Handbook of Biomaterial Properties - Murphy, William
Handbook of Biomaterial Properties - Murphy, William
Handbook of Biomaterial Properties - Murphy, William
Murphy · Jonathan Black
Garth Hastings Editors
Handbook of
Biomaterial
Properties
Second Edition
Handbook of Biomaterial Properties
William Murphy • Jonathan Black
Garth Hastings
Editors
Handbook of Biomaterial
Properties
Second Edition
Editors
William Murphy Jonathan Black
Department of Biomedical Engineering Principal: IMN Biomaterials
University of Wisconsin King of Prussia, PA, USA
Madison, WI, USA
Garth Hastings
Staffordshire University (Emeritus)
Lyme, Staffordshire, UK
Progress in the development of surgical implant materials has been hindered by the
lack of basic information on the nature of the tissues, organs and systems being
repaired or replaced. Materials’ properties of living systems, whose study has been
conducted largely under the rubric of tissue mechanics, has tended to be more
descriptive than quantitative. In the early days of the modern surgical implant era,
this deficiency was not critical. However, as implants continue to improve and both
longer service life and higher reliability are sought, the inability to predict the
behavior of implanted manufactured materials has revealed the relative lack of
knowledge of the materials properties of the supporting or host system, either in
health or disease. Such a situation is unacceptable in more conventional engineering
practice: the success of new designs for aeronautical and marine applications
depends exquisitely upon a detailed, disciplined and quantitative knowledge of ser-
vice environments, including the properties of materials which will be encountered
and interacted with. Thus the knowledge of the myriad physical properties of ocean
ice makes possible the design and development of icebreakers without the need for
trial and error. In contrast, the development period for a new surgical implant, incor-
porating new materials, may well exceed a decade and even then only short term
performance predictions can be made.
Is it possible to construct an adequate data base of materials properties of both
manufactured materials and biological tissues and fluids such that in vitro simula-
tions can be used to validate future implant designs before in vivo service? While
there are no apparent intellectual barriers to attaining such a goal, it clearly lies in
the distant future, given the complexity of possible interactions between manufac-
tured materials and living systems.
However, a great body of data has accumulated concerning the materials aspects
of both implantable materials and natural tissues and fluids. Unfortunately, these
data are broadly distributed in many forms of publication and have been gained
from experimental observations of varying degrees of accuracy and precision. This
is a situation very similar to that in general engineering in the early phases of the
Industrial Revolution. The response then was the publication of engineering hand-
books, drawing together, first in general publication and later in specialty versions,
v
vi Foreword
the known and accepted data of the time. In this spirit, we offer this 2nd Edition of
the Handbook of Biomaterial Properties
Biomaterials, as manufactured for use in implants, do not exist usefully out of
context with their applications. Thus, a material satisfactory in one application can
be wholly unsuccessful in another. In this spirit, the Editors have given direction to
the experts responsible for each part of this Handbook to consider not merely the
intrinsic and interactive properties of biomaterials but also their appropriate (and in
some cases inappropriate) applications as well as their historical context. The
experts have in some cases added significant content specific to each class of mate-
rial. For example, content is included on not just bulk properties but also surface
modifications of titanium, which has become quite important in orthopedic implant
design. It is hoped that the results will prove valuable, although in different ways, to
the student, the researcher, the engineer and the practicing physician who uses
implants.
A handbook like this necessarily becomes incomplete immediately upon publi-
cation, since it will be seen to contain errors of both omission and commission.
Such has been the case with previous engineering handbooks: the problem can only
be dealt with by providing new, revised editions. This 2nd Edition provides updated
insights, data, citations, and topic areas not found in the 1st Edition. The Editors
would appreciate any contributions and/or criticisms which the users of this hand-
book may make and promise to take account of them in future revisions.
Introduction
It is a feature of any developing science and its accompanying technology that infor-
mation relating to different aspects is scattered throughout the relevant, and sometimes
not so relevant literature. As the subject becomes more mature, a body of information
can be categorized and brought together for the use of practitioners. In providing this
Handbook of Biomaterial Properties the Editors believe that the latter stage has been
reached in several parts of the vast field of biomaterials science and engineering. This
2nd Edition of the Handbook provides an updated body of information for a subset of
chapters, as well as new chapters that represent biomaterials fields that have matured
in the intervening time between the 1st and 2nd Editions.
Many of the properties of the synthetic materials have been available for some
time, for example those of the various metallic alloys used in clinical practice have
been specified in various International, European and National Standards and can be
found by searching. In the case of polymeric materials, while the information is in
commercial product literature and various proprietary handbooks, it is diverse by
the nature of the wide range of materials commercially available and the search for
it can be time consuming. The situation is much the same for ceramic and composite
materials: there the challenge is finding the appropriate properties for the specific
compositions and grades in use as biomaterials.
However, when information is sought for on materials properties of human tis-
sues, the problem is more acute as such data are even more scattered and the meth-
ods for determination are not always stated or clearly defined. For the established
worker this presents a major task. For the new researcher it may make establishing
a project area a needlessly time consuming activity. The biomaterials bulletin boards
(on the Internet) frequently display requests for help in finding characterization
methods and/or reliable properties of natural materials, and sometimes the informa-
tion is actually not available. Even when it is available, the original source of it is
not always generally known.
In approaching their task, the Editors have tried to bring together into one source
book the information that is available. To do this they have asked for the help of
many colleagues worldwide to be contributors to the Handbook. It has not been pos-
sible to cover all the areas the Editors had hoped. Some topics could not be covered,
vii
viii Introduction
Foreword ........................................................................................................... v
Introduction ....................................................................................................... vii
Contributors ...................................................................................................... xv
PART I
A1 Cortical Bone ........................................................................................... 3
J. Currey
A2 Cancellous Bone ...................................................................................... 15
Christopher J. Hernandez
A3 Dentin and Enamel ................................................................................. 23
K.E. Healy
B1 Cartilage................................................................................................... 37
J.R. Parsons
B2 Fibrocartilage .......................................................................................... 45
V.M. Gharpuray
B3 Ligament and Tendon ............................................................................. 55
Connie S. Chamberlain and Ray Vanderby
B4 Skin and Muscle ...................................................................................... 63
A.F.T. Mak and M. Zhang
B5 Brain Tissues ........................................................................................... 67
S.S. Margulies and D. F. Meaney
B6 Arteries, Veins and Lymphatic Vessels .................................................. 77
X. Deng and R. Guidoin
xi
xii Contents
PART II
PART III
1 General Concepts of Biocompatibility .................................................. 563
D.F. Williams
2 Soft Tissue Response ............................................................................... 571
J.M. Anderson
3 Hard Tissue Response............................................................................. 581
T.O. Albrektsson
4 Immune Response ................................................................................... 593
K. Merritt
5 Cancer ...................................................................................................... 607
M. Rock
6 Blood–material Interactions .................................................................. 621
S.R. Hanson
7 Soft Tissue Response to Silicones........................................................... 631
S.E. Gabriel
8 Vocal Folds ............................................................................................... 645
Joel Gaston and Susan L. Thibeault
xv
xvi Contributors
J. Currey
A1.1 Composition
A1.1.1 Overall
The main constituents are the mineral hydroxyapatite, the fibrous protein collagen,
and water. There is some non-collagenous organic material.
Highly mineralized bone (petrosal bones of some non-human mammals) has
little organic material (8% in the horse petrosal to 3% in the tympanic bulla) [3].
(Almost certainly human ear bones will be somewhere near or in this region, though
they seem not to have been studied.)
A1.1.2 Organic
The main organic component is collagen. Most is Type I, but there are small amounts
of Type III and Type VI, found in restricted locations [4]. Slowly heated collagen
shrinks at a particular temperature, giving an indication of the stability of the mol-
ecules. Bone collagen in men has a shrinkage temperature of about 61.5°–63.5°C up
to the age of about 60, but about 60°C over that age. Bone from women showed
much greater variability [5]. About 10% of the bone organic material is non-
collagenous, mainly non-collagenous protein, NCP. The main ones are listed below.
They have supposed functions that change rapidly.
• Osteocalcin (OC), or bone Gla protein (BGP)
• Osteonectin (ON), or SPARC
• Osteopontin (OPN) or secreted phosphoprotein I (SPPI)
J. Currey (*)
Department of Biology, York University, York YOl 5DD, UK
A1.1.3 Mineral
The mineral has a plate-like habit, the crystals being extremely small, about 4 nm by
50 nm by 50 nm. The mineral is a variant of hydroxyapatite, itself a variant of cal-
cium phosphate: Ca10(PO4)6(OH)2 [7]. The crystals are impure. In particular there is
about 4–6% of carbonate replacing the phosphate groups, making the mineral tech-
nically a carbonate apatite, dahllite, and various other substitutions take place [8].
The cement line round Haversian systems (secondary osteons) contains less calcium
and phosphorus, and more sulphur than nearby parts of bone. This may indicate the
presence of more sulphated mucosubstances, making the cement line viscous [9].
A1.2.1 Density
various anomalies with the potentials generated, which did not always accord with
theory. The consensus now is that ‘SGPs’ (stress-generated potentials) are over-
whelmingly caused by streaming potentials [10, 11]. Scott and Korostoff [12] deter-
mined, amongst other things, the relaxation time constants of the stress generated
potentials, which varied greatly as a function of the conductivity and viscosity of the
permeating fluid. As an example of their findings: a step-imposed loading moment
which produced a peak strain of 4 × 10-4 induced an SGP of 1.8 mV, yielding a value
of the SGP/strain ratio of 4500 mV. The SGP decayed rapidly at constant strain,
reaching zero within about one second. For more detail, the complex original paper
must be consulted.
Behari [10] gives a useful general review of many ‘solid state’ properties of bone,
both human and non-human, many of which are not dealt with here. These proper-
ties include the Hall effect, photo-electric effects, electron paramagnetic resonance
effects and so on.
A1.3.1 General
There is a great range for values in the literature for many reasons. Amongst
these are:
(a) Different treatment of specimens Drying bone and then re-wetting it pro-
duces some small differences [13], as does formalin fixation [14]. Testing bone dry
produces results quite different from those in wet bone; dry bone is stiffer, stronger,
and considerably more brittle. Very small samples produce values for stiffness and
strength less than those from larger samples [15, 16]. High strain rates generally
produce a higher modulus of elasticity, a higher strength [17], and a greater strain to
failure than specimens tested at low strain rate.
(b) Different age and health of donors Age may affect intrinsic properties.
Osteoporotic bone may differ from ‘normal’ bone in ways other than the fact that it
is more porous; there is evidence that the collagen is different from that in similar-
aged non-osteoporotic subjects [18]. Bone from osteogenesis imperfecta patients
has a higher proportion of Type III and Type V collagen compared with Type I col-
lagen, than bone from normal subjects [19]. Bone collagen from osteopetrotic sub-
jects is in general older than that from normal subjects, and has correspondingly
different properties [5].
6 J. Currey
(c) Differences between bones, and sites in the bones The ear bones (ossicles)
and portions of the temporal bones (petrosals) are highly mineralized, and will
undoubtedly be stiffer and more brittle than others (though they seem not to have
been investigated in humans). Long bones differ along their length and around their
circumference. The distal femur is less highly mineralized and weaker in tensile and
compressive static loading, and at any level the posterior part is similarly less min-
eralized and weaker [20].
The values reported below should be considered paradigmatic, that is, to be valid
for a well-performed test on bone obtained from a middle aged person with no dis-
ease. Other values are reported in such a way as to make it clear how some property
is a function of other features of the specimen.
A1.3.2 Stiffness
(a) General There are two ways of testing bone: mechanically by relating stresses
to strains; ultrasonically, by subjecting the bone to ultrasound and measuring the
velocity of the sound. From a knowledge of the density one can then obtain a stiff-
ness matrix. If this is inverted it becomes a compliance matrix, the reciprocal of the
individual terms of which are equivalent to the so-called technical moduli derived
by mechanical testing [21]. Reilly and Burstein [22] give mechanical values, and
Ashman et al. [23] give ultrasonic measurements. Reilly and Burstein [22] assumed
transverse isotropy (that is, symmetry around the longitudinal axis of the bone),
while Ashman et al. [23] assumed orthotropy (that is, that the values for stiffness
could be different in the longitudinal, radial and tangential directions).
Reilly and Burstein [22] give values for Young’s modulus at a number of inter-
mediate angular orientations, but they do not form a very uniform set.
(b) Tensile modulus versus compressive modulus Reilly et al. [24] tested femo-
ral specimens specifically to determine whether the value for Young’s modulus was
different in tension and compression. A paired Student’s ‘t’ test showed no signifi-
cant difference between the compressive and tensile moduli at the 95% confidence
level. Calculations on their data show the the 95% confidence interval ranged from
compression modulus 1.72 GPa higher to tension modulus 0.27 GPa higher. The
load-deformation traces showed no change of slope going from compression into
tension and vice versa.
(c) Very small specimens The bending modulus of very small specimens was 6.62
GPa [5].
(d) Locational variations: Metaphysis versus diaphysis Young’s modulus has
been determined in three-point bending for extremely small plates (7 mm by 5 mm
by (about) 0.3 mm) from the femoral metaphyseal shell and from the diaphysis of
the same bones [16].
A1 Cortical Bone 7
The differences between these values and those reported by Reilly and Burstein
[22] are probably attributable not to the difference in testing mode, since bending
and tension tests from the same bone generally give similar values for Young’s
modulus, but to the very small size of the specimen, and to the rather low density of
the specimens.
(e) Compression; effect of mineral The compressive behavior of cubes, relating
the properties to the density of the specimens gives, using ρa (fat-free mass divided
by anatomical volume, g cm-3) as the explanatory variable:
Young’s modulus (GPa) = 3.3ρa2.4 for compact bone [25].
The higher values of ρa were of the order of 1.8 g cm-3(=1800 kg m-3); this equa-
tion [25] predicts a value of 13.5 GPa for such a specimen. Multiple regression
analysis showed that the dependence of Young’s modulus on density was caused by
8 J. Currey
the effect of porosity on density, and that, in these specimens, the effect of mineral
content was insignificant.
