neighbor neighbor
Mesh system Eulerian Lagrangian Lagrangian
Poisson’s ratio flexible fixed flexible
Time process static dynamic based on dynamic based on
Newton’s second law Newton’s second law
Force- Displacement (strain) Force interpreted Force interpreted from
displacement interpreted from force from displacement displacement
relation (stress) (distance between (distance between
particles) particles)
The principle of HLPM can be described as follows: the particle-particle interaction is

derived from lattice modeling (LM) theory whereas the computational scheme follows
particle modeling (PM) technique. The current demonstration of the HLPM is based on a
linear elastic model with an ultimate translational strength (i.e., tension/compression). Once
the translational strength is exceeded, the spring is broken and a fracture is created. The linear
model is created by using a quadratic form instead of the Lennard-Jones or polynomial
potential. The nonlinear constitutive laws are not implemented at the present stage because
the lack of independent measurement of the nonlinear material properties of the material
tested. Once such constants are available, the implementation of nonlinear constitutive law is
inherent in the Lennard-Jones or polynomial potential approach of the PM, and is not a
difficulty for the HLPM at all.
The above-introduced Figures 4 and 5 are the successful outcomes of HLPM that
eliminate mesh bias problem after employing a two-layer neighboring particle interaction
scheme [Holnicki-Szulc and Rogula, 1979a] shown in Figure 17.
Holnicki-Szulc and Rogula, 1979a.
Figure 17. Schematic of a two-layer neighboring particle interaction scheme.
Inputting the spring constants α1 , α 2 and α 3 of Eqs. (20) and (21) into HLPM, a
preliminary modeling result of the analogous case to Figure 4 while with a large Poisson
ratioν = 0.443 is obtained as illustrated in Figure 18.
Figure 18. A preliminary HLPM result using α1 − α 2 − α 3 scheme. Poisson ratioν = 0.443 .
Figure 19. A nonlinear constitutive polynomial for HLPM.
As has been addressed, HLPM can be easily extended to inelastic considerations if a

nonlinear constitutive polynomial is alternatively employed, instead of a Lennard-Jones type,
as a definition of particle-particle interactions in Figure 19. A preliminary inelastic numerical
result of HLPM is given herein. Adopting a 3-order polynomial into HLPM shown in Figure
20(a) to present a material with weak inelasticity, the modeling result of the analogous case to
Figure 4 is predicted in Figure 20(b).
(a)

(b)
Figure 20. A weakly nonlinear modeling result of HLPM, (a) constitutive equation, (b) HLPM result.
4. Conclusion
This paper systematically introduce a novel particulate dynamics based modeling
approach, particle modeling (PM) technique, and its successful applications in a number of
fracture problems of solids with dynamic fragmentation under various extreme loading
conditions. These loading conditions can include situations of collapse, impact, blasting or
high strain rate tension/compression, as well as thermally-induced breakage problems.
Meanwhile, the deficiencies of PM are also addressed and the associated solution, a hybrid
lattice particle modeling (HLPM) scheme, is proposed.
PM is a numerical technique similar to the molecular dynamic (MD) simulation; but
rather than simulating actual atoms, it is based on lumped mass particles distributed on a grid
to allow macro scale modeling. The PM utilizes an equivalent Lennard-Jones or polynomial
potential to model the nonlinear constitutive law at the continuum, macroscopic level. The
mass has inertia that obeys Newton’s second law of motion. It is a Lagrangian model that
keeps track of particle location and velocity. The advantages of PM over the existing discrete
element based methods can be summarized as follows:
1) Sample in theory. Four conservative/equivalent rules (mass, potential energy,

Young’s modulus and tensile/compression strength) are applied to preserve the
equivalent material properties.
2) Easy for implementation. Since the physical size of each particle is ignored other
than its equivalent mass, the algorithm of coding a PM computation is fairly easy.
Despite its success, the PM has a few deficiencies: (i) a bias in the fracture propagation
direction inherent to the geometry of the lattice network employed, say, along the 60°
direction for the equilateral triangular, nearest neighborhood network; (ii) a fixed Poisson
ratio. In the modeling of solid with PM, the potential type formulation allows only one elastic
constant to be modeled, which is typically selected as the bulk modulus, or Young’s modulus.
The second elastic constant present in an isotropic material, say, Poisson’s ratio, becomes the
property of the grid system used, such as the triangular or rectangular networks; and (iii) an
isotropic solver, as a result of the geometry of the adopted lattice network.
To overcome the three major above-mentioned shortcomings, a hybrid lattice particle
modeling (HLPM) approach is proposed for the simulation of dynamic fracture phenomena in
homogeneous and heterogeneous materials at macro-scales with a variable Poisson’s ratio
effect and anisotropic properties. It is concerned with the mathematical derivations of
employing elastic interaction formula between contiguous particles in 2-D lattice networks
accounting for different linkage mechanisms and different shapes of lattice. For instance,
axial ( α ) and combined axial-angular ( α - β ) models are considered coupling with different
lattices (triangular and rectangular, etc.) as well as multiple neighborhoods for the particle
dynamic interactions. The principle of HLPM can be described as follows: the particle-
particle interaction is derived from lattice modeling (LM) theory whereas the computational
scheme follows particle modeling (PM) technique. The newly proposed HLPM is free from
the above-mentioned deficiencies and can be applied to a wide range of impact and dynamic
fracture failure problems.
An outlook of future PM development will be outlined as follows:
1) Implementation of HLPM with all the above-mentioned linear schemes from lattice
modeling (LM) and other nonlinear constitutive laws.
2) Validation of HLPM with more real tests.
3) Improvement of PM by feedbacks from modeling applications.
Acknowledgement
This work was partially supported by Natural Sciences and Engineering Research
Council of Canada (NSERC), COREM (Quebec Mining Technology Research Institute), and
also by the funding received under a subcontract from the Department of Homeland Security-
sponsored Southeast Region Research Initiative (SERRI) at the Department of Energy's Oak
Ridge National Laboratory, USA.
References
[1] Al-Ostaz, A. and Jasiuk, I. (1997), Crack initiation and propagation in materials with
randomly distributed holes, Engineering Fracture Mechanics, 58, 395-420.
[2] Ashby, M.F. and Jones, D.R.H. (1980), Engineering Materials 1: An Introduction to
Their Properties and Applications, Pergamon Press.
[3] Greenspan, D. (1997), Particle Modeling, Birkhäuser Publishing.
[4] Greenspan, D. (1981), Computer-Oriented Mathematical Physics, University of Texas
at Arlington, Pergamon Press.
[5] Hockney, R.W. and Eastwood, J.W. (1999), Computer Simulation Using Particles,
Institute of Physics Publishing.
[6] Holnicki-Szulc, J. & Rogula, D. (1979a), Non-local, continuum models of large
engineering structures, Arch. Mech. 31(6), 793-802.
[7] Jirasek, M., and Bazant, Z. P. (1995a), Macroscopic fracture characteristics of random
particle systems, Int. J. Fract., 69(3), 201-228.
[8] Jirasek, M., and Bazant, Z. P. (1995b), Particle model for quasi-brittle fracture and
application to sea ice, J. Eng. Mech., 121(9), 1016-1025.
[9] Meguro, M. and Tagel-Din, H. (2000), Applied element method for structural analysis:
theory and application for linear materials, Structural Eng./Earthquake Eng., JSCE,
17(1), 1-14.
[10] Monaghan, J. (2005), Smoothed particle hydrodynamics, Rep. Prog. Phys, 68(1), 1703-
1759.
[11] Oñate, E., Idelsohn, S.R., Pin, F. D, and Aubry., R. (2004), The particle finite element
method. An Overview, International Journal Computational Method, 1(2), 267-307.
[12] Ostoja-Starzewski, M. (2007), Microstructural Randomness and Scaling in Mechanics
of Materials, in Modern Mechanics and Mathematics Series, Chapman &
Hall/CRC/Taylor & Francis.
[13] Ostoja-Starzewski, M. and Wang, G. (2006), Particle modeling of random crack patterns
in epoxy plates, Probabilistic Engineering Mechanics, 21, 267-275.
[14] Ostoja-Starzewski, M. and Wang, G. (2005), Probability and materials: from nano- to
macro-scale, NSF Workshop, Baltimore, MD, USA, January 5-7.
[15] Ostoja-Starzewski, M. and Wang, G. (2004), Particle modeling of dynamic
fragmentation, International Symposium on Developments in Plasticity and Fracture:
Centenary of M. T. Huber Criterion, Cracow, Poland, August 12-14.
[16] Ostoja-Starzewski, M., Lattice models in micromechanics, Appl. Mech. Rev., 2002,
55(1), 35-60.
[17] Schlangen, E. (1995), Computational aspects of fracture simulations with lattice models,
Fracture Mechanics of concrete structures (Proc. FraMCoS-2, Zurich), F. H. Wittmann,
ed., Aedificatio, Freiburg, Germany, 913-928.
[18] Wang, G., Al-Ostaz, A., Cheng, A.H.-D. and Mantena, P.R. (2008a), Particle modeling
of a polymeric material (nylon-6, 6) due to the impact of a rigid indenter, Computational
Materials Science (in press).
[19] Wang, G., Radziszewski, P. and Ouellet, J. (2008b), Particle modeling simulation of
thermal effects on ore breakage, Computational Materials Science (In press).
[20] Wang, G., Al-Ostaz, A., Cheng, A.H.-D. and Mantena, P.R. (2008c), A hybrid particle-
lattice model for dynamic fracture problems, AAM’08, Louisiana, June 17-20.
[21] Wang, G., Al-Ostaz, A., Cheng, A.H.-D. and Mantena, P.R. (2008d), Hybrid lattice
particle modeling: theoretical considerations for a 2-D elastic spring network for
dynamic fracture simulations, Computational Materials Science. (revised).
[22] Wang, G., Al-Ostaz, A., Cheng, A.H.-D. and Mantena, P.R. (2008e), Particle modeling
of crack propagation in materials at a macroscopic level, AAM’08, New Orleans,
Louisiana, June 17-20.
[23] Wang, G., Al-Ostaz, A., Cheng, A.H.-D. and Mantena, P.R. (2008f), Particle modeling
of dynamic fracture simulations of a 2D polymeric material (nylon-6,6) subject to the
impact of a rigid indenter, EM’08, Minneapolis, Minnesota, May 18-21.
[24] Wang, G., Radziszewski, P. and Caron, S. (2007a), Particle modeling of material
comminution, Discrete Element Methods (DEM) ’07, Brisbane, Australia, August 27-29,
8pp.
[25] Wang, G., Radziszewski, P. and Ouellet, J. (2007b), Exploring DEM model
development for the simulation of thermal effects on ore breakage, Discrete Element
Methods (DEM) ’07, Brisbane, Australia, August 27-29, 8pp.
[26] Wang, G., Radziszewski, P. and Ouellet, J. (2007c), Particle modeling approach for the
simulation of thermal effects on ore breakage, Mid-South Area Engineering and
Sciences Conference (MAESC) ’07, the University of Mississippi, Oxfod, May 17-18.
[27] Wang, G., Radziszewski, P. and Caron, S. (2007d), Particle modeling of material
comminution process, Mid-South Area Engineering and Sciences Conference (MAESC)
’07, the University of Mississippi, Oxfod, May 17-18.
[28] Wang, G., Ostoja-Starzewski, M., Radziszewski, P. and Ourriban, M. (2006), Particle
modeling of dynamic fragmentation – II: fracture in single- and multi-phase materials,
Computational Materials Science, 35, 116-133.
[29] Wang, G. and Ostoja-Starzewski, M. (2005a), Particle modeling of dynamic
fragmentation – I: theoretical considerations, Computational Materials Science, 33, 429-
442.
[30] Wang, G. (2005b), Particle modeling of dynamic fragmentation, Ph.D. Dissertation,
McGill University, Canada.
[31] Wang, G., Ostoja-Starzewski, M. and Radziszewski, P. (2005c), Particle modeling of
comminution, 20th Canadian Congress of Applied Mechanics, May 30th to June 2nd,
Montreal, Canada, 149-150.
[32] Synder, K.A., Garboczi, E.J., and Day, A.R., The elastic moduli of simple two-
dimensional composites: computer simulation and effective medium theory, J. Mech.
Phys. Solids, 1992, 72, 5948-5955.
Chapter 6
NON-ORIENTED ELECTRICAL STEELS: MATERIALS

FOR SAVING ENERGY AND CONSERVING
THE ENVIRONMENT
Taisei Nakayama*
Sumitomo Metal Industries, Ltd.
Wakayama Steel Works
1850 Minato, Wakayama, 640-8555 Japan
Abstract
Electrical steels are the core materials for electrical motors or transformers. Those materials
for motors are played an energy conversion roll from electricity to motion. However, energy
losses are accompanied with this conversion. To minimize these losses is a key technology to
conserve our environment.
Numerous researches on the grain-oriented electrical steels reported. Those researches
especially for transformers are focused on the reducing the losses at supplying the electricity
from power plants. On the other hand, home or industrial appliances are the power consuming
devices, and the most effective point on the energy loss reduction. These home appliances are
used small motors using non-oriented electrical steels.
In this review, several researches on the non-oriented electrical steels are discussed and
focused on the metallurgical control of the steels to reduce the core loss for generating waste
heats and motor building innovation technologies for decreasing the building factor of the
core losses.
In the metallurgical part, some additive elements as phosphorus, aluminum and
manganese for improving magnetic properties reviewed. Moreover some contaminating
elements as vanadium, titanium and zirconium are discussed especially for precipitation
studies in the steels have been done. These precipitations are inhibited the grain growth at
final annealing or stress relief annealing. These inhibited small grains increase the core losses.
For studying motor building technologies, compression stress effect, shearing stress
effect are discussed. Even though the best core materials are used for manufacturing motors,
those building deteriorations make worse for the motor efficiency. Therefore, those
technologies are also important for reducing the carbon dioxide emission.
*
E-mail address: [email protected]. Phone:+81-73-451-2400. Fax:+81-73-451-2412.
184 Taisei Nakayama
Introduction
Electrical steels are the core materials for electrical motors or transformers. Those
materials for motors are played an energy conversion roll from electricity to motion.
However, energy losses are accompanied with this conversion. To minimize these losses is a
key technology to conserve our environment. Reducing total carbon dioxide emission is an
important issue identified since the 3rd Conference of Parties (COP3) of United Nation
Framework Convention on Climate Change (UNFCCC) in Kyoto in 1997. Thus, much
attention is now paid to the motor or transformer efficiency.
Numerous researches on the grain-oriented electrical steels have been reported [1-5].
Those researches especially for transformers are focused on the reducing the losses at
supplying the electricity from power plants [1,4]. Metallurgical studies on grain oriented
electrical steels such as the domain control technologies [2] or secondary recrystallization
[3,5] to improve the magnetic properties for reducing core loss.
On the other hand, electrical motors are the power consuming devices. To reduce the
carbon dioxide emission, these devices should be performed more effectively. Most of the
electrical motors are used for home appliances. They are installed for motion controlling or
power devices. These small motors are made of non-oriented electrical steels as core
materials. Recent eco-applications need to install such high efficient motors [6]. Same as
home appliances, large motors in industrial use should be focused on the motor efficiency.
Moreover, hybrid electrical vehicles [7,8], or trains and trams for transportation [9] are used
some small size high power high efficient motors.
Thus, the core loss reductions of electrical motors is necessary for reducing carbon
dioxide emission, and to adopt more energy effective materials is a key to solve the problem.
Much attention is now paid to developing magnetic properties through the control of
chemistry [10], grain size [11,12] and texture [13,14], processes [15]. Several new products
with low core loss and high induction have been developed　 [16-18]. These improvements
have been achieved with the advances in steelmaking technology, which turn out to be the
relationship between structure and magnetic properties.
In this paper, several researches on the non-oriented electrical steels are reviewed. There
are two parts of the major topics. The first part refers to the effects of additive and
contaminating trump elements in the silicon steels, and the second part is reviewed the
deteriorations by the core manufacturing process
In the first part, several effects of some positive elements for improving magnetic
properties such as manganese [19] or aluminum [20] or phosphorus [21,22], and effects of
trump elements as contaminants for deteriorating magnetic properties such as vanadium [23],
titanium [24] or zirconium [25] are reviewed.
The key technology of using positive additives, such as manganese or aluminum is how
to prevent to form harmful inclusions and precipitates. These inclusions or precipitates inhibit
to the grain growth during final annealing. Small grains and precipitates themselves lead to
the increase of the core loss, especially hysteresis loss. The effect of phosphorus is the grain
orientation control. It assists magnetically favorable textures during final annealing.
Contaminating elements such as vanadium, titanium and zirconium easily form the
precipitations with carbon or nitrogen in the steels. These precipitations inhibit the grain
growth during final annealing or stress relief annealing. These inhibited small grains increase
Non-Oriented Electrical Steels: Materials for Saving Energy and Conserving… 185
the core losses. The contaminating trump element controls are important technologies to
reduce the core loss. To clarify the maximum limit of these contaminants in the steels is
valuable for the commercial production.
Even though using the superior core material for manufacturing small motors, the core
building affect the core loss. In this review two deterioration cases, the magnetically
deterioration of interlocking lamination stacking [26,27] and the deterioration by the
compression force [28] after the core heat-shrunk into the case are referred.
To solve those deteriorations of core loss, some special steels with excellent stampability
and machinability controlled by chemistry are introduced [29,30].
Part 1. Effects of Additive or Contaminating Elements in the

Silicon Steels
Non-oriented electrical steels are classified as fully processed type and semi-processed
type [31,32]. Among fully processed type, some improved grades such as high magnetic
induction type are developed recently. Figure 1 [16,31] shows the magnetic properties of fully
processed steels. Magnetic inductions of the improved steels are higher than that of regular
one. Therefore, the improved steels make the motor or transformer size smaller than regular
steels with the same performances.
Figure 1. Non-oriented electrical steels manufactured by Sumitomo Metals.
Semi-processed steels are used to be manufactured with final skin-pass process. These
skin-passed steels relatively non-sensitive to the contaminant or inclusions. However,
innovations of the steel making technologies make the semi-processed electrical steels free
from skin-pass process. Figure 2 [31] shows the magnetic properties of semi-processed steels.
186 Taisei Nakayama
Controlling the chemistry in the steels especially, prevention from the contaminants makes
the semi-processed steels same process as fully processed steels.
Figure 2. Magnetic properties of semi-processed steels.
1.1. Additive Elements for Improving Magnetic Properties
1.1.1. Effects of Phosphorus [21 ,22]
Non-oriented electrical steels with low silicon (less than 1.0mass %) usually contain
phosphorus as additives for controlling mechanical properties. Low silicon steels with ultra-
low carbon lead to the low yielding strength if they do not have any other additive elements.
Due to prevent the magnetic aging carbon cannot be added more than 0.003 mass % [33].
However core making such as high speed stamping requires the high yielding strength
materials. To solve this problem phosphorus is widely used as an element for increasing the
yielding strength in low costs.
However higher silicon products more than 1% are usually non-phosphorus bearing
steels, because their hardness or tensile strength increases by silicon itself. If the steels
contain both silicon and phosphorus, the mechanical properties of the steels are high strength
and they are harmful for stamping with die systems.
Marvelous die systems [34] enable to make high speed core stamping with such high
strength silicon steels. Recently some high efficient interior permanent magnet synchronous
motors whose rotor core shapes are sophisticated are developed [6,35], and these high
efficient motors require the special die system with special core materials.
In this case, core materials for high efficient motor with low core loss and high magnetic
induction are realized not only by the phosphorus content but also by the process especially
for the hot band annealing heat cycles.
In this study, two different level phosphorus bearing steels with 0.002 mass % carbon, 2.0
mass % silicon, 0.3 mass % aluminum and 0.2 mass % manganese are used.
Figure 3 shows the relationship between cold rolled reduction and core loss at 1.0 T and
400Hz (W10/400) and magnetic induction at 5000 A/m (B50) in comparison with phosphorus
bearing steel (P=0.1mass %) and regular steel (P=0.01mass %).
Cold rolled reduction (%)

77 84 87
W10/400 (W/kg)
35
Core loss 30
25
20
15
1.74
Induction
1.73
B50 (T)
1.72 P (mass %)
1.71 0.1
1.70 0.01
1.69
0.6 0.5 0.4 0.3 0.2
Sheet gauge (mm)
Figure 3. Effect of cold rolling reduction on the magnetic properties of phosphorus bearing steels.
In the case of regular silicon steel (P=0.01mass %) the magnetic induction is decreased as
increasing the cold rolled reduction for reducing core loss especially for eddy current loss,
while the magnetic induction of phosphorus bearing silicon steel (P=0.1mass %) is less
decreasing than that of regular silicon steel.
This less magnetic induction decrease with thin gauge (or high cold rolled reduction)
sheet requires a suitable hot band annealing condition for the texture control during the final
annealing Figure 4 (A) and (B) show the texture diagrams in the final products in comparing
with regular silicon steel and phosphorus bearing steel. The grains in the regular silicon steel
are oriented the magnetically unfavorable texture such as {111}<112> in the case of high
cold roll reduction (thin gauge) sheet, while that of phosphorus bearing steel are not.
Even though the phosphorus bearing steel without adequate annealing, this unfavorable
texture evolution occurs during the final annealing. Figure 4 (B) and (C) show the texture
diagrams comparison between with and without hot band annealing. Hot band annealing
affects the grain orientation of the final products. In this case, hot band annealing induced the
magnetic favorable texture after cold rolling and final annealing.
188 Taisei Nakayama
(A)regular steel (0.01 mass% P) with batch type hot band annealing; (B)improved steel (0.1 mass % P) with
batch type hot band annealing ; (C)improved steel without hot band annealing.
Figure 4. Texture comparison between phosphorus bearing steel and a regular steel.
1.1.2. Effects of Aluminum [20]
Aluminum plays a roll to improve the magnetic properties as an additive element. It