(f) Single secondary osteons Ascenzi and co-workers [26–29] distinguish two
types of secondary osteon: ‘longitudinal’ osteons, whose collagen fibres have a basi-
cally longitudinal orientation, and ‘alternate’ osteons, whose fibres have markedly
different courses in neighboring lamellae. (This difference is a contentious issue.)
N.B.: These studies of Ascenzi and co-workers [26–29] are widely quoted, so
beware of some apparent anomalies (apart from changes in nomenclature between
papers). The bending modulus is remarkably low compared with the tension and
compression moduli. The torsional (shear) modulus is remarkably high, compared
both with the shear modulus values obtained by others (above), and with the tension
and compression values. Torsional moduli are expected, on theoretical grounds, to
be less than the tension and compression moduli. Furthermore, the large differences
between the tension and compression moduli have not been reported elsewhere.]
(g) Strain rate effects Calculations [30], incorporating data from non-human as
well as human material, predict that Young’s modulus is very modestly dependent
upon strain rate:
[N.B. statements about strain rate effects in bone are suspect unless it is clear that
the workers have taken machine compliance into account!]
Î (t ) = b1 exp[to - t t ] + b 2
where the betas are parameters, t is time (s), to is time at which the specimen is held
at a constant stress below the creep threshold: 6.1 s [31]. For reference, its value in
bovine bone: 3.6 s.
A1 Cortical Bone 9
A1.3.3 Strength
(a) Overall
(b) Combined loading Cezayirlioglu et al. [32] tested human bone under com-
bined axial and torsional loading. The results are too complex to tabulate, but should
be consulted by readers interested in complex loading phenomena.
(c) Metaphysis versus diaphysis Same specimens as reported for modulus above
(Table A1.4) [16]. ‘Tensile’ strength calculated from the bending moment, using a
‘rupture factor’ to take account of the non-uniform distribution of strain in the
specimen.
(d) Effect of mineral Keller [25], using the same specimens as above, provides the
following relationship:
(e) Single secondary osteons The same nomenclature applies as for moduli of
osteons (Table A1.5).
[N.B. The bending strengths and torsional strengths seem very high, even bearing
in mind that no allowance has been made in bending for non-elastic effects.]
(f) Strain rate effects Bone will bear a higher stress if it is loaded at a higher strain
(or stress) rate. Carter and Caler [17] found an empirical relationship that failure
.
stress (σf (MPa)) was a function of either stress rate((s ) ) or strain rate (Î) :
s f = 87(s )0.053
.
s f = 87 (Î) 0.055
N.B. These relationships imply an increase of 44% in the failure stress if the
stress rate is increased one thousandfold. This relationship has been found to be
roughly the same in other, non-human, mammals.
(g) Creep Creep threshold (the stress below which no creep occurs): 73 MPa [31].
The equivalent value for bovine bone is 117 MPa [31]. Specimens in tension or
compression were held at particular stresses [33]. The time (seconds) to failure is
given as a function of normalized stress (stress/Young’s modulus (MPa/MPa)):
(h) Fatigue Some workers report the log of the number of cycles as a function of
the applied stress levels, some report the log cycle number as a function of log stress
levels, and some report log stress levels as a function of log cycle number. [The last
seems wrong, since the applied stress can hardly be a function of the number of
cycles the specimen is going to bear, but it is frequently used in fatigue studies. It is
not possible simply to reverse the dependent and independent axes because the
equations are derived from regressions with associated uncertainty.] The variation
between the results for different testing modes is considerable.
Carter et al. [34] report on the effect of Young’s modulus of elasticity and poros-
ity in their specimens. They find that Young’s modulus is positively associated with
fatigue life, and porosity is negatively associated:
A1 Cortical Bone 11
(i) Effect of remodeling Vincentelli and Grigorov [35] examined the effect of
Haversian remodelling on the tibia. The specimens they reported were almost
entirely primary or Haversian, with few specimens having a scattering of secondary
osteons. [Unfortunately they probably (it is not clear) allowed their specimens to
dry out, so it is not sure that bone in vivo would show the same behavior. However,
their results are similar to those found in nonhuman specimens.]
Additional Reading
Cowin, S.C. (ed.)(1989) Bone Mechanics Boca Raton: CRC Press.
A more rigorous, less chatty and less biologically, oriented approach than the
following books by Currey and by Martin and Burr. The chapters on mechanics (2,
6 and 7), written by Cowin himself, are particularly authoritative.
Currey, J.D. (1984) The Mechanical Adaptation of Bones Princeton: University
Press.
Out of print, new edition in preparation. Tries to deal with all aspects of mechani-
cal properties of bone as a material and of whole bones. Not overly technical.
Written from a general biological perspective, thus, does not concentrate on human
material.
12 J. Currey
Martin, R.B. and Burr, D.B. (1989) Structure, Function and Adaptation of
Compact Bone New York: Raven Press.
There are not many values of mechanical properties here, but the treatment of the
biology of bone, and of fatigue of bone tissue, is excellent and the discussion of
remodeling, although now somewhat out of date, is a very good introduction to this
intellectually taxing topic.
Nigg, B.M. and Herzog, W. (eds)(1994) Biomechanics of the Musculoskeletal
System John Wiley: Chichester.
Deals with many aspects of biomechanics, including locomotion, with an empha-
sis on human material. There is a full treatment of the measurement of many biome-
chanical properties.
References
1. Gong, J.K., Arnold, J.S. and Cohn, S.H. (1964) Composition of trabecular and cortical bone.
Anat. Rec., 149, 325–331.
2. Biltz, R.M. and Pellegrino, E.D. (1969) The chemical anatomy of bone I. A comparative study
of bone composition in sixteen vertebrates. J. Bone Joint Surg., 51A, 456–466.
3. Lees, S. and Escoubes, M. (1987) Vapor pressure isotherms, composition and density of
hyperdense bones of horse, whale and porpoise. Con. Tiss. Res., 16, 305–322.
4. Keene, D.R., Sakai, L.Y. and Burgeson, R.E. (1991) Human bone contains type III collagen,
type VI collagen, and fibrillin: Type III collagen is present on specific fibers that may mediate
attachment of tendons, ligaments, and periosteum to calcified bone cortex. J. Histochem.
Cytochem., 39, 59–69.
5. Danielsen, C.C., Mosekilde, Li., Bollerslev, J. et al. (1994) Thermal stability of cortical bone
collagen in relation to age in normal individuals and in individuals with osteopetrosis. Bone,
15, 91–96.
6. Ninomiya, J.T., Tracy, R.P., Calore, J.D., et al. (1990) Heterogeneity of human bone. J. Bone
Min. Res., 5, 933–938.
7. Lowenstam, H.A. and Weiner, S. (1989) On Biomineralization, Oxford University Press,
New York.
8. McConnell, D. (1962) The crystal structure of bone. Clin. Orthop. Rel. Res., 23, 253–68.
9. Burr, D.B., Schaffler, M.B. and Frederickson, R.G. (1988) Composition of the cement line and
its possible mechanical role as a local interface in human compact bone. J. Biomech., 21,
939–945.
10. Behari, J. (1991) Solid state bone behavior. Prog. Biophys. Mol. Biol., 56, 1–41.
11. Martin, R.B. and Burr, D.B. (1989) Structure, Function, and Adaptation of Compact Bone,
Raven Press, New York.
12. Scott, G.C. and Korostoff, E. (1990) Oscillatory and step response electromechanical phenom-
ena in human and bovine bone. J. Biomech., 23, 127–143.
13. Currey, J.D. (1988) The effects of drying and re-wetting on some mechanical properties of
cortical bone. J. Biomech., 21, 439–441.
14. Sedlin, E.D. (1967) A rheological model for cortical bone. Acta Orthop. Scand., (Suppl. 83),
1–78.
15. Choi, K. and Goldstein, S.A. (1992) A comparison of the fatigue behavior of human trabecular
and cortical bone tissue. J. Biomech., 25, 1371–1381.
16. Lotz, J.C., Gerhart, T.N. and Hayes, W.C. (1991) Mechanical properties of metaphyseal bone
in the proximal femur. J. Biomech., 24, 317–329.
A1 Cortical Bone 13
17. Carter, D.R. and Caler, W.E. (1985) A cumulative damage model for bone fracture. J. Orthop.
Res., 3, 84–90.
18. Bailey, A.J., Wotton, S.F., Sims, T.J. et al. (1993) Biochemical changes in the collagen of
human osteoporotic bone matrix. Con. Tiss. Res., 29, 119–132.
19. Bateman, J.F., Chan, D., Mascara, T. et al. (1986) Collagen defects in lethal perinatal osteo-
genesis imperfecta. Biochem. J., 240, 699–708.
20. Saito, S. (1983) (Distribution of the X-ray density, compressive and tensile breaking strength
in the human femoral shaft) Die Verteilung von Dichte, Druck und Festigkeit im menslichen
Femurschaft. Anat. Anzeiger Jena, 154, 365–376.
21. Cowin, S.C. (1989) Bone Mechanics, CRC Press, Boca Raton.
22. Reilly, D.T. and Burstein, A.H. (1975) The elastic and ultimate properties of compact bone
tissue. J. Biomech., 8, 393–405.
23. Ashman, R.B., Cowin, S.C., Van Buskirk, W.C. (1984) A continuous wave technique for the
measurement of the elastic properties of cortical bone. J. Biomech., 17, 349–361.
24. Reilly, D.T. and Burstein, A.H. (1974) The mechanical properties of cortical bone. J. Bone
Joint Surg., 56A, 1001-1022
25. Keller, T.S. (1994) Predicting the compressive mechanical behavior of bone. J. Biomech., 27,
1159–1168.
26. Ascenzi, A. and Bonucci, E. (1967) The tensile properties of single osteons. Anat. Rec., 158,
375–386.
27. Ascenzi, A. and Bonucci, E. (1968) The compressive properties of single osteons. Anat. Rec.,
161, 377–391.
28. Ascenzi, A., Baschieri, P., and Benvenuti, A. (1990) The bending properties of single osteons.
J. Biomech., 23, 763–771.
29. Ascenzi, A., Baschieri, P., and Benvenuti, A. (1994) The torsional properties of single selected
osteons. J. Biomech., 27, 875–884.
30. Carter, D.R. and Caler, W.E. (1983) Cycle-dependent and time-dependent bone fracture with
repeated loading. J. Biomech. Eng., 105, 166–170.
31. Fondrk, M., Bahniuk, E., Davy, D.T. et al. (1988) Some viscoplastic characteristics of bovine
and human cortical bone. J. Biomech., 21, 623–630.
32. Cezayirlioglu, H., Bahniuk, E., Davy, D.T. et al. (1985) Anisotropic yield behavior of bone
under combined axial force and torque. J. Biomech., 18, 61–69.
33. Caler, W.E. and Carter, D.R. (1989) Bone creep-fatigue damage accumulation J. Biomech., 22,
625–635.
34. Carter, D.R., Caler, W.E., Spengler, D.M. et al. (1981) Uniaxial fatigue of human bone. The
influence of tissue physical characteristics. J. Biomech., 14, 461–70.
35. Vincentelli, R., and Grigorov, M. (1985) The effect of haversian remodeling on the tensile
properties of human cortical bone. J. Biomech., 18, 201–207.
Chapter A2
Cancellous Bone
Christopher J. Hernandez
A2.1 Microstructure
0.70
BS æ BV ö
= 8.84 ç ÷
TV è TV ø
At the tissue level cancellous bone is a ceramic polymer composite with composi-
tion similar to that of cortical bone. By mass, bone tissue is 65 % mineral (primarily
an impure hydroxyapatite), 25 % organic (primarily type I collagen but also includ-
ing non-collagenous proteins), and 10 % water. By volume bone tissue is 36 %
mineral, 55 % organic, and 9 % water [3]. These proportions vary with age of the
individual as well as length of time the tissue has been present in the body (the “tis-
sue age”) and differ among species. In general tissue density is highly conserved
and shows a small range within a species (1.6–2.0 g/cm3 in humans) [3, 4]. Hetero
geneity within bone tissue is present, however, and is caused primarily by bone
remodeling. Bone remodeling is a process in which bone cells remove and replace
bone tissue at discrete locations on bone surfaces. Structures known as cement lines
mark the boundaries of previous remodeling activity within trabeculae and divide
more recently remodeled (younger) tissue from non-remodeled (older) regions of
the tissue that are typically stiffer and more mineralized. The large surface-
to-volume ratio of cancellous bone (as compared to cortical bone) enables greater
turnover of bone tissue during remodeling, reducing the proportion of older tissue,
possibly explaining why cancellous bone has been reported to have 10–14 % lower
tissue density than cortical bone [5].
The stiffness and strength of cancellous bone are determined primarily by apparent
density [6, 7]. When loaded in uniaxial compression or tension cancellous bone
displays a slightly nonlinear stress-strain curve [8] and yield is commonly deter-
mined using the 0.2 % offset criteria. Cancellous bone is anisotropic due to both
microstructure and tissue anisotropy.
The Young’s modulus and shear modulus of cancellous bone are determined
primarily by tissue apparent density (Table A2.1) and can vary over tenfold among
regions of the skeleton within the same individual. Hence, there is no one value for
an elastic modulus of cancellous bone and cancellous bone mechanical properties
must therefore be estimated from apparent density. Empirically derived power law
Table A2.1 Mechanical properties of cancellous bone are strongly influenced by density. Regression models relating cancellous bone mechanical properties
to apparent density (wet) are provided. Mechanical properties are reported on axis (in the direction of primary trabecular orientation). Results are provided as
mean ± SD and (95 % CI) when available. Regression equations were achieved using linear regression on log-transformed data
Y = A × XB
3
Mechanical property n Age Apparent density (g/cm ) A B r2
A2 Cancellous Bone
models are useful for general prediction of specimen Young’s modulus and shear
modulus, although linear regressions are appropriate when the range in apparent
density is small (for example, if only one region of the skeleton is considered). The
Poisson’s ratio of cancellous bone is difficult to measure and poorly understood and
is typically estimated (common estimates range from 0.1 to 0.3). Cancellous bone
strength is strongly correlated with Young’s modulus. The ultimate strength of can-
cellous bone is strongly correlated with yield strength (in vertebral cancellous bone
ultimate strength is 20 % greater than yield strength [14]). Yield strain and ultimate
strain are not correlated with apparent density. Yield strain of cancellous bone within
a region of the skeleton shows little interindividual variability, but differences among
skeletal sites have been noted. Compressive yield strains of human cancellous bone
have been reported to range from 0.70 to 0.85 (across different regions of the skele-
ton). Yield strains in tension are always lower than those in compression (reported to
range from 0.60 to 0.70). Ultimate strain is more variable for unknown reasons.
Trabecular alignment and microstructure have been shown to influence mechani-
cal properties of cancellous bone. Trabecular alignment is the primary cause of
cancellous bone anisotropy. When loading is applied at 90° to the primary trabecu-
lar orientation, cancellous bone Young’s modulus is 40–60 % smaller and the ulti-
mate strength is 30–45 % smaller [15]. Interestingly yield strain in the transverse
directions is not different from that on axis. The effects of trabecular alignment on
Young’s modulus can be described using a fabric tensor and a quadratic Tsai-Wu
criteria has been used successfully to describe a multiaxial failure envelope for can-
cellous bone [16]. More complicated multiaxial failure criteria have been used to
address some shortcomings of the Tsai-Wu criteria [17, 18].
Tissue material properties also influence cancellous bone Young’s modulus and
strength. Compressive Young’s modulus and strength of bone (cancellous and corti-
cal bone pooled) is related to bone volume fraction (BV/TV) and the tissue degree
of mineralization (α, inorganic mass/bone mass) in the following manner [19]:
2.58
æ BV ö
E ( GPa ) = 84 ç ÷ a 2.74
è TV ø
1.92
æ BV ö
s ult ( MPa ) = 794 ç ÷ a 2.79 .
è TV ø
The regression exponents applied to tissue degree of mineralization are greater than
those on bone volume fraction, demonstrating that bone strength is more sensitive to
variation in tissue degree of mineralization than to variation in bone volume fraction.
Cancellous bone displays viscoelastic properties. The strain rate has only a small
effect on the Young’s modulus and strength of cancellous bone under uniaxial load-
ing; Young’s modulus and strength are related to strain rate to the power of 0.06 [6].
Hydraulic stiffening of the marrow does not influence Young’s modulus and strength
A2 Cancellous Bone 19
until strain rates exceed 1 s−1 [6]. Creep deformation in cancellous bone follows the
pattern of a rapid primary phase, slow secondary phase, and rapid tertiary phase
[20]. The creep rate of bovine cancellous bone is related to normalized stress (σ/E0,
where E0 is the initial Young’s modulus) as follows [20]:
17.65
de c æs ö
= 2.21 ´ 1033 ç ÷ .
dt è E0 ø
Human vertebral cancellous bone submitted to low-magnitude, compressive creep
loading (σ/E0 = 1500 με or less) has been shown to have nonlinear viscoelastic prop-
erties such that residual strains persist up to ten times longer than the period of
constant loading [21].
Under fatigue loading cancellous bone displays an S-N curve such that the num-
ber of cycles to failure (Nf) is related to normalized stress (σ/E0) as follows [22]:
N f = 4.57 ´ 10 -18 ( s / E0 )
-8.54
.
following loading are typically small; upon removal of load cancellous bone recov-
ers as much as 70–94 % of the applied strain [24, 25, 28].
Additional Reading
Cowin S. Bone Mechanics Handbook. In. 2 ed. Boca Raton: CRC Press; 2001.
This book provides a complete review of bone mechanical properties and inter-
actions with bone cell biology.
Currey JD. Bones: Structure and Mechanics. Princeton, NJ, USA: Princeton
University Press; 2002.
This is a review that includes thorough discussion of non-human bone mechani-
cal properties and function.
References
1. Liu XS, Sajda P, Saha PK, Wehrli FW, Bevill G, Keaveny TM, Guo XE (2008) Complete volu-
metric decomposition of individual trabecular plates and rods and its morphological correla-
tions with anisotropic elastic moduli in human trabecular bone. J Bone Miner Res 23:223–235
2. Fyhrie DP, Fazzalari NL, Goulet R, Goldstein SA (1993) Direct calculation of the surface-to-
volume ratio for human cancellous bone. J Biomech 26:955–967
3. Robinson RA (1975) Physicochemical structure of bone. Clin Orthop 53:263–315
4. Galante J, Rostoker W, Ray RD (1970) Physical properties of trabecular bone. Calcif Tissue
Res 5:236–246
5. Gong JK, Arnold JS, Cohn SH (1964) Composition of trabecular and cortical bone. Anat Rec
149:325–332
6. Carter DR, Hayes WC (1977) The compressive behavior of bone as a two-phase porous struc-
ture. J Bone Joint Surg 59-A:954–962
7. Keaveny TM, Morgan EF, Niebur GL, Yeh OC (2001) Biomechanics of trabecular bone. Annu
Rev Biomed Eng 3:307–333
8. Morgan EF, Yeh OC, Chang WC, Keaveny TM (2001) Nonlinear behavior of trabecular bone
at small strains. J Biomech Eng 123:1–9
A2 Cancellous Bone 21
9. Morgan EF, Bayraktar HH, Keaveny TM (2003) Trabecular bone modulus-density relation-
ships depend on anatomic site. J Biomech 36:897–904
10. Morgan EF, Keaveny TM (2001) Dependence of yield strain of human trabecular bone on
anatomic site. J Biomech 34:569–577
11. Garrison JG, Gargac JA, Niebur GL (2011) Shear strength and toughness of trabecular bone
are more sensitive to density than damage. J Biomech 44:2747–2754
12. Kopperdahl DL, Keaveny TM (1998) Yield strain behavior of trabecular bone. J Biomech
31:601–608
13. Cook RB, Zioupos P (2009) The fracture toughness of cancellous bone. J Biomech
42:2054–2060
14. Crawford RP, Cann CE, Keaveny TM (2003) Finite element models predict in vitro vertebral
body compressive strength better than quantitative computed tomography. Bone 33:744–750
15. Goulet RW, Goldstein SA, Ciarelli MJ, Kuhn JL, Brown MB, Feldkamp LA (1994) The rela-
tionship between the structural and orthogonal compressive properties of trabecular bone.
J Biomech 27:375–389
16. Cowin SC (1985) The relationship between the elasticity tensor and the fabric tensor. Mech
Mater 4:137–147
17. Bayraktar HH, Gupta A, Kwon RY, Papadopoulos P, Keaveny TM (2004) The modified super-
ellipsoid yield criterion for human trabecular bone. J Biomech Eng 126:677–684
18. Wolfram U, Gross T, Pahr DH, Schwiedrzik J, Wilke HJ, Zysset PK (2012) Fabric-based
Tsai-Wu yield criteria for vertebral trabecular bone in stress and strain space. J Mech Behav
Biomed Mater 15:218–228
19. Hernandez CJ, Beaupre GS, Keller TS, Carter DR (2001) The influence of bone volume frac-
tion and ash fraction on bone strength and modulus. Bone 29:74–78
20. Bowman SM, Keaveny TM, Gibson LJ, Hayes WC, McMahon TA (1994) Compressive creep
behavior of bovine trabecular bone. J Biomech 27:301–310
21. Yamamoto E, Paul Crawford R, Chan DD, Keaveny TM (2006) Development of residual
strains in human vertebral trabecular bone after prolonged static and cyclic loading at low load
levels. J Biomech 39:1812–1818
22. Haddock SM, Yeh OC, Mummaneni PV, Rosenberg WS, Keaveny TM (2004) Similarity in the
fatigue behavior of trabecular bone across site and species. J Biomech 37:181–187
23. Morgan EF, Yeh OC, Keaveny TM (2005) Damage in trabecular bone at small strains. Eur
J Morphol 42:13–21
24. Hernandez CJ, Lambers FM, Widjaja J, Chapa C, Rimnac CM (2014) Quantitative relation-
ships between microdamage and cancellous bone strength and stiffness. Bone 66:205–213
25. Keaveny TM, Wachtel EF, Kopperdahl DL (1999) Mechanical behavior of human trabecular
bone after overloading. J Orthop Res 17:346–353
26. Lambers FM, Bouman AR, Rimnac CM, Hernandez CJ (2013) Microdamage caused by
fatigue loading in human cancellous bone: relationship to reductions in bone biomechanical
performance. PLoS One 8, e83662
27. Wang X, Niebur GL (2006) Microdamage propagation in trabecular bone due to changes in
loading mode. J Biomech 39:781–790
28. Fyhrie DP, Schaffler MB (1994) Failure mechanisms in human vertebral cancellous bone.
Bone 15:105–109
29. Guo XE (2001) Mechanical properties of cortical bone and cancellous tissue. In: Cowin SC
(ed) Bone Mechanics Handbook. CRC Press, Boca Raton, pp 10.11–10.23
30. Kim G, Cole JH, Boskey AL, Baker SP, van der Meulen MC (2014) Reduced tissue-level stiff-
ness and mineralization in osteoporotic cancellous bone. Calcif Tissue Int 95:125–131
31. Norman J, Shapter JG, Short K, Smith LJ, Fazzalari NL (2008) Micromechanical properties of
human trabecular bone: a hierarchical investigation using nanoindentation. J Biomed Mater
Res A 87:196–202
32. Rho JY, Roy ME, Tsui TY, Pharr GM (1999) Elastic properties of microstructural components
of human bone tissue as measured by nanoindentation. J Biomed Mater Res 45:48–54
33. Bayraktar HH, Morgan EF, Niebur GL, Morris GE, Wong EK, Keaveny TM (2004) Comparison
of the elastic and yield properties of human femoral trabecular and cortical bone tissue.
J Biomech 37:27–35
Chapter A3
Dentin and Enamel
K.E. Healy
A3.1 Introduction
The permanent adult human dentition normally consists of 32 teeth, of which 16 are
located in the mandible and 16 in the maxilla. There are 4 incisors, 2 canines, 4
premolars and 6 molars for the upper and lower dentition. The incisors are used for
cutting food, the canines for tearing, the premolars for grasping, and the molars for
grinding (i.e., masticating). There is a generic heterogeneous structure for these
teeth, where enamel forms an exterior layer over the underlying dentin. From the
cervix to the apex of the root, the exterior of the dentin is covered by cementum to
which the periodontal ligament attaches the tooth to alveolar bone. Dental enamel is
dense, highly mineralized, hard, and brittle. It contains prism-like structures that
span from the enamel surface to the junction of enamel and dentin, the dentino-
enamel junction (DEJ). The prisms are comprised of hydroxyapatite crystallites and
contain very little organic matrix. These properties make dental enamel an excellent
material for cutting and masticating food (i.e., processes that involve friction and
wear). In contrast, dentin is not as hard as enamel, but it is tougher. Dentin is a
heterogeneous material and can be thought of as a composite structure containing
four major components: dentin matrix; dentinal tubules; mineral (i.e., carbonate
containing hydroxyapatite); and, dentinal fluid. The dentinal tubules (~45 000 per
mm2) are formed during development of the dentin matrix and are distributed
throughout the dentin matrix in a somewhat uniform manner. The dentin matrix
mineralizes in an anisotropic fashion, where a highly mineralized tissue, peritubular
dentin, surrounds the dentinal tubules. The mineralized tissue between the dentinal
tubules and peritubular dentin is referred to as intertubular dentin. Histological
examination has revealed that intertubular dentin is less mineralized than peritubular
dentin. Furthermore, the matrix and mineral content of root dentin is different from
coronal dentin. A good review of the structure of teeth can be found in Waters [1].
A3.2 Composition
Table A3.2 Major Elemental Composition of Surface and Bulk Dental Enamel
Enamel Mean wt% (range or Dentin Mean wt% (range or Source,
standard deviation, ±) standard deviation, ±) Comments
Ca 37.4 ± 1.0 – [4]#
37.1 ± 0.2 (26.7–47.9) 26.9 ± 0.2 (21.8–31.3) [5]#
36.3 ± 0.1 (27.7–42.0) 27.6 ± 0.1 (24.7–31.5) [5]‡
P 17.8 ± 0.2 13.5 ± 0.1 [5]●, age >25yrs
17.68 ± 0.2 [6]¤
Na 0.72 ± 0.008 (0.42–1.03) 0.72 ± 0.008 (0.26–0.87) [5]#
0.72 ± 0.008 (0.49–0.88) 0.64 ± 0.001 (0.55–0.75) [5]‡
Cl 0.28 ± 0.01 0.05 ± 0.004 [5]#, age >25yrs
0.32 ± 0.01 0.072 ± 0.022 [7]#
K 0.026 ± 0.001 0.02 ± 0.001 [5]‡, age <25yrs
Mg 0.39 ± 0.02 (0.13–0.77) 0.74 ± 0.02 (0.25–0.94) [5]#
0.39 ± 0.004 (0.24–0.48) 0.76 ± 0.004 (0.58–0.89) [5]‡
CO3 3.2 (2.4–4.2) 4.6 (4–5) [2,3]†
# Neutron activated gamma-ray spectrometric analysis.
‡ Atomic absorption spectrophotometry.
● Colorimetic assay.
† Average compiled from the literature.
* Neutron activated gamma-ray spectrometric analysis (Na, Cl, Al, Mn, Ca, and P), atomic
absorption spectrophotometry (K, Mg, Zn, Cu, and Fe), or a fluoride-specific electrode (F).
¤ Atomic absorption spectrophotometry (Ca), and colorimetric method (P).