increase the resistivity of steels same as silicon. However, aluminum combines easily with
nitrogen, forms aluminum nitrides. Thus, controlling this precipitates is important to improve
the magnetic properties.
In steels containing　 0.001mass % carbon, 0.3 mass % silicon, 0.3 mass % manganese,
0.08 mass % phosphorus, 0.005 mass % sulfur and 0.0005-0.0042 mass % nitrogen with
0.001-0.30 mass % aluminum, Figure 5 shows the relationship between core loss at 1.5 T and
50 Hz (W15/50), and aluminum and nitrogen content after stress relief annealing. The core
loss of the steels with <0.001 mass % nitrogen keeps low in the range of aluminum < 0.01
mass %, and it increases with an increase of the aluminum content toward 0.03 mass %, then
it peaks at 0.03 mass %, over there, the core loss decreases with an increase of the aluminum.
In the case of the steels with the nitrogen > 0.002 mass %, the core loss increase in the range
of aluminum <0.01 mass %, then it peaks at 0.01 mass %, over there, it decrease with an
increase of aluminum. Hou et al.[36] investigated the effects of aluminum on the magnetic
properties of lamination steels and concluded that the aluminum addition to the steels
decreased the core loss through increasing the electrical resistivity to reduce the eddy current
loss and coarsening grain size to decrease the hysteresis loss. However in that study, any
comment of the effect of nitrogen did not mention, despite of studying the effect of aluminum
nitride precipitates.
Figure 5. Effect of aluminum and nitrogen contents on core loss W15/50 after stress relief annealing.
(□: N=0.0005-0.0009 mass %; ○:N=0.0021-0.0028 mass %;　 △:N=0.0035-0.0043 mass %;).
On the point of the magnetic induction, Figure 6 supported the Hou’s study. As
increasing the aluminum content, the magnetic induction at 5000 A/m (B50) decreased in any
nitrogen level. Yashiki and Okamoto [37] investigated the effect of hot band annealing and
concluded that the grain size of the hot band affect the magnetic induction, however in this
study as shown in Figure 6, hot band grain size does not affect the magnetic induction in the
range of the nitrogen content in this study (<0.0043 mass %). However magnetic induction
B50 is slightly decreasing as an increase of aluminum contents over 0.1 mass % in steels with
<0.0043 mass % nitrogen. This decreasing magnetic induction is affected by a decrease of the
saturated magnetic induction as increasing aluminum content.
The deterioration of the core loss was caused by the small grains pinned down by the
precipitation of small aluminum nitrides (< 0.5 μm). To compare with 0.001 mass %, 0.078
mass % and 0.30 mass % aluminum steels with < 0.001 mass % nitrogen, grain size in these 3
steels are different. Steels with 0.001 mass % and 0.30 mass % aluminum are almost same
grain size (50-80 μm), while steel with 0.078 mass % aluminum is small (20-30 μm). To
make the reason why steel with 0.078mass % aluminum has small grains clear, precipitates in
these steels were investigated by TEM. In the steel with 0.001 mass % aluminum, the major
190 Taisei Nakayama
inclusions or precipitates are MnS, and in the steel with 0.078 mass % aluminum, many
‘harmful AlN’ (< 0.5μm in size) were observed along the grain boundaries, while in the steel
with 0.30 mass % aluminum, some ‘harmless large AlN’ (>1.0μm in size) were observed in
the grain. This means the small AlN around 0.5 μm were pinned down the grain boundaries,
therefore, the grains were not coarsened at all at the stress relief annealing.
Figure 6. Effect of aluminum and nitrogen contents on induction B50 after stress relief annealing. ( □:
N=0.0005-0.0009 mass %; ○:N=0.0021-0.0028 mass %;　 △:N=0.0035-0.0043 mass %;).
As Sawamura and Mori [38] showed the AlN solubility of 1.0 mass % Si steel with
0.01mass % Al and 0.005 mass % nitrogen and concluded that the all Al and N were
precipitated as AlN at the temperatures < 900 K. Following that study, aluminum and
nitrogen should be precipitated as no ‘harmful AlN’ on the magnetic properties with < 0.0024
mass % AlN. Darken et al. [39] studied the behavior of AlN in austenite steel and introduced
the equation of AlN solubility. According to the Darken’s equation, all AlN were soluted with
< 0.005 mass % nitrogen and < 0.01 mass % aluminum at the slab reheating temperature
(1450 K). The magnetic properties and the precipitates studies indicated that the ‘harmful
AlN’ (< 0.5μm in size) in the steels with > 0.0024 mass % AlN during the hot rolling process.
These results follow Darken’s equation statistically. Moreover, Darken studied nitriding in
aluminum killed steel and pointed out the nitrogen diffusivity was low and formed ‘subscale
layer’ as the surface nitrided band, however, this layer was not observed in the hot band. In
the case of precipitating AlN during the hot rolling process, the ‘harmful AlN’ in the slab
were soluted, while ‘harmless large AlN’ (>1.0μm) were not. In steels with Al > 0.1 mass %
in Figure 11, this ‘harmless large AlN’ (>1.0μm) which keep large in size at slab reheating
temperatures (1450 K) with following hot rolling process, did not pinned down the grain
boundaries, and did not affect on the magnetic properties. These ‘harmless large AlN’
precipitates were observed with > 0.1 % Al in this study.
In steels with excess nitrogen, MnSiN2 were observed (Figure 7), this is why the core loss
of 0.001 mass % aluminum and high nitrogen steel, which have no ’harmful AlN’, were
higher than steel with < 0.001 mass % nitrogen. These manganese silicon nitrides were the
same as Yashiki and Kaneko [40] reported in the electrical steels.
Figure 7. Precipitates in 0.3 mass % silicon steels with aluminum.
1.1.3. Effects of Manganese [19]
Same as aluminum, manganese is also an element for increasing a resistivity of the steels.
However, its resistivity increase per weight addition is about a half of the silicon’s or
aluminum’s [33]. Moreover, manganese also forms harmful precipitates as manganese
sulfides. It affects the magnetic properties. In steels containing　 0.001 mass % carbon, 0.3
mass % silicon, 0.3 mass % aluminum, 0.08 mass % phosphorus, and 0.001 mass % nitrogen
with 0.12-0.91 mass % manganese and 0.0005-0.038 mass % sulfur, Figure 8 shows the
relationship between core loss at 1.5 T and 50 Hz (W15/50), and manganese and sulfur
content after stress relief annealing. The core loss of the steels increases with an increase of
the sulfur in each steel with different manganese content. The core loss decreases with an
increase of manganese. These phenomena are explained that the decrease of core loss with an
decrease of sulfur mainly are derived from an decrease of hysteresis loss by an decrease of
MnS inclusions and coarsened grains, and the decrease of core loss with an increase of
manganese are from a decrease of eddy current loss by increasing resistivity. For instance the
resistivity of the steels containing 0.3 mass % Si, 0.12% Mn, and 0.3 mass % Al is 20.9Ωm
×10-8, while that of steels cintaining 0.3 mass % Si, 0.93% Mn, 0.3 mass % Al is 25.9Ωm
×10-8, and the calculated core loss of steels with 0.93 mass % Mn decreased to 94% that of
0.1 mass % Mn, on the basis of eddy current decrease by the increase of resistivity. However,
observed core loss of the steels with 0.93 mass % Mn is 90% that of the 0.12 mass % Mn, and
the difference between the calculation and observation supposed to be a decrease the
hysteresis loss by coarsen grains in the steel with 0.93 mass % Mn [41].
192 Taisei Nakayama
Figure 8. Effect of manganese and sulfur on core loss W15/50 after stress relief annealing.
Figure 9. Effect of manganese and sulfur on induction B50 after stress relief annealing.
On the point of the magnetic induction, Figure 9 shows the effects of manganese and
sulfur. As increasing the sulfur content in each manganese bearing steel, the magnetic
induction B50 is almost no change in each manganese level. Yashiki and Okamoto [37]
investigated the effect of hot band annealing and concluded that the grain size of the hot band
affect the magnetic induction. However, the grain size of the hot band is slightly decreased as
increasing the sulfur content in the steels bearing 0.31% manganese, while magnetic
induction at 5000 A/m (B50) is slightly increasing as increasing the sulfur content. This might
be another effect, such as textures, on the magnetic induction B50.
In these MnS precipitations, slab reheating temperatures (SRT) strongly affect on the
magnetic properties. To clarify the precipitation, TEM observation was carried out in those
different SRT with different sulfur-bearing steels (Figure 10). MnS is observed very seldom
in steels with S<0.001mass % processed any SRT, however, steels with >0.001 mass % sulfur
processed SRT at 1273 K has some ‘coarse MnS’ (ca. 0.5μm), while those steels processed
SRT at 1423, or 1523 K many ‘fine MnS’ (ca. 0.1μm) were observed. This ‘fine MnS’ retard
the grain growth in recrystalization annealing and deteriorated the core loss.
Figure 10. Effect of sulfur content and slab reheating temperatures (SRT) on core loss W15/50 in 0.31
mass % Mn steels. (□:SRT at 1273 K; ○: SRT at 1423 K, : SRT at 1523 K).
1.2. Trump Elements for Deteriorating Magnetic Properties
1.2.1. Effects of Vanadium [23]
To control the contaminants such as vanadium, titanium or zirconium is important to

improve the magnetic properties. In the case of vanadium as contaminants in steels
containing　 0.001mass % carbon, 0.3 mass % silicon, 0.3 mass % manganese, 0.07 mass %
phosphorus, 0.005 mass % sulfur and 0.001 mass % nitrogen with 0.001-0.124 mass %
vanadium, vanadium nitride precipitates influenced the final grain size. The grain size
decreases with an increase in vanadium content < 0.016 mass %. However, when the
vanadium content is > 0.016 mass %, the steel grains again growth towards the lower
vanadium content size. It is well known that the grain size affects magnetic properties
[42,43]. In steel containing vanadium 0.016 mass %, vanadium carbonitrides were observed
along the grain boundary (Figure 11) when it was pinned down. Therefore, the effect of
vanadium carbonitrides　 in steel with 0.016 mass % vanadium on the retardation of grain
growth is more pronounced than that of compound precipitates in 0.001 mass % vanadium
steel. However, the pinning effect is weak in steels with vanadium contents > 0.016 mass %,
whose carbonitrides are larger than those with lower vanadium contents. This behavior is very
194 Taisei Nakayama
similar to that with aluminum nitrides. Nakayama and Honjou [19] or Hou et al.. [36] studied
the effect of aluminum on the magnetic properties of lamination steels and pointed out that
the size and distribution of aluminum nitride precipitates in hot-rolled plates influenced the
final grain size.
　　　V=0.016　　　　　V=0.124
　　　　　　　　　　　　　　　　　　(m a ss %)
　　　　　　　　　　　　　　 1μｍ
Figure 11. Precipitations of steels with vanadium.
Figure 12. Effect of vanadium addition on core loss W15/50 after stress relief annealing.
Figure 12 shows the relationship between core loss at 1.5 T and 50 Hz (W15/50) and
vanadium content after stress-relief annealing. The core loss is low at 0.001 mass %
vanadium content, as there is no vanadium carbonitrides region; at 0.124 mass % vanadium
content there is a large vanadium carbonitrides region in the precipitates study. By the
contrast with the core loss, magnetic induction at 5000 A/m (B50) decreased with an increase
in vanadium content < 0.016 mass %, but increased at contents > 0.016 mass % (Figure 13).
Magnetic induction or permeability is affected by the grain size and texture of hot bands.
Transmission electron microscope (TEM) study (Figure 11) made it clear that the vanadium
carbonitrides precipitated during hot rolling pinned down the grain boundary in the steel
containing vanadium 0.016 mass %. Therefore, these small hot band grains lead to a low
magnetic induction B50.
Figure 13. Effect of vanadium addition on induction B50 after stress relief annealing.
1.2.2. Effects of Titanium [24]
Same as vanadium, titanium also forms harmful precipitates to deteriorate the magnetic
properties. In the case of titanium as contaminants in steels containing　 0.001mass % carbon,
0.3 mass % silicon, 0.3 mass % manganese, 0.1 mass % phosphorus, 0.004 mass % sulfur, 0.3
mass % aluminum and 0.001 mass % nitrogen with 0.001-0.11 mass % titanium, the grain
size decreases with the increase of the titanium content. Although in the steel contains less
than 0.016 mass % titanium, the shape of grains was equiaxed, while in the range of the
titanium content over 0.016mass %, the grains were elongated. Vanderschueren [44] studied
the mechanism of recrystallization of IF steel and pointed out that the deformed ferrite grain
was very difficult to recrystallize. Moreover, Park et al. [45] investigated phosphorus in
titanium stabilized IF steel and summarized that phosphorus made a retardation on the
recrystallization of cold-rolled steel. Furthermore, Brun et al.[46] and Jeong and Chung [47]
investigated this deterioration and concluded that the precipitation of phosphides as (Fe, Ti)P
was inevitable in the steel containing much titanium and phosphorus unless coiling
temperatures was lower than 773 K on the hot rolling.
Figure 14 shows the TEM replica of the various titanium containing steel sheets after
continuous annealing. In steels containing 0.006 mass % titanium, titanium carbonitride,
pinning down the grain boundary, were observed along it. The effect of titanium carbonitride
on the retardation of grain growth is more pronounced than that of compound inclusions in
196 Taisei Nakayama
steel containing 0.002 mass % titanium. However, in steels containing more titanium (> 0.016
mass %) numerous iron-titanium phosphides (Fe,Ti)P precipitates were observed both in
grains and along grain boundaries. In the case of non-oriented semi-processed steels, some
phosphorus ( ca.0.1 mass %) is added to improve the stampability. Therefore, in these steels,
the precipitating behavior turns at the point between 0.006 mass % titanium, and major
precipitates change from titanium carbonitrides to iron-titanium phosphides.
　　　Ti=0.006　　　　　Ti=0.110
　　　　　　　　　　　　　　　　　　(m a ss %)
　　　　　　　　　　　　　　 1μm
Figure 14. Precipitations of steels with titanium.
Figure 15. Effect of titanium addition on core loss W15/50 after stress relief annealing.
titanium contents after stress relief annealing. Same as vanadium carbonitrides, titanium
carbonitrides affect the core loss drastically. Matsumura et al.[41] reviewed the non-
oriented electrical steel sheet and mentioned that the grain size affected the core loss and the
lowest core loss was achieved at the size about 150μm. The grain size smaller than 150 μm,
core loss was increasing with the decease of the grain size. In these case, the grain size was
less than 40μm, so that the core loss is increasing with the increase of titanium content.
Furthermore, Matsumura et al. [41] pointed out that core loss increase with increase of the
number of inclusions. In this case, steels containing lower than 0.016 mass % titanium have
less inclusions such as titanium carbonitrides, while steels containing over 0.016 mass %
titanium have numerous inclusions as (Fe,Ti)P. This is the reason why the core loss increases
dramatically in the steel containing over 0.016 mass % titanium.
By contrast with the core loss, magnetic induction at 5000 A/m (B50) was decreasing
with an increase in titanium content < 0.016 mass % and was dramatically dropped at
contents > 0.016 mass % (Figure 16). Magnetic induction or permeability is affected by the
grain size and texture of hot bands. Steels containing <0.016 mass % titanium were equiaxed
grain, while steels containing >0.016 mass % titanium recrystallized partially. Therefore,
these unrecrystallized hot bands lead the low magnetic induction B50. Moreover,{222}
texture, which is not easy to magnetize, developed in the finished sheets. This is another
factor of the deterioration.
Figure 16. Effect of titanium addition on induction B50 after stress relief annealing.
1.2.3. Effects of Zirconium [25]
Zirconium is also well-known element to form nitrides or carbonitrides. In the case of

zirconium as contaminants in steels containing　 0.001mass % carbon, 0.3 mass % silicon, 0.3
198 Taisei Nakayama
mass % manganese, 0.08 mass % phosphorus, 0.003-0.005 mass % sulfur, 0.3 mass %
aluminum and 0.0011-0.0036 mass % nitrogen with 0.001-0.133 mass % zirconium, the grain
size in steels containing >0.04 mass % zirconium are smaller than that in steels containing <
0.01 mass % shown in Figure 17. Numerous precipitates were observed both along the grain
boundaries and in grains. These precipitates were analyzed by the TEM replica method with
EDAX and proved that they are Zr3Fe. These precipitates did not inhibit the grain growth
compared to those of V(C,N) and Ti(C,N). It is well known that zirconium is one of the
easiest elements to combine with nitrogen and form zirconium nitrides, or zirconium
carbonitrides, but any zirconium nitrides or carbonitrides were observed in these steels [42].
Therefore, in steels with zirconium < 0.01 mass % , no zirconium containing precipitates but
aluminum nitrides and manganese sulfides were observed in this range , while in steels with
zirconium > 0.01 mass % , numerous Zr3Fe were observed and increased the core loss.
Zr=0.040　　　　　Zr=0.093
　　　　　　　　　　　　　　　　　　(m a ss %)
　　　　　　　　　　　　　　 1μm
Figure 17. Precipitations of steels with zirconium.
zirconium content after stress relief annealing. Core loss is almost no change with an increase
in steels containing zirconium < 0.01 mass %, while in steels containing zirconium > 0.01
mass %, the core loss is increasing drastically by the numerous precipitates both along the
grain boundaries and in grains. The effect of nitrogen content on the core loss is independent
of that of zirconium content, however, the iron loss in the higher nitrogen containing steel
was more deteriorated than that of lower one, due to the increasing hysterisis loss by
aluminum nitrides precipitates. On the other hand, the induction at 5000 A/m (B50) is very
little change by zirconium addition (Figure 19), due to no effects of zirconium addition on the
grain size of hot bands. As increasing the zirconium content, there is no change on both grain
size and texture, therefore, the magnetic induction B50 is no change in steels containing
zirconium < 0.133 mass %. Contrarily, as shown in Figure 18, the core loss is deteriorated in
steels containing zirconium > 0.01 mass %.
Figure 18. Effect of zirconium addition on core loss W15/50 after stress relief annealing. (○:N=0.0011-
0.0020 mass %; □: N=0.0027-0.0036 mass %).
Figure 19. Effect of zirconium addition on induction B50 after stress relief annealing. (○:N=0.0011-
0.0020 mass %; □: N=0.0027-0.0036 mass %).
200 Taisei Nakayama
Part 2. Core Manufacturing Technologies

The other problem on using electrical steels is core manufacturing. Even though using the
steels with the excellent magnetic properties, poor manufacturing leads the poor results.
Figure 20 shows a diagram of making motor cores. The following part is some key
technologies on the core manufacturing topics.
Figure 20. A process flow of motor core manufacturing.
2.1. Magnetic Properties Deterioration by Interlocking Lamination [26]
Figure 21. Effect of surface insulation coating thickness on the stamping performance with stamping
conditions; evaluated number of strokes : stamped number of strokes when burrs height reach to 50
μm; blanking shape: 17mm X 17mm square core; die materials: SKD11; knife clearance: 5- 7% of the
sheet thickness; stamping speed: 350 strokes/min; lubrication oil: kerosene; core material: 50SX1300)
Electrical steels sheets are usually coated with organic, inorganic or organic and
inorganic mixture insulation on the surface. Stamping performance or die worn rate of the
electrical steels with organic material as the surface insulation is better than that without
organic insulations. Japanese Industrial Standard (JIS) C-2552 classified the surface
insulation to CS-1(inorganic) and CS-2(organic and inorganic mixture), and CS-2 designated
to better performance in high speed continuous stamping. This means that the organic
insulation plays a lubrication role between steels and die at stamping. Figure 21 shows a
relationship between insulation coating thickness and number of strokes as a result of die
worn in continuous stamping. As increasing the coating thickness, stamping performance was
improved due to the highly lubrication by thick organic layer.
However this lubrication cause the interlocking performance to reduce the fastened
strength. Figure 22 shows the relationship between coating thickness and fastened strength.
As increasing the coating thickness, the fastened strength was decreased. In order to check the
lubricating role, we investigate the cross-section of sheared surface. Figure 23 shows the Cr
distribution map on the cross-section of stamped edges by electron probe micro-analyzer
(EPMA). To compare the cross-section after stamped, the electrical steel sheet with thick
coating (ca. 0.6μm thick) is more widely rubbed the coating fragments (detected as Cr) over
the sheared area than that with thin coating (ca. 0.3μm thick). These coating fragments play a
rubricating role and affect the fastened strength.
Figure 22. Effect of surface insulation coating thickness on the interlocking fastened strength of the ring
core with 4 V-cut bottom rectangular shape protuberances (VR) illustrated in fig.25 (material:
50SX1300).
Figure 23. Cr distribution on the cross-section of interlocking protuberances by electron probe micro-
analyzer (EPMA).
202 Taisei Nakayama
Since interlocking is a fastening by the friction at the cross-section, thin gauge sheets are
disadvantage on the stacking strength. Figure 24 shows the relationship between sheet gauge
and stacking strength. Thin gauge sheet such as 0.2 mm thick was so weak fastened that it
needs careful handling on core making or coil winding. On this handling-easy point of view,
0.27mm thick sheet is the thinnest gauge on the usual high-speed stamping for building a
small size motor. Kabasawa et al. [8] reported to apply the 0.27 mm thick non-oriented
electrical steel sheet for hybrid electrical vehicles and mentioned that the 0.27 mm thickness
sheet has the optimal balance between stacking strength and low core loss at high frequency.
Figure 24. Gauge effect on interlocking fastened strength of the ring core with 4 V-cut bottom circle
shape protuberances (VC) illustrated in fig.25 (core material: 2.0 % Si steels).
The deterioration of magnetic properties caused by not only simple obstacle of magnetic
flux flow but by flow from lamination to lamination through interlocking protuberances and
by compression or expansion stress between two fastened protuberances.
Fastened by flat bottom type interlocking (FC and FR in Figure 25) is weaker than that by
V-cut type (VC and VR). However, the magnetic deterioration of flat bottom interlocking is
less than that of V-cut one (Figure 26). Although the size of interlocking protuberance affects
the deterioration of core loss strongly, obstacle area on the core magnetic circuit is estimated
as no flow in protuberances as the model in Figure 27. Magnetic flux is concentrated in the
narrowed cross-sectional area beside protuberances. As increasing magnetic flux density, the
core loss is increased at the narrowed point. In the case of using the circular protuberances
and average magnetization at 1.0T, the narrowed area near protuberance, magnetic flux
density is calculated about 1.78 T, while in the case of using the rectangular type, it is about
1.2 T. The calculated core loss deterioration in Figures 28 and 29 is based on this core loss
increase at 4 protuberance points.
Fujimura et al.[27] analyzed and classified the core loss and concluded that the hysteresis
loss and anonymous eddy current loss increase by compressing or expanding stress. Therefore
in this study, to clarify the core loss deterioration by interlocking, all cores were measured
after stress relief annealing at 1023 K to eliminate the deterioration by all stress. The
calculated core loss is blank core loss without interlocking and deterioration by increasing
Figure 25. Dimension of ring core and dimensions of interlocking protuberances.
Figure 26. Comparison with protuberance types on fastened strength of the ring core (core material:
50SX400).
magnetic flux density near protuberance. Figure 28 shows the magnetic deterioration by
interlocking among several protuberance shapes on the ring cores made of 50SX400. Core
loss though laminations flux is estimated by the subtraction the calculated core loss from the
measured core loss that annealed for stress relieving. Basically low silicon containing steels
such as 50SX1300 is grown the grains and improved its magnetic properties by stress relief
annealing at 1023 K. Although 50SX400, which contains 2.0 mass % silicon, is hardly
growth grain and less improved its core loss by stress relief annealing at 1023 K without any
204 Taisei Nakayama
stress, occasionally it occurred a suitable stress for grain growth at 1023 K annealing (FC). In
that case, measured core loss is lower than the calculated one by improving its blank core loss
after annealed.
Figure 27. Magnetic flux model in the ring core.
Figure 28. Comparison with calculated and measured core loss deteriorated by magnetic flux flow
obstruction by interlocking protuberances of the ring core (core material: 50SX400).
Figure 29 shows the core loss comparison among 4 type protuberances on 50SX1300.
This steel is grown the grain easily and is improved core loss by stress relief annealing at
1023 K. Figure 30 shows the near protuberance microstructures before and after annealing.
Measured core loss is less than calculated one on 3 type protuberances. This type of low
silicon steel is unsuitable to investigate the deterioration effects.
3.1
Core loss W10/50 (W/kg)

50SX1300
3.0
2.9
2.8
2.7
2.6
2.5
FC VC FR VR Blank
Type of interlocking
Measured core loss with interlocking

Calculated deterioration by interlocking
Measured core loss without interlocking (blank)
Figure 29. Comparison with calculated and measured core loss deteriorated by magnetic flux flow
obstruction by interlocking protuberances of the ring core (core material: 50SX1300).
Figure 30. Cross-sectional microstructures of protuberances (A. before annealing; B. after stress relief
annealing at 1023 K in nitrogen atmosphere for 2 hours, both core materials: 50SX1300).
V-cut interlocking protuberances are rammed into next two laminations, while flat
bottoms are rammed into a half thickness of the next lamination. However, the core loss
deterioration through lamination is hardly separated, due to unexpected grain growth near
protuberances. This result indicates that the core loss increase by interlocking through
laminations depends on not the overlap area ratio of the protuberance, but the cross-sectional
obstacle by the interlocking protuberances and the stress.
206 Taisei Nakayama
2.2. Magnetic Properties Deterioration by Compressive Elastic Stress [27]
The compressive elastic stress deteriorates magnetic properties such as core loss or
magnetic induction [47,48]. On manufacturing electrical motor, the core fits into the case by
heat-shrunk. This shrink fitting introduces the compressive elastic stress into motor cores. For
instance, industrial motors are fit by heat-shrunk into the die-casting frame. Or compressor
motors for air conditioners are fit into the compressor shells. Those two cases are usually
faced on the deterioration from the uncased motor to the cased motor as a final product.
To make this effect clear, single sheet magnetization test under several compression
forces investigated. Figure 31 shows change of magnetic properties by compressive stress
using 35SX230 as a core material. The magnetic induction deceases as increasing the
compressive stress in the range of the lower magnetic force under 1500A/m. In the range over
1500 A/m, the affection by compressive force to the magnetic properties diminish.
Figure 31. Effect of compressive elastic stress on the magnetic properties of 35SX230 at 400 Hz
alternative current.
The influence on the core loss is shown in Figures 32 and 33. The core losses are
separated to the hysteresis loss and the eddy current loss. The core loss deterioration by
compressive stress is mainly on the hysteresis loss in the range of the magnetic induction
lower than 1.2T and is mainly on the eddy current loss in the range of that over 1.2 T. This
means compressive stress deteriorates not only the hysteresis but also the eddy current loss.
This range of magnetic induction is most important range in operating the motors. The
magnetic induction of the stator core yoke usually designed around 1.2-1.5 T, where is the
most deteriorated range by the compressive stress. This eddy current loss increase derives
from the domain structure change illustrated in Figure 34. Compressive stress generates the
90-degree domains, which increase the core loss [50].
Figure 32. Compressive stress influence on the hysteresis loss of 35SX230 at 400 Hz alternative
current.
Figure 33. Compressive stress influence on the eddy current loss of 35SX230 at 400 Hz alternative
current.
To reduce this deterioration the core shape especially stator core are designed by the
calculations of stress analyses. Some stator cores have outer spins to intense the compression
stress out of the main magnetic flux flows in the yoke. Other cores have widened the yoke
where the stress is focused. Those straggle means how the compression stress makes worse to
the motor efficiency.
208 Taisei Nakayama
Figure 34. Domain structure under conditions of compressive stress and a magnetic field (a)without
stress (b) with stress; (c) and (d) in a magnetic field.
2.3. Excellent Productivity Silicon Steel [29,30,32]
On manufacturing the motor core, productivity is very important factor for both
production cost and product quality.
Motor cores are stamped from the rolled hoops or cut sheets continuously at high speed
around 200 – 1500 strokes per minutes (s/min). Thus, stampability of the non-oriented
electrical steels sheets is important factor for building the motors [35].
By controlling the chemistry, especially sulfur content, stampability of the steels is
improved as shown in Figure 35. Regular steel 50SX1300 has 0.003 mass % carbon, 0.1 mass
% silicon, 0.2 mass % manganese, and 0.08 mass % phosphorus with 0.005 mass % sulfur,
while improved steel 50SXK1300 has 0.003 mass % carbon, 0.1 mass % silicon, 0.2 mass %
manganese and 0.08 mass % phosphorus with 0.018 mass % sulfur. Figure 36 shows the SEM
images of cut edge The improved steel has wider ductile fracture than the regular steel. This
means knifes on the stamping dies touched less time and areas with the steels Therefore, the
burr height / numbers of stamping as die-worn rate of the improved steel is less than that of
the regular steel.
Figure 35. Stampability comparison between a regular steel 50SX1300 and the improved steel
50SXK1300.
Figure 36. SEM images of the cut edge (A: regular steel 50SX1300, B: improved steel 50SXK1300).
This less sheared area of the improved steel leads to the less magnetic deterioration by
the stamping stress. Figure 37 shows the magnetic deterioration by the shearing stress.
210 Taisei Nakayama
Introducing the shearing stress as a half cut or 3-pieces or 6-pieces cut to the improved steel
magnetically less deteriorates than that to the regular one. Same as the results in Figure 36,
Baudouin et al. [51] studied the relationship between the knife clearance and coercive force.
The coercive force decreases as increasing the knife clearance, and the clearance increase
makes the fracture increase. These deterioration occur mainly by the sheared zone stress,
therefore, improved steel as 50SXK1300 has less deteriorations than regular steel 50SX1300.
Figure 37. Deterioration of the magnetic induction at 300 A/m (B3) by the sheared stress. (A: regular
steel 50SX1300, B: improved steel 50SXK1300).
Same as stampability, machinability is another important factor. Small motors with high-
speed rotation require the balancing of the rotor. This balancing is usually done by curving of
the rotor core surface. Figure 38 shows the bit worn rate comparison with regular steel
(50SX1000) and improved one (50SXK1000). Same as high speed stamping die worn rate,
improved steel bit worn rate is less than that of regular one.
The other characteristic of the improved steels is size deviations on stamping. For motor
core stamping, inner diameter of the stator core and outer diameter of the rotor core is very
important. The core size after stamping affects the air gap of the motor, which leads to the
motor efficiency. Figures 39 and 40 show the stamped and interlocked ring core or disk
diameter deviations. Both interlocked ring core simulated as a stator and stamped disk
simulated as a rotor, reached the same result that the improved steel (50SXK700) is less
deviated from the exact circle than the regular steel (50SX700). These phenomena derive
from not only the mechanical properties such as elongation on the tensile test, but deviations
of the area ratio between sheared area and ductile fracture (S/F). S/F ratio of the regular steel
more deviated form the rolling direction than that of improved steel. Figure 41 shows the
inclusions as manganese sulfides in the regular steel and improved steel. The manganese
sulfides in the regular steel are liner shapes, which elongated to the rolling direction, while
that of improved steels semispherical or rectangular shapes. These non-liner inclusions affect
the mechanical properties and S/F ratio deviations from the rolling direction.
Figure 38. Bit worn rate comparison (A: regular steel 50SX1000, B: improved steel 50SXK1000).
Figure 39. Interlocked ring core inner diameter deviations from the real circle. (A: regular steel
50SX700, B: improved steel 50SXK700).
212 Taisei Nakayama
Figure 40. Stamped disk outer diameter deviations from the real circle. (A: regular steel 50SX700, B:
improved steel 50SXK700).
Figure 41. Manganese sulfides inclusions in regular steel(50SX1300) and improved steel(50SXK1300).
Conclusion
The effects of the additive elements as phosphorus, aluminum and manganese are
improve the magnetic properties.
In the case of regular silicon steel (P=0.01mass %) the magnetic induction is decreased as
increasing the cold rolled reduction for reducing core loss especially for eddy current loss,
while the magnetic induction of phosphorus bearing silicon steel (P=0.1mass %) is less
decreasing than that of regular silicon steel. The grains in the regular silicon steel are oriented
the magnetically unfavorable texture such as {111}<112> in the case of high cold roll
reduction (thin gauge) sheet, while that of phosphorus bearing steel are not.
The magnetic properties such as the core loss is not affected by the content of aluminum
or nitrogen in the range of AlN <0.0024 mass %. In the case of steels with AlN > 0.0024
mass % and Al < 0.1mass %, the core loss was deteriorated by small grains which pinned
down the grain boundaries by the ‘harmful AlN’ (< 0.5 μm in size), but not deteriorated the
magnetic induction, due to the little effect on the grain size by those precipitates. In the case
of steels with AlN > 0.0024 mass % and Al > 1.0 mass %, core loss was not deteriorated, but
observed the ‘harmless large AlN’ (> 1.0μm) in steels. MnSiN2 were formed in the steels with
excess nitrogen and deteriorated the core loss.
The magnetic properties as the core loss are affected by the manganese and sulfur
contents. The core loss of the steels increases with an increase of the sulfur in each steel with
different manganese content. Among the same sulfur range steels, the core loss decreases
with an increase of manganese. As increasing the sulfur content in each manganese level
steels, the magnetic induction B50 decreased in any manganese level steels. To compare the
steels processed in the condition of the SRT at 1273, 1423, and 1523 K, the core loss of the
steels processed SRT at 1273 K is the lowest in each steels, and grains in each steels are same
trend as decreasing in size as increasing the SRT. The deterioration of the core loss caused by
the small grains pinned down by the ‘fine MnS’ (ca. 0.1 μm).
The grain size decrease with an increase in vanadium content < 0.016 mass %, and grow
towards the lower vanadium content > 0.016 mass %. These phenomena caused by vanadium
carbonitride along the grain boundary as pinning precipitates. In steel containing 0.001 mass
% vanadium, no vanadium carbonitrides were observed, while large vanadium carbonitrides
were observed in steel 0.124 mass % vanadium content steel. No pinning effects were
observed in steels 0.001 mass % and 0.124 mass % containing vanadium. Magnetic properties
were affected the vanadium content. The highest core loss was observed in 0.016 mass %
vanadium containing steel, due to the smallest grain size after stress-relief annealing. The
lowest induction was observed in steels containing vanadium 0.016 mass %, due to the
smallest grain size after hot rolling, and magnetically unfavorable texture after stress-relief
annealing.
The grain size decrease with an increase in titanium content < 0.11 mass %, and this
phenomenon caused by titanium carbonitride along the grain boundary as pinning precipitates
in steel containing < 0.016 mass % titanium, and numerous (Fe,Ti)P were observed in steels
with titanium >0.016 mass % in grains and along the grain boundaries. Magnetic properties
were affected the titanium content. The core loss increased as increasing titanium content, and
induction B50 decreased as increasing the titanium content by the precipitates above.
The magnetic properties are not affected by the zirconium addition in the range of Zr
<0.01 mass %. In steels containing Zr 0.01-0.13 mass %, the core loss are increasing with an
increase of zirconium content by the numerous precipitates of Zr3Fe, however, there is less
effect on the magnetic induction, because of the no effect on the hot band grain size and
texture by those precipitates.
Electrical steels sheets are usually coated with organic, inorganic or organic and
inorganic mixture insulation on the surface. Stamping performance or die worn rate of the
electrical steels with organic material as the surface insulation is better than that without
214 Taisei Nakayama
organic insulations. Coating fragments pasted on the sheared area play a rubricating role at
interlocking and affect the fastened strength. As increasing the coating thickness, stamping
performance was improved due to the highly lubrication by thick organic layer. However this
lubrication cause the interlocking performance to reduce the fastened strength. Magnetic flux
is concentrated in the narrowed cross-sectional area beside protuberances. As increasing
magnetic flux density, the core loss is increased at the narrowed point.
The core loss deterioration by compressive stress is mainly on the hysteresis loss in the
range of the magnetic induction lower than 1.2T and is mainly on the eddy current loss in the
range of that over 1.2 T. This means compressive stress deteriorates not only the hysteresis
but also the eddy current loss. The anonymous eddy current loss increase derives from the
domain structure change by compressive stress, which generates the 90-degree domains.
Stampability of the steels is improved by controlling the sulfur content. SEM images of
sheared cross section show that the improved steel has wider ductile fracture than the regular
steel. The improved steel magnetically less deteriorates by sheared stress and less bit worn
rate than that to the regular one. Both interlocked ring core simulated as a stator and stamped
disk simulated as a rotor, reached the same result that the improved steel is less deviated from
the exact circle than the regular steel.
References
[1] Enokizono, M.; Tanabe, I.; Takahashi, K.; Fukunaga, H.; Yoshida Y.; Kubota, T. J.
Magnetism Magnetic Materials 1999, 196-197, 335-337.
[2] Salsgiver,J.A. Physica Scripta 1989, 39, 775-779
[3] Yashiki ,H ; Kaneko ,H J. Appl. Phys. 1993, 73, 6606-6608.
[4] Beckley, P. Electrical steels; European Electrical Steels: New Port South Wales, 2000;
pp1-307
[5] Hayakawa,Y.; Szpunar, J.A. Acta Mater. 1997, 45, 1285-1295.
[6] Oyama,K; J. Inst. Electrical Engineers Jpn Trans 1A 2003, 123, 63-66
[7] Yaguchi, H.; Sasaki, S; SAE 2005,01-0271, 11-15
[8] Kabasawa, A; Tahakashi, H; SAE 2005, 01-0276,49-54
[9] Hamada, M; Nikkei Monozukuri 2006, 8, 111-115.
[10] Lorenzo,J.B.; Ros-Yanes, T; De Wulf, M; Houbaert Y; IEEE Trans Mag. 2004, 40,
2739-2741.
[11] Chong, S.W.; Hilinski, E.J.; Rolett, A.D.; Metallugical Mater. Trans. A 2003, 34A,
1321-1327.
[12] Dzubinsky,M; Sidor, Y; Kovac,F; Mater. Sci. Eng. A 2004, 385, 449-454.
[13] Moseley, D.; Hu, Y.; Randle, V.; Irons, T; Mater. Sci. Eng. A 2005 , 392, 282-291.
[14] Park, J-T.; Szpunar J.A; Cha, S-Y.; ISIJ International 2003, 43, 1611-1614.
[15] de Campos, M.F.; Landgraf, F.J.G.; Tahanohashi, R.; Chagas, F.C.; Filleiros,I.G.S.;
Fronzaglia, G.C.; Kahn, H.; ISIJ International 2004, 44, 591-597.
[16] Nakayama, T; in Recent Research Development in Material Science ; Pandalai S.G.;Ed.,
Research Signpost, Trivandrum, India, 2002; Vol.3-Part II, pp.811-824.
[17] Lindenmo, M; J. Magnetism Magnetic Materials, 2006, 304, 178-182.
[18] Kubota, T.; Steel Research Int. 2005, 76, 464-470.
[19] Nakayama, T., Honjou, N., Minaga, T. and Yashiki, H. J. Mag. Mag. Mater. 2001, 234,
55-61.
[20] Nakayama, T. and Honjou, N. J. Mag. Mag. Mater.:2000,213, 87-94.
[21] Tanaka, I.; Nitomi, Y.; Yashiki, H.; Kojima, H.; Nakayama,T.; Materia Japan 2006, 45,
225-227
[22] Tanaka, I.;Yashiki, Y.; ISIJ International 2007, 47, 1666-1671
[23] Nakayama, T. ; Takahashi, M.;J. Mater. Sci. 1995,30, 5979-5984
[24] Nakayama, T. ; Tanaka, T.; J. Mater. Sci. ,;1997, 32, 1055-1059.
[25] Nakayama, T. ; Honjou, N.; J. Mater. Engineering Performance, 2000, 9, 552-556
[26] Nakayama,T. ; Kojima, H.; J. Mater. Engineering Performance, 2007,16, 7-11.
[27] Fujimura, H.; Yashiki, H.; Kojima, H.; Nakayama, T.; Paper Tech. Meeting Magnetics
IEE Jpn. 2003, MAG-03-190,9-14.
[28] Fujimura, H.; Yashiki, H.; Kojima, H.; Nakayama, T.; Paper Tech. Meeting Magnetics
IEE Jpn. 2003, MAG-05-103,1-6.
[29] Nakayama, T; Honjou, N.; Yashiki, H.; Paper Tech. Meeting Magnetics IEE Jpn. 1998,
MAG-98-43,23-26
[30] Yashiki, H.; Doi, M.; Ishikawa, H; Honjou, N.; Nakayama, T. Materia Jpn. 1999,
38,169-171.
[31] Nakayama, T. Honjou, N.; Nagai, A; Yashiki, H.; Sumitomo Kinzoku, 1996, 48, 39-44.
[32] Nakayama, T; Honjou, N; Sumitomo Kinzoku, 1998, 50, 58-63.
[33] Yensen, T.D.; Metals Alloys 1930, 1,493.
[34] Iguchi, N.; Sakanishi, S.; Werkstatt Betribe 1987, 120, 323-328
[35] EL-Refaire, A.M.; Manzke, R.; Jahns.M.; IEEE Trans. Industrial Application 2004,
40,717-725.
[36] Hou, C-K.; Hu,C-T. ; Lee, S.; IEEE Trans Mag. 1991, MAG-27, 4305-4309
[37] Yashiki, H.; Okamoto, A. ; IEEE Trans. Mag. 1987,MAG-23, 3086-3088.
[38] Sawamura, H.; Mori, T.; Tetsu-to-hagane, 1955, 41,1082.
[39] Darken, L.S.; Smith, R.P.; Filer, E.W.; J Metals Trans AIME 1951,3, 1174-1179..
[40] Yashiki, H.; Kaneko, T.; Tetsu-toHagane, 1994, 80,79-83.
[41] Matsumura, K.; Fukuda, B. ; IEEE Trans. Mag. 1984, MAG-20, 1553-1538.
[42] Shimoyama, Y.; Miyishi, K.; Tanino. M.; Wada, T.; IEEE Trans. Mag. 1983, MAG-19,
2013-2015.
[43] Stephenson, E.T.; Marder, A.R.; 1986, IEEE Trans. Mag. 1986, MAG-22, 101-106.
[44] Vanderschueren, J.; in Physical metallurgy of IF steel, ISIJ, Tokyo, 1994,pp.145-148
[45] Park, Y.B.; Kang, H.J.; Chang, S.K.; in Physical metallurgy of IF steel, ISIJ, Tokyo,
1994,pp.245-248
[46] Brun, C.; Patou, P.; Parniere, P.; in 'Metallurgy of continuous annealed sheet steel',
Bramfitt, B.L. and Mangonon, P.L.; Eds., TMS-AIME, Warrendale, PA ,1982, pp173-
197.
[47] Jeong, W.-C.; Chung, J.-H.; Tetsu-to-Hagane, 1989, 75 , 74-81
[48] Nakata, T; Nakano,M.; Kawahara, K.; IEEE Trans. Mag. 1992, MAG-7, 453-457.
[49] LoBue, M; Sasso, C; Bosso, V Fiorillo, F.; Bertotti ,S; , J. Mag. Mag. Mater. 2000, 215-
216.
[50] Pry, R.H.; Bean, C.P; J. Appl. Phys. 1958, 29, 532-533.
[51] Baudouin, P; DeWulf, M.; Kestens, L.; Houbaert, Y. , J. Mag. Mag. Mater. 2003, 32-40.
Chapter 7
INFLUENCE OF LUTING CEMENT APPLICATION

TECHNIQUE ON QUARTZ FIBER POST REGIONAL
BOND STRENGTHS
Camillo D’Arcangeloa,*, Francesco De Angelisa,

Maurizio D’Amariob, Simone Zazzeronia, Mirco Vadinia
and Sergio Caputia
a
Department of Restorative Dentistry, School of Dentistry, University G. D’Annunzio -
Chieti, Italy
b
Department of Restorative Dentistry, Dental Clinic, University of L’Aquila, Italy
Abstract
The aim of this study was to investigate regional root canal push-out bond strengths for a
fiber-reinforced post system varying the application method of the luting agent.
Recently extracted maxillary incisors (n=30) were sectioned transversally at the labial
cemento-enamel junction, and the roots treated endodontically. Following post space
preparations, fiber-reinforced posts (Endo Light-Post; RTD) were placed using adhesive
system and resin cement provided by the manufacturer. Three equal groups (n=10) were
assessed according to the technique used to place the luting agent into post space: using a
lentulo spiral, applying the cement onto the post surface, injecting the material with a specific
syringe. Each root was sliced into three discs (2 mm thick) representing the coronal, middle
and apical part of the bonded fiber post. Push-out tests were performed for each specimen to
measure regional bond strengths. Results were statistically analyzed using two-way ANOVA
and Tukey tests (α = 0.05). All fractured specimens were observed using a scanning electron
microscope to identify the types of failure.
The results indicated that bond strength values were significantly affected by the
application method of the resin cement (p < 0.05). The "syringe technique" and the "lentulo
technique" showed higher bond strength values compared with the "post technique". No
*
E-mail address: [email protected]. Phone: +39(0)85.4549652. Fax: +39(0)85.4541279. Corresponding author:
Prof. Camillo D’Arcangelo, Department of Restorative Dentistry, Dental School, University “G.
D’Annunzio”, Via dei Vestini 31, 66100, Chieti, Italy.
218 Camillo D’Arcangelo, Francesco De Angelis, Maurizio D’Amario et al.
significant differences were recorded among the post space thirds. Microscopic analysis
revealed a prevalence of post/cement and mixed failures.
The best performance in terms of push-out bond strengths for the post system tested was
obtained when the luting agent was applied into the post space either with a specific syringe or
using a lentulo spiral. There were not differences in bond strength among root thirds.
Introduction
The increasing popularity and widespread use of fiber-reinforced (FRC) posts is changing
the restorative procedures for endodontically treated teeth. Fibre-reinforced composite posts
are commonly used for the restoration of endodontically treated teeth with reduced crown
structure [1]. Since fiber-reinforced posts (18-22 GPa) have a modulus of elasticity (E)
similar to the dentin (18 Gpa) [2], they produce a stress field similar to that of natural teeth,
thereby reducing the risk of root fractures [3]. FRC posts contain a high percentage of
continuous reinforcing fibers embedded in a polymer matrix. Matrix polymers are commonly
epoxy resins or other polymers with a high degree of conversion and a highly cross-linked
structure [4],[5]. The benefits of adhesive techniques used for dental restoration are well
documented, so the use of adhesive resin cements has been proposed for cementing
endodontic fibre posts in non-vital teeth. FRC post and resin cement have similar moduli of
elasticity to dentine enabling loading forces to be transferred consistently from the restoration
to tooth structure [6]. The loss of bond at the fibre post/resin cement/root dentine interfaces
still represents the main reason for which endodontically treated teeth reconstructed with fibre
posts show clinical failures [7]-[9]. Retention of fibre-reinforced composite (FRC) posts
within root canals is affected by several factors: type of post, its adaptation into the post
space, type of adhesive and operative procedures [10]-[14]. The distribution of resin cement
into the post space during the luting procedure and the anatomical and histological
characteristics of the root dentine seemed to influence bond strength between resin luting
agent and root canal regions [15]. An adequate polymerisation of luting agent is necessary to
provide its mechanical properties, that clinically ensure post retention. Many current resin
luting agents polymerise through a dual-curing process that requires light exposure to initiate
the reaction. It has been reported that the mechanical properties of dual-cure type resin agents
appear improved after photo-activation compared with chemical-activation alone [16]. Dual-
cure resin cements are different in their handling characteristics, compositions and properties
(such as polymerisation ability, flexural strength, hardness). The quality of adhesion to root
dentine is also affected by the density and orientation of dentine tubules at different levels of
the root canal walls [17] and the accessibility of the coronal, middle and apical third of the
root during handling of the materials [18]. It has been demonstrated that the control of
moisture after the application and removal of phosphoric acid as well as incomplete
infiltration of the resin into the dentine significantly affect bond strengths [19].
A resin luting agent may create polymerization shrinkage stresses within the post space.
Shrinkage stresses of luting materials in root canals are especially relevant due to the
unfavourable factor of configuration (C-Factor) that restricts the flow of resin cement, which
may affect the integrity of the adhesive interface at different root levels [8]. The contraction
stress of resin cements in confined spaces depends upon the thickness of the cement layer
[20], but this thickness changes with different root canal morphologies. Moreover, the post
can be closer to the dentine on one side, which also influences cement behaviour [21] and
Influence of Luting Cement Application Technique… 219
explains the importance of inserting posts with identical pressure to achieve a standard
cement thickness. Although the chemical-physical properties of resin cements have been
evaluated, little information is available on the role of the application methods of dual-cure
resin cements to the post space and their effect on regional bond strength of fiber post.
Aim of this study was to evaluate the bond strengths of quartz-fiber posts to coronal,
middle and apical thirds of post space dentine varying the application method of the luting
agent.
Materials and Methods

Specimen Preparation
Thirty freshly extracted human maxillary incisors were selected. External debris was
removed (Suprasson P-max; Satelec/Acteon Equipment, Merignac, France). Selected
specimens were stored in 0.5% chloramine T aqueous solution at 4°C. Crown surfaces of
each tooth were sectioned at the labial cemento-enamel junction (CEJ) using a cylindrical
diamond rotary cutting instrument (Intensiv 314, Ø ISO 014, L.8.0 mm; Intensiv, Grancia,
Switzerland) mounted on a high-speed hand-piece (Bora L; Bien-Air, Bienne, Switzerland)
with water-spray cooling.
Root canals were mechanically enlarged to ISO size 25, 0.06 taper (MTwo; VDW
GmbH, Munich, Germany). Irrigants used were 5% sodium hypochlorite (Ogna, Muggiò,
Milan, Italy) and 17% EDTA (Pulpdent, Watertown, MA). Enlarged canals were rinsed with
distilled water, dried with paper points (Roeko, Langenau, Germany) and sealed with gutta-
percha (Lexicon Gutta Percha Points; Dentsply Tulsa Dental, Tulsa, Okla) using the System-
B HeatSource (Analytic Technology, Redwood City, CA) and endodontic sealer (Pulp Canal
Sealer EWT; Kerr, Romulus, MI). Backfilling was performed with Obtura II (Spartan,
Fenton, MO).
Bonding of Fiber Posts
After 24 hours, gutta-percha was removed with warm endodontic pluggers (Sybron
Dental Specialties, Romulus, MI). Post spaces were prepared to a depth of 10 mm measured
from the sectioned surfaces using Torpan drills ISO 100 Yellow (batch no. 042190611)
provided by the manufacturer (RTD, St. Egrève, France). Post space preparations were rinsed
with 5% NaOCl. A final irrigation was accomplished with distilled water, and post spaces
were dried with paper points. Before cementation procedures, each post was marked at a
distance of 10 mm from the apical end corresponding to the length of the post space
preparation and sectioned horizontally with a water-cooled diamond rotary cutting instrument
(R879.014; Diaswiss, Geneva, Switzerland). In this way, the complete seating of the posts has
been verified. The root canal walls were etched for 60 s with 36% phosphoric acid,
Conditioner 36, (batch no. 0507002142; Dentsply DeTrey, Konstanz, Germany), introduced
into the spaces with a needle, rinsed using a water syringe and then gently dried with paper
points. Bonding procedures were performed following the instructions provided by the
manufacturers. XP Bond (batch no. 065001399; Dentsply DeTrey) and SelfCure Activator
(batch no. 0510061; Dentsply DeTrey) were mixed for 2 s and applied to the root canal for 30
s with a microbrush (Microbrush X; Microbrush Corp, Grafton, WI). After 20 s, the excessive
adhesive solution was removed with a paper point and then gently air-dried for 5 s.
The specimens were randomly divided into 3 groups (n=10) according to the technique
used to place the luting agent (FluoroCore 2, batch no. 0610021; Dentsply DeTrey) into root
canal: using a #30 lentulo spiral instrument (Dentsply Maillefer, Ballaigues, Switzerland) for
4 seconds before setting the post; applying the cement onto the post surface; injecting the
material with a tube with needle and the appropriate plug (KerrHawe SA, Bioggio,
Switzerland) using a specific Composite-Gun (KerrHawe SA). The posts used were Endo
Light-Posts, size 2, 2% taper, with a maximum cross-section diameter of 1.36 mm and a
minimum diameter at radicular end of 1.0 mm (batch no. 049520702; RTD). Endo Light-
Posts are made of unidirectional pre-tensed quartz fibres (60% volume) embedded in an
epoxy resin matrix.
The posts were seated to full depth in the prepared spaces using finger pressure. Excess
of luting agent was immediately removed with a small brush. A constant axial load of 5 kg
was applied for 60 seconds to stabilize the fiber posts in the post spaces. After the initial
chemical polymerization, the resin luting agents were light polymerized (L.E. Demetron I,
Sybron/Kerr, Orange, CA, with a 1200 mW/cm2 output) for 40 seconds. Thirty minutes after
the cementation procedures, all root specimens were stored in distilled water for 24 hours.
Then, specimens underwent 10,000 thermal cycles between 5°C and 55°C, with a 30-second
dwell time and a 5 second transfer between temperature baths. Specimens were also subjected
to 300,000 cycles of mechanical loading parallel to the long axis of the post with a
masticatory simulator (Willytec, Munich, Germany) at 30 N force and 1.6 Hz. The
mechanical loading pattern was equivalent to 1 year of clinical function [22],[23]. Specimens
were then preserved in a saline solution at room temperature for 1 week.
Push-out Testing
Specimens were then fixed to phenolic ring forms filled with a autopolymerizing acrylic
resin (Technovit 4000; Heraeus Kulzer, Wehrheim, Germany). Specimens were attached to
the arm of a low-speed saw (Micromet M; Remet S.p.A., Casalecchio di Reno, Italy) and
sectioned perpendicular to the long axis under water cooling. Three slices per each root
(Figures 1, 2), containing cross sections of coronal, middle and apical part of the bonded fiber
post, were obtained by sectioning the root under distilled water coolant. The sections were 2.0
± 0.1 mm thick. Each slice was marked on its apical side with an indelible marker. The
thickness of each specimen was measured and recorded by a digital caliper with an accuracy
of 0.001 mm (Figure 3). The sections were stored individually in black film canisters with
sterile water. Push-out-test was performed by applying a compressive load to the apical aspect
of each slice via a cylindrical plunger mounted on a Universal Testing Machine (Lloyd LR
30K; Lloyd Instruments Ltd., Fareham, UK) managed by PC software (Nexygen-Ondio
Version 4.0; Lloyd Instruments Ltd.). With regard to the tapered design of the post, three
different sizes of punch pins were used for the push-out testing. The diameter of the punch
pin was 1.2 mm for the coronal slices, 1.0 mm for the middle slices and 0.8 for the apical
slices [24]. Punch pin was positioned to contact only the post, without stressing the
surrounding root canal walls [25] (Figure 4). Care was also taken to ensure that the contact
between the punch tip and the post section occurred over the most extended area, to avoid
Figure 1. Root section performed by the Micromet M machine. The numbers express (millimeters) the
distance from the arm of the machine and, thus, from the cervical end of the root. The diamond saw is
0.6 mm thick. The sections directed to the push-out-test are 2 mm thick.
Figure 2. Coronal, middle and apical slices obtained from the same root and addressed to the push-out
test.
Figure 3. The thickness of each sample was measured using a digital calliper with an accuracy of 0.001
mm.
Figure 4. Test specimen mounted on the Universal Testing Machine for push-out bond strength test.
notching of the punch tip into the post surface. The load was applied to the apical aspect of
the root slice and in an apical-coronal direction, so as to push the post toward the larger part
of the root slice, thus avoiding any limitation to the post movement. Loading was performed
at a crosshead speed of 0.5 mm/min until the post segment was dislodged from the root slide
[14]. Maximum failure load values was recorded (N) and converted into MPa, considering the
bonding area (mm2) of the post segments. Post diameters were measured on each surface of
the post/dentine sections using the digital caliper and the total bonding area for each post
segment was calculated using the formula: π(R+r)[(h2+(R-r)2]0.5, where π = 3.14, R represents
the coronal post radius (mm), r the apical post radius (mm), and h the thickness of the slice
(mm).
Figure 5. SEM micrograph showing the cervical root portion of a sample from the POST group
(original magnification 100x). The arrow indicates a bubble embedded in the resin cement.
All fractured specimens were carefully removed and observed under stereomicroscope
(Zeiss MC 80 DX; Zeiss, Jena, Germany) at 20X and 50X magnification from the coronal as
well as from the apical direction to determine, for each root third, the mode of failure, which
were classified into 5 types [26]: (1) adhesive between post and resin cement (no cement
visible around the post); (2) mixed, with resin cement covering 0-50% of the post diameter;
(3) mixed, with resin cement covering 50-100% of post surface; (4) adhesive between resin
cement and root canal (post enveloped by resin cement); (5) cohesive in dentine.
Furthermore, representative specimens from each group were analysed (Figures 5, 6, 7) using
a scanning electron microscope (LEO 435 vp; LEO Electron Microscopy Ltd, Cambridge,
UK).
Figure 6. SEM view of interface between dentin (left side) and resin cement (right side) in cervical root
segment of a specimen from the LENTULO group (original magnification 1000x).
Figure 7. SEM view of interface between dentin (upper side) and resin cement (lower side) in cervical
root segment of a specimen from the POST group (original magnification 1500x).
Results
Push-out test results are shown in Table 1. Statistical analysis displayed that the
application method of the resin cement significantly affected the bond strength values (p <
0.05). No significant differences were recorded among the section levels (root thirds) (p >
0.05). The interaction between these two factors was not significant (p > 0.05). For the resin
cement application method factor, the “syringe technique” showed the highest bond strength
(13.51 ± 3.11 MPa) compared with the others methods. The “lentulo technique” exhibited
high retentive strength (11.49 ± 2.33 MPa), while the “post technique” revealed significantly
lower bond strength value (7.88 ± 2.08 MPa). For the section level factor, the coronal third
showed the highest retentive strength (11.87 ± 3.21 MPa) but no statistically significant
differences were found with the middle (10.79 ± 3.72 MPa) and the apical third (10.24 ± 3.27
MPa). Regarding failure types of tested specimens, the majority failed adhesively with a
prevalence of post/cement interface and mixed failures (Figure 8). There were no cohesive
failures in dentine (Table 2). The fracture pattern observed was very similar among the
specimens of the tested groups.
Table 1. Mean push-out bond strengths (MPa) and Standard deviation for experimental
groups according to the root thirds
RTD ENDO LIGHT POST