A3 Dentin and Enamel 25
Table A3.3 Trace Elemental Composition of Surface and Bulk Dental Enamel
Surface Enamel Whole Enamel
Mean (range) Median Median Source,
At.# μg/g μg/g Mean (range) μg/g μg/g comments
S 16 281 (530–130) 270 [8]†,[9]‡
F 9 752 (1948–25) 666 293 (730–95) 200 [8]†,[9]‡
123.8 ± 7.9 [10]◊
Zn 30 893 (5400–61) 576 199 (400–91) 190 [8]†,[9]‡
276 ± 106 [4]#
263.42 ± 14.8 [10]◊
Mg 12 745 (3600–115) 576 1,670 (3,000–470) 1,550 [8]†,[9]‡
AI 13 343 (2304–16) 202 12.5 (70–1.5) 5.6 [8]†,[9]‡
Sr 38 204 (7632–9) 36 81 (280–26) 56 [8]†,[9]‡
93.5 ± 21.9 [4]#
111.19 ± 9.86 [10]◊
Fe 26 138 (1404–18) 68 4.4 (21–0.8) 2.6 [8]†,[9]‡
2.77 [7]*
Si 14 70 (504–1.3) 40 [8]†,[9]‡
Mn 25 59 (468–2.6) 33 0.28 (0.64–0.08) 0.26 [8]†,[9]‡
0.54 ± 0.08 [4]#
0.59 ± 0.04 [10]◊
Ag 47 32 (396–0.2) 2 0.35 (1.3–0.03) 0.16 [8]†,[9]‡
0.56 ± 0.29 [10]◊
Pb 82 24 (79–1.2) 18 3.6 (6.5–1.3) 3.6 [8]†,[9]‡
Ni 28 23 (270–0.4) 9 [8]†,[9]‡
Ba 56 22 (432–0.8) 7 4.2 (13–0.8) 3.4 [8]†,[9]‡
Se 34 18 (72–2.9) 16 0.27 (0.5–0.12) 0.22 [8]†,[9]‡
Li 3 14 (58–0.3) 10 1.13 (3.4–0.23) 0.93 [8]†,[9]‡
Sb 51 8 (90–0) 3 0.13 (0.34–0.02) 0.11 [8]†,[9]‡
Ga 31 6 (32–0) 5 [8]†,[9]‡
Sn 50 9.3 (72–0.9) 5.8 0.21 (0.92–0.03) 0.14 [8]†,[9]‡
Ge 32 7.6 (39.6–0.5) 4.0 [8]†,[9]‡
B 5 5.3 (13.0–0.8) 3.6 5.0 (39–0.5) 2.4 [8]†,[9]‡
Cu 29 4.20 (81–0.1) 0.45 [8]†,[9]‡
0.26 ± 0.11 [4]#
1.38 [7]*
Br 35 3.1 (14.0–0.4) 4.1 1.12 (2.6–0.32) 0.93 [8]†,[9]‡
4.6 ± 1.1 [4]#
Cd 48 2.7 (7.6–0.6) 1.8 0.51 (2.4–0.03) 0.22 [8]†,[9]‡
Y 39 1.8 (9.3–0) 0.9 0.007 (0.17–<0.01) <0.01 [8]†,[9]‡
Ti 22 1.6 (24.5–0.1) 0.6 0.19 (4.4–<0.1) <0.1 [8]†,[9]‡
V 23 1.4 (14.4–0.1) 0.5 0.017 (0.03–0.01) 0.02 [8]†,[9]‡
La 57 1.4 (7.2–0) 0.8 [8]†,[9]‡
Be 4 1.3 (6.1–0) 1.2 [8]†,[9]‡
Cr 24 1.1 (4.7–0.2) 0.7 3.2 (18–0.1) 1.5 [8]†,[9]‡
26 K.E. Healy
Table A3.4 Significant Differences in Trace Element Composition of Whole Human Enamel for
High and Low Caries Populationst†
High Caries Low Caries
At.# (Mean ± SE), μg/g (Mean ± SE), μg/g Source
F 9 82.1 ± 7.99 125.7 ± 11.23 [11]
Sr 38 104.1 ± 9.14 184.0 ± 14.68 [11]
Mn 25 1.57 ± 0.24 0.87 ± 0.15 [12]
Zr 40 0.27 ± 0.1 0.16 ± 0.09 [11]
Cu 26 0.71 ± 0.2 0.17 ± 0.04 [12]
† Determined by spark source mass spectroscopy
A3 Dentin and Enamel 27
Table A3.6 Crystallite Size and Lattice Parameters of the Apatite in Human Enamel and Dentin*
a-axis (nm) c-axis (nm) Width (nm) Thickness (nm) Source, Comments
Enamel
0.9445 0.6885 [2]#
0.9440 0.6872 [15]†
0.9441 0.6880 68.4 ± 3.4 26.3 ± 2.2 [6]†, ● ±S.D.
0.9446 0.6886 [16]†
68.3 ± 13.4 26.3 ± 2.19 [17]‡,● ± S.E.
Dentin
0.9434 ± 0.6868 ± [18]‡
0.0007 0.0009 29.6 ± 3.7 3.2 ± 0.5 [19]●, intertubular
dentin
36.55 ± 1.45 10.33 ± 7.91 [20]●, mixed carious
and sound dentin
* Asymmetric hexagonal crystal with the thickness of the crystal less than the width.
† X-ray diffraction method of determination.
● High resolution transmission electron microscopy.
# Data from [2], average compiled from the literature.
28 K.E. Healy
Table A3.7 Elastic Moduli and Viscoelastic Properties of Human Dentin and Enamel
Incisors Canine Pre-molars Molars Source, Comments
E: Dentin
11.0 (5.8) [21]t,†,‡
13 (4) 14 (6) 14 (0.7) 12 (2) [22]Crown, c,†
9.7 (2) 12 (3) 9.0 (2) 7.6 (3) [22]Root, c,†
10.16 [23]b,||
10.87 [23] b,dehyd., ||
9.49 [23] b, re-hyd, ||
E: Enamel
84.3 (8.9) [24]Cusp, c, ||
77.9 (4.8) [24]Side, c, ||
48 (6) 46 (5) [22]Cusp, c,‡
33 (2) 32 (4) [22]Axial (side), c, ^
9.7 (3) [22]Axial (side), c, ||
12 (3) [22]Occlusal, c, ||
E,(∞): Dentin
12 [25]c, constant strain,
hydrated,^,‡
H1(t): Dentin
0.38 [25] c, constant
(0.136) strain, hydrated,^,‡
E: modulus of elasticity (GPa); Er (∞): relaxed modulus (GPa); H1(t): distribution of relaxation
times (GPa); c: compression; t: tension; b: three-point bending.
|| Applied load approximately parallel to either the long axis of the enamel rods or dentinal tubules.
^ Applied load approximately perpendicular to either the long axis of the enamel rods or dentinal
tubules.
† Applied load with respect to either the long axis of the enamel rods or dentinal tubules was variable.
‡ Type of tooth unknown or various teeth used for measurement; data are tabulated under molar.
Note: standard deviations are given in parentheses.
Table A3.10 Toughness, Fracture Toughness, and Work of Fracture of Human Dentin and Enamel
Incisors Canine Pre-molars Molars Source, comments
Fracture
Toughness,
Kc (MNm-3/2)
[28]*
Enamel 0.97(0.09) 1.00(0.23) Maxillary,
cervical, †
1.27(0.09) 0.7(0.08) Mandibular,
cervical, †
Toughness (MJm-3)
Dentin 62.7 (6.2)† [27] Root, shear,
oil storage, ^, ‡
2.4 (1.1) [17]Tension,
crown, hydr., ||
Work of Fracture
(102 Jm-2) Dentine 2.7 (1.6) [29],^
5.5 (1.7) [29] ||
Enamel 1.9(0.56) [29], ^
0.13(.065) [29] ||
|| Applied load approximately parallel to either the long axis of the enamel rods or dentinal
tubules.
^ Applied load approximately perpendicular to either the long axis of the enamel rods or dentinal
tubules.
† Applied load with respect to either the long axis of the enamel rods or dentinal tubules was
variable.
‡ Type of tooth unknown or various teeth used for measurement; data are tabulated under
molar.
* Microindentation method used. Load was 500 g with a Vickers’ indenter.
Note: standard deviations are given in parentheses.
A3 Dentin and Enamel 31
Table A3.11 Hardness of Fracture of Human Dentin and Enamel (see notes for units)
Incisor Pre-molar Molar Source, comments
Enamel 365 (35) [30] >90% incisors,®, †
393 (50) [30]‡, molars and premolars,®, †
385 (5.8) [31]⨂ † ‡
367 (17) [32] ||,⨂, incisors, premolars
327 (34) [32]^,⨂, incisors, premolars
Dentin
25–81.7 [33]∆, ||, [34]a
97.8 [33]a, calculated for zero tubule density
44.5–80.9 [14]◊, ||, [34]a
100 [14]a, calculated for zero tubule density
75 (0.8) [31]⨂,†,‡
a
Inverse correlation between hardness and dentinal tubule density.
|| Applied load approximately parallel to either the long axis of the enamel rods or dentinal tubules.
^ Applied load approximately perpendicular to either the long axis of the enamel rods or dentinal
tubules.
† Applied load with respect to either the long axis of the enamel rods or dentinal tubules was variable.
‡ Type of tooth unknown or various teeth used for measurement; data are tabulated under molar.
* Microindentation method used. Load was 500 g with a Vickers’ indenter.
® Knoop hardness test using 500 g load.
⨂ Knoop microhardness test using 50 g load.
∆ Knoop microhardness test using 100 g load.
◊ Micromdentation method used. Load was 50 g with a Vickers’ indenter.
Table A3.15 Critical Surface Tensions (γc) of Human Enamel and Dentin
Critical Surface Tension, γc
(dynes cm-1) Source, Comments
Enamel 46.1 (40.0 – 55.6)a [38]*, calculated from polar and
Ground surface non-polar liquids
In situ enamel, γcd 45.3 ± 70.2b [39]∆, calculated from polar liquids,
In situ enamel, γcd 32.9 ± 4.7 [38]∆, calculated from non-polar liquids
In situ enamel, γcd 32 [37]†, calculated from non-polar liquids
Dentin 45.1 (40.7 – 51.1)a [38]*, calculated from polar and
non-polar liquids
a
Range of values from different test liquids.
b
Standard deviation.
* Plane ground dentin surfaces, measurements from 46 erupted and unerupted teeth, mixed loca-
tion (molars, premolars, incisors). Parentheses: standard error.
∆ In situ measurements from 76 test subjects: 29 female and 47 male. Measurements made on
teeth with intact pellicle (i.e., biofilm). γcp only calculated from glycerol and thiodiglycol.
† Average of 4 teeth from 2 subjects. γcd calculated from non-polar liquids.
The quality of data presented can be inferred from the standard deviations or stan-
dard error associated with the mean values. In some cases the error can be attributed
to either small sample populations or specimen preparation. Where possible, either
the number of specimens used or the number of replications of a measurement was
reported. The reader should use this information as a guideline of the quality of
data. When data are reported for small sample populations, then these data were
usually the only source for a given physical property. In review of the literature,
specimen preparation appears to have had the most influence on the precision and
accuracy of data. Sample collection and storage conditions (e.g., dehydration, cross-
linking agents, exogenous contamination) need to be taken into consideration when
34 K.E. Healy
utilizing the information tabulated. Additional sources of error are dependent on the
analytical technique or test method used to make the measurement. It is more diffi-
cult to discern the influence of the instrumentation on the reliability of the
measurements. However, confidence of the accuracy was judged based on the use
of adequate control samples with known physical properties (e.g., correction of
mechanical data). In light of these comments, data in the literature were deemed
most accurate and appropriate for this handbook when the following conditions
were met: the sample population was large; non-destructive specimen preparation
and storage conditions were used; and, multiple replications of measurements on a
single sample were performed.
There are significant omissions in the data available in the literature. Most nota-
ble, is the lack of quantitative analysis of the organic phase of dentin and enamel,
and determination of the viscoelastic properties of dentin. The lack of data is attrib-
uted to the technical difficulty required to make such measurements and the hetero-
geneous nature of the dentin, which imparts large variations in these data depending
on anatomical location. Other significance absences are the lack of electrical and
thermal properties. Finally, vacancies in the tables provided demonstrate omissions
in available data.
Additional Reading
Carter, J.M., Sorensen, S.E., Johnson, R.R., Teitelbaum, R.L. and Levine, M.S.
(1983) Punch Shear Testing of Extracted Vital and Endodontically Treated Teeth.
J. Biomechanics 16(10), 841–848.
Utilized a miniature punch shear apparatus to determine shear strength and
toughness perpendicular to the direction of dentinal tubules. Dentin harvested from
the cemento-enamel junction to one-third the distance to the root apex. Strengths:
novel measurements, precise measurements, defined specimen location, defined ori-
entation of testing. Limitations: tooth type not defined for ‘constrained’ tests, teeth
stored in mineral oil prior to testing.
Driessens, F.C.M., and Verbeeck, R.M.H. (1990a) The Mineral in Tooth Enamel
and Dental Carries. In Biominerals, F.C.M and Verbeeck, R.M.H. (eds), CRC Press,
Boca Raton, Florida, pp. 105–161.
Driessens, F.C.M., and Verbeeck, R.M.H. (1990b) Dentin, Its Mineral and
Caries, In Biominerals, F.C.M and Verbeeck, R.M.H. (eds), CRC Press, Boca Raton,
Florida, pp. 163–178.
An authoritative text on biominerals with an excellent review of the properties of
enamel and dentin. An excellent supplement to this handbook.
Glantz, P-O. (1969) On Wetability and Adhesiveness. Odontologisk Revy, 20
supp. 17, 1–132.
Comprehensive assessment of the wetability of human enamel and dentin.
Strengths include using multiple probe liquids on numerous teeth.
A3 Dentin and Enamel 35
Korostoff, E., Pollack, S.R., and Duncanson, M.G. (1975) Viscoelastic Properties
of Human Dentin. J. Biomedical Materials Res., 9, 661–674.
Measured some viscoelastic properties of human radicular dentin under constant
strain. Linear viscoelastic theory applied. Strengths: unique examination of visco-
elastic properties, defined orientation of dentinal tubules, storage conditions and
testing environment well controlled. Limitations: large scatter in H1(t), mixed data
for different teeth.
Marshall, G.W. (1993) Dentin: Microstructure and Characterization. Quintessence
International, 24(9), 606–616.
A Review of the microstructure and characterization of dentin.
Waters, N.E. (1980) Some Mechanical and Physical Properties of Teeth.
Symposia of the Society for Experimental Biology, 34, 99–135.