TOTAL
LENTULO POST SYRINGE
CORONAL 12.62 (1.63) 8.70 ( 0.39) 14.28 ( 3.50) 11.871 ( 3.21)
MIDDLE 11.33 ( 3.09) 8.04 ( 2.95) 12.99 ( 3.50) 10.791 ( 3.72)
APICAL 10.53 ( 1.68) 6.91 ( 1.81) 13.28 ( 2.39) 10.241 ( 3.27)
TOTAL 11.49b ( 2.33) 7.88c ( 2.08) 13.51a ( 3.11)
Same numbers in pedex indicate not significant differences among the levels of the factor section-level.
Different lower case letters represent significant differences with regard to the factor luting technique.
Figure 8. SEM view of interface between dentin (lower side) and resin cement (upper side) in cervical
root segment of a specimen from the SYRINGE group (original magnification 1000x).
Table 2. Failure mode for experimental groups
1 2 3 4
5
Groups Adesive : Mixed : Mixed : Adhesive:
Cohesive
post-cement 0-50% 50-100% cement- dentin
LENTULO
Coronal 4 2 3 1 0
Middle 3 3 3 1 0
Apical 4 3 2 1 0
POST
Coronal 4 2 3 1 0
Middle 3 4 3 0 0
Apical 2 3 4 1 0
SYRINGE
Coronal 3 4 2 1 0
Middle 4 2 4 0 0
Apical 4 2 3 1 0
Discussion
This investigation was performed to evaluate if the application methods of the luting
agent can influence bond strengths of a quartz post system to root canal dentin, using a push-
out model. Push-out tests result in a shear stress at the interface between dentine and cement
as well as between post and cement [27] and is comparable with the stress under clinical
conditions. The push-out design is characterised by polymerisation stresses that would
happen in the clinical situation. It has been suggested that, due to the small size of specimens,
microtensile test permits a uniform stress distribution along the bonded interface [28].
Nevertheless, as observed previously [25], push-out test is a more reliable method for
determining bond strengths between fiber posts and post space dentine because of the high
number of premature failures occurring during specimen preparation and large data
distribution spread associated with microtensile testing.
The effect of different resin-based luting agents on post retention has been investigated
extensively, and various conclusions have been drawn [29],[30]. Regarding the adhesive/resin
cement/fiber post system tested, the retentive strength was significantly affected by the luting
technique. Irrespective of post luting technique, the interfacial strength was not significantly
affected by the region of the root canal. Various in vitro researches revealed controversial
results concerning bond strength values of different luting agents to FRC posts and root canal
dentin [24],[25],[26]. Shear bond strengths depend on the degree and stability of interfacial
micromechanical interlocking and chemical adhesion between root canal dentin, dentin
bonding agent/resin-based luting cement and fibre post. Recent studies highlighted that
retention of bonded fiber posts was contributed predominantly by friction [31],[32].
Bond strengths in the present study were not significantly affected by root canal thirds.
This result confirms two previous studies [25],[33] that observed no influence of root canal
region on fibre post retention. Gaston et al. [34] recording no significant differences in
microtensile bond strength values between coronal and middle thirds of the post space,
concluded that retentive strength may be related more to the area of solid dentine than to the
density of the dentinal tubules. In contrast, Perdigao et al. [35] found that the coronal third
resulted in statistically higher bond strengths than the apical third, while the middle and the
apical thirds had no statistically significant different bond strengths. For the authors, the
lower bond strength values found in the apical zone could be expected due to more difficult
access to this third and the possible limitations of cement flow, but also due to a high and
unfavourable C-factor. In a previous study [36], the most likely explanation for the higher
resistance to post dislodgement in the coronal region of the root canal was identified in the
decreasing effectiveness of light curing at greater distances from the light source. Moreover,
the coronal portion of the canal seems to be the most accessible part of the canal space,
making it easier to etch and more thoroughly apply the adhesive agents. Rinsing with water
during the etching procedure, the difficulties of moisture control in the apical third of the post
space probably result in the retention of remnant water within the dentine tubules, causing an
incomplete infiltration of the resin agent. A reduction in strength in middle and apical thirds
was also be related to the more difficult distribution of resin cement with voids formation
[8],[37] or to traces of gutta-percha and endodontic sealer that may remain in these thirds
after post space preparation. Bouillaguet et al. [8] reported that when endodontic posts are
cemented inside root canals, the C-factor may exceed 200 (ratio of the bonded to the
unbonded area). This is because there is a large area of resin cement bonded to the dental
substrate and endodontic post, and there is little free area to allow for polymerization
contraction. These findings seem to suggest that lack of direct viewing and luting agent
application techniques may affect the bond strength in the apical region of the post space
which will be inevitably lower.
Despite the promising results of dual-cure resin luting agents, existing literature regarding
the role of the application methods of resin cements to the root canal and their effect on
retentive strength of fiber post is scarce. It was reported that the application of luting agent
with a lentulo spiral instrument permits a favourable distribution of resin cement throughout
the post space and a formation of uniform, continuous cement layer [38]. Moreover this
technique may guarantee the reduction of voids and bubbles within the luting agent (Figure
9). In this situation, if a dual-cure resin was the selected cementing material, the major
recommendation is to avoid partial polymerisation before the adequate post seating. Even if it
was reported that the presence of some porosity in the luting agent is not per se
disadvantageous [11],some authors suggested that voids and air bubbles can impede an
appropriate cementation of the post, thus causing its debonding [40]. The injection technique
used for application of the resin cements is also reported as an effective technique for
reducing voids and bubbles within the luting agent [6]. Fonseca et al. [41] in vitro evaluated
the retention within root canals of posts cemented with dual-cure resin varying the application
method of the primer/adhesive solution and luting agent. They reported that when the luting
agent is placed into the root canal using only a lentulo spiral or when it is placed both using a
lentulo and applying the cement on the post surface, post retention is increased. In a recent
study [42], it was shown that bond strengths of three post systems in the apical post space
third seemed to be not affected by luting agent application techniques. However, further
researches were suggested to find out which luting protocol could be more suitable for each
individual clinical situation. The present study revealed that bond strength to root canal
dentine is significantly influenced by the application technique of luting agent for
adhesive/resin cement/fiber post system tested. Statistical analysis showed that “syringe
technique” exhibited the highest retentive strength to root canal dentin compared with the
others luting methods. The “post technique” resulted in statistically lower bond strength
value. The complete post system (post, adhesive, resin cement) supplied by the manufacturer
was tested. Since it was demonstrated that the resin cement thickness significantly influences
the pullout strengths of fibre-reinforced posts [13], the post spaces were prepared using the
appropriate drill from the respective post manufacturer. Further researches would be
necessary to clarify the influence of each component in the retentive values of the respective
system group.
Figure 9. Representative SEM micrograph of a mixed (type 3) failure between post and cement from
the apical slice of a LENTULO group specimen (Magnification 100x).
Analyses of failure mode demonstrated that most failures occurred at the cement-post
interface or in a mixed mode. Polymerization shrinkage stresses that were generated because
of the highly unfavourable cavity configuration factor (C-factor) of the post space probably
accounted for the relatively higher percentage of mixed failure [8]. This finding suggested
that the nature of the dentine surface of the canal wall or the tubule density might not be the
basis for the difference in bond strengths between the coronal and middle/apical regions,
which finding lends support to those investigations that aimed at improving the retention
through various surface pretreatment procedures for the post [23],[43],[44].
Conclusion
Based on these findings, and within the limitations of an in vitro study, it may be
concluded that, for the quarz fiber post system tested, the best performance was obtained
when the luting agent was taken into the post space with a specific syringe. Moreover, push-
out bond strengths were not statistically influenced by the root regions.
References
[1] D’Arcangelo, C; Prosperi, GD; Passariello, P; Caputi, S; and Malagnino, VA. Capacity
of coronal dentin to increase fiberglass post retention: a pull-out test. Am J Dent. 2005,
18, 307-310.
[2] Asmussen, E; Peutzfeldt, A; and Heitmann, T. Stiffness, elastic limit, and strength of
newer types of endodontic post. J Dent. 1999, 27, 275-278.
[3] Sirimai, S; Riis, DN; and Morgano, SM. An in vitro study of the fracture resistance and
the incidence of vertical root fracture of pulpless teeth restored with six post-and-core
systems. J Prosthet Dent. 1999, 81, 262-269.
[4] Terry, DA; Triolo, PT; and Swift, EJ. Fabrication of direct fiber-reinforced posts: a
structural design concept. J Esthet Restor Dent. 2001, 13, 228-240.
[5] Grandini, S; Goracci, C; Monticelli, F; Tay, FR; and Ferrari, M. Fatigue resistance and
structural characteristics of fiber posts: three-point bending test and SEM evaluation.
Dent Mater. 2005, 21, 75-82.
[6] Pest, LB; Cavalli, G; Bertani, P; and Gagliani, M. Adhesive post-endodontic restorations
with fiber posts: push-out tests and SEM observations. Dent Mater. 2002, 18, 596-602.
[7] Ferrari, M; Vichi, A; Mannocci, F; and Mason, PN. Retrospective study of the clinical
performance of fiber posts. Am J Dent. 2000, 13, 9B-13B.
[8] Bouillaguet, S; Troesch, S; Wataha, JC; Krejci, I; and Meyer, JM; and Pashley DH.
Microtensile bond strength between adhesive cements and root canal dentin. Dent
Mater. 2003, 19, 199-205.
[9] Mannocci, F; Bertelli, E; Watson, TF; and Ford, TP. Resin-dentin interfaces of
endodontically-treated restored teeth. Am J Dent. 2003, 16, 28-32.
[10] Peters, J; Zyman, G; Kogan, E; Kuttler, S; and García-Godoy, F. Retention of three
endodontic post systems. Am J Dent. 2007, 20, 198-200.
[11] Vichi, A; Grandini, S; Davidson, CL; and Ferrari, M. An SEM evaluation of several
adhesive systems used for bonding fiber posts under clinical conditions. Dent Mater.
2002, 18, 495-502.
[12] Vichi, A; Grandini, S; and Ferrari, M. Comparison between two clinical procedures for
bonding fiber posts into a root canal: a microscopic investigation. J Endod. 2002, 28,
355-360.
[13] D'Arcangelo, C; Cinelli, M; De Angelis, F; and D'Amario, M. The effect of resin cement
film thickness on the pullout strength of a fiber-reinforced post system. J Prosthet Dent,
2007, 98, 193-198.
[14] Vano, M; Cury, AH; Goracci, C; Chieffi, N; Gabriele, M; Tay, FR; and Ferrari M. The
effect of immediate versus delayed cementation on the retention of different types of
fiber post in canals obturated using a eugenol sealer. J Endod. 2006, 32, 882-885.
[15] Mannocci, F; Pilecki, P; Bertelli, E; W and atson, T. Density of dentinal tubules affects
the tensile strength of root dentin. Dent Mater. 2004, 20, 293-296.
[16] Hofmann, N; Papsthart, G; Hugo, B; and Klaiber, B. Comparison of photoactivation
versus chemical or dual-curing of resin-based luting cements regarding flexural strength,
modulus and surface hardness. J Oral Rehabil. 2001, 28, 1022-1028.
[17] Ferrari, M; Mannocci, F; Vichi, A; Cagidiaco, MC; and Mjör IA. Bonding to root canal:
structural characteristics of the substrate. Am J Dent. 2000, 13, 255-260.
[18] Ferrari, M; Vichi, A; Grandini, S. Efficacy of different adhesive techniques on bonding
to root canal walls: an SEM investigation. Dent Mater, 2001, 17:422-429.
[19] Tay, FR; Gwinnett, AJ; and Wei, SH. The overwet phenomenon: an optical,
micromorphological study of surface moisture in the acid-conditioned, resin-dentine
interface. Am J Dent. 1996, 9, 43-48.
[20] Alster, D; Feilzer, AJ; De Gee, AJ; and Davison, CL. Polymerization contraction stress
in thin resin composite layers as a function of layer thickness. Dent Mater. 1997, 13,
146-150.
[21] Ceballos, L; Garrido, MA; Fuentes, V; and Rodríguez, J. Mechanical characterization of
resin cements used for luting fiber posts by nanoindentation. Dent Mater. 2007, 23, 100-
105.
[22] Stegaroiu, R; Yamada, H; Kusakari, H; and Miyakawa, O. Retention and failure mode
after cyclic loading in two post and core systems. J Prosthet Dent. 1996, 75, 506-511.
[23] D’Arcangelo, C; D’Amario, M; Prosperi, GD; Cinelli, M; Giannoni, M; and Caputi S.
Effect of surface treatments on tensile bond strength and on morphology of quartz-fiber
posts. J Endod. 2007, 33, 264-267.
[24] Bitter, K; Meyer-Lueckel, H; Priehn, K; Kanjuparambil, JP; Neumann, K; Kielbassa,
AM. Effects of luting agent and thermocycling on bond strengths to root canal dentine.
Int Endod J, 2006, 39, 809-818.
[25] Goracci, C; Tavares, AU; Monticelli, F; Raffaelli, O; Cardoso, PC; Tay, F; et al. The
adhesion between fiber posts and root canal walls: comparison between microtensile and
push-out bond strengths measurements. Eur J Oral Sci. 2004, 112, 353-361.
[26] Perdigao, J; Gomes, G; and Lee, IK. The effect of silane on the bond strengths of fiber
posts. Dent Mater. 2006, 22, 752-758.
[27] Van Meerbeek, B; De Munck, J; Yoshida, Y; et al. Buonocore memorial lecture.
Adhesion to enamel and dentin: current status and future challenges. Oper Dent,
2003, 28, 215-235.
[28] Cardoso, PE; Sadek, FT; Goracci, C; and Ferrari, M. Adhesion testing with the
microtensile method: effects of dental substrate and adhesive system on bond strength
measurement. J Adhes Dent. 2002, 4, 291-297.
[29] Cohen, BI; Pagnillo, MK; Newman, I; Musikant, BL; and Deutsch, AS. Retention of
three endodontic post cemented with five dental cements. J Prosthet Dent. 1998, 79,
520-525.
[30] O’Keefe, KL; Miller, BH; and Powers, JM. In vitro tensile bond strength of adhesive
cements to new post materials. Int J Prosthodont. 2000, 13, 47-51.
[31] Goracci, C; Fabianelli, A; Sadek, FT; Papacchini, F; Tay, FR; and Ferrari M. The
contribution of friction to the dislocation resistance of bonded fiber posts. J Endod.
2005, 31, 608-612.
[32] Pirani, C; Chersoni, S; Foschi, F; Piana, G; Loushine, RJ; Tay, FR; and Prati, C. Does
hybridization of intraradicular dentin really improve fiber post retention in
endodontically treated teeth? J Endod. 2005, 31, 891-894.
[33] D'Arcangelo, C; D'Amario, M; Vadini, M; Zazzeroni, S; De Angelis, F; and Caputi, S.
An evaluation of luting agent application technique effect on fibre post retention. J Dent.
2008, 36, 235-240.
[34] Gaston, BA; West, LA; Liewehr, FR; Fernandes, C; Pashley, DH. Evaluation of regional
bond strength of resin cement to endodontic surfaces. J Endod. 2001, 27, 321-324.
[35] Perdigao, J; Geraldeli, S; and Lee, IK. Push-out bond strengths of tooth-colored posts
bonded with different adhesive systems. Am J Dent. 2004, 17, 422-426.
[36] D'Arcangelo, C; Zazzeroni, S; D'Amario, M; Vadini, M; De Angelis, F; Trubiani, O;
and Caputi, S. Bond strengths of three types of fibre-reinforced post systems in various
regions of root canals. Int Endod J. 2008, 41, 322-328.
[37] Mallmann, A; Jacques, LB; Valandro, LF; Mathias, P; and Muench, A. Microtensile
bond strength of light- and self-cured adhesive systems to intraradicular dentin using a
translucent fiber post. Oper Dent. 2005, 30, 500-506.
[38] Akgungor, G; and Akkayan, B. Influence of dentin bonding agents and polymerization
modes on the bond strength between translucent fiber posts and three dentin regions
within a post space. J Prosthet Dent. 2006, 95, 368-378.
[39] Alster, D; Feilzer, AJ; De Gee, AJ; Mol, A; and Davidson, CL. The dependence of
shrinkage stress reduction on porosity concentration in thin resin layers. J Dent Res,
1992, 71, 1619-1622.
[40] Ferrari, M; Vichi, A; Grandini, S; Goracci, C. Efficacy of a self-curing adhesive/resin
cement system on luting glass-fiber posts into root canals: An SEM investigation. Int J
Prosthod,.2001, 14, 543-549.
[41] Fonseca, TS; Alfredo, E; Vansan, LP; Silva, RG; Sousa, CS; Saquy, PC; and Sousa-
Neto, MD. Retention of radicular posts varying the application technique of the adhesive
system and luting agent. Braz Oral Res. 2006, 20, 347-352.
[42] D'Arcangelo, C; D'Amario, M; De Angelis, F; Zazzeroni, S; Vadini, M; and Caputi, S.
Effect of application technique of luting agent on the retention of three types of fiber-
reinforced post systems. J Endod. 2007, 33, 1378-1382.
[43] D'Arcangelo, C; D'Amario, M; Vadini, M; De Angelis, F; and Caputi, S. Influence of
surface treatments on the flexural properties of fiber posts. J Endod. 2007, 33, 864-867.
[44] Monticelli, F; Osorio, R; Albaladejo, A; Aquilera, FS; Tay, FR; Ferrari, M; and
Toledano, M. Effect of adhesive systems and surface treatment of methacrylate resin-
based fiber posts on post-resin-dentin bonds. Am J Dent. 2007, 20, 231-234.
Chapter 8
MICROSTRUCTURAL INFLUENCE ON FLEXURE

STRENGTH OF A CEROMER REINFORCED
BY TWO TYPES OF FIBERS
(POLYETHYLENE AND GLASS)
Silvana Marques Miranda Spyrides1,*

and Fernando Luiz Bastian2,**
*Department of Prosthesis and Dental Materials, School of Dentistry,
Federal University of Rio de Janeiro, Rio de Janeiro, RJ, Brazil
**Program of Metallurgical and Materials Engineering,
COPPE, Federal University of Rio de Janeiro, Caixa Postal 68505,
Rio de Janeiro, RJ, Brazil, CEP 21941-972COPPE/UFRJ
Abstract
In this study, the microstructures of a ceromer (Artglass®) reinforced by either glass fibers
(GlasSpan®) or polyethylene fibers (Connect®) were characterized and compared and the
influence of the fiber reinforcement on the flexural strength of the resulting products
evaluated. With this objective, seven bars of each material were produced. One bar of each
material was separated for microstructural analysis. The microstructural samples were
subjected to metallographic polishing and finishing, and then analyzed using optical
microscopy at different magnifications. The images obtained were treated using an image
processing computer program (Image Pro Plus) in order to quantify the microstructure by
calculating the mean diameter and mean volume fraction of fibers. The flexure tests were
made by three-point bending, using six samples of each material. After statistical analysis, the
results showed that the mean diameter of the glass fibers (4μm) was smaller than the
polyethylene ones (23.6 μm). The mean volume fraction of glass fibers (0.42) was larger than
that of the polyethylene fibers (0.28) and the mean center-to-center distance between fibers
was smaller in the glass fibers material (33 μm) than in the polyethylene fibers material (61
μm). The flexural strength of both glass and polyethylene fiber-reinforced materials was
1
E-mail address: Av. N. S. de Copacabana, 195/310, Copacabana CEP 22020-000 Rio de Janeiro – RJ – Brasil.
2
E-mail address: [email protected].
234 Silvana Marques Miranda Spyrides and Fernando Luiz Bastian
material.
Keywords: Fiber Reinforced Composites, Ceromer, Microstructure, Mechanical Properties,

Metal Free Fixed Prosthesis
1. Introduction
Fiber-reinforced composites were developed as an alternative to conventional fixed
partial prosthodontic, metaloplastic or metaloceramics, to fit small edentulous spaces. This
system, comprised of a fiber-reinforced ceromer allows a conservative, less invasive and
more esthetic preparation of the supporting teeth [1,2].
These prostheses have a substructure formed by long fibers impregnated with resin
matrix which provide strength and rigidity, and a particulate composite cover that improves
wear resistance and esthetics [3].
Several types of fibers have been used and studied for dental purpose, and it seems that
glass and polyethylene fibers are the most widely used and studied due to their transparency,
and ability to keep the original color of the covering composite [4,5,6,7].
Polyethylene fibers have shown high mechanical properties if the load is applied in
tension; however, if the load is in compression it shows lower resistance. On the other hand,
the glass reinforced material have shown the same properties regardless of the direction of
loading so their bending properties are higher than those of polyethylene fibers [6].
Fibers act as reinforcement, increasing the flexural strength, the fracture resistance and
the tensile strength of the polymer matrix, but they do not increase the compressive strength
[8,9]. Failures of fiber-reinforced ceromers are due to their poor adhesion or cohesion at the
fiber/matrix interface.
Mechanical properties of the fiber-reinforced composites are influenced by some factors
listed by BEHR [10] as follows: the orientation and the quantity of fibers, the impregnation of
fibers within the polymeric matrix, the adhesion of fibers to the polymeric matrix and the
properties of fibers versus properties of the polymeric matrix. Another factor that may affect
the properties is the position of the fiber reinforcement [11].
The properties of the reinforced fibers differ according to the type of fibers and their
architecture [12]. The properties of the covering composites vary according to their
composition, not only in the organic matrix but also in the dispersed inorganic phase (type,
size and distribution of particles). According to FREILICH et al. [12], ELLAKWA et al., [13]
this variation in the composite composition and the type and architecture of fibers affects the
properties of fiber-reinforced composites.
ELLAKWA et al., [13] verified that different composite compositions did not produce
significant differences in flexure strength of polyethylene fiber-reinforced composites. In a
previous work, SPYRIDES and BASTIAN [14] observed that the variation between glass and
polyethylene fibers did not produce significant changes in the flexure strength of composites
reinforced with these two types of fibers.
The impregnation of the fibers is required to provide close union of the fiber to the
polymeric matrix and thus for the strength of the composite [15]. Poor impregnation leads to
failure in transferring the load from the matrix to the fiber. On the other hand, spaces between
Microstructural Influence n Flexure Strength… 235
them may occur, which may lead to a considerable decrease of tensile strength and the
elasticity modulus of the fiber composites when compared with the theoretical strength and
modulus of such composites [16]. These regions increase the water absorption and, thus,
decrease the mechanical strength of the composite [17]. Porosity between matrix and fibers
diminishes the capacity of load support of fiber-reinforced composites [5,13,15].
According to FREILICH [12], due the necessity of having all fibers fully wetted by the
resins, the fiber volumes are generally limited to less than 50%. Pre-impregnated
unidirectional dental fiber composites incorporating approximately 45% of glass fibers have
flexure strength in the range of 600 and 1,000 MPa. These values are about ten times higher
than those of composites without fiber reinforcement, and this represents the primary
mechanical benefit of using fiber reinforcement in a prosthetic framework for dental purpose.
VALLITTU and NARVA [5], VALLITTU [15] demonstrated that there is a relationship
between the amount of fibers within the polymeric matrix and the increase of flexure strength
on the fiber reinforcements.
Larger amounts of fibers result in better properties. However, an excess can lead to an
insufficient wetting of the resin, which represents less resistance and poor characteristics of
manipulation. To reach an optimum strength, it’s necessary to have a high volume fraction of
fibers, thorough wetting and a uniform distribution of fibers within the matrix resin [6].
Due to the mechanical demands that fiber-reinforced fixed partial dentures are submitted
during the masticators function, it is known that their clinical behavior depends directly on
their microstructure and components. As a consequence, the purpose of this study was to
evaluate, quantify and compare the microstructure and its influence on flexure strength of a
ceromer reinforced by either glass or polyethylene fibers.
2. Materials and Methods

A microstructural and flexure strength chacterization of fiber reinforced ceromer was
performed. Manufacturers and compositional information of the tested materials are
summarized in table 1.
Table 1. Investigated materials
Material Composition Manufacturer

Artglass Second-generation laboratory composite resin. Multifunctional Heraeus Kulzer
Dentine* metacrilic ester matrix (30% by weight) filled with glass particles Germany
of silicium, barium and aluminum oxides and silanized (70% by
weight with 1 μm medium particle width) and photoiniciators.
Connect* Braided polyethylene fiber (UHMW-PE) 2mm width, coated by Kerr Corporation
plasma gas. USA
GlasSpan* Silicium, aluminum and boro oxide braided silanized glass fiber GlasSpan Inc.
with 2mm width. USA
Connect Fluid resin composed of non cured metacrilic ester monomer, Kerr Corporation
Resin* photoiniciators, inorganic fillers and additives for stabilization. USA
*Brand name.
Fourteen rectangular specimens (30 mm length, 4 mm height and 4 mm width) of

Artglass, seven reinforced by polyethylene fibers and seven by glass fiber were produced by
inserting the material into a silicone mold (Stern Tek, Sterngold) in three increments. One
impregnated fiber layer was placed between the first and second layer of material. Each layer
was photocured in the UniXS unit for 90 seconds and after the last cure the bar was removed
from the silicone mold and photocured for 180 seconds.
2.1. Microstructural Characterization
2.1.1. Samples Preparation
Two bars previously prepared, one with glass fiber and the other with polyethylene
reinforcement were randomly selected and cut in small blocks each measuring 5 X 4 X 4 X
mm. These blocks were embedded in epoxy resin (Ciba Geisy 311 with 10% of 24 catalizer)
with the transversal, longitudinal and longitudinal base surface exposed as shows Figure 1.
Figure 1. Schematic representation of produced material bar showing the position of the fibers layer and
the cuts: transversal (1), longitudinal (2) and longitudinal base (3).
After the resin cure, the samples was finished in an industrial polishing machine (DP -
NAP - Struers model Knut Rotor) with a series of emery papers (600, 800 and 1,200 grit - 3M
Company), under current tap water refrigeration.
After that, fine metallographic polishing was carried through by pan cloth with 1μm and
0,5μm diamond paste (AP paste F - Struers Metallographic Equipment) in industrial polishing
machine (Prazis APL – 4). Final cleanness after the polishing process was performed by an
ultrasonic device.
2.1.2. Acquisition, Treatment of the Images and Quantitative Characterization
After polishing and finishing, samples of Artglass/Connect and Artglass/GlasSpan were

observed at the fiber region using optical microscopy (Olympus BX60M) in three sections
with magnifications of 100X, 200X and 500X. Digital photographs in JPEG format were
acquired with a photographic camera (CCD COMU/Snappy), with the intention of obtaining
the mean volume fraction, diameter and centre-to-centre spacing of the fiber reinforcement
region of those materials.
The acquired pictures were then treated with Adobe Photoshop 5.0 program
(Image/Adjustments/Equalize) in order to compare shape, size and distribution of fibers, and
also detect if there was any void or porosity next to fibers or in the fiber/matrix interface.
Six acquired pictures (500X) of the glass fibers region and another six (200X) of the
polyethylene fiber region were subjected to a new treatment with Corel Photo Paint program
(Magic wand, brush, Eraser, Fill tool and Zoom) with the objective of leaving the fibers with
white and the matrix with black color. After that these pictures were processed with an Image
Pro Plus program with the purpose of obtaining statistically the average diameter and volume
fraction of each type of fiber.
Using the obtained values of fibers mean diameter (μm) and the fibers mean volume
fraction (%) the mean centre to centre distance between fibers (λ) were calculate with the
following equation [18]:
λ = L x 1-Vv
Vv
where λ is the mean centre to centre distance between fibers, L the fiber mean diameter (μm)
and Vv the fibers mean volume fraction (%).
2.2. Flexure Strength
Flexure testing was performed according to ANSI/ADA No. 27 – 1993 specification

7.8.1.1. Six bars of each material were used. The bars were subjected to manual grinding with
800 grade emery paper (3M Co.). The tests were performed in an Instron testing machine,
model 4204, at a crosshead speed of 0.5 mm/min. The specimens were placed on a three-point
bending test device with 20mm between the supports as illustrated in figure 2.
Figure 2. Diagrammatic representation of three point flexure testing device.
From the results of the flexure tests it was possible to calculate the flexural strength,
deflexion and elasticity modulus of the test specimens using the following equation:
3Pl
σ = ____
2bh2
where: σ is the flexure strength, l the span between the supports, b the width, h the height of
the sample and P the fracture load.
Means, standard deviations and coefficients of variation were computed (Excel 2000
Microsoft Office). All data were statistically analyzed with t student test for averages
comparison between the groups. The level of significance was set at α = 0.05.
3. Results and Discussion

3.1. Microstructural Characterization
The observations of the acquired images in transversal, longitudinal and longitudinal base
cuts (figures 3 and 4) shows that the shape, size and distribution of fibers were different
between the two studied materials and that the fibers had a reasonable impregnation by the
polymeric matrix in both glass and polyethylene reinforced composites.
In transversal sectional views, figures 3a, 4a 5a, 5b, 6a and 6b, the glass fibers showed
circular transversal shape, a relative uniform distribution and appearance of being totally
involved by polymeric matrix, meanwhile the polyethylene fibers showed irregular and
elliptical shapes, lager size, less uniform distribution, with some areas without fibers
alternating with areas of high concentration of fibers. This irregular distribution of
polyethylene and glass fibers can be attributed to the braid architecture of both. In places
where fibers cross each other, greater proximity between fibers occurs.
In present study, the shape and distribution of fibers in transverse section were similar to
the observed in the study of GOLDBERG and BURSTONE [19], BEHR et al [10] and BAE
et al [20]. It also appeared that wetting of glass fibers to the polymeric matrix was similar to
that of study of BAE et al [20] which used GlasSpan fibers impregnated with Aelitefil.
In longitudinal section views, figures 3b and 4b, the glass fibers showed a range of
ovalized shapes. This happened due to the position of cutting. In this same section, the
polyethylene fibers presented varied shapes with tendency to elongated forms.
(a) (b) (c)
Figure 3. Optical microscopy views of Artglass/GlasSpan material (a) transversal section X 500, (b)
longitudinal section X 200 and (c) longitudinal base section X 200.
In longitudinal base section, figure 3c and 4c, it was possible to observe the braid
architecture of glass fibers. The polyethylene fibers didn’t show the braid architecture
probably due to the larger diameter of its fibers in relation to the glass fibers.
(a) (b) (c)

Figure 4. Optical microscopy views of Artglass/Connect material (a) transversal section X 100, (b)
longitudinal section X 200 and (c) longitudinal base section X 200.
(a) (b)
Figure 5. Optical microscopy views of Artglass/Connect material (fibers region) with magnification X
200 in transversal section, (a) treated image using Photoshop 5.0 program and (b) treated image using
the Corel Photo Paint program.
(a) (b)
Figure 6. Optical microscopy views of Artglass/GlasSpan material (fibers region) with magnification X
500 in transversal section (a) treated image using 5.0 Photoshop program and (b) treated image using
the Corel Photo Paint program.
The values of fiber diameter, fiber number (per square mm) and fiber volume fraction are
shown in table 2.
The Student t-test exhibited significant differences (p < 0.05) for all means values.
Table 2. Values of fiber diameter (μm), fiber number (per mm2) and fiber volume
fraction (%). Means values, standard deviations (sd), p-values and coefficients of
variation (CV)
Material Mean p-value s.d. C.V.

Diameter GlasSpan 4.0 0.5 12.5
(μm) 0.000004*
Connect 23.6 6.0 25.4
Number GlasSpan 31385.2 2051.2 6.53
in 1 mm2 0.00000028*
Connect 640.8 141.9 22.1
Volume GlasSpan 42 3.44 8.2
fraction (%) 0.0002*
Connect 28 4.84 17.3
* significant difference (α = 0.05).
According to CHAWLA [21], smaller diameter of the fibers leads to lower probability of
defects. In present study (Table 2) the mean diameter of the glass fibers (4μm) was smaller
than the polyethylene ones (23.6 μm).On the other hand, the mean centre to centre distance
between fibers was smaller in the glass fibers material (33 μm) than in the polyethylene fiber
one (61 μm). Also, the mean volume fraction of glass fibers 0.42 (42%) was larger than that
of the polyethylene fibers 0.28 (28%)
GOLDBERG and BURSTONE [19], VALLITU [15], BEHR et al. [10] performed
microstructural evaluation by the incinerating or dissolution methods. The weight was
calculated before and after the incinerating or dissolution process in order to get the total
weight of fibers and thus the total fibers weight fraction. These methods seem to be
imprecise, due to the fact that by incinerating or dissolving the materials there is a possibility
of finding particles and resin rests together with the fibers themselves. Besides, these two
methods do not allow observing inhomogeneous regions in the composite materials. Another
problem is that these two methods consider that all fibers have an uniform distribution in the
composite matrix which is not always true.
In our investigation, the fiber glass content (42%) was very close to the reported by
GOLDBERG and BURSTONE [19] (43% and 45%) which used the dissolution method.
However, the present results (42%) were completely different from the results of BERH et al.
[10] (28% and 12%) which used the incineration method.
The results of BAE et al. [20] of mean volume fraction for both glass and polyethylene
fibers was far smaller than the results found in present study. This probably happened due to
the method used in the BAE et al [20] study. In that work, the authors weighed the material as
a whole, the fiber reinforcement plus the composite. As a result, the mean volume fraction of
fibers was very small. However, this method is not used in engineering because the mean
volume fraction must represent the structural region of the reinforcement and not the whole
material itself.
One of the main reasons for different fiber content in the literature is the method of
measurement. However, different methods of sample preparation, other types of fibers or
fiber wetting also can cause differences in fiber content [10].
A problem with the obtained values of volume fraction of fibers, as already pointed by
BEHR et al [10], is that the fibers are not cut completely axially or transversally during the
preparation of the samples. As a result, “the sectional view of the fiber is not circular but
instead elliptical, which results in incorrect volume content of fibers”. The same authors
(BEHR et al) also pointed that “the distribution of fibers in the matrix is not always
homogeneous. A sectional view with a representative low or high fiber content could falsify
the fiber volume calculation”. The first point could not be avoided in present work, so there is
some error in the obtained values of volume fraction, mainly in the polymeric fiber material.
The second point was avoided through measurements on different regions of the samples and
subsequent averaging of the values.
3.2. Mechanical Behavior Characterization - Flexure Strength
The values of flexure strength obtained are listed in table 3.

The t student test exhibited no statistical differences for flexure strength (p = 0.20)
between groups.
Artglass/Connect presented the largest coefficient of variability in all tests indicating a
higher dispersion.
During the flexure tests Artglass/GlasSpan presented an elastic behavior with only one
maximum load peak before breaking, while Artglass/Connect presented an elastoplastic
behavior and the occurrence of two load peaks before breaking as can be seen in Figure 7.
The load used to calculate the flexure strength of this material was the first load peak from
which the material was already in process of cracking by delamination process.
During the tests, the bars showed different behavior: The glass fiber bars (Figure 8a)
developed cracks in the tensile part of the central region of the bar with propagation into de
compression side; while the polyethylene fiber bars (Figure 8b) showed not only these cracks
in the central portion but also fracture in other parts of the bar, mainly next to one of the
supports.
Table 3. Values of flexure strength (σ) of Artglass/GlasSpan and Artglass/Connect.

Mean values, standard deviations (sd), p-values and coefficients of variation (CV)
Material Mean p-value s.d. C.V.

Flexure Artglass/GlasSpan 131.02 9.95 7.59
Strength 0.20
(σ) (MPa) Artglass/Connect 116.81 23.51 20.12
* significant difference (α = 0.05).
After the tests, it was observed that Artglass/GlasSpan did not show delamination
between the particulate composite and fibers, which means that this material probably
reached its maximum and possible flexure strength. On the other hand, Artglass/Connect
showed cracks at the interface with delamination that could had resulted from porosities,
impurities, air captured or even lack of adhesion at the interface. This can indicate that this
material had reached its limit of shear bond strength at this interface, with consequent
cracking, as shown in Figure 9
Figure 7. Typical load vs. displacement records in the flexure test a) Artglass/GlasSpan and b)
Artglass/Connect.
(a) (b)
Figure 8. Flexure testing (a) fiber glass reinforced bar (b) polyethylene fiber reinforced bar.
(a) (b) (c)
Figure 9. Cracked bars after flexure testing (a) glass fiber reinforced, (b) and (c) polyethylene fiber
reinforced.
The flexure strength of the two materials was considered statistically equal despite of
Artglass/Connect had presented delamination and supported a slightly larger load. In terms of
material development, it can be presumed that this material still presents potential for
improvement through further studies on its fiber/matrix interface.
The mean values of flexure strength of ceromer reinforced with polyethylene fiber
(116.81 MPa) found in present work was smaller than the mean value of ELLAKWA et. al.
[22] work for the same material (261.1 MPa). This probably happened due to their fibers
position which was also in the tensile side of the bars but closer to the base.
In BEHR et. al. [10] work the flexure strength values found for composites reinforced
with previously impregnated glass fibers Vectris, 618.0 MPa, and manually impregnated
Fibrekor, 585.0 MPa, are very high not only in relation to the values found in present work
but also to the values found by ELLAKWA et al.[22] and BAE et al [20]. This great
difference may be due to the fact that in BEHR et al.[10] study the authors tested bars with
fibers impregnated without coverage of particulate composite. This fact can be evidenced
comparing the values found by BAE et al. [20] and CHONG and CHAI [23] for the fiber
glass Vectris with coverage of composite Targis, respectively 296.0 MPa and 84,0 MPa, and
for fiber glass Fibrekor with coverage of composite Sculpture, 203.0 MPa and 165.0 MPa.
VALLITU [5,12] demonstrated that there is a relationship between fiber content and the
flexure strength. The same fact had been reported by GOLDBERG and FREILICH [6] which
found that higher fiber content caused higher strength. Despite this, FREILICH [12] limited
the fibers volume fraction to 50% in order of having all fiber wetted by the resin.
GOLDBERG and BURSTONE [19] also found that the highest fibers content lead to an
increase of flexure strength contradicts the results of BEHR et al. [10] and BAE et al. [20]
which did not obtain any increase in flexure strength when the amount of fibers had
increased.
In the present study the mean volume fraction of glass fibers 0.42 (42%) was larger than
that of the polyethylene fibers 0.28 (28%). Moreover, when the flexural strength results of
these two types of fiber reinforced materials are compared, we can see that despite the larger
mean volume fraction of the glass fiber material and the higher tensile strength of the glass
fibers according to FREILICH et al. [12], the flexural strength of both glass and polyethylene
fiber reinforced materials was statistically equal.
6. Conclusion
The flexural strength of both glass and polyethylene fiber-reinforced materials was
material. The elastic behavior and the type of cracking, without any delamination observed in
the glass fiber reinforced ceromer, indicates that the fibers had a good distribution and
adhesion within the ceromer, probably reaching the maximum possible strength. On the other
hand, the elastoplastic behavior and cracking by delamination of the polyethylene reinforced
ceromer indicates that the development of this material needs improvement, by increasing the
fiber adhesion within the ceromer, better distribution of fibers and a higher mean volume
fraction of fibers.
Acknowledgements
The authors thank Mr. Renan Hufnagel Bella for the session of his dental prostheses
laboratory. Also, to CNPq for continous support and IMA-Macromolecule Institute for the
use of equipments.
References
[1] Shannon, A. (1997). Fiber-Reinforced Composite Bridge: Inlay-to-Inlay Technique.
Dentistry Today, 48-53.
[2] Winters, K.L. (1999). Using a fiber-reinforced ceromer for fixed restorations. Dent.
Today, 70-73.
[3] Freilich, M.A; Duncan, J.P; Meiers, J.C; Goldberg, A.J. (1998). Preimpregnated, fiber-
reinforced prostheses. Part I Basic rational and complete-coverage and intracoronal ficed
partial denture designs. Quint. Int. v.29, n.11, 689-696.
[4] ramos, v; runyan, d.a; christensen, L.C. (1996). The effect of plasma-treated
polyethylene fiber on the fracture strength of polymethyl methacrylate. J. Prosthet.
Dent., v..76, n.1, 94-96.
[5] Vallitu, P.K and Narva, K. (1997). Impact strength of a modified continuous glass fiber-
polymethyl methacrylate. Int. J. Prosthod., v.10, n.2, 142-148.
[6] Goldberg, A.J., Freilich M.A., (1998, march).Fiber-reinforced Composites. Dental
Advisor, v.15, n.2, 2-5.
[7] Honda, M.I; Vieira, D; Guedes, A.P.P. (2001, nov.) Restaurações com reforço de fibras.
Dental Special [on-line serial]. 1-8, www.Dentalspecial.com.br
[8] Stipho, H.D. (1998). Effect of glass fiber reinforcement on some mechanical properties
of autopolimerizing polymethyl methacrylate. J. Prosthet. Dent., v.79, n.5, 580-584.
[9] Belvedere, P.C. (1998). Single-sitting, fiber-reinforced fixed bridges for the missing
lateral or central incisors in adolescent patients. Dent. Clin.of Nor. Am., v.42, n.4, 665-
682.
[10] Behr, M; Rosentritt, M; Lang,R; Handel,G. (2000). Flexural properties of fiber
reinforced composite using a vacuum/pressure or a manual adaptation manufturing
process. Journal of Dentistry, v.28, n.7, 509-514.
[11] Ellakwa, A.E; Shorthall, A.C; Shehata, M.K; Marquis, P.M. (2001). The influence of
fibre placement and position on the efficiency of reinforcement of fibre reinforced
composite bridgework. J. Oral Rehabil., v.28, n.8, 785-91.
[12] Freilich, M.A., Meiers, J.C., Duncan, J.P.,Goldberg,A.J.(2000), Fiber-Reinforced
Composites, Hong Kong, ed. Quintessence Publishing Co, Inc.
[13] Ellakwa, A.E., Shorthall, A.C., Shehata, M.K., Marquis, P.M. (2001), Influence of
veneering composite composition on the efficacy of fiber-reinforced restorations (FRR).
Oper. Dent, v. 26, n.5, 467-75.
[14] Spyrides, S.M.M. And Bastian F.L. (2004). In vitro comparative study of the
mechanical behavior of a composite matrix reinforced by two types of fibers
(polyethylene and glass). Materials Science and Engineering C, v. 24, 671-677.
[15] Vallitu, P.K. (1999).Flexural properties of acrylic resin polymers reinforced with
unidirectional and woven glass fibers. J. Prosthet. Dent., v.81, n.3, 318-326.
[16] Vallitu, P.K. (1998). Some aspects of the tensile strength of unidirectional glass fiber
polymethyl methacrylate composite used in dentures. J. Oral Rehabil., v.25, 100-5.
[17] Vallitu, P.K; Ruyter, I.E; Ekstrand, K. (1998). Effect of water storage on the flexural
propertires of E-glass, and silica fibers acrylic resin composite. Int. J. Prosthodont, v.11,
340-350.
[18] Underwood, E.E. (1970). Quantitative Stereology. Massachusetts, Addison-Wesley

Publishing Company.
[19] Goldberg, A.J. and Burstone, C.J. (1992).The use of continuous fiber reinforcement in
dentistry. D. Material, v.8, 197-202.
[20] Bae, J. M; Kim, K. N; Hattori, M; Hasegawa, K; Yoshinari, M; Kawada,E; Oda, Y.
(2001). The flexural properties of fiber-reinforced composite with light-polimerized
polymer matrix. Int. J. Prosthodont, v.14, n.1, 33-39.
[21] CHAWLA, K. K. (1987). Composite Materials Science and Engineering. Germany. ed.
Springer-Verlag New York Inc.
[22] Ellakwa, A.E; Shorthall, A.C; Shehata, M.K; Marquis, P.M. (2001). Influence of
veneering composite composition on the efficacy of fiber-reinforced restorations (FRR).
Oper. Dent., v. 26, n.5, 467-75.
[23] Chong, K.H. and CHAI, J. (2003). Strength and mode failure of unidirectional and
bidirectional glass fiber-reinforced composite material. The International Journal of
Prosthodontics, v.16, n.2, 161-166.
Chapter 9
INFLUENCE ON STRENGTH PROPERTIES

OF ANISOTROPY PLANES IN SLATES SAMPLES
IN THE NW OF SPAIN
M.A. Rodríguez-Sastre*, M. Gutiérrez-Claverol,

M. Torres-Alonso and L. Calleja
Geology Department, Oviedo University. c/Jesús Arias de Velasco, s/n, Oviedo 33013
Asturias, Spain
Abstract
The purpose of this paper is to describe the influence of anisotropy on the geomechanical
strength properties of two Spanish slates with different chemical and physical characteristics.
From laboratory testing results of slates under point load and uniaxial compression and the use
of indirect methods, as it is the measurement of P velocities, principal parameters were
calculated for this rock material. As it is well known under uniaxial compressive strength
slates are strong and also very strong rock when loading is parallel (90o) o perpendicular (0o)
to the main anisotropic planes. In contrast it is a weak rock with minimum strength values for
angles between 45 to 60o of inclination of anisotropy planes. The correlation equations were
calculated between different parameters. Despite weak correlation between different
geotechnical properties were found and when all lithologies are considered together
correlation of geomechanical properties is weak. However when each lithology is considered
separately the geomechanical properties can be coherently defined. Linear and polynomial
equations were found for the point load and uniaxial strength correlations with the inclination
of the anisotropy. Different strength fields were calculated when uniaxial strength and point
load test plot and its comparison include the inclination of the anisotropy planes on slates.
Uniaxial compressive strength and P wave velocity appears to be strongly influenced by
uniaxial strength and good polynomial correlations resulted. Plots of slates with other
sedimentary type of rocks from Cantabrian Zone, CZ, revealed the hardness and highest
strength of slates when loading is perpendicular to the main anisotropy planes.
Keywords: slates, anisotropic rocks, rock strength, uniaxial compressive strength, point load test.
*
Corresponding author: Geology Department, Oviedo University, c/Jesus Arias de Velasco, s/n, Oviedo 33005
Asturias, Spain
248 M.A. Rodríguez-Sastre, M. Gutiérrez-Claverol, M. Torres-Alonso et al.
1. Introduction
Slates under investigation in this study are one of the problematic geological materials
due to the presence of anisotropic planes, which represent weak planes, were the failure of the
rocks could suddenly occur.
In previous studies, researchers had been focused on core samples from boreholes and
had investigated most of the important aspects of rock strength related with the anisotropy
with the loading axis perpendicular to any planes of weakness and mainly considerations on
rock strength were provided in relation with different conditions of saturation, cored
diameters, different sample shapes, and different geotechnical tests (Rodríguez-Bouzo, 1993;
González-Buelga, 1995; Hawkins, 1998; Bell, 2000; Tsiambaos & Sabatakakis, 2004;
Vásárhelyi & Ván, 2005). These studies have also described the influence of strata orientation
when measuring the strength under unconfined compressive conditions.
Another extensive work was based on the influence of these cleavage planes on strength
properties and it was supported in factual information on metamorphic foliated rocks and
demonstrated that cores cut perpendicular to the cleavage planes possessed the highest
strength whilst those cores cut at 30 or 60o exhibit the lowest (Griggs, 1951; Donath, 1961;
Brown et al, 1977; Goshtasbi, et al. 2006). All studies were related to engineering works
developed in a metamorphic area where the influences of the anisotropy under triaxial
conditions were calculated.
In other way, the failure of anisotropic rocks had been analyzed between others by
Sheorey (1997), Mustschler & Natau (1991) and Natau et al (1995), and an exhaustive
description of anisotropy and its influence on uniaxial compressive strength, modulus of
Young and confining pressure was described by Ramamurthy (1993). Although Ramamurthy
had established the classification of the main types of anisotropy for different types of rocks,
during his investigation all properties were related with the inclination of the anisotropy
planes under confined conditions and a rock anisotropy index was proposed, in which the
maximum and minimum values of one geotechnical property were resolved.
Further investigations on artificial anisotropy rock samples were proposed by Singh et al.
(2002) who had differentiated several types of failures depending on the cleavage anisotropy
and the influence of different space between two anisotropy planes with normal disposition
between both planes and foliation.
Truchas slates are a planar anisotropic rock due to the millimetre spaced cleavage planes
developed during Variscan Orogeny on these Ordovician rocks within a geological structure
known as Truchas syncline in the NW of Spain. Casaio and Rozadais formations are an
abundant geological sequence in this syncline dominated by silty slates, fine grain slates and
quartzite rocks that are characterized by wide variations in their engineering properties. These
rocks are problematic material because of the presence of cleavage planes. There is a little
data on the engineering performance of Casio and Rozadais slates, and therefore the current
study is focused on the determination of the strength related to the inclination of the
anisotropy planes on this rocks. This formation is a main source of roof slate. These slates
were extensively investigated by Barros-Lorenzo (1989) and Taboada (1993) and related to
the organization of mining works Taboada et al., (1997, 1998 and 2006) or to a new design
and planning techniques of mining as was proposed by Bastante et al. (2004). The
Influence on Strength Properties of Anisotropy Planes in Slates Samples… 249
investigation only covered the geology, some geotechnical data and mining aspects of the
rock.
The geomechanical properties discussed are essential to guide future geotechnical
developments as noted by Pine & Harrison (2003). These data have a great value for the
investigations of anisotropic rocks, for example when selecting different engineering works in
Truchas area that can be designed on these slate materials or related to new methods of
mining extraction.
The geomechanical properties of these slates for the entire range of inclination of
anisotropy related to the load application are the main concern of this paper. We compare
with the strength values of other data from factual reports in the NW of Spain (Labrada-Rubio
et al., 1982).
Values of the uniaxial compression tests (UCS) on these rocks are presented because this
is undoubtedly the main geotechnical property used for determining the strength of intact rock
ISRM (1981). These test results are directly applicable to different engineering works as slope
excavation and tunnelling and also it is included as a main input parameter for rock mass
characterization and classification. More often the point load testing (PLT) is used to
determine rock strength indexes in geotechnical practice, following the proceed of the ISRM
(1985), and the results of this test are to its correlation with the UCS values on these rock and
to establish the influence of the inclination of the anisotropic rocks on it.
The main objective of this study is to determine main strength properties to investigate
the influence of cleavage anisotropy and the control of the lithological compositions on the
geomechanical properties of the slates in the two main formations of the Truchas syncline.
2. Slates Under Study