Concise review of mechanical and physical properties of teeth. Good paper for
anatomy of enamel and dentin.
References
1. Waters, N.E. (1980) Some mechanical and physical properties of teeth. Symp. Soc. Exp. Biol.,
34, 99–135
2. Driessens, F.C.M. and Verbeeck R.M.H. (1990) The mineral in tooth enamel and dental caries.
In: Biominerals, F.C.M. and Verbeeck, R.M.H. (eds), CRC Press, Boca Raton, Florida,
pp. 105–161
3. Driessens, F.C.M. and Verbeeck, R.M.H. (1990) Dentin, its mineral and caries, In: Biominerals,
Driessens, F.C.M. and Verbeeck, R.M.H. (eds), CRC Press, Boca Raton, Florida, pp 163–178
4. Söremark, R. and Samsahl, K. (1961) Gamma-ray spectrometric analysis of elements in nor-
mal human enamel. Arch. Oral. Bio., Special Suppl., 6, 275–283.
5. Derise, N.L., Ritchey, S.J. and Furr, A.K. (1974) Mineral composition of normal human
enamel and dentin and the relation of composition to dental caries: I Macrominerals and com-
parison of methods of analyses. J. Dental Res., 53(4),847–852
6. LeGeros, R.Z., Silverstone, L.M., Daculsi, G. et al. (1983) In vitro caries-like lesion formation
in F-containing tooth enamel. J. Dental Res., 62(2), 138–144
7. Lakomaa, E-L. and Rytömaa, I. (1977) Mineral composition of enamel and dentin of primary
and permanent teeth in Finland. Scand. 1. Dent. Res., 85, 89–95.
8. Cutress, T.W. (1979) A preliminary study of the microelement composition of the outer layer
of dental enamel. Caries Res., 13, 73–79.
9. Losee, F.L., Cutress, T.W. and Brown, R (1974) Natural elements of the periodic table in
human dental enamel. Caries Res., 8, 123–134.
10. Retief, D.H., Cleaton-Jones, P.E., Turkstra, J. et al. (1971) The quantitative analysis of sixteen
elements in normal human enamel and dentine by neutron activation analysis and high-
resolution gamma-spectrometry. Arch. Oral Bio., 16, 1257–1267.
11. Curzon, M.E.J. and Losee, F.L. (1977) Dental caries and trace element composition of whole
human enamel: Eastern United States. J. Amer. Dental Assoc., 94, 1146–1150.
12. Curzon, M.E.J. and Losee, F.L. (1978) Dental caries and trace element composition of whole
human enamel: Western United States. J. Amer. Dental Assoc., 96, 819–822.
13. Kodaka, T., Debari, K., Yamada, M. et al. (1992) Correlation between microhardness and
mineral content in sound human enamel. Caries Res., 26, 139–141.
14. Panighi, M. and G’Sell, C. (1992) Influence of calcium concentration on the dentin wetability
of an adhesive. J. Biomed. Mater. Res., 26, 1081–1089.
36 K.E. Healy
15. Holcomb, D.W. and Young, R.A. (1980) Thermal decomposition of human tooth enamel.
Calcif Tiss. Intern., 31, 189–201
16. Sakae, T. (1988) X-Ray diffraction and thermal studies of crystals from the outer and inner
layers of human dental enamel. Archs. Oral Bio., 33(10), 707–713.
17. Huang, T.-J.G., Schilder, H. and Nathanson, D. (1992) Effects of moisture content and end-
odontic treatment on some mechanical properties of human dentin. J. Endodontics, 18(5),
209–215
18. Kerebel, B, Daculsi, G. and Kerebel, L.M. (1979) Ultrastructure studies of enamel crystallites.
J. Dental Res., 58(B), 844–851.
19. Jervøe, P. and Madsen, H.E.L. (1974) Calcium phosphates with apatite structure. I. Precipitation
at different temperatures. Acta Chem. Scand., A28, 477–481.
20. Daculsi, G., Kerebel, B. and Verbaere, A (1978). (Méthode de mesure descristaux d’apatite de
la dentine humanie en microscopie électronique en transmission de Haute Résolution)(Fr.)
(Method of measurement of apatite crystals in human dentin by high resolution transmission
electron microscopy), Comptes Rendu Acad. Sci. Paris, Sér. D., 286, 1439.
21. Voegel, J.C. and Frank, R.M. (1977) Ultrastructural study of apatite crystal dissolution in
human dentine and bone. Jour. Bioi. Buccale, 5, 181–194.
22. Lehman, M.L. (1963) Tensile strength of human dentin. J. Dent. Res., 46(1), 197–201
23. Stanford, J.W., Weigel, K.V., Paffenbarger, G.C. et al. (1960) Compressive properties of hard
tooth tissues and some restorative materials. J. American Dental Assoc., 60, 746–756.
24. Jameson, M.W., Hood, J.A.A. and Tidmarsh, B.G. (1993) The effects of dehydration and rehy-
dration on some mechanical properties of human dentine. J. Biomech., 26(9), 1055–1065.
25. Craig, R.G., Peyton, F.A. and Johnson, D.W. (1961) Compressive properties of enamel, dental
cements, and gold. J. Dent. Res., 40(5), 936–945.
26. Korostoff, E., Pollack, S.R and Duncanson, M.G. (1975) Viscoelastic properties of human
dentin. J. Biomed. Mater. Res., 9, 661–674.
27. Bowen, R.L. and Rodriguez, M.S. (1962) Tensile strength and modulus of elasticity of Tooth
Structure and Several Restorative Materials. J. American Dental Assoc., 64, 378–387.
28. Carter, J.M., Sorensen, S.E., Johnson, R.R., et al. (1983) Punch shear testing of extracted vital
and endodontically treated teeth. J. Biomech., 16(10), 841–848.
29. Hassan, R, Caputo, A.A. and Bunshah, R.F. (1981) Fracture toughness of human enamel.
J. Dent. Res., 60(4), 820–827.
30. Rasmussen, S.T., Patchin, R.E., Scott, D.B. et al. (1976) Fracture properties of human enamel
and dentin. J. Dent. Res., 55(1), 154–164.
31. Caldwell, R.C., Muntz, M.L., Gilmore, R.W. et al. (1957) Microhardness studies of intact
surface enamel. J. Dent. Res., 36(5), 732–738.
32. Remizov, S.M., Prujansky, L.Y. and Matveevsky, R.M. (1991) Wear resistance and microhard-
ness of human teeth. Proc. Inst. Mech. Eng., Part H: J. Eng. in Med., 205(3), 201–202.
33. Davidson, C.L., Hoekstra, I.S. and Arends, J. (1974) Microhardness of sound, decalcified and
etched tooth enamel related to the calcium content. Caries Res., 8, 135–144.
34. Pashley, D.H., Andringa, H.J., Derkson, G.D. et al. (1987) Regional variability in the perme-
ability of human dentin. Arch. Oral Biol., 32(7), 519–523.
35. Pashley, D.H., Okabe, A. and Parham, P. (1985) The Relationship between dentin microhard-
ness and tubule density. Endod. Dent. Traumatol., 1, 176–179.
36. Baier, R.E. and Zisman, W.A. (1975) Wetting properties of collagen and gelatin surfaces, in
‘Applied Chemistry at Protein Interfaces’, vol. 145, Advances in Chemistry series (ed.
R.F. Gould), American Chemical Society, Washington DC, pp. 155–174.
37. Jendresen, M.D., Baier, R.E. and Glantz, P-O. (1984) Contact angles in a biological setting:
Measurements in the human oral cavity. J. Coli. Interface Sci., 100(1), 233–238.
38. Baier, R.E. (1973) Occurrence, nature, and extent of cohesive and adhesive forces in dental
integuments. in: Surface Chemistry and Dental Integument’s. Lasslo, A. and Quintana, R.P.
(eds), Thomas, Springfield, IL pp. 337–391.
39. Glantz, P-O. (1969) On wetability and adhesiveness. Odontologisk Revy, 20 supp. 17, 1–132.
40. Jendresen, M.D. and Glantz, P-O. (1980) Clinical adhesiveness of the tooth surface. Acta
Odontol. Scand., 38, 379–383.
Chapter B1
Cartilage
J.R. Parsons
B1.1 Introduction
Articular or hyaline cartilage forms the bearing surfaces of the movable joints of the
body. Hyaline cartilage also exists in tissues of the larynx, tracheal tube rings, rib
and costral cartilage, nasal septum and in the growth plates of long bones. As a
bearing surface, this tough, resilient tissue displays exceptional mechanical and
tribologic properties due exclusively to the unique interaction of the constituents of
the tissue extracellular matrix. Usually, the phenotypic cells (chondrocytes) of car-
tilage make up less than 10% of the total volume of the tissue and have not been
considered to contribute to the mechanical properties of the tissue. The extracellular
matrix consists of a tight collagen fiber network which contains and constrains a
highly hydrophilic gel of aggregated proteoglycan macromolecules. Collagen
accounts for approximately 50% of the dry weight of the tissue, the remainder being
proteoglycans and cellular material. In the fully hydrated state, water contributes
60% to 80% of the wet weight of the tissue. Mechanically, intact normal articular
cartilage behaves as a linear viscoelastic solid. This behavior is the result of viscous
drag of fluid through the tissue in concert with the intrinsic properties of the extra-
cellular matrix. Further, fluid exudation across the cartilage surface in response to
physiologic loading is thought to play a significant role in the lubrication of joints.
The importance of articular cartilage as a bearing surface has led to extensive
mechanical and tribologic studies of this tissue.
B1.1.2 Fibrocartilage
Fibrocartilage contains a higher dry weight percentage of collagen and less proteo-
glycan than does articular cartilage. Consequently, in the hydrated state, fibrocarti-
lage contains less water. Fibrocartilage is generally considered to be tougher and
somewhat less resilient than articular cartilage. In humans, fibrocartilage is found in
the meniscus of the knee joint, the annulus fibrosus of the intervertebral disc and in
the temperomandibular joint. The mechanical behavior of fibrocartilage has not
been studied to the extent of articular cartilage; however, it is not as satisfactory a
load bearing material.
Elastic cartilage is more elastic and resilient in nature than is either hyaline or fibro-
cartilage. This is the result of lower collagen content coupled with the presence of
elastin fibers. Elastic cartilage is found in the human epiglottis, external ear struc-
tures and eustachian tube. As elastic cartilage is not a major structural component
of the musculoskeletal system, the mechanical properties of this tissue have been
largely ignored.
B1.2 Composition
Table B1.1 Cartilage Composition
Water (wt%) Organic (wt%) Source
Articular Cartilage 60–80 20–40 [1]
Type II Collagen — 15–20
Other collagen — <2
Proteoglycan — 10
Fibrocartilage 74 26 [2]
Type I Collagen — 20
Other collagen — <1
Proteoglycan <1
Compressive cartilage properties have often been examined using creep indenta-
tion, confined compression or free/unconfined compression methods. Indentation
techniques permit in situ testing without the necessity of special specimen prepara-
tion as with tensile testing. However, the extraction of intrinsic mechanical
B1 Cartilage 39
parameters from creep indentation data is analytically complex [3, 4]. Confined
compression or unconfined compression tests require preparation of cylindrical
cored specimens of tissue and underlying bone. With unconfined compression, the
free draining tissue edges and low aspect ratio, layered nature of the test specimen
may introduce error. Compression of a laterally confined specimen by a porous
plunger produces uniaxial deformation and fluid flow. Confined compression creep
data has been analyzed to yield an aggregate equilibrium compressive modulus and
permeability coefficient [5] and uniaxial creep compliance [6].
Tensile properties for human articular have been determined by cutting standard
tensile specimens from the cartilage surface and performing constant strain rate,
creep or stress relaxation tensile tests. Test results are strongly influenced by
40 J.R. Parsons
collagen volume fraction and orientation and are largely insensitive to proteoglycan
content [7] Collagen volume fraction and orientation is highest in the cartilage sur-
face layer. Collagen content and orientation diminishes in subsequent lower layers.
Shear properties for articular cartilage have been determined through torsional
creep, stress relaxation and torsional dynamic tests of excised cartilage disks. Creep
and dynamic shearing of rectangular cartilage specimens between plates has been
conducted on animal tissue (usually bovine) but not human tissue. When torsional
shear strains remain small, the observed shear properties are flow independent. That
is, under small strain conditions, fluid flow is negligible and viscoelastic behavior
can be attributed strictly to the collagen/proteoglycan extracellular matrix.
B1 Cartilage 41
Poisson’s ratio has been calculated directly from tensile tests (v = 0.37–0.50) [10]
and indirectly from torsional shear and confined compression creep data (v = 0.37–
0.47) [6, 9]. More recently, the relationship between Poisson’s ratio, n, aggregate
modulus, Ha, and permeability, k, have been established for cartilage indentation
testing based on biphasic (fluid and porous solid) constitutive theory [15]. Using a
complex numerical solution and curve fitting scheme, Poisson’s ratio can be
extracted from indentation data, resulting in values of v = 0.00–0.30 [11, 12, 16].
However, care must be exercised in interpreting such indirect measures of Poisson’s
ratio as unexpected results can arise; e.g. v = 0.0.
B1.3.5 Permeability
The porous solid matrix of articular cartilage permits the movement of interstitial
water in response to a pressure gradient. Flow of water through and across the tissue
is largely responsible for the viscoelastic character of cartilage. Flow is related to
tissue permeability through the hydraulic permeability coefficient, k, as defined by
Darcy’s law. The permeability coefficient has been measured in flow chambers
where a known pressure gradient produces flow across a cartilage layer of known
thickness and area (k = 4.0 – 17.0 x 10-16m4/Ns)[7]. However, such experiments have
demonstrated significant decreases in permeability coefficient with increasing pres-
sure gradient, increasing compressive tissue strain and with increasing proteoglycan
content. Evoking biphasic constitutive theory with numerical solutions and/or curve
fitting routines permits an indirect determination of the permeability coefficient
from confined compression creep data and creep indentation data (k = 5.2 – 21.7 ×
10-16m4/Ns)[11, 12].
Healthy articular cartilage has remarkable tribologic properties. Under high load
conditions the tissue displays extremely low frictional coefficients and virtually
undetectable wear. The dynamic coefficient of friction, μd, has been measured in
whole joints using Stanton pendulum or other pendulum techniques where the joint
forms the pendulum pivot (μd = 0.015–0.04)[17–19]. The coefficients of friction of
cartilage plugs bearing on other materials has been determined for human and ani-
mal tissue but these sorts of experiments have little relevance for actual in situ car-
tilage behavior and are not reported here.
The lubrication of articular cartilage remains a subject of continuing debate and
no one lubrication mechanism can be clearly identified. Both fluid film and bound-
ary lubrication are thought to play primary roles in joint lubrication and the domi-
nance of one or the other probably depends on loading and velocity conditions.
Further as cartilage is a relatively soft viscoelastic material, elastohydrodynamics
42 J.R. Parsons
may discourage fluid film breakdown and thus promote hydrodynamic lubrication.
Exudation of fluid across the cartilage surface in response to an advancing load has
also been suggested to aid lubrication.
No reliable wear tests have been performed on human articular cartilage bearing
surfaces under physiologic conditions.
Additional Reading
Freeman, MAR (ed.) (1979) Adult Articular Cartilage, 2nd ed., Pitman Medical
Publishing Co, Kent, UK.
Although now somewhat out of date, this classic text forms the basis for current
thinking on cartilage biochemistry, physiochemistry, biomechanics and tribology.
The volume of original data, found nowhere else, is truly impressive.
Mow V.C., Holmes, M.H. and Lai, W.M. (1984): Fluid transport and mechanical
properties of articular cartilage. A review. J. Biomech., 17:377–394.
This survey article provides an historical perspective of cartilage mechanics
research and leads the reader through the modern biphasic theory of cartilage
mechanics at the material level. References provided are particularly useful in
developing a bibliography of the important classic studies in this field.
Mow, V.C. and Ratcliffe, A (eds)(1993): Structure and Function of Articular
Cartilage, CRC Press, Boca Raton.
This up-to-date monograph is perhaps the best current work on the subject.
Details from many of the references in this section (below) can be found in the sec-
tion on cartilage biomechanics.
B1 Cartilage 43
References
V.M. Gharpuray
B2.1 Introduction
The human menisci and intervertebral discs perform several important mechanical
functions in the human body. The ability to perform these functions and conse-
quently their intrinsic biomechanical properties are dependent on the interaction of
the constituents of these structures. Both the menisci and intervertebral discs have a
fibrocartilaginous structure that consists of two distinct phases: a fluid phase con-
sisting of mainly water and dissolved electrolytes, and a solid phase composed of
highly oriented collagen fibers, cells, proteoglycans and other proteins. As with all
other biological materials, both menisci and discs exhibit non-linear viscoelastic
and anisotropic properties. The non-linear stiffness or elasticity of the structure is
imparted by the collagen fibers and to a lesser extent by osmotic pressures within
the tissue which are generated by the degree of hydration [1, 2]. The viscoelastic or
energy dissipation properties are a result of fluid flow within and through the struc-
tures and also of molecular relaxation effects from the motion of long chains of
collagen and proteoglycans [3]. Anisotropy is a consequence of the orientation and
concentration of collagen fibers within the proteoglycan gel.
A normal adult human knee contains two menisci – the lateral and the medial,
whose average lengths are 38 and 45 mm, and average volumes are 2.9 and 3.45 cm3
respectively [4]. At the femoral articulating surface of each meniscus, for a depth of
approximately 100 µm, fine collagen fibrils (mainly Type II) are randomly oriented
to form a woven mesh [5, 6]. Beneath this surface layer, larger rope-like bundles
(approximately 100 µm in diameter) of Type I collagen are arranged predominantly
in the circumferential direction. In the posterior half of the medial meniscus how-
ever, the fibers are not as highly oriented [6, 7]. A few radial fibers may also be seen
interspersed within the circumferential fibers [6–8]. At the tibial surface is another
articulating layer, in which the principal orientation of the fibers is radial.
Collagen and other proteins make up the organic content of the meniscus, while
dissolved electrolytes make up the inorganic content (which is negligibly small).
The collagen is primarily Type I (98%) with small amounts of Type II, III and
Type V [9], and comprises up to 25% of the wet weight and 90% of the dry weight
in human material (Table B2.1). 10% of the dry weight and up to 2% of the wet
weight are due to non-collagenous proteins, which consist predominantly of proteo-
glycans, and smaller amounts of structural glycoproteins, cell membrane bound
receptors and intercalated membrane glycoproteins. These proportions do not
appear to vary with location in the menisci [10].
There are 23, approximately cylindrical, intervertebral discs in the human spine
that account for 20–30% of its overall length [13]. The cross sectional shape varies
with level: roughly elliptical, rounded triangular and kidney shaped in the cervical,
thoracic and lumbar regions respectively [14]. The cross sectional shape is some-
times quantified by a shape index (S) defined by S = 4πA/C2 where A= cross sec-
tional area, and C =circumference of the cross section (Table B2.2). Most discs are
wedge-shaped in sagittal section with the anterior height greater that the posterior
height. The cross sectional area increases with level such that lumbar discs are
larger than cervical discs.
The intervertebral disc contains two distinct regions, the nucleus pulposus and
the annulus fibrosus. The nucleus occupies about 50% of the volume of the disc, and
contains mostly water and small amounts of randomly oriented collagen fibers, cells
and non-collagenous proteins [19]. The annulus is a tough ring-like structure that
surrounds the nucleus. Highly oriented collagen fibers (primarily Types I and II)
form a laminate structure in the annulus, with approximately 20–25 laminae ori-
ented alternately at approximately +(60–70)° and –(60–70)° to the spinal axis [14,
20, 21]). Type IX collagen is believed to cross link the Types I and II fibers, and
provide some resistance to circumferential tears. Collagen content and fiber orienta-
tion is highest in the outermost layers, and both decrease as the nucleus is approached
from the periphery of the disc (Table B2.3) The water content in the disc varies with
position in the disc: it is highest in the nucleus and lowest in the outermost layers of
the annulus [22, 23]. It has also been shown that the water content varies with cir-
cumferential position [4], and is higher in the posterior of the disc. The nucleus
loses water and becomes more fibrous and desiccated with age, causing the bound-
ary between the annulus and the nucleus to become less clear [19, 23].
Experimental data suggest that water is capable of flowing through both meniscal
and discal tissues, and is dependent on a material property of the tissue called
hydraulic permeability [25], which may be modeled by Darcy’s Law as:
48 V.M. Gharpuray
AD P
Q=k (B2.1)
h
(f f ) 2
K= (B2.2)
k
where Φf is the porosity of the tissue and is defined as the ratio of interstitial fluid
volume to total tissue volume.
For human meniscal (annulus) tissue, k, the permeability is 2.5 × 10-16 m4/Ns
[22], about one third of that reported for bovine tissue: 8.1 ×10-16 m4/Ns [26]. There
is no significant variation in the permeability coefficient with location of the speci-
men. The porosity (Φf) of both tissues is approximately 0.75, and therefore the drag
coefficient, K, is very high and ranges from 1014 to 1015 Ns/m4.
s = A[e Be - 1] (B2.3)
where A and B are constants for the given material. The constant B is proportional
to the tangent modulus (i.e., dσ/dϵ), and sometimes a third constant C is defined as
C = A*B, and is the tangent modulus as σ → 0 [8].
The macroscopic tensile strength of the entire meniscus was studied by Mathur
et al. [4] by gripping the horns of the meniscus, and stretching it to failure. The
results suggested that the medial meniscus was significantly weaker than the lateral
meniscus (ultimate loads of 247 N and 329 N respectively), and that the mode of
failure was not by transverse cracking, but predominantly by oblique (medial) or
spiral (lateral) tearing.
Strength and modulus of the meniscus vary with different locations and with dif-
ferent orientations of the specimen due to structural and compositional changes
(Table B2.4). For loading parallel to the fibers, it appears that the meniscus may be
stronger in the anterior location, and that the lateral meniscus may be stronger than
the medial meniscus. This may be explained in part by the fact that the fiber orienta-
tion is more random in the posterior part of the medial meniscus.
B2 Fibrocartilage 49
the phase angle shift between the applied displacement and the measured force. The
dynamic modulus is than obtained as
F F0 i d
G* = = e = G¢ + iG¢¢ (B2.4)
u u0
where G′ and G″ are the loss and storage moduli respectively. In some cases, it may
be more convenient to express viscoelastic properties in terms of the magnitude of
the dynamic shear modulus and the phase angle shift as
B2 Fibrocartilage 51
¢¢ G¢¢
| G | * = (G¢ 2 + G 2 ); d = tan -1 ( ) (B2.5)
G¢
The anisotropic viscoelastic properties in shear of the meniscus have been deter-
mined by subjecting discs of meniscal tissue to sinusoidal torsional loading [35] (Table
B2.8). The specimens were cut in the three directions of orthotropicsymmetry, i.e.
circumferential, axial and radial. A definite correlation is seen with the orientation of
the fibers and both the magnitude of the dynamic modulus |G*| and the phase angle δ.
The viscoelastic properties of the human intervertebral disc have been modeled [36,
37] using the three-parameter solid. The parameters were obtained by fitting experimen-
tally obtained creep curves to analytical equations using linear regression (Table B2.9).
B2.6 Discussion
Since it is nearly impossible to carry out meaningful experiments in vivo on the human
disc or meniscus, the properties reported above have been obtained from cadaveric tis-
sue. Test specimens were obtained from autopsy material (10–48 hours after death), and
were either tested immediately or stored frozen for varying periods of time before test-
ing. Statistical analyses of these data show high standard deviations and errors may be a
consequence of aging and degeneration, diurnal changes or surgical interventions) cause
a subsequent change in mechanical properties. Further, it has been well documented that
the mechanical properties of collagenous tissues change with storage medium, storage
temperature, time after death and ‘preconditioning’ state [38, 39]. Both the disc and the
meniscus contain highly oriented collagen fibers, and location and orientation of the test
specimen can cause significant changes in the test results. Additional factors such as sex,
diet, and level of activity also play a relatively minor role in this variation.
Additional Reading
Ghosh, P. (ed.) (1988) The Biology of the Intervertebral Disc, Vol I and II, CRC
Press, Boca Raton.
52 V.M. Gharpuray
One of the most comprehensive texts available about the intervertebral disc. It is
written from the biological perspective, and contains exhaustive information about
each component of the disc. Volume I includes chapters on disc structure and
development, vasculature, innervation, collagen and non-collagenous proteins.
Volume II contains information on nutrition and metabolism, mechanics, pathology
and disease states.
Mow, V.C., Arnoczky, S.P. and Jackson, D.W. (1992) Knee Meniscus: Basic and
Clinical Foundations. Raven Press, New York.
This monograph is designed to serve as a comprehensive reference for clinicians
and researchers interested in the meniscus. It includes chapters on gross anatomy,
structure and function of the menisci and their mechanical behavior, pathological
disorders, clinical and surgical methods of treatment and meniscal disorders.
Mow, V.C. and Hayes, W.C. (1991) Basic Orthopaedic Biomechanics, Raven
Press, New York.
This book is aimed at teaching senior engineering students or orthopaedic resi-
dents the fundamental principles of biomechanics of the musculoskeletal system.
The book contains several chapters on the mechanics of joints, and the properties
and functions of joint tissues. The chapter devoted to articular cartilage and the
meniscus includes a review of collagen-proteoglycan interactions, and how these
directly affect the mechanical behavior of the tissue. The biphasic and the triphasic
theories for the viscoelastic properties are also discussed.
White, A.A. and Panjabi, M.M., (1990) Clinical Biomechanics of the Spine,
J.B. Lippincott Company, Philadelphia.
An excellent reference book for an engineer or a physician interested in the spine.
Each topic is written from the viewpoint of a biomechanician and the topics covered
include kinetics and kinematics of vertebral joints, pathological disorders of the
spine and their surgical management. Chapter 1 contains an introductory section on
the intervertebral disc that describes its structure, function and biomechanics.
References
1. Armstrong, C.G. and Mow, V.C. (1982) Variations in the intrinsic mechanical properties of
human articular cartilage with age, degeneration and water content. J. Bone Joint Surg., 64A,
88–94.
2. Mow, V.C., Holmes, M.H. and Lai, W.M. (1984) Fluid transport and mechanical properties of
articular cartilage: A review. J. Biomech., 102, 73–84.
3. Hayes, W.C. and Bodine, A.J. (1978) Flow independent viscoelastic properties of articular
cartilage matrix. J. Biomech., 11, 407–419.
4. Mathur, P.D., McDonald, J.R. and Ghormley, R.K. (1949) A study of the tensile strength of the
menisci of the knee. J. Bone Joint Surg., 32A, 650–654.
5. Aspden, R.M., Yarker, Y.E. and Hukins, D.W.L. (1985) Collagen Orientations in the meniscus
of the knee joint. J. Anat., 140, 371–380.
6 Bullough, P.G., Munuera, L., Murphy, J. and Weinstein, A.M. (1970) The strength of the
menisci of the knee as it relates to their fine structure. J. Bone Joint Surg., 52B, 564–570.
B2 Fibrocartilage 53
7 Fithian, D.C., Kelly, M.A. and Mow, V.C. (1990) Material properties and structure-function
relationships in the menisci. Clin. Orthop. Rei. Res., 252, 19–31.
8 Mow, V.C., Zhu, W. and Ratcliffe, A. (1991) Structure and function of articular cartilage and
meniscus, in Basic Orthopaedic Biomechanics, (eds V.C. Mow and W.C. Hayes), Raven Press,
New York, pp. 143–198.
9 Eyre, D.R. and Wu, J.J. (1983) Collagen of fibrocartilage: A distinctive molecular phenotype
in bovine meniscus. F. E. B. S. Letters, 158, 265–270.
10 Fithian, D.C., Zhu, W.B., Ratcliffe, A., Kelly, M.A. and Mow, V.C. (1989b) Exponential law
representation of tensile properties of human meniscus. Proceedings of the Institute of
Mechanical Engineers. The Changing Role of Orthopaedics, Mechanical Engineering
Publications Limited, London. pp. 85–90
11 Ghosh, P. and Taylor, T.K.F. (1987) The knee joint meniscus: A fibrocartilage of some distinc-
tion. Clin. Orthop. Rei. Res., 224, 52–63.