The slates evaluated in this study are located in the Centro Iberian Zone (CIZ) in the NW
of Spain, and belong to the Truchas syncline area that has been described as a D1
synclinorium, with more than 20 km wide, and consists of many individual recumbent folds
(Martínez-Catalán et al., 2004). They belong to two different types: Casaio (C) and Rozadais
(R) slates. From chemical analysis on slates the main differences between Casasio and
Rozadais slates have been established as it is shown in Table1. Those slates showing a planar
anisotropy are Ordovician metapelite rocks metamorphosed under low grade conditions with
a fine or silty grain with a slaty cleavage microstructure and a lepidoblastic texture, having
some of them intercalated millimetre sandstone layers (Rodríguez-Sastre, 2003), main
petrographical data for the fine grain slates are given in Table 2. Chlorite and opaque minerals
are slightly more abundant in Rozadais than in Casaio slates.
Table 1. General chemical analyses characteristics of the slates studied
Chemical composition (%)

Mayor Elements
Casaio Rozadais
SiO2 76-57 56-52
Al2 O3 20.8-8.6 22-20
Fe2O3 7.7-3.2 9.8-8.9
Table 2. General petrographical characteristics of the fine grain slates studied
Minerals Microe- Minor Opaque

Sample Texture Matrix
(% Estimated) structure minerals minerals
Quartz (45%)
phyllosilicate
C2
(40%)
Chlorite (15%) Pyrhotite
Chlorite+
Others (<1%) Tourmaline Chalcopyrite
Lepido- Slaty muscovite+o
Quartz (25%) Rutile Magnetite
blastic cleavage paque
phyllosilicate Zircon Illmenite
minerals
(40%) Pyrite
R2
Chlorite (20%)
Calcite (10%)
Others (5 %)
Three main types of slate have been recognised in Casaio Formations, fine grain slates
(C2) comprise a fine grained slate with variable quartz content and abundance of sulphurous
minerals concentrated in zones, varying in size from millimetre to centimetre scale and
concentrated on the cleavage planes, silty slates (C3 and C4) are a grey slate with 1-2 cm
nodules of quartz and sulphurous or sulphide minerals related with the cleavage formation
and with the presence of thin millimetre laminations, and sandstone laminated slates (C1a and
b) include well foliated slates with millimetre interlayer sandstones of 3-5 cm thick, which are
related with the initial stratification planes at the original sedimentary basin deposit. Sulphur
minerals often are found on the slaty cleavage planes. Two types have been identified and
related to the millimetre interlayer sandstone and slaty cleavage planes. C1a type presents a
sub horizontal orientation of the interlayer sandstone (So) that shows an angle of 20o dip in
relation with the cleavage planes, whereas C1b shows angles between interlayer sandstones
(So) and cleavage planes (Sp) of 50o. From field observations the C1a type appears as a more
homogenous rock than the C1b type, which has weaker cleavage planes. In Rozadais
Formation two main types were identified, fine grained (R2) is a bluish grey slates with
widespread sulphurous minerals of a millimetre scale, and siltly slates (R1 and R5) samples
which consist of silty grey slate with bluish colours and sulphur minerals widespread.
3. Test Procedure
Rock samples were obtained from unweathered quarry outcrops. Also all these samples
were kept submerged in water a minimum of 24 hours to ensure saturation or wet conditions.
Thus tests were carried out under water saturation conditions, which are the natural conditions
for rocks under the water level surface in the outcrop. Therefore these results will have shown
the lowest values of strength for these slates, because as demonstrated by Hawkins (1998), the
strength values decrease with the water content in samples. Thus laboratory core drill for NX
cylindrical samples and saw machines were used to cut the samples with a length-to-diameter
ratio of 2.0 to 2.5 and end faces were ground in order to provide specimens with size, shape
and ends geometries according to ISRM (1981) specifications. The test specimens were
obtained from the same block of rock in order to decrease the effect of natural variations in
the intrinsic characteristics of the slates and so as to compare the values for the various
properties in the different slate types differentiated, to allow a coherent interpretation to be
made.
The rock samples examined in this study were prepared and tested in uniaxial
compression and point loading with the loading axis from parallel to perpendicular through
intermediate positions of the plane of weakness. The fracture created by the point load test
was always fresh and through the rock material at parallel and perpendicular positions of
loading whereas at the intermediate positions fracture had a little interference of the planes of
weakness. Axial and diametral tests were conducted on rock core samples and during the
axial test, for cores with inclination of the anisotropy from 0 to 60o, the core is loaded parallel
to the longitudinal axis of the core and this test it most comparable to a UCS test. In the
diametral test, for cores with inclination of the anisotropy of 90o, the core is loaded
perpendicular to the longitudinal axis of the core and in case of anisotropic rocks it is parallel
to the anisotropy planes and this test is most comparable to an indirect tensile strength test
and also in this study it is compared with the UCS test.
To finish and to characterize as a preview method the influence of the anisotropy planes
on the properties of this rocks ultrasonic velocities were recorded using an ultrasonic NEW
Sonic Viewer model 5217A from OYO (Japan) and correlated with the UCS values.
The better comparison of the results was as follows: PLT- inclination of anisotropy,
UCS- inclination of anisotropy and UCS-PLT, UCS-Vp and Vp-inclination of anisotropy.
These results have to be considered as an input parameter to rock classification rather than as
a means of predicting mechanical properties of rocks because these parameters are not well
correlated as it has been shown by Hawkins (1998) and Tsiambaos & Sabatakakis (2004).
4. Physico-Mechanical Properties
Physical properties for slates were calculated following the procedures for laboratory
tests in the UNE (Aenor, 1999) and ISRM suggested methods (1981) from 5 unweathered
specimens. Uniaxial compressive strength (UCS) and the Point load test (PLT) were
determine on 26 samples. The propagation rate of seismic waves was calculated on the basis
of determination performed on 65 specimens of each rock type identical to those used in the
uniaxial compressive strength test.
The studied slates have very low natural moisture content (<0.004%) with very low void
index ratio and very low porosity (<0.01%), which it is characteristic in this type of rocks.
The bulk density calculated varies between 2.47 and 2.89 kN/m3. Rozadais slates had
recorded the lowest values in porosity and in the saturation of moisture content and also in the
natural moisture content when it is related to Casaio slates.
The values of UCS and Is(50) for the 26 slate rocks which we tested are given in Table 3.
The Casaio slates show a uniaxial strength varying with the lithological type of slate and
classified as strong to very strong. Thus these highest values of strength were observed on the
slates with millimetre interlayer sandstones up to 80 MPa, whilst intermediate uniaxial values
between 60 and 80 MPa were found in the fine grain slates and the lowest values in silty
slates were recorder under 60 MPa. Those values are consistent with the values measured in
Schist, slate and phyllite rocks from the Toros Mountains tectonic group in Turkey (Őzsan &
Karpuz, 1996).
The point load strength index values ranged from 14.23 for the axial test perpendicular to
the cleavage planes on slates to 0.34 MPa for the diametral test. To compare these results with
the values from other sedimentary rocks in the Cantabrian Zone (CZ) from Labrada-Rubio et
al. (1982), and all these data were plotted together from the results recorded in P-wave
velocity results (Vp) range from 2.691 (perpendicular to anisotropy planes) to 6.991 m/s
(parallel to anisotropy planes). These data are consistent with the values obtained by several
authors for slates in the NW of Spain (Pernia et al., 1986). On the other hand these ranges are
wider than those recorded of P wave velocity in granites from the NE of Portugal (Sousa et
al., 2005). In other cases measurements made along and across the cleavage planes in
metamorphic rocks like quartzitic phyllite, carbonaceous phyllite and micaceous phyllite
varies in same range or even lower values than found in our study (Bell, 2000; Johnson & De
Graff, 1994).
Table 3. Data from laboratory testing of slates from ZCI and compared with WALZ
values of sedimentary rocks
Numbers of Is (50)
Zone Formation lithology σc (MPa)
samples (MPa)
ZCI
Casaio Slates 20 143.2-18.42 14.23-0.34
(Truchas (C1a, C1b, C2, C3, C4)
Syncline slates)
loading at 0, 10, Rozadais Slates 6 161.1-36.6 6.07-0.78
25 and 90 (R1, R2, R5)
Herreria, Lancara,
Ermita, Griotte,
CZ Montaña, Villamanin,
Limestones,
Oville, San Pedro,
sandstones,
loading Santa Lucia, La vid, 83 152.8-1.8 15.1-0.4
quartzites and
perpendicular at Nocedo Portilla, Pola
volcanics
any anisotropy de Gordon, Levinco-
plane Llanon-Tendeyon,
Volcanic rocks
Huergas and Formigoso Shales 10 22.9-2.6 2.53-0.64
5. The Influence of Anisotropy on Strength and Other Properties

of Slates
As shown in Figure 1, expected uniaxial compressive strength decreases from high values
for anisotropy inclination of zero to a minimum for an inclination of 60o. The UCS then
increases with further increases in inclination to 90o. These findings concur with those of
Griggs, 1951; Donath, 1961; Brown et al., 1977; Ramamurthy, 1993; Goshtasbi et al., 2006
and others. In terms of the ISRM (1981) classification the values varies betweenstrong and
very strong for inclination of 0 and 90o to weak at 60o.
160
100
140
C1b
90
Poly. (C1b)
C1a
MPa
(MPa)
2
120
80 c = 0,028 - 2,2 + 60,5
R2 = 0,9
Poly. (C1a)
c
c
2
c= 0,05 - 5,0 + 135,9 70
Uniaxial compressive strength

100
Uniaxial compressive strength,
R2 = 0,98
60
80
50
40
60
30
40 20
10
20
0
0 10 20 30 40 50 60 70 80 90
0
0 10 20 30 40 50 60 70 80 90 Inclination of anisotropy, (º)
Inclination of anisotropy, º
70
60 90
80
(MPa)
C4
50 P oly. (C 4)
(MPa)
70
c
C3
c

60 Poly. (C3)
2
40
2 c = 0,02 - 2,2 + 73,8
c = 0,01 - 1 ,5 + 52,8 50
2
2
R = 0,8 R = 0,9
40
30
30
20
20
10
10 0
0 10 20 30 40 50 60 70 80 90
Inclination of anisotropy, (º)
0
0 10 20 30 40 50 60 70 80 90
120
100
(MPa)
C2
c
80 Poly. (C2)
2
c = 0,03 - 3,1 + 89,9
60 R2 = 0,7
40
20
0
0 10 20 30 40 50 60 70 80 90
Figure 1. Polynomial function relationship between uniaxial compressive strength and inclination of
anisotropy for Casio slate Formation.
As shown in Figure 2, two trends were found relating P wave velocities to the anisotropy
inclination, namely approximately linear and polynomial. In both formations it was found that
maximum P wave velocities occur along the cleavage planes, parallel to the maximum grain
length growth direction. However when waves travel along the cleavage planes but
perpendicular to the grain length growth direction the velocities are reduced by 6% in Casaio
Slate and 2% in Rozadais Slate. The lowest values were obtained perpendicular to the cleavage
planes where velocity drops around 55-36% in Casaio Slates and 34% in Rozadais Slate.
y = 0.334x2 - 2.1951x + 3977.4

n=65 R2 = 0.7559
P wave velocities, Vp (m/s) 6000
5000
n=65
Vp (m/s)
4000 Linear (Vp (m/s))
Poly. (Vp (m/s))
y = 22.765x + 3721.3
R2 = 0.6215
3000
0 10 20 30 40 50 60 70 80 90
Inclination of anisotropy, β ( )
0
Figure 2. Linear and polynomial function relationship between P wave velocities and the inclination of
anisotropy.
Figure 3. Polynomial function relationship between uniaxial compressive strength and P wave
velocities for Casaio slate Formation.
The relationship between uniaxial compressive strength and compressional P-wave

velocity appears in Fig. 3. Although the P-wave velocity increases linearly with the angle of
inclination of the anisotropy the UCS shows a good polynomial correlation as described
above (Table 4), which means that polynomial functions are found for the different slate
types. The P wave velocity increases and strength decreases as the inclination of the
anisotropy increases from 0o to 60o. As the inclination increases to a maximum at 90o but the
strength decreases.
Good linear correlation was found between P-wave velocity and cleavage inclination
where the P-wave velocity varies between 2,000 and 4,000 m/s for an inclination of 0o, 4,000
and 5,000 m/s for inclinations between 25 and 45o and over 5,500 m/s or an inclination of
90o.
Good correlation was also found between point load strength and inclination of
anisotropy for all the Casaio slate types (Fig. 4). As seen in Fig. 4, the lowest correlation was
found for the C1b type, which is the most anisotropic specimen, whereas good correlations
were obtained for the other types.
12
C1a c1b
C3 C2
C4 Linear (C1a)
10 y = -0.10 x + 10.5
R 2 = 0.90
Linear (c1b) Linear (C3)
Linear (C2) Linear (C4)
Fin
8 y = -0.07x + 8.1 eg
rai
R 2 = 0.96 n
Is (50) MPa
s la
te
Si lty
y = -0.06 x + 7.2 sl at
6 e
R 2 = 0.94 y = -0.0719x + 8.0507
R 2 = 0.84
4 Slate w ith millimetric sandstone layers
y = -0.05 x + 5.99
R 2 = 0.78
2
0
0 10 20 30 40 50 60 70 80 90
o
Inclination of anistotropy, β ( )
Figure 4. Linear function relationship between point load index and inclination of anisotropy in Casaio
slates.
In Figure 5, Point Load Strength is plotted against UCS for different inclinations. Thus it
would appear that the relationship between these to parameters depends on the inclination of
the cleavage.
160

140
0o
0o
0o
120
σc MPa 90o
0o
100
10o C1a C1b
90o
80 10o
90o 0o 0o C3 C2
60o
10o 0o
90 90o
25o C4 R2
10o
40
o
90o 25
25o
20 25o 10o R1 R5
25o
0
1 10 100
Is (50) MPa
Figure 5. Fields correlating point load index with uniaxial compressive strength and inclination of
anisotropy in slates from Truchas syncline.
Figure 6. Comparison of plots values of different rock class and slates in correlation between uniaxial
compressive strength and point load index.
Figure 6 shows the relationship between UCS and PLS for various rocks, including
limestones and slates that were loaded perpendicularly to the cleavage. These data
demonstrate that the metamorphic rocks have higher strength. The difference in behaviour
between Casaio and Rozadais formations appears to be due to variation of porosity and
mineralogy.
6. Conclusion
It has been confirmed that anisotropy orientation strongly influences the strength
properties of the slates studied. Polynomial correlations for relationships between uniaxial
compressive strength and inclination of anisotropy, and also with P- wave velocity have been
established. The strength is lowest for an inclination of 60o, and highest at 0 and 90o.
As would be expected for rocks of this type, PLS varies over a very range although good
linear correlations were found with inclination of the foliation. In other words the PLS
decreases with increasing angle of cleavage inclination between 60 and 90o. However, it was
also influenced by lithology. Thus, Point load tests on foliated rocks would be able to
discriminate between differences on lithological compositions of slate sequences and tectonic
setting could have an influence on this test results. This could help to interpret sequences and
tectonic setting at outcrop scale.
Good linear correlations are found between P-wave velocity and the inclination of
anisotropy planes. However relatively high values of Uniaxial Compressive strength may
occur for either high or low velocities values and are at a minimum for intermediate
velocities.
Although poor correlation was obtained between Point Load Strength and UCS data,
different rock types group in different areas of this graph.
As expected, the geomechanical properties of the slates under study leads us to conclude
that the anisotropic fabric gives rise to relatively weak planes and which influence the
strength and elastic properties of the material.
Acknowledgements
Many thanks go to H. Stoll who is acknowledged for their critical reviews and correction
of the manuscript. J. Cripps kindly read through the text and provided useful comments.
References
[1] AENOR (1999): Geotecnia. Ensayos de campo y de laboratorio. Ed. AENOR, 385 pp.
[2] Barros-Lorenzo, J. C. (1989): Nuevos datos geológicos y cartográficos sobre el flanco
Sur del Sinclinorio de Truchas (Ourense-León, NW de España). Cuadernos do
Laboratorio Xeolóxico de Laxe, 14: 93-116.
[3] Bastante, F. G., Taboada, J. & Ordoñez, J. (2004): Design and planning for slate mining
using optimization algorithms. Engineering Geology, 73: 93-103.
[4] Bell, F. G. (2000): Engineering properties of Soils and Rocks. Ed Blackwell, London.
[5] Brown, E. T., Richards, L. R. & Barr, M. V. (1977): Shear strength characteristics of the
Delabole Slates. In: Proc Conf Rock Engineering Newcastle upon Tyne: 35-51.
[6] Davis, J. C. (1986): Statistics and data analysis in Geology. John Wiley and Sons, New
York.
[7] González Buelga, M. T. (1995): Caracterización geomecánica y valoración geotécnica
de materiales rocosos del Paleozoico inferior en la cuenca alta del Río Bernesga (León,
Noroeste de España). Thesis. Geology Department, University of Oviedo, 445 pp.
[8] Goshtasbi, K., Ahmadi, M. & Seyedi, J. (2006): Anisotropic strength behaviour of
slates in the Sirjan-Sanandaj zone. South African Institute Mining and Metallurgy, 106
(1): 71-75.
[9] Hawkins, A. B. (1998): Aspects of rock strength. Bull. Engineering Geology Env. 57:
17-30.
[10] ISRM SUGGESTED METHODS (1981): Description of Rock Masses. Int. Jour Rock
Mech. and Min. Sci. y Geomech. Abst. Pergamon Press, 18 (1): 85-110.
[11] ISRM SUGGESTED METHODS (1985): Determining Point Load Strength. Int. Jour
Rock Mech. and Min. Sci. y Geomech. Abst. Pergamon Press, 22 (2): 51-60.
[12] Johnson, R. B. & Degraff, J. V. (1994): Engineering Geology. Ed. Macmillan, 190 pp.
[13] Labrada-Rubio, J. A., Martínez- Álvarez, J. A., Torres M. & Gutiérrez-Claverol, M.
(1982): Estudio geológico-geotécnico el paso por ferrocarril de la divisoria Astur-
Leonesa (La Robla-Ujo- Pola de Lena). Factual Report. E.T.S.I.M.O, University of
Oviedo.
[14] Martínez-Catalán, J. R., Fernández-Suárez, J., Jenner, G. A., Belousova, E. & Díez-
Montes, A. (2004): Provenance constraints from detrital zircon U-Pb ages in the NW
Iberian Massif: implications for Paleozoic plate configuration and Variscan evolution.
Journal of the Geological Society, 161: 463-476.
[15] Natau, O., Bühler, M., Keller, S. & Mustchler, Th. (1995): Large scale triaxial tests in
combination with a FEM analysis for the determination of the properties of a transversal
isotropic rock mass. Proc. 8th Int. Congr. ISRM: 635-643, Tokio.
[16] Őzsan, A. & Karpuz, C. (1996): Geotechnical rock-mass evaluation of the Anamur dam
site, Turkey. Engineering Geology, 42: 65-70.
[17] Pernia, J. M., Fresno, F., Mena, J. M., Lucena, A. & Lechosa, R. (1986): Registro de
datos en sondeos de reconocimiento. Ed. IGME, 144 pp.
[18] Pine, R. J. & Harrison, J. P. (2003): Rock mass properties for engineering design.
Quarterly Journal of Engineering Geology and Hydrogeology, 36: 5-16.
[19] Ramamurthy, T. (1993): Strength and modulus responses of anisotropic rocks. In:
Comprehensive rock engineering, principles practice and projects. vol.3.
[20] Rodríguez Bouzo, L. F. (1993): Caracterización geomecánica de materiales
carbonatados del Carbonífero en la zona de Pajares-Villamanín. Thesis, Explotación y
Prospección de Minas Department, University of Oviedo, 416 pp.
[21] Rodríguez-Sastre, M. A. (2003): Caracterización geomecánica de materiales pizarrosos
del Sinclinal de Truchas (León-Orense). Thesis, Geology Department, University of
Oviedo, 388 pp.
[22] Sheorey, P.R. (1997): Empirical Rock Failure Criteria. Ed. A. A. Balkema, Rotterdam.
176 pp.
[23] Singh, M., Rao, K. S. & Ramamurthy, T. (2002): Strength and Deformational Behaviour
of a Jointed Rock Mass. Rock Mechanics and Rock Engineering, 35 (1): 45-64.
[24] Sousa, L. M. O., Suarez del Río, L. M., Calleja, L., Ruíz de Argandoña, V. G. &
Rodríguez Rey, A. (2005): Influence of microsfractures and pososity on the physico-
mechanical properties and weathering of ornamental granites. Engineering Geology, 77:
153-168.
[25] Taboada, J. (1993): Diseño de minas subterráneas de pizarra. Thesis, Explotación y
Prospección de Minas Department, University of Oviedo, 173 pp.
[26] Taboada, J., Vaamonde, A., Saavedra, A. & Alejano, L. (1997): Application of
geostatistical techniques to exploitation planning in slate quarries. Engineering Geology,
47: 269-277.
[27] Taboada, J., Vaamonde, A., Saavedra, A. & Argüelles, A. (1998): Quality index for
ornamental slate deposits. Engineering Geology, 50: 203-210.
[28] Taboada, J., Matias, J. M., Araujo, M. & Ordoñez, C. (2006): Assessing the viability of
underground slate mining by combining an expert system with a GIS. Engineering
Geology, 87: 75-84.
[29] Tsiambaos, G. & Sabastakakis, N. (2004): Considerations on strength of intact
sedimentary rocks. Engineering Geology, 72: 261-273.
INDEX
alcohol, 11, 12, 84, 96

algorithm, ix, 107, 157, 179
A alkaline, 58, 72
alloys, 95, 96
absorption, 64, 84, 235 alpha, 36
access, 227 alternative, 83, 98, 104, 112, 158, 206, 207, 234
accessibility, 218 aluminum, ix, 26, 34, 56, 122, 183, 184, 186, 188,
accounting, 180 189, 190, 191, 194, 195, 198, 212, 213, 235
accuracy, vii, ix, 8, 21, 98, 107, 108, 109, 113, 118, aluminum oxide, 235
122, 127, 158, 164, 220, 222 amino, 68
acetate, 87 amino acid, 68
acetone, 13 amino acids, 68
achievement, 108 ammonium, 76, 80, 94
acid, 62, 65, 66, 67, 68, 69, 73, 78, 79, 80, 88, 93, amorphous, 4, 14, 23, 25, 28, 33, 44, 45, 50
94, 95, 96, 218, 219 amplitude, 81, 114
acrylate, 83 anaerobic, 5
ACS, 93, 94 analog, 66, 72, 73, 74, 75, 79, 81, 83, 87
activation, 45, 46, 47, 48, 49, 50, 53, 54, 94, 148, anisotropy, x, xi, 247, 248, 249, 251, 252, 253, 254,
151, 152 255, 256, 257, 259
activation energy, 46, 47, 48, 49, 50, 53, 151, 152 annealing, ix, 24, 28, 31, 32, 33, 34, 37, 38, 58, 155,
ADA, 237 183, 184, 186, 187, 188, 189, 190, 191, 192, 193,
adaptation, 218, 244 194, 195, 196, 197, 198, 199, 202, 203, 204, 205,
additives, 4, 33, 82, 184, 186, 235 213
adhesion, 82, 87, 218, 226, 230, 234, 241, 243 anode, 11
adjustment, 66 ANOVA, x, 217
adolescent patients, 244 ants, 219
adsorption, 148 APL, 236
aerospace, 2 application, vii, x, 1, 2, 4, 24, 45, 53, 54, 83, 86, 90,
AFM, 94 131, 135, 165, 169, 181, 217, 218, 219, 225, 226,
age, 5, 218 227, 231, 249
ageing, 154 aqueous solution, 219
agent, x, 65, 83, 86, 87, 90, 217, 218, 219, 220, 226, argon, 26
227, 229, 230, 231 ART, 133
agents, 74, 82, 83, 92, 218, 220, 226, 227, 231 arthroplasty, viii, 133
aggregates, 68, 72, 81 aspect ratio, 79, 137
aggregation, 63, 72 assessment, 12
aging, viii, 87, 133, 134, 135, 148, 149, 150, 151, assumptions, 9, 47, 48, 104
152, 153, 154, 186 ASTM, 6, 55, 135
aging studies, 87 atmosphere, 27, 31, 36, 37, 38, 40, 53, 57, 205
aid, 11, 85, 174 atoms, 25, 158, 162, 179
aiding, 13 attachment, 8
air, 8, 31, 32, 35, 36, 38, 39, 48, 50, 206, 210, 227, Australia, 182
241 availability, 83, 92
air-dried, 220 averaging, 98, 109, 111, 112, 241
262 Index
cell, viii, 6, 61, 91, 159, 175, 176