12 Ingman, A.M., Ghosh, P. and Taylor, T.K.F. (1974) Variation of collagenous and non-
collagenous proteins of human knee joint menisci with age and degeneration. Gerontology, 20,
212–223.
13 Pooni, J.S., Hukins, D.W.L., Harris, P.F., Hilton, R.C. and Davies, K.E. (1986) Comparison of
the structure of human intervertebral discs in the cervical, thoracic and lumbar regions of the
spine. Surg. Radiol. Anat., 8, 175–182.
14 Hirsch, C., Inglemark, B-H. and Miller, M. (1963) The anatomical basis of back pain. Acta
Orthop. Scand., 33, 2–17.
15 Lin, H.S., Liu, Y.K. and Adams, K.H. (1978) Mechanical response of the lumbar intervertebral
joint under physiological (complex) loading. J. Bone Joint Surg., 60A, 41–55.
16 Lin, H.S., Liu, Y.K., Ray, G. and Nikravesh, P. (1978) System identification for material prop-
erties of the intervertebral joint. J. Biomech., 11, 1–14.
17 Taylor, J.R. (1975) Growth of human intervertebral discs and vertebral bodies. J. Anat., 120,
49–68.
18 Panjabi, M.M., Summers, D.J., Pelker, R.R., Videman, T., Friedlander, G.E. and Southwick,
W.O. (1986) Three-dimensional load-displacement curves due to forces on the cervical spine.
J. Ortho. Res., 4, 152–161.
19 Inoue, H. (1981) Three-dimensional architecture of lumbar intervertebral discs. Spine, 6,
139–146.
20 Panagiotacopulos, N.D., Knauss, W.G. and Bloch, R. (1979) On the mechanical properties of
human intervertebral disc materials. Biorheology, 16, 317–330.
21 Marchand, F. and Ahmed, A.M. (1988) Investigation of the laminate structure of lumbar disc
annulus fibrosus. Trans. Orthop. Res. Soc., 13, 271.
22 Best B.A., Guilak, F., Setton, L.A., Zhu, W., Saed-Nejad, F., Ratcliffe, A., Weidenbaum, M.
and Mow, V.C. (1994) Compressive mechanical properties of the human annulus fibrosus and
their relationship to biochemical composition. Spine, 19, 212–221.
23 Gower, W.E. and Pedrini, V. (1969) Age-related variations in protein-polysaccharide from
human nucleus pulposus, annulus fibrosus and costal cartilage. J. Bone Joint Surg., 51A,
1154–1162.
24 Lyons, G., Eisenstein, S.M. and Sweet, M.B.E. (1981) Biochemical changes in intervertebral
disc degeneration, Biophysics Acta, 673, 443.
25 Lai, W.M. and Mow, V.C. (1980) Drag-induced compression of articular cartilage during a
permeation experiment. Biorheology, 17, 111–123.
26 Proctor, C.S., Schmidt, M.B., Whipple, R.R., Kelly, M.A. and Mow, V.C. (1989) Material
Properties of the normal medial bovine meniscus. J. Ortho. Res., 7, 771–782.
27 Eberhardt, A.W., Keer, L.M., Lewis, J.L. and Vithoontien, V. (1990) An analytical model of
joint contact. J. Biomech. Eng., 112, 407–413.
28 Fithian D.C., Schmidt, M.B., Ratcliffe, A. and Mow, V.C. (1989a) Human meniscus tensile
properties: Regional variation and biochemical correlation. Trans. Orthop. Res. Soc., 14, 205.
54 V.M. Gharpuray
29 Brown, T., Hansen, R.J. and Torra, A.J. (1957) Some mechanical tests on the lumbosacral spine
with particular reference to the intervertebral discs. J. Bone Joint Surg., 39A, 1135–1164.
30 Hirsch, C. and Nachemson, A. (1954) New observations on the mechanical behavior of the
lumbar discs. Acta Orthop. Scand., 23, 254–283.
31 Schultz, A.B., Warwick, D.N., Berkson, M.H. and Nachemson, A.L. (1979) Mechanical prop-
erties of human lumbar spine motion segments Part I. J. Biomech. Eng., 101, 46–52.
32 Berkson, M.H., Nachemson, A. and Schultz, A.B. (1979) Mechanical properties of human
lumbar spine motion segments Part I. J. Biomech. Eng., 101,53–57.
33 Marchand, F. and Ahmed, A.M. (1989) Mechanical properties and failure mechanisms of the
lumbar disc annulus. Trans. Orthop. Res. Soc., 14, 355.
34 Galante, J.O. (1967) Tensile properties of the human lumbar annulus fibrosus. Acta Orthop.
Scand., (Suppl. 100), 1–91.
35 Chern, K.Y., Zhu, W.B. and Mow, V.C. (1989) Anisotropic viscoelastic shear properties of
meniscus. Adv. Bioeng., BED-15, 105–106.
36 Burns, M.L. et al. (1984) Analysis of compressive creep behavior of the vertebral unit sub-
jected to uniform axial loading using exact parametric solution equations of Kelvin solid
models Part I. J. Biomech., 17, 113–130.
37 Kazarian, L.E. and Kaleps, I. (1979) Mechanical and physical properties of the human inter-
vertebral joint. Technical Report AMRL-TR-79-3, Aerospace Medical Research Laboratory,
Wright Patterson Air Force Base, OH
38 Black, J. (1976) Dead or alive: The problem of in vitro tissue mechanics. J. Biomed. Mats.
Res., 10, 377–389.
39 Black, J. (1984) Tissue properties: Relation of in vitro studies to in vivo behavior, in Natural
and Living Biomaterials, Ed. G.W. Hastings and P. Ducheyne, CRC Press, Boca Raton,
pp. 5–26.
Chapter B3
Ligament and Tendon
B3.1 Structure
Fig. B3.1 Hierarchical structure of the tendon/ligament. The tropocollagen cross-linked mole-
cules are arranged in microfibrils, subfibrils, fibrils, fibers, and fascicles. The fascicles are sur-
rounded by the epitenon/epiligament and paratenon
B3.2 Composition
Ligaments and tendons consist of few cells and abundant extracellular matrix. The
cellular component, primarily consisting of fibroblasts/tenocytes, comprises 20 %
of the total tissue volume whereas the ECM accounts for the remaining 80 %
(Fig. B3.2) [1, 2]. Approximately 70 % of the ECM consists of water, which con-
tributes to modulating cellular function and viscoelastic (i.e., time dependent)
behavior. The remaining 30 % of the ECM is comprised of solids, such as collagen
and ground substance. The ground substance (non-fibrous component of the matrix)
primarily comprises hyaluronan, glycoproteins, and proteoglycans and modulates
tissue metabolism, provides shock absorption, decreases internal friction, and binds
water. The fibrous component, comprising 75–80 % of T/L dry weight, provides
B3 Ligament and Tendon 57
strength and a support framework and consists primarily of type I collagen. Lesser
amounts of types III, VI, V, XI, and XIV collagens are also present.
Fig. B3.3 Example of force-elongation (a) and stress-strain (b) curves of the ligament tested to
failure. (b) Stress-strain curve of a ligament contains the three primary regions (in red), including
the toe, linear, and failure regions. The toe region indicates a nonlinear increase in load as the tissue
elongates. The linear region indicates the area of greatest stiffness whereas the failure region indi-
cates that the collagen fibers have disrupted and failed. F = force; A0 = ligament cross-sectional
area; Li = instantaneous length after loading; L0 = original length
58 C.S. Chamberlain and R. Vanderby Jr.
stretched ligament length (Li) minus its original length (Lo) and divided by its original
length (Lo). Often this result is multiplied by 100 to be expressed as a percent. The
calculated stress-strain curve further identifies the commonly reported elastic modu-
lus (slope of the linear region), ultimate tensile strength (σu), and strain (εu).
The resulting stress-strain or force-elongation curves separate the tissue behavior
into three regions, called the toe, linear, and failure regions (Fig. B3.3b). Within the
toe region, collagen fibers of unstressed T/L are composed in a sinusoidal (“crimp”)
pattern. The mechanism causing collagen fiber crimp is not well understood but the
crimped fibers provide minimal resistance to movement over an initial range of
strains. The crimp region is very compliant. As the T/L becomes loaded, the fibers
straighten and the crimp pattern is lost [3, 4]. As loading continues, T/L stiffness
increases and greater force is required to produce an equivalent increment of strain.
Once the strain increases to 1.5–4 % [5], a nearly linear stiffness is observed (i.e.,
the linear region). As strain is further increased beyond the linear range (8–10 %
strain), fiber disruption and ultimately complete and irreversible failure occur [4, 6].
Irreversible damage begins approximately half way through the linear region, after
which the original length (Lo) can no longer be recovered with unloading, and fibril-
lar damage is observed with electron microscopy [7]. Once entering the linear
region, ultimate failure occurs after a predictable increment of strain (~9.3 %) [8].
Viscoelastic behavior of T/Ls aids and protects joint function. Viscoelasticity
adds time dependence to the above elastic phenomena. It manifests itself by allow-
ing easier joint motion after stretching, by increasing the T/L strength to resist
impact forces, and by absorbing some of the energy in dynamic loadings [5]. These
behaviors can be quantified via creep and relaxation testing. If a T/L is held at a
constant deformation (or strain), it exhibits relaxation which means that the load (or
stress) decreases over time (Fig. B3.4a). T/L also exhibits creep where deformation
Fig. B3.4 Viscoelastic properties of ligaments are demonstrated by load relaxation (a) and creep
(b). Load relaxation is demonstrated by loading the ligament (below the linear region) and main-
taining constant strain over an extended period. Load decreases rapidly the first 6–8 h, and then
continues at a low rate (a). Creep takes place when the ligament is loaded (below the linear region)
at a constant level over time. As a result, deformation increases quickly at first and then continues
at a low rate
B3 Ligament and Tendon 59
(or strain) increases over time under constant tensile force (Fig. B3.4b). As a T/L is
subjected to increased rate of loading, the linear portion of the stress-strain curve
becomes steeper, indicating greater stiffness and a greater tissue load before reach-
ing the failure strain.
In vitro ligament and tendon testing is often conducted as a composite of bone-
ligament-bone and muscle-tendon-bone, respectively, because bone blocks are easy
to grip and this minimizes gripping artifacts in the tissue. However, the maximum
load, stiffness, and elongation values can then represent the average properties of a
composite and may vary depending on the material behavior of the bone and enthe-
sis, as well as the size, shape, and orientation of the sample [9]. Mechanical proper-
ties of the human ligaments and tendons from various joints are included in
Table B3.1.
Table B3.1 Mechanical properties of the lower and upper limb ligaments and tendons
Tissue Modulus (MPa) UTS (MPa) Strain at UTS (%) Source
Knee ligament
Anterior cruciate 65–541 13–46 9–44 [10, 11]
Posterior cruciate 109-248 24–36 10–29 [12, 13]
Knee tendon
Patellar 143–660 24–69 14–27 [10, 14–17]
Ankle ligament
Lateral collateral 216–512 24–46 13–17 [18]
Medial collateral 54–321 16–34 10–33 [18, 19]
Ankle tendon
Achilles 65 24–61 24–59 [20]
Palmaris longus 2310 ± 620 91 ± 15 NA [19]
Other
Semintendinous 362 ± 22 89 ± 5 52 ± 3 [10]
Gracilis 613 ± 41 112 ± 4 34 ± 2 [10]
Shoulder ligament
Inferior glenohumoral 30–42 5–6 8–15 [21, 22]
Capsule 32–67 8–21 NA [22]
Spine ligament
Posterior longitudinal NA 21–28 11–44 [23]
Ligamentum flavum NA 1–15 21–102 [23]
Anterior longitudinal 286–724 8–37 10–57 [23, 24]
Supraspinal NA 9–16 39–115 [23]
Interspinal NA 2–9 39–120 [23]
Forearm ligament
Carpal joint 23–119 NA NA [25]
Palmar radioulnar 39 ± 18 5.7 ± 1.7 51 ± 24 [26]
Dorsal radioulnar 52 ± 33 8±5 61 ± 29 [26]
Tissues separated into bundles or tested in multiple studies are listed as a range of values. Data
determined by one bundle or one study are expressed at mean ± S.E.M.
60 C.S. Chamberlain and R. Vanderby Jr.
T/L injury results from sudden, prolonged, or excessive forces that exceed the
elastic or recoverable limits of the tissue. When subfailure damage occurs, the tis-
sue does not recover its original length (Lo) after removal of the deforming load.
Ligaments incapable of returning to their original length after elongation are con-
sidered lax and no longer provide the same passive restraints to joint motion and
the same proprioception. If tendon length changes, optimal muscle strength occurs
at different joint positions. Once the ultimate strain is exceeded, the collagen fibers
reach a point of failure and all will rupture. The functional recovery of T/Ls after
injury is a slow and incomplete process. After 2 years of healing, MCLs only
reached 80 % of their control values for strength [8]. During the remodeling phase
of healing, viscoelastic properties are significantly compromised, and scars tend to
stress-relax and maintain load less efficiently than normal ligament [27–29].
Remodeling ligament scars also creep twice the amount as normal ligaments dur-
ing cyclic and static loads [30]. This would increase joint laxity and decrease pro-
prioception after healing. Long-term biomechanical recovery of ligament is
dependent on a number of factors, including the size of the initial gap, proximity
of the torn ligament ends, and joint movement during healing. Tendon’s functional
recovery is also dependent on length, but the bigger concern is scarring down dur-
ing healing to restrict joint motion. Tendon movement is necessary early to prevent
adhesions but load must be very low because of mechanical compromise. Many
strategies have attempted to recapitulate original properties after healing, including
surgical repair, physical therapy protocols, grafting, gene therapy, and tissue engi-
neering, but thus far, healing does not regenerate the native tissue. This issue con-
tinues as a focus for future investigations.
B3.5 Conclusions
Ligament and tendon injuries are a frequent occurrence with significant morbidity.