B cement, x, 217, 218, 219, 220, 223, 224, 225, 226,
227, 228, 229, 231
barium, 235 ceramic, vii, viii, 2, 3, 4, 7, 15, 19, 21, 44, 47, 49, 51,
barrier, 38, 86, 96 54, 55, 56, 57, 58, 133, 134, 135, 136, 137, 141,
baths, 220 144, 145, 146, 147, 150, 152
beams, 11 ceramics, 2, 3, 15, 21, 23, 50, 58, 59, 60, 139, 144,
behavior, 3, 7, 10, 19, 28, 39, 43, 45, 46, 48, 50, 57, 145, 148, 154
58, 59, 60, 63, 73, 94, 95, 96, 101, 103, 108, 114, chain mobility, 69
131, 146, 149, 168, 190, 193, 196, 235, 241, 243, chemical composition, 4, 152
244 chemical stability, 3
benchmark, 113 China, 1, 54
bending, x, 7, 9, 44, 47, 136, 145, 146, 153, 229, chromium, 137
233, 234, 237 classes, 97
benefits, 63, 65, 73, 75, 86, 92, 158, 218 classical, 7, 98, 168
bias, 98, 174, 175, 177, 179 classification, 248, 249, 251, 252
binary blends, 82, 93 clay, 75, 79, 80
binding, 103, 129 cleavage, 248, 249, 250, 252, 253, 255, 257
binding energy, 129 clusters, 62, 64, 65, 70, 93
biomaterial, 154 CMC, 4
biomaterials, 135 codes, 159
biomedical applications, 134, 149 coding, ix, 157, 179
blends, viii, 61, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, cohesion, 67, 234
92, 95, 96 coil, 202
blocks, 168, 169, 236 colors, 169
boiling, 13 combined effect, 40, 41
bonding, 6, 7, 12, 25, 130, 223, 226, 229, 230, 231 combustion, 58
bonds, 158, 159, 161, 164, 231 combustion environment, 58
boundary conditions, 113 commercialization, 65
Brazil, 233 communication, 150
breakdown, 104 compatibility, 82, 83, 90, 100
brittle polymer, 69 compatibilizing agents, 74, 82, 83, 92
BSR, 8, 9, 10, 11, 29, 34, 37, 39, 45, 46, 47, 48, 50, complexity, 31
51, 52, 54, 59 compliance, 48, 115
bubble, 223 components, viii, 54, 83, 85, 87, 90, 100, 108, 110,
bubbles, 34, 37, 227 133, 134, 137, 138, 145, 153, 235
bulk materials, 25 composites, vii, viii, 1, 2, 4, 54, 55, 56, 58, 59, 60,
82, 95, 97, 98, 99, 100, 113, 122, 125, 127, 128,
129, 135, 149, 182, 234, 235, 238, 243
C composition, 4, 18, 25, 26, 28, 29, 32, 37, 44, 47, 53,
57, 134, 135, 137, 138, 149, 150, 152, 234, 244,
Ca2+, 70, 71, 72, 73, 74
245, 249
calibration, 98, 99, 103, 104, 108, 109, 113, 118, 127
compounds, 103
Canada, 157, 180, 182
compressive strength, xi, 234, 247, 248, 251, 252,
canals, 218, 219, 227, 229, 231
253, 254, 255, 256, 257
candidates, 75, 79, 80
computation, ix, 58, 157, 159, 179
capacity, 229, 235
computer technology, 158
caprolactone, 66
concentration, 18, 19, 21, 31, 37, 44, 45, 62, 64, 66,
carbide, 3, 25, 35, 43, 51, 55, 56, 57, 58, 59, 60
70, 71, 73, 75, 77, 79, 80, 82, 84, 85, 87, 88, 89,
carbon, x, 4, 5, 6, 12, 13, 14, 25, 26, 27, 28, 29, 31,
90, 91, 231, 238
44, 45, 46, 50, 51, 53, 54, 57, 86, 183, 184, 186,
conceptual model, 159
188, 191, 193, 195, 197, 208
concrete, 181
carbon dioxide, x, 28, 44, 86, 183, 184
conductivity, 3, 135, 136
carboxyl, 67, 68
confidence, 167
carboxyl groups, 68
configuration, 100, 153, 218, 228, 258
carboxylic, 68
Congress, 182
carboxylic groups, 68
conservation, 159, 161, 162
cardboard, 6, 7
constraints, 106, 110, 112, 258
cation, 93
consumption, 173
Index 263
contaminant, 185 definition, 21, 164, 178

contaminants, 29, 184, 185, 186, 193, 195, 197 deformation, 10, 59, 101, 106, 146
control, ix, 66, 183, 184, 187, 193, 218, 227, 249 degradation, vii, 1, 4, 25, 26, 27, 28, 29, 31, 38, 40,
convergence, 113, 118 44, 45, 53, 55, 127, 128, 149
conversion, ix, 63, 65, 183, 184, 218 degradation mechanism, vii, 1, 4, 29, 31, 40, 44, 53
cooling, 36, 219, 220 demand, 173, 175
coordination, 62 density, 18, 21, 65, 67, 100, 114, 122, 125, 136, 141,
copolymer, 69, 96 144, 159, 164, 168, 172, 173, 202, 203, 214, 218,
copolymerization, 69 227, 228, 251
copolymers, 65, 93, 95, 96 dentin, 218, 224, 225, 226, 228, 229, 230, 231
copper, 12, 158 dentistry, 92, 245
core-shell, 64 denture, 244
correlation, xi, 20, 24, 31, 35, 47, 139, 150, 247, 249, dentures, 235, 244
255, 256, 257 Department of Energy, 180
correlations, xi, 145, 247, 255, 257 Department of Homeland Security, 180
corrosion, 43, 44, 45, 54, 57, 58 deposition, 27, 31
corrosive, 54 deposits, 259
costs, 186 derivatives, 110
couples, 111, 147 deviation, 144, 153, 225
coupling, 101, 147, 180 dew, 31
covalent, 63 diamond, 152, 219, 221, 236
coverage, 243 diffraction, 11, 144, 151
covering, 223, 234 diffusion, 25, 27, 44, 46, 50, 53, 54, 60
crack, vii, ix, 1, 2, 3, 4, 15, 18, 23, 43, 44, 45, 70, 85, diffusion process, 53
103, 106, 129, 130, 139, 140, 154, 158, 165, 168, diffusivity, 25, 44, 190
170, 171, 174, 181 dilation, 163
cracking, 2, 46, 98, 241, 243 disaster, 150, 170, 172
CRC, 154, 181 discontinuity, 100, 101, 102, 103, 104, 105, 114,
creep, 3, 5, 6, 8, 9, 10, 29, 30, 31, 32, 33, 34, 35, 37, 115, 116, 158, 168
39, 44, 45, 46, 47, 48, 50, 51, 52, 53, 54, 55, 57, discretization, 107
58, 59, 60, 72, 90 discs, x, 217
creep tests, 30, 51 dislocation, 46, 53, 230
critical temperature, 148 dispersion, 7, 74, 75, 83, 125, 241
crosslinking, 64, 67, 68, 70 displacement, 6, 98, 99, 100, 101, 103, 104, 106,
cross-sectional, 202, 205, 214 107, 114, 115, 116, 118, 126, 127, 163, 176, 242
crystal growth, 4 disposition, 144, 248
crystal structure, 11 distilled water, 219, 220
crystalline, 11, 33, 73, 74, 75, 79 distribution, viii, 7, 12, 13, 15, 55, 56, 67, 109, 110,
crystallinity, vii, 1, 4, 5, 6, 31, 37, 44, 53, 63, 65, 69, 111, 133, 146, 149, 194, 201, 218, 226, 227, 234,
73, 83, 87, 88, 90 235, 237, 238, 240, 241, 243
crystallites, 93 distribution function, 55
crystallization, 14, 21, 25, 29, 33, 50, 53, 63, 69, 73, divergence, 98
75, 87, 94, 95 domain structure, 206, 214
crystals, 11, 25, 28, 34, 45, 60, 69, 137 dominance, 94
curing, 5, 55, 227 doped, 137, 138
CVD, 46, 50, 51 doping, 142
cycles, 70, 147, 153, 154, 186, 220 draft, 134
cycling, 153, 154 drying, 6, 11
ductility, 81
DuPont, 65
D durability, vii, 1, 3, 4, 5, 53, 54
duration, 48, 150, 151, 153
data analysis, 258
data generation, 143
data set, 20 E
decay, 7
decomposition, 5, 25, 26, 28, 31, 44, 45, 50 earth, 72, 73, 172, 174
defects, 6, 18, 19, 145, 160, 240 elasticity, 23, 82, 101, 161, 218, 235, 237
deficiency, 174 elasticity modulus, 235, 237
264 Index
electricity, ix, 183, 184 fiber, vii, viii, x, 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13,
electron, x, 5, 11, 31, 55, 195, 201, 217, 223 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26,
electron beam, 5, 55 27, 28, 29, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
electron microscopy, 11 41, 42, 43, 44, 45, 46, 47, 48, 50, 51, 53, 54, 55,
elongation, 8, 66, 67, 68, 74, 76, 79, 80, 81, 84, 89, 56, 57, 58, 59, 60, 97, 99, 127, 128, 131, 217, 219,
90, 91, 210 220, 226, 227, 229, 230, 231, 233, 234, 235, 236,
emission, x, 183, 184 237, 239, 240, 241, 242, 243, 244, 245
energy, viii, ix, 2, 3, 7, 11, 22, 23, 30, 31, 46, 53, 56, fiber content, 240, 241, 243
75, 79, 80, 87, 97, 103, 104, 105, 142, 157, 158, fibers, vii, x, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13,
159, 161, 162, 167, 168, 173, 174, 179, 183, 184 14, 15, 16, 18, 19, 20, 21, 23, 24, 25, 26, 27, 28,
energy consumption, 173 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41,
engines, 2, 54 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 53, 54, 55,
England, 57 56, 57, 58, 59, 218, 233, 234, 235, 236, 237, 238,
entanglements, 72 239, 240, 241, 243, 244
entropy, 82 field-emission, 31
environment, ix, 4, 6, 8, 9, 31, 85, 94, 145, 148, 149, filament, 13, 14
150, 151, 183, 184 filler particles, 92
environmental conditions, 10 filler surface, 74
epoxy, ix, 122, 158, 165, 181, 218, 220, 236 fillers, 63, 65, 74, 75, 81, 92
epoxy resins, 218 film, 7, 32, 34, 36, 43, 44, 45, 77, 220, 229
equality, 159 film thickness, 229
equilibrium, 45, 100, 104, 161, 162, 163, 170 films, 74, 95
ester, ix, 158, 165, 167, 235 filters, 128, 129
esthetics, 234 finite element method, 106, 129, 130, 158, 159
estimating, 21 first generation, 4
estimator, 131 fission, 54
estimators, 129 flexibility, 4
etching, 227 flexural strength, x, 218, 230, 233, 234, 237, 243
ethylene, 73, 93, 95, 96 flow, 18, 43, 56, 169, 200, 202, 204, 205, 218, 227
ethylene vinyl alcohol, 96 flow rate, 43
ethylenediamine, 67 Ford, 229
Eulerian, 176 forecasting, 165
Euro, 56, 59 fracture, vii, ix, 1, 2, 3, 4, 7, 12, 15, 16, 17, 19, 21,
EVOH, 84, 85, 86, 96 22, 23, 27, 28, 29, 30, 31, 38, 42, 43, 44, 45, 56,
evolution, 4, 26, 28, 37, 57, 59, 98, 108, 114, 115, 70, 79, 80, 85, 87, 94, 96, 103, 104, 105, 128, 129,
116, 118, 122, 187, 258 134, 139, 141, 144, 157, 158, 159, 164, 165, 168,
EXAFS, 72 171, 173, 174, 176, 179, 180, 181, 182, 208, 210,
examinations, 39 214, 225, 229, 234, 238, 241, 244, 251
exfoliation, 74, 75, 76, 79, 80, 81, 92 fractures, 8, 41, 218
expansions, 112, 113 fragmentation, vii, ix, 157, 158, 163, 165, 179, 181,
expert, 259 182
exploitation, 259 France, 150, 217, 218, 219, 220, 222, 224, 226, 228,
exposure, 9, 10, 25, 26, 28, 31, 35, 40, 41, 53, 57, 230
149, 218 FRC, 218, 226
extraction, 249 friction, 202, 226, 230
fuel, viii, 61, 91
fuel cell, viii, 61, 91
F functionalization, 65, 92
funding, 180
fabric, 257 fusion, 54, 56
fabricate, 171
fabrication, 2, 3, 4, 18, 24, 25, 26, 53
failure, vii, viii, x, 2, 6, 7, 8, 23, 31, 43, 46, 73, 80, G
81, 82, 97, 98, 113, 114, 115, 116, 127, 136, 139,
144, 149, 153, 158, 163, 164, 166, 170, 172, 180, Gamma, 8
217, 223, 225, 228, 230, 234, 245, 248 gas, 2, 3, 26, 28, 29, 37, 43, 44, 45, 47, 49, 57, 58,
fatigue, 63, 69, 70, 71, 90, 136, 154 84, 86, 151, 235
FEM, 158, 159, 258 gas barrier, 84, 86
ferrite, 195 gas turbine, 2, 3
Index 265
gases, 28, 45 histological, 218

gauge, 6, 8, 10, 15, 187, 202, 213 Homeland Security, 180
Gaussian, 108, 111, 113, 129 homogenized, 101
Gaussian random variables, 113 homogenous, 250
generation, 5 Honda, 46, 59, 244
Geneva, 219 Hong Kong, 244
geology, 249 hot-rolled, 194
Germany, 1, 181, 219, 220, 223, 235, 245 human, 148, 149, 174, 219
GIS, 259 humidity, 84
glass, viii, x, 11, 15, 21, 56, 63, 94, 95, 97, 99, 127, hybrid, 175, 176, 179, 180, 181, 184, 202
131, 233, 234, 235, 236, 237, 238, 239, 240, 241, hybridization, 231
242, 243, 244, 245 hydrocarbon, 68
glass transition, 63, 94, 95 hydrodynamics, 159, 181
glass transition temperature, 63 hydrothermal, viii, 133, 134, 135, 148, 149, 150,
glass-fiber, 231 151, 152, 153
glycerin, 7 hydroxide, 148, 149
glycol, 66 hysteresis, 184, 189, 191, 202, 206, 207, 214
goals, 62, 83, 90
GPA, 52
grades, 79, 81, 185 I
grain, ix, 5, 6, 9, 14, 19, 24, 25, 26, 27, 28, 29, 31,
34, 38, 39, 42, 43, 44, 45, 46, 47, 48, 50, 53, 54, identification, 11, 24, 46, 97, 122, 127, 128, 130
137, 140, 141, 144, 148, 172, 173, 183, 184, 187, image analysis, 12, 13
189, 190, 192, 193, 194, 195, 196, 197, 198, 203, images, x, 42, 43, 57, 208, 209, 214, 233, 238
204, 205, 213, 248, 249, 250, 251, 253 imagination, 92
grain boundaries, 14, 25, 29, 39, 44, 53, 172, 173, impact energy, 100, 127, 165, 166
190, 196, 198, 213 impact strength, 87, 88, 89, 90
grains, ix, 6, 25, 26, 27, 28, 29, 34, 42, 50, 183, 184, implementation, ix, 113, 137, 157, 159, 177, 179
187, 189, 190, 191, 193, 195, 196, 198, 203, 213 impregnation, 234, 238
granites, 252, 259 impurities, 18, 29, 46, 57, 241
graph, 257 in situ, 59, 74, 95
graphite, 26, 28 in vitro, 226, 227, 229
groups, viii, x, 61, 63, 64, 65, 66, 67, 68, 69, 70, 72, in vivo, 146
73, 75, 76, 79, 95, 217, 220, 225, 226, 238, 241 incentives, 74
growth, ix, 4, 25, 27, 28, 29, 34, 42, 43, 44, 45, 50, incidence, 229
54, 104, 106, 127, 129, 130, 144, 150, 154, 183, incineration, 240
184, 193, 195, 198, 203, 204, 205, 253 inclusion, 16, 17, 18, 38
India, 214
induction, 184, 185, 186, 187, 189, 190, 192, 194,
H 195, 197, 198, 199, 206, 210, 212, 213, 214
industrial, viii, ix, 61, 63, 83, 92, 183, 184, 206, 236
handling, 13, 29, 202, 218 industry, 81
hardness, xi, 67, 89, 135, 136, 137, 141, 144, 146, inelastic, 178
147, 186, 218, 230, 247 inert, 53
HDPE, 81, 82, 83, 84, 85, 95, 96 inertia, 179
healing, 104 infinite, 37
heat, 3, 24, 25, 26, 27, 28, 29, 30, 31, 48, 49, 50, 53, infrastructure, 169, 172
57, 58, 90, 172, 174, 186 inhomogeneities, 125
heating, 172, 173 initial state, 154
height, 200, 208, 236, 238 initiation, 3, 15, 139, 140, 180
heterogeneity, 169 injection, 227
heterogeneous, 173, 180 innovation, ix, 183
high temperature, vii, viii, 1, 2, 3, 4, 5, 10, 24, 25, inorganic, 93, 135, 200, 213, 234, 235
28, 31, 36, 38, 39, 40, 43, 49, 50, 53, 54, 55, 57, inorganic filler, 235
58, 59, 60, 133, 142 inorganic fillers, 235
high-speed, 202, 219 insight, 10, 69, 173
hip, vii, viii, 133, 135 instabilities, 108
hip arthroplasty, 133 instability, 5
hip joint, vii, 133 Instron, 6, 7, 237
266 Index
insulation, 200, 201, 213

integration, 99, 107, 130 L
integrity, 218
intensity, 2, 11, 140, 149, 168 Lagrangian, 176, 179
interaction, 53, 76, 80, 101, 103, 104, 114, 148, 152, lamina, 18, 99, 101, 114, 115, 122, 127
159, 160, 162, 170, 171, 175, 176, 177, 180, 225 laminar, 18, 127
interactions, ix, 62, 66, 74, 75, 94, 157, 159, 175, laminated, 128, 129, 250
176, 178, 180 laminated composites, 128
intercalation, 74 lamination, 185, 189, 194, 202, 205
interface, viii, 2, 56, 97, 98, 100, 101, 103, 104, 105, lattice, ix, 46, 53, 148, 149, 157, 159, 161, 170, 174,
108, 114, 115, 116, 117, 118, 120, 121, 124, 126, 175, 176, 179, 180, 181
127, 128, 131, 173, 218, 224, 225, 226, 228, 230, lattices, 180
234, 237, 241, 242 law, 11, 19, 20, 50, 101, 102, 103, 104, 105, 113,
interfacial adhesion, 82, 87 114, 116, 117, 118, 119, 120, 121, 123, 126, 131,
interfacial tension, 82, 88 176, 177, 179
interference, 251 laws, viii, 97, 98, 101, 104, 108, 113, 115, 116, 118,
interphase, 3, 101, 103 127, 128, 159, 176, 180
interpretation, 251 lead, 18, 46, 63, 72, 76, 80, 81, 83, 84, 85, 86, 91,
interval, 11, 13, 14, 107, 108, 110, 118 92, 104, 118, 184, 186, 195, 197, 235, 243
intrinsic, 6, 73, 74, 144, 145, 146, 158, 251 learning, 91
intrinsic viscosity, 73 LEO, 223
invasive, 234 liberation, 172, 173
ionic, viii, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 72, lifetime, 71
73, 75, 83, 88, 92, 93, 94 limestones, 257
ionic forces, 63 limitation, 4, 222
ionization, 66, 67 limitations, 227, 229
Ionomer, 91, 96 linear, 8, 9, 20, 21, 23, 46, 47, 64, 101, 102, 103,
ions, 25, 62, 64, 65, 69, 72, 73, 75, 79, 80, 83, 94, 104, 107, 108, 116, 117, 119, 120, 121, 123, 131,
106, 148, 149, 180 159, 163, 176, 180, 181, 253, 255, 257
iron, 198 linear law, 102, 117, 119, 120, 121, 123
irradiation, 55 linear polymers, 64
irrigation, 219 linear systems, 108, 131
ISO, 134, 135, 136, 145, 149, 150, 152, 219 linkage, 54, 159, 174, 180
isotropic, 9, 114, 174, 175, 180, 258 liquid crystalline polymers, 73
Italy, 97, 217, 219, 220 liquid phase, 58
lithium, 56, 95
location, 7, 12, 22, 64, 109, 179
J London, 56, 95, 130, 257
long-term, 46
January, 181 losses, ix, 183, 184, 185, 206
Japan, 4, 6, 183, 215, 251 Louisiana, 181
Japanese, 54, 200 low density polyethylene, 95
low molecular weight, 73
low temperatures, 53, 150, 152
K low-density, 96
lubrication, 200, 201, 214
Kalman filter, viii, 97, 98, 99, 108, 109, 127, 128,
129, 130, 131
Kalman Filtering, 108 M
kernel, 159
kerosene, 200 machines, 250
kinematics, 100, 101, 105 magnet, 186
kinetic energy, 23 magnetic, ix, 183, 184, 185, 186, 187, 188, 189, 190,
kinetics, 31, 33, 50, 57, 60, 148, 149, 150, 153, 154 191, 192, 193, 194, 195, 197, 198, 200, 202, 203,
King, 93, 94, 95 204, 205, 206, 207, 208, 209, 210, 212, 213, 214
magnetic field, 208
magnetic properties, ix, 183, 184, 185, 187, 188,
190, 191, 193, 194, 200, 202, 203, 206, 212, 213
magnetization, 202, 206
Index 267
manganese, ix, 183, 184, 186, 188, 191, 192, 193, mixing, 82, 83, 88
195, 198, 208, 210, 212, 213 MMT, 74, 75, 76, 77, 78, 79, 80, 81, 92
manipulation, 235 mobility, 62, 63, 64, 65, 69
manufacturer, x, 14, 142, 217, 219, 228 modeling, vii, ix, 97, 100, 127, 129, 145, 157, 158,
manufacturing, x, 6, 145, 183, 184, 185, 200, 206, 159, 160, 165, 167, 174, 175, 176, 177, 178, 179,
208 180, 181, 182
mapping, 108, 109, 110, 113 models, 99, 104, 113, 114, 118, 128, 129, 131, 148,
market, 92 158, 159, 162, 168, 175, 180, 181
Mars, 172 modulus, ix, 2, 5, 8, 12, 15, 23, 40, 41, 47, 48, 55,
Massachusetts, 245 67, 68, 70, 71, 75, 76, 78, 79, 80, 81, 82, 83, 84,
masticatory, 220 85, 86, 87, 88, 93, 114, 135, 136, 145, 157, 159,
material degradation, 149 161, 162, 163, 171, 175, 179, 180, 218, 230, 235,
mathematics, 58 248, 258
matrix, 2, 3, 4, 46, 54, 56, 60, 65, 78, 79, 80, 81, 99, moisture, 64, 218, 227, 230, 251
101, 104, 106, 107, 110, 111, 112, 118, 129, 136, moisture content, 251
137, 139, 142, 144, 172, 173, 218, 220, 234, 235, mold, 236
237, 238, 240, 241, 242, 244, 245 molecular dynamics, ix, 157, 158, 159
maxillary, x, 217, 219 molecular weight, 66, 67, 68, 73
MDA, 67, 68 molecular weight distribution, 67, 68
meanings, 62 molecules, 148
measurement, x, 21, 45, 47, 108, 118, 150, 176, 230, monolithic, 2, 3
240, 247 monomer, 75, 235
mechanical behavior, 96, 101, 244 monomeric, 79
mechanical energy, 174 monomers, 63
mechanical properties, vii, 1, 3, 4, 6, 11, 25, 40, 57, montmorillonite, 95
69, 77, 82, 86, 93, 95, 96, 114, 140, 141, 144, 186, Moon, 172
210, 211, 218, 234, 244, 251 morphological, 96
mechanical stress, 140, 154 morphology, 3, 34, 42, 43, 57, 76, 93, 94, 95, 96, 230
melt, 5, 63, 74, 81, 83, 88 motion, ix, 23, 99, 106, 107, 108, 161, 179, 183, 184
melting, 3, 34, 69, 73, 74, 95 motion control, 184
melts, 58 motors, ix, x, 183, 184, 185, 186, 206, 208, 210
membranes, 91, 93 movement, 100, 222
metallurgy, 215 MWD, 67
metals, 103, 129, 130
methacrylic acid, 62, 65, 78, 79, 80, 93, 95
methyl methacrylate, 90, 95, 96 N
methylene, 66
Mg2+, 71, 72, 73, 74 nanocomposites, viii, 61, 65, 74, 75, 76, 77, 78, 79,
microcracking, 172, 173 80, 81, 82, 92, 95
micrometer, 158 nanocrystals, 5, 6, 25
microscope, x, 31, 195, 217, 223 nanoindentation, 230
microscopy, x, 11, 233, 236, 238, 239 nanotubes, 92
Microsoft, 238 naphthalene, 95
microstructure, vii, viii, x, 1, 4, 9, 11, 15, 25, 27, 28, natural, 145, 153, 218, 250, 251
29, 31, 34, 35, 36, 38, 40, 41, 43, 45, 47, 48, 50, NCS, 48, 50, 51, 52, 54
53, 54, 60, 63, 133, 135, 137, 139, 140, 141, 144, needles, 98
145, 233, 235, 249 network, ix, 157, 159, 174, 179, 180, 181
microstructure features, 31, 41 neural network, 98, 130, 131
microstructures, x, 6, 51, 204, 205, 233 neural networks, 98, 130, 131
microwave, 172, 173, 174 neutralization, 69
microwave radiation, 173 New Jersey, 95
migration, 33, 44, 148 New Orleans, 181
mimicking, ix, 157, 176 New York, 93, 94, 95, 245
mineralogy, 257 Newton, 6, 176, 179
minerals, 172, 173, 249, 250 nickel, 70, 93
mining, 174, 248, 249, 257, 259 nitride, 58, 59, 189, 193, 194
Minnesota, 181 nitrides, 188, 189, 191, 194, 197, 198
mirror, 15, 16, 17, 20, 21, 22, 23, 27, 28, 42, 44 nitrile rubber, 88, 96
Mississippi, 157, 165, 167, 182
268 Index
nitrogen, 67, 184, 188, 189, 190, 191, 193, 195, 198, parameter, viii, 7, 8, 9, 10, 39, 47, 48, 49, 50, 85, 97,
205, 213 98, 103, 104, 107, 110, 111, 118, 127, 128, 131,
nodes, ix, 157, 159 249, 251
nodules, 250 parameter estimates, 98, 107, 110, 118
noise, 108, 113 parameter estimation, viii, 97
nonionic, viii, 61, 73 particles, viii, ix, 74, 80, 81, 92, 133, 137, 139, 140,
nonlinear, 98, 103, 108, 109, 110, 129, 131, 159, 142, 144, 157, 159, 160, 161, 163, 164, 176, 179,
160, 161, 168, 176, 177, 178, 179, 180 180, 234, 235, 240
nonlinear systems, 108, 129, 131 passive, 29, 31, 34, 35, 36, 37, 38, 43, 44, 53
nonlinearities, 98, 109, 127 patents, 61
normal, 7, 22, 99, 114, 141, 142, 144, 145, 146, 147, PBT, 75
248 PCS, 5, 26
nuclear, 2, 3, 54 perforation, 98
nuclear energy, 2 performance, x, 2, 3, 15, 24, 26, 31, 40, 54, 109, 113,
nuclear reactor, 54 128, 129, 133, 139, 140, 141, 145, 146, 147, 200,
nucleation, 44, 53, 75, 150 201, 213, 214, 218, 229, 248
numerical analysis, 128 periodic, 159
nylon, ix, 96, 158 permeability, 34, 44, 195, 197
nylons, 96 permeation, 36
PET, 73, 74, 75, 86, 87
phase transformation, 26, 134, 139, 140, 141, 142,
O 149, 150, 151, 152, 153, 154
phenolic, 220
observations, ix, 25, 27, 29, 108, 158, 229, 238, 250 phosphides, 195, 196
obstruction, 204, 205 phosphorus, ix, 183, 184, 186, 187, 188, 191, 193,
Ohio, 61 195, 196, 198, 208, 212, 213
oil, 79, 84, 200 photographs, 27, 236
on-line, 244 physical properties, 67, 69, 73, 87, 96, 135, 136
operator, 100 physics, 55, 98
optical, x, 11, 63, 69, 73, 84, 230, 233, 236 pinning effect, 193, 213
optical microscopy, 11, 236 planar, 11, 248, 249
optimization, 2, 45, 257 plane waves, 113
Ordovician, 248, 249 planets, 172
ores, 173, 174 planning, 248, 257, 259
organ, 57, 78 plasma, 235
organic, 93, 200, 201, 213, 214, 234 plastic, viii, 7, 88, 97, 99, 127, 168
organization, 25, 248 plastics, 81
organoclay, 78 platelet, 137
organoclays, 95 platelets, 81, 137
organometallic, 55, 57 play, 3, 25, 201, 214
orientation, 99, 101, 184, 187, 218, 234, 248, 250, PLS, 257
257 Poisson, 114, 161, 174, 175, 176, 177, 180
oscillations, 107 Poisson ratio, 161, 174, 175, 177
oxidation, 4, 5, 29, 31, 32, 33, 34, 35, 36, 37, 38, 39, Poland, 181
40, 43, 44, 45, 50, 53, 57, 58, 59 polarity, 62, 63, 64, 68
oxidative, vii, 1, 31, 38, 54, 55 poly(ethylene terephthalate), 73, 95, 96
oxide, 4, 29, 33, 76, 96, 137, 235 polyamide, 96
oxides, 137 polycrystalline, 5, 9, 10, 23, 46, 47, 59, 60
oxygen, vii, 1, 4, 5, 6, 25, 28, 29, 31, 34, 35, 36, 37, polyester, 73, 95
38, 39, 40, 43, 44, 45, 50, 53, 55, 56, 57, 58, 60, polyethylene, x, 65, 87, 95, 96, 233, 234, 235, 236,
86, 148, 149, 151 237, 238, 239, 240, 241, 242, 243, 244
polyisoprene, 71, 72, 94
polymer, vii, viii, 29, 61, 62, 63, 64, 65, 66, 69, 72,
P 73, 74, 75, 78, 79, 81, 82, 83, 85, 92, 93, 129, 131,
218, 234, 245
paper, 6, 160, 179, 184, 215, 219, 220, 237, 247, 249 polymer blends, 82, 83, 92
parabolic, 33, 50 polymer chains, 63, 64, 72
polymer composites, 131
polymer matrix, 79, 129, 218, 234, 245
Index 269
polymer structure, 64
polymeric materials, ix, 56, 158, 165 R
polymerization, 63, 74, 218, 220, 227, 231
polymers, vii, viii, 61, 62, 63, 64, 65, 66, 68, 70, 72, radiation, 5, 11, 55, 173
73, 74, 75, 79, 80, 82, 83, 84, 85, 86, 88, 92, 218, radius, 9, 16, 20, 22, 23, 223
244 Raman, 25, 26, 57, 140, 147, 149, 155
polynomial, xi, 159, 176, 177, 178, 179, 247, 253, Raman spectroscopy, 26
254, 255 random, 6, 64, 73, 94, 110, 111, 113, 181
polynomial functions, 255 range, 11, 21, 31, 36, 44, 46, 57, 58, 61, 63, 64, 65,
polyolefin, 96 79, 135, 145, 148, 158, 163, 180, 188, 189, 195,
polypropylene, 66, 85, 96 198, 206, 213, 214, 235, 238, 249, 252, 257
polystyrene, 65, 68, 69, 70, 71, 93, 94 rare earth, 72, 73
polyurethane, 66, 76, 94, 95 Rayleigh, 57
polyurethanes, 66, 67, 68, 76, 94 reality, 168
poor, 7, 15, 73, 118, 200, 234, 235, 257 reasoning, 85
population, 6, 38, 42, 45 recovery, 59
pore, 16, 17, 38, 41, 42, 43 recrystallization, 184, 195
pores, 42 recrystallized, 197
porosity, 227, 231, 237, 251, 257 reduction, ix, 13, 63, 66, 69, 80, 81, 84, 101, 127,
porous, 27, 28, 34, 55 183, 187, 212, 213, 227, 231
Portugal, 252 reference frame, 104
potential energy, ix, 23, 157, 162, 179 reflection, 67
powder, 11, 142, 143 refrigeration, 236
power, ix, 3, 9, 10, 19, 20, 46, 47, 158, 168, 172, regional, x, 217, 219, 231
173, 183, 184 regression, 23
power plant, ix, 183, 184 regression analysis, 23
power plants, ix, 183, 184 regular, 134, 141, 144, 145, 152, 174, 185, 187, 188,
powers, 173 208, 209, 210, 211, 212, 213, 214
precipitation, ix, 183, 189, 193, 195 reinforcement, x, 2, 3, 4, 31, 53, 56, 81, 95, 134, 140,
prediction, 51, 59, 127, 131, 150 147, 149, 233, 234, 235, 236, 237, 240, 244, 245
preference, 21 reinforcing fibers, 3, 218
pressure, 29, 31, 34, 35, 36, 37, 38, 39, 40, 43, 44, relationship, 8, 21, 46, 48, 49, 53, 101, 184, 187,
45, 57, 58, 64, 152, 219, 220, 244, 248 188, 191, 194, 197, 198, 201, 202, 210, 235, 243,
prevention, 186 253, 254, 255, 257
probability, 7, 8, 109, 110, 111, 131, 159, 240 relationships, 91, 94, 96, 102, 257
probability distribution, 109, 110, 111, 131 relative size, 22
probe, 201 relaxation, 8, 9, 10, 30, 39, 44, 47, 48, 49, 50, 54, 59,
production, 18, 26, 134, 140, 142, 144, 185, 208 64, 72, 81, 93
productivity, 208 reliability, vii, viii, 1, 4, 133, 153
program, x, 233, 237, 239 research, vii, viii, ix, 54, 61, 62, 63, 65, 66, 83, 91,
propagation, vii, 1, 2, 4, 15, 23, 31, 44, 45, 85, 98, 92, 158, 171
100, 107, 113, 114, 130, 139, 140, 154, 165, 170, research and development, 83
171, 172, 174, 179, 180, 181, 241, 251 researchers, 44, 47, 51, 66, 73, 75, 79, 80, 81, 248
property, vii, viii, 61, 70, 84, 93, 94, 144, 161, 180, resin, x, 80, 217, 218, 219, 220, 223, 224, 225, 226,
248, 249 227, 228, 229, 230, 231, 234, 235, 236, 240, 243,
propulsion, vii, 1, 2, 53 244
propylene, 96 resins, 101, 235
prostheses, 154, 234, 243, 244 resistance, 3, 4, 5, 6, 8, 29, 39, 40, 44, 45, 47, 50, 53,
protocol, 227 57, 58, 63, 65, 69, 70, 71, 79, 84, 90, 139, 154,
pseudo, 97, 127 169, 227, 229, 230, 234, 235
pyrite, 172, 173, 174 resistivity, 188, 189, 191
pyrolysis, 5 resolution, 11
restorations, 229, 244, 245
retardation, 193, 195
Q retention, vii, 1, 4, 26, 29, 37, 38, 40, 218, 226, 227,
228, 229, 231
quartz, 220, 226, 250 rheology, 96
quaternary ammonium, 76, 94 rigidity, 234
Quebec, 180 risk, vii, 134, 145, 147, 218
270 Index
robustness, 99 SiO2, 32, 34, 36, 38, 39, 43, 82, 137, 249
rolling, 187, 190, 195, 210, 211, 213 sites, 62, 64, 66, 67
room temperature, 9, 10, 24, 26, 40, 47, 64, 142, 220 sodium, 75, 93, 95, 96, 219
rubber, 70, 88 software, 220
solid phase, 142
solubility, 190
S solutions, 160
solvent, 90
SAE, 214 South Africa, 258
safety, 146 space exploration, 174
saline, 220 space-time, 108, 114, 122, 125, 129
salt, 65, 67, 93 Spain, 247, 248, 249, 252
salts, 72, 80, 93, 95, 96 spatial, 11
sample, 8, 11, 16, 17, 110, 113, 174, 222, 223, 238, species, 4, 5, 29, 44, 45, 53
240, 248 spectroscopy, 18, 149
sampling, 11 speed, ix, 107, 122, 125, 158, 167, 168, 169, 170,
sandstones, 250, 251, 252 172, 186, 200, 208, 210, 222, 237
saturation, 150, 248, 250, 251 spin, 27
scaling, 111 springs, ix, 157, 175
scatter, viii, 12, 19, 20, 21, 133, 144, 145, 153 SPT, 111, 112
scattering, 11, 21, 51, 93, 118 SRT, 193, 213
scientists, 158 S-shaped, 48
SCS, 46 stability, vii, 1, 3, 4, 5, 6, 19, 25, 27, 28, 29, 31, 40,
sea ice, 181 53, 54, 55, 56, 57, 58, 59, 60, 75, 87, 98, 107, 108,
search, 158 113, 118, 131, 226
Seattle, 131 stabilization, viii, 133, 135, 137, 141, 142, 148, 235
seismic, 251 stabilize, 29, 70, 220
selecting, 249 stages, 9, 33
SEM, 7, 11, 12, 13, 16, 18, 21, 27, 31, 42, 43, 50, 73, standard deviation, 12, 13, 118, 238, 240, 241
208, 209, 214, 223, 224, 225, 228, 229, 230, 231 standards, 134, 136, 152
SEM micrographs, 16, 21 statistical analysis, x, 153, 233
semi-crystalline polymers, 63 statistics, 19, 98, 108, 109
sensitivity, 144 steel, 185, 187, 188, 189, 190, 191, 192, 193, 195,
sensors, 10 196, 197, 198, 201, 202, 204, 205, 208, 209, 210,
separation, 13, 66, 103, 147 211, 212, 213, 214, 215
series, 72, 79, 80, 82, 108, 110, 112, 113, 154, 236 sterile, 220
serum, 150 stiffness, vii, 1, 4, 5, 80, 81, 101, 103, 104, 106, 118,
shape, 3, 4, 22, 42, 44, 56, 106, 113, 114, 116, 131, 136, 163
139, 159, 195, 200, 201, 202, 207, 237, 238, 250 Stochastic, 56, 98, 131
sharing, 50 stoichiometry, 4, 5, 6, 53
shear, vii, 82, 226, 241 storage, 71, 81, 244
shock, 7, 98, 107, 114, 115, 116 strain, vii, ix, 7, 8, 9, 10, 23, 31, 46, 47, 48, 50, 51,
shock waves, 98, 107, 116 53, 54, 73, 100, 142, 157, 169, 176, 179
Si3N4, 46, 55 strains, 9, 10, 51
signs, 81 stratification, 250
silane, 230 strength, vii, viii, ix, x, xi, 1, 2, 3, 4, 5, 6, 7, 8, 12,
silica, 29, 32, 33, 34, 36, 37, 39, 43, 44, 45, 50, 57, 15, 16, 19, 20, 21, 22, 23, 24, 25, 26, 28, 29, 37,
80, 81, 244 38, 40, 41, 44, 45, 53, 55, 56, 57, 58, 61, 63, 66,
silicate, 56 67, 68, 69, 78, 79, 80, 81, 83, 84, 85, 86, 89, 90,
silicates, 74 91, 101, 103, 113, 116, 117, 125, 127, 130, 131,
silicon, 3, 25, 35, 43, 46, 50, 51, 53, 54, 55, 56, 57, 133, 134, 136, 137, 139, 141, 144, 145, 146, 149,
58, 59, 60, 184, 186, 187, 188, 191, 193, 195, 197, 152, 153, 154, 157, 158, 159, 161, 162, 163, 171,
203, 205, 208, 212, 213 176, 179, 186, 201, 202, 203, 214, 217, 218, 219,
Silicon carbide, 57 222, 225, 226, 227, 228, 229, 230, 231, 233, 234,
simulation, 128, 129, 130, 146, 158, 165, 166, 167, 235, 237, 238, 241, 242, 243, 244, 247, 248, 249,
172, 174, 179, 180, 181, 182 250, 251, 252, 253, 254, 255, 256, 257, 258, 259
simulations, ix, 107, 127, 129, 145, 158, 164, 169, stress, vii, viii, ix, x, 2, 3, 8, 9, 10, 18, 19, 21, 26, 27,
181 28, 30, 31, 34, 36, 37, 39, 43, 44, 45, 46, 47, 48,
sintering, viii, 19, 26, 27, 28, 133, 144 49, 50, 51, 53, 54, 59, 64, 72, 76, 81, 86, 87, 88,
Index 271
97, 100, 101, 115, 139, 140, 146, 147, 148, 153, test procedure, 8
154, 161, 168, 172, 176, 183, 184, 188, 189, 190, Texas, 180
191, 192, 194, 195, 196, 197, 198, 199, 202, 203, theory, ix, 2, 7, 21, 22, 23, 129, 131, 157, 159, 160,
204, 205, 206, 207, 208, 209, 210, 214, 218, 226, 174, 176, 179, 180, 181, 182
230, 231 thermal activation, 149, 150
stress level, 140, 153, 154 thermal decomposition, 28, 44, 45
stretching, 101 thermal expansion, 19, 26, 36, 173
strokes, 200, 201, 208 thermal properties, 83, 87
strontium, 137 thermal stability, vii, 4, 5, 6, 25, 27, 31, 40, 59, 60,
structural characteristics, 229, 230 75, 87
styrene, 69 thermal treatment, 11, 138
substances, 135 thermoplastic, 63, 64
subtraction, 203 thermoplastics, 86
sulfur, 188, 191, 192, 193, 195, 198, 208, 213, 214 thin film, 69
sulphur, 250 thin films, 69
surface treatment, 230, 231 third order, 112, 113
Surgery, 134, 135, 147 three-dimensional, 60, 99
surgical, 153 threshold, 104, 148
surviving, 153 time, viii, ix, 8, 9, 10, 11, 39, 40, 41, 44, 45, 46, 47,
Switzerland, 219, 220 48, 49, 50, 51, 53, 54, 63, 64, 72, 83, 84, 92, 98,
symbols, 99 99, 100, 107, 108, 109, 110, 111, 112, 114, 115,
symmetry, 101 116, 118, 120, 121, 122, 123, 124, 126, 130, 131,
synchronous, 186 133, 149, 150, 151, 152, 153, 158, 159, 164, 167,
synergistic, 54 170, 173, 208, 220
synthesis, 75, 83 time use, 150
systems, vii, 1, 2, 53, 54, 56, 63, 64, 65, 66, 68, 69, TiO2, 137
79, 81, 92, 127, 159, 175, 181, 186, 227, 229, 230, tissue, 147
231 titanium, ix, 183, 184, 193, 195, 196, 197, 213
Tokyo, 215
tolerance, 147
T toughness, vii, viii, 1, 2, 3, 4, 22, 30, 31, 65, 69, 70,
79, 82, 86, 113, 116, 117, 130, 133, 134, 135, 137,
target response, 118 139, 141, 144, 149, 163
targets, 86, 92 tracking, viii, 97, 99, 108, 118
Taylor series, 108, 110, 113 traction, 2, 99, 100, 101, 103, 104, 105, 114, 115,
technology, viii, ix, 61, 183, 184 116
teeth, 218, 229, 231, 234 trans, 72, 75
TEM, 25, 26, 75, 76, 81, 189, 193, 195, 198 transfer, 46, 149, 220
temperature, vii, viii, 1, 2, 3, 4, 5, 8, 9, 10, 19, 24, transformation, viii, 26, 33, 110, 111, 112, 129, 131,
25, 26, 27, 28, 29, 31, 34, 35, 38, 39, 40, 41, 42, 133, 134, 138, 139, 140, 141, 142, 144, 147, 148,
43, 44, 46, 47, 48, 49, 50, 51, 53, 54, 55, 56, 58, 149, 150, 151, 152, 153, 154
59, 73, 133, 142, 148, 150, 151, 152, 154, 190, transformations, 131
220 transition, 29, 35, 36, 38, 43, 44, 53, 57, 58, 93, 101,
temperature dependence, 47, 51, 54, 151 103, 130
tensile, vii, ix, 2, 6, 7, 8, 10, 15, 16, 19, 20, 21, 22, transition temperature, 63
23, 24, 26, 28, 34, 36, 37, 40, 41, 42, 45, 46, 47, transitions, 65
48, 51, 52, 54, 55, 57, 58, 59, 63, 66, 67, 68, 69, translational, 168, 176
72, 73, 75, 76, 79, 80, 81, 82, 83, 84, 85, 86, 87, transparency, 63, 234
89, 90, 91, 95, 101, 102, 104, 113, 114, 115, 116, transport, 25, 46, 57, 93
118, 139, 157, 159, 161, 162, 163, 171, 179, 186, transportation, 184
210, 230, 234, 235, 241, 243, 244, 251 transpose, 99
tensile strength, 2, 7, 15, 16, 19, 20, 21, 22, 23, 24, transverse section, 238
28, 37, 40, 41, 55, 57, 58, 66, 67, 68, 69, 72, 79, travel, 253
80, 81, 85, 89, 90, 91, 101, 159, 161, 162, 163, trend, 21, 163, 213
186, 230, 234, 235, 243, 244, 251 trial, 104
tensile stress, 72, 101, 115, 139, 161 triggers, 154
tension, vii, ix, 18, 26, 88, 118, 147, 157, 158, 164, Turkey, 252, 258
165, 169, 176, 179, 234 two-way, x, 217
ternary blends, 82, 85, 90
272 Index
U W
UNFCCC, 184 Wales, 214
uniform, 9, 10, 12, 47, 48, 51, 163, 226, 227, 235, water, 13, 64, 67, 84, 85, 87, 148, 149, 150, 151,
238, 240 152, 153, 219, 220, 227, 235, 236, 244, 250
universal gas constant, 47, 151 water absorption, 64, 84, 235
universal law, 150 WAXS, 76
updating, 159 weakness, 248, 251
urethane, 67 wealth, 65
wear, 134, 146, 147, 234
weathering, 259
V Weibull, 7, 8, 12, 15, 40, 41, 55, 136, 144, 145
Weibull distribution, 7, 55
vacancies, 148, 152 wetting, 235, 238, 240
vacuum, 244 wood, 81, 82, 95
valence, 148
values, x, xi, 10, 12, 15, 20, 21, 36, 73, 82, 84, 107,
112, 114, 116, 119, 122, 123, 126, 127, 136, 146, X
147, 163, 170, 171, 217, 223, 225, 226, 227, 228,
235, 237, 239, 240, 241, 242, 243, 247, 248, 249, X-ray analysis, 88
250, 251, 252, 253, 256, 257 X-ray diffraction (XRD), 11, 14, 144, 153
vanadium, ix, 183, 184, 193, 194, 195, 197, 213
vapor, 56
variability, 241 Y
variable, 180, 250
variables, 108, 110 yarn, 6, 12, 40
variance, 113, 118 yield, 5, 44, 84, 85, 86, 153
variation, 12, 13, 14, 18, 19, 53, 104, 150, 234, 238, yttrium, 137
240, 241, 257
vector, 99, 100, 106, 108, 109, 110, 111, 112, 161
vehicles, 184, 202 Z
velocity, xi, 23, 97, 98, 107, 114, 115, 116, 118, 120,
121, 122, 124, 125, 126, 127, 130, 164, 166, 167, zinc, 83, 84, 93, 95, 96
179, 247, 252, 253, 255, 257 zirconia, viii, 133, 134, 135, 136, 137, 139, 140, 142,
victims, 170 144, 148, 149, 150, 151, 152, 154
viscosity, 33, 60, 63, 74, 81 zirconium, ix, 183, 184, 193, 197, 198, 199, 213
visible, 223 ZnO, 81, 88
voids, 18, 227

Strength of Materials

Uploaded by

Copyright:

Available Formats

Strength of Materials

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Strength of Materials

Uploaded by

Copyright:

Available Formats

https://boilersinfo.

Nova Science Publishers, Inc.

NOTICE TO THE READER

LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA

Strength of materials / edited by Gustavo Mendes and Bruno Lago.

Published by Nova Science Publishers, Inc. New York

Chapter 2 Ionomers as Candidates for Structural Materials 61

Chapter 9 Influence on Strength Properties of Anisotropy Planes in Slates 247

Chapter 5 studies particle modeling (PM), which is an innovative particulate dynamics

HIGH TEMPERATURE MECHANICAL PROPERTIES

2. Materials System and Characterization Technique

Reduced oxygen Stoichiometric

Low thermal stability Improved thermal stability Increased elastic modulus

Amorphous Si-C-O Fibers Low-Oxygen Si-C Fibers Crystalline SiC Fibers

Figure 3. Illustration of R&D of SiC-based Fibers.

To avoid the thermal instability caused by the decomposition of oxycarbide phase

Oxygen Modulus Diameter

HNL 1.39 0.5 2.8 270 2.74 14

2.2.1. Single Fiber Tensile Test Technique

Single filament Single filament

2.2.2. Bending Stress Relaxation Test

tR=tR(T, σ, L, G). (5)

Conventional Method Modified method

m (t, T)=1- R0/Ra (7)

Initial loop at room temperature Broken loop after stress relaxation

2.2.3. Microstructural Characterization

Accurate determination of fiber diameter is necessary for the estimation of fiber’s

3.1.1. Fiber Diameter Variation within a Tow

Diameter of selected filament (um)

Diameter of selected filament (um)

Fiber Types Hi-NicalonTM Hi-NicalonTM Type S TyrannoTM SA

Number of fibers counted 80 56 112

3.1.2. Fiber Diameter Variation along the Fiber Length

D ev. From A vg. D ia. (um )

3.2. XRD Patterns

TySA, DSiC(111) =22.7 nm

HNLS, DSiC(111) =11.4 nm

Figure 9. XRD patterns of as-received SiC-based fibers, (β-SiC).

3.3. Tensile Properties and Fracture Surface

3.4. Correlation between Tensile Strength and Fiber Diameter

d0=14.5 d0=12.9 d0=7.4

3.5. Correlation between Tensile Strength and Mirror Size

(a) HNL (b) HNLS

3.6. Fracture Toughness and Critical Fracture Energy

3.6.1. Fracture Toughness

where Am is the mirror constant.

5 (a) HNL (b) HNLS

3.6.2 Critical Fracture Energy

where Y is a geometric factor, E the modulus of elasticity. By substituting the fracture

4. Mechanical Properties and Microstructure Under Various

4.1.1. Correlation between Tensile Strength, Crystal Size and Heat

▲・△ TySA 40.0

(a):1780 °C (b): 1780 ° C (c): 1900° C

Obvious differences were found in subsequent observations of fracture surfaces.

SiCx Oy → SiC ( s) + C ( s) + SiO( g ) + CO( g ) (17)

SiO ( g ) + 3CO ( g ) → SiC ( s ) + 2CO2 ( g ) (18)

SiO ( g ) + 2C ( s ) → SiC ( s ) + CO ( g ) (19)

CO2 ( g ) + 2C ( s ) → 2CO( g ) (20)

4.1.3. BSR Creep Resistance

Stress relaxation ratio, m