Appropriate clinical management requires an understanding of damaged T/L
mechanical function and its capacity for self-repair. The composition and hierar-
chical organization of these tissues facilitate its important function. One-
dimensional, elastic properties such as the ultimate tensile strength, modulus, and
strain are commonly gathered and pertinent for comparing normal and compro-
mised tissue and evaluating the efficacy of different healing treatments. In addition,
viscoelastic properties provide metrics to further characterize tissue behavior [31].
These are typically one-dimensional, mean behaviors of normal or healing tissues.
Defects, scaring, and tissue flaws are really three-dimensional, localized phenom-
ena. Alternative methods are needed that locally characterize these regions of com-
promise both in vitro and in vivo. Recently, ultrasound-based methods (example
[32]) have been explored as a possible approach to measure local tendon stress and
B3 Ligament and Tendon 61
strain in a dynamic, time-dependent manner. Such data would further enhance our
knowledge of T/L behavior and provide a tool to evaluate new treatment
strategies.
References
20. Paulos LE, Cawley PW, France EP (1991) Impact biomechanics of lateral knee bracing. The
anterior cruciate ligament. Am J Sports Med 19(4):337–342
21. Bigliani LU, Pollock RG, Soslowsky LJ et al (1992) Tensile properties of the inferior
glenohumeral ligament. J Orthop Res 10(2):187–197
22. Itoi E, Grabowski JJ, Morrey BF et al (1993) Capsular properties of the shoulder. Tohoku
J Exp Med 171(3):203–210
23. Pintar FA, Yoganandan N, Myers T et al (1992) Biomechanical properties of human lumbar
spine ligaments. J Biomech 25(11):1351–1356
24. Neumann P, Keller TS, Ekstrom L et al (1992) Mechanical properties of the human lumbar
anterior longitudinal ligament. J Biomech 25(10):1185–1194
25. Savelberg HH, Kooloos JG, Huiskes R et al (1992) Stiffness of the ligaments of the human
wrist joint. J Biomech 25(4):369–376
26. Schuind F, An KN, Berglund L et al (1991) The distal radioulnar ligaments: a biomechanical
study. J Hand Surg Am 16(6):1106–1114
27. Woo SLY, Inoue M, Mcgurkburleson E et al (1987) Treatment of the medial collateral ligament
injury. 2. Structure and function of canine knees in response to differing treatment regimens.
Am J Sports Med 15(1):22–29
28. Clayton ML, Miles JS, Abdulla M (1968) Experimental investigations of ligamentous healing.
Clin Orthop Relat Res 61:146–153
29. Chimich D, Frank C, Shrive N et al (1991) The effects of initial end contact on medial collat-
eral ligament healing—a morphological and biomechanical study in a rabbit model. J Orthop
Res 9(1):37–47
30. Thornton GM, Leask GP, Shrive NG et al (2000) Early medial collateral ligament scars have
inferior creep behaviour. J Orthop Res 18(2):238–246
31. Duenwald SE, Vanderby R Jr, Lakes RS (2009) Viscoelastic relaxation and recovery of tendon.
Ann Biomed Eng 37(6):1131–1140
32. Duenwald S, Kobayashi H, Frisch K et al (2011) Ultrasound echo is related to stress and strain
in tendon. J Biomech 44(3):424–429
Chapter B4
Skin and Muscle
B4.1 Introduction
Early studies [11] of the material properties of human skin and muscle are largely
suspect due to problems of inappropriate tissue handling, preservation and speci-
men preparation. Recent efforts have focused on methods which can determine
properties in situ in living individuals or on very freshly excised tissues. Among the
in vivo testing methodologies, indentation has proven to be the most popular,
although it sums up the contributions of various tissue layers [1, 3, 4, 6, 7, 9]. The
load-displacement curve obtained during indentation depends in decreasing degree
upon each of the tissues beneath the indentor. The derived properties, in addition,
can be expected to vary with anatomical site, subject age and external environmen-
tal conditions (temperature, relative humidity, etc.). Additional results have been
obtained in vivo through the use of Doppler ultrasound techniques [2, 5].
Krouskop et al. [5, 8] applied Doppler ultrasound techniques to measure the point-
to-point biomechanical property of the human skin and subcutaneous musculatures.
Tests of the forearms and legs suggested that the elastic moduli are strongly
dependent on the contraction status of the muscles. A 16-fold increase (from 6.2 kPa
Data are provided from indentation and ultrasound measurement techniques only.
A.F.T. Mak (*) • M. Zhang
Rehabilitation Engineering Centre, Hong Kong Polytechnic, Hunghom, Kowloon, Hong Kong
Table B4.1 Apparent Elastic Anatomical Location Elastic Modulus (kPa)(std dev.)
Moduli of Relaxed Above
Anterior 57.9 (31.1)
Knee Tissues [5]*
Lateral 53.2 (30.5)
Posterior 141.4 (79.1)
Medial 72.3 (45.5)
Superficial 117.6 (63.0)
Underlying 59.1 (74.0)
* Determined by Doppler ultrasonic techniques.
to 109 kPa) in the modulus was reported at a 10% strain as muscle contraction
changed from minimal to maximum. Malinauskas et al. [5] used the same technique
to examine the stump tissues of above-knee amputees and found that the average
modulus was significantly higher in posterior tissues than in other locations (Table
B4.1). They found, additionally, that superficial tissues were stiffer than deeper
structures. Note that the ages and sex of the subjects of these two studies [2, 5] were
not reported.
Ziegert and Lewis [9] measured the in vivo indentation properties of the soft tissues
covering the anterior-medial tibiae. A preload of 22.4 N was used with indentors of
6 to 25 mm in diameter. The observed load displacement relationship were essen-
tially linearly elastic. The structural stiffness was noted to vary by up to 70%
between sites in one individual and up to 300% between individuals. Unfortunately,
the thicknesses of overlying tissues were not determined at the different sites for the
individuals studied.
Lanir et al. [3] measured the in vivo indentation behavior of human forehead skin
with pressures up to 5 kPa. The observed behavior was linearly elastic and calcu-
lated stiffnesses were 4 to 12 kPa.
Bader and Bowker [1] studied the in vivo indentation properties of soft tissues on
the anterior aspects of human forearms and thighs by applying constant pressures of
11.7 and 7 kPa respectively. Tissue thickness was measured by using a skinfold cali-
per and Poisson’s rato was assumed to be 0.3. With these data, the stiffness of fore-
arm and thigh tissue were calculated to be, respectively, 1.99 and 1.51 kPa.
Vannah and Chlidress [7] applied similar techniques to measure the human calf,
but confined the limb within a shell. They noted that stress relaxation occurred
within one second of load application and no preconditioning effect was noted.
Torres-Morenos et al. [6] performed a similar study, working through ports in quad-
rilateral sockets of three above-knee amputees. However, they found the mechanical
properties of the tissues to be significantly non-linear, with site and rate dependen-
cies, as well as being strongly influenced by muscular activity.
Mak et al. [4] (Table B4.2) measured the in-vivo indentation properties of
the below knee tissues of young adults (N=6) between the ages of 25 and 35.
B4 Skin and Muscle 65
Table B4.2 Initial and Relaxed Elastic Moduli of Tissues Around Proximal Human Tibiae [4]*
Initial Elastic Modulus Relaxed Elastic Modulus
Anatomical Location (Ein) (kPa)(std dev.) (Eeq) (kPa)(std dev.)
Medial
Relaxed 102.6 (8.6) 99.8 (9.2)
Contracted 147.3** (15.8) 142.9** (16.7)
Con./Relax. 1.44 1.43
Lateral†
Relaxed 132.9 (7.2) 130.1 (7.9)
Contracted 194.3** (24,7) 188.4** (23.0)
Con./Relax. 1.46 1.45
* By indentation; ** Different from relaxed (p <0.001); † Between tibia and fibula.
Additional Reading
Bader, D.L. and Bowker, P. (1983) Mechanical characteristics of skin and underly-
ing tissues in vivo. Biomaterials, 4, 305–8
This paper describes an indentation experiment to investigate in vivo the bulk
mechanical properties of the composite of skin and underlying tissues on the ante-
rior aspects of human forearms and thighs by applying constant pressures. Significant
variations in tissue stiffness with sex, age and body site were also demonstrated.
Malinauskas, M., Krouskop, T.A. and Barry, P.A. (1989) Noninvasive measure-
ment of the stiffness of tissue in the above-knee amputation limb. J. Rehab. Res.
Dev., 26(3), 45–52
The paper reports a noninvasive technique to measure the mechanical properties
of the bulk soft tissues by a pulsed ultrasonic Doppler system. An ultrasonic trans-
ducer was used to measure internal displacement resulting from external acoustical
perturbations. Measurements were made at four sites of 8 aboveknee residual limbs.
The Young’s moduli were found in a range of 53–141 kPa. Superficial tissue had a
significantly higher modulus than the tissue beneath.
The repeatability test indicated an acceptable repeatibility. An improved device
can possibly be a useful tool in prosthetic fitting and CAD socket design.
Rab, G.T. (1994) Muscle, in Human Walking (2nd ed.) (eds J. Rose, J.C. Gamble),
Williams & Wilkins, Baltimore, pp. 101–122.
66 A.F.T. Mak and M. Zhang
A concise description of the active properties of muscle tissue, with direct application
to the development of forces within the human gait cycle.
Reynolds, D. and Lord, L. (1992) Interface load for computer-aided design of
below-knee prosthetic sockets. Med. & Biol. Eng. & Comput., 30, 419–426.
The authors investigated the bulk tissue behaviour of the below-knee amputee’s
residual limb. An assessment of Young’s modulus was made by matching the inden-
tation experimental curves with the curves produced by the finite element modelling
of the indentation into a layer of tissue with idealized mechanical properties. In vivo
tests, conducted at four sites of a below-knee amputee’s limb (patella tendon, pop-
liteal, and anterolateral regions) found the local moduli to be 145, 50, 50 and 120
kPa respectively. The effect of muscle tension on the measured indentation response
was also investigated. The results showed that the stiffness increased with muscle
contraction.
References
1. Bader D.L. Bowker, P. (1983) Mechanical characteristics of skin and underlying tissues in vivo.
Biomaterials 4:305–8
2. Krouskop, T.A, Dougherty. D.R and Vonson, F.S. (1987) A pulsed Doppler ultrasonic system
for making noninvasive measurements of the mechanical properties of soft tissue. I. Rehab.
Res. Dev., 24(1), 1–8.
3. Lanir, Y., Dikstein, S., Hartzshtark, A., et al. (1990) In-vivo indentation of human skin. Trans.
ASME (J. Biomech. Eng.), 112, 63–69.
4. Mak, A.F.T., Liu, G.H.W. and Lee, S.Y. (1994) Biomechanical assessment of below-knee
residual limb tissue, J. Rehab. Res. Dev., 31, 188–198.
5. Malinauskas, M., Krouskop, T.A and Barry, P.A (4) (1989) Noninvasive measurement of the
stiffness of tissue in the above-knee amputation limb. I.Rehab. Res. Dev., 26(3), 45–52.
6. Torres-Morenos, R., Solomonidis, S.E. and Jones, D. (1992) Geometric and mechanical charac-
teristics of the above-knee residual limb. Proc. 7th. World Cong. Int. Soc. Prosthetics Orthotics,
pp. 149.
7. Vannah, W.M. and Chlidress, D.S. (1988) An investigation of the three dimensional mechani-
cal response of bulk muscular tissue: Experimental methods and results, in Computational
Methods in Bioengineering (eds R.L. Spilker and B.R Simon), Amer. Soc. Mech. Eng.,
New York, pp. 493–503
8. Yamada, H. (1970) (ed. F.G. Evans) Strength of Biological Tissues, Williams & Wilkins,
Baltimore.
9. Ziegert, J.C. and Lewis, J.L. (1978) In-vivo mechanical properties of soft tissue covering bony
prominence, Trans ASME (J. Biomech. Eng.), 100, 194–201.
Chapter B5
Brain Tissues
S.S. Margulies and D.F. Meaney
B5.1 Introduction
The brain is organized into the cerebrum, brain stem, and cerebellum. The cerebrum
consists of two cerebral hemispheres, basal ganglia, and the diencephalon. The
hemispheres contain the cerebral cortex and underlying white matter, and are asso-
ciated with higher order functioning, including memory, cognition, and fine motor
control. The basal ganglia, contained within the hemispheres, controls gross motor
function. The diencephalon is much smaller than the cerebrum, contains the thala-
mus and hypothalamus, and is associated with relaying sensory information and
controlling the autonomic nervous system. The brainstem contains the mesence
phalon, pons and the medulla oblangata. The smallest segment of the brain, the
mesencephalon, is located below the diencephalon and is thought to play a role in
consciousness. Muscle activation, tone and equilibrium is controlled in the pons and
cerebellum located below the mesencephalon, and respiratory and cardiac processes
are governed by the medulla oblongata, located directly beneath the pons.
The brain contains grey matter and white matter substances that are easily distin-
guished upon gross examination. Grey matter contains a densely packed network of
neural cell bodies and associated glial cells, whereas white matter contains myelin-
ated axonal tracts, relatively few neuronal cell bodies, and a supporting environment
of glial cells. The entire brain is surrounded by cerebrospinal fluid contained within
an extensive ventricular system that occupies approximately one-tenth of the total
brain volume. The ventricular system supports the brain as well as the spinal cord, and
provides nutrients to and removes waste products from the central nervous system.
B5.2 Composition
Table B5.1 Brain Tissue Composition
Water (wt%) Ash (wt%) Lipid (wt%) Protein (wt%)
Whole brain 76.3–78.5 (77.4) 1.4–2 (1.5) 9–17 8–12
Grey matter 83–86 1.5 5.3 8–12
White matter 68–77 1.4 18 11–12
Approximate overall ranges given. Values in parentheses indicate averages.
Source: [1–3].
The adult brain with the brain stem is approximately half an ellipsoid.
Table B5.2 Elastic Modulus and Shear Modulus of Normal Brain at 37°C (typical values, not averages)
Species
Testing technique (post mortuum time) Frequency (Hz) Analysis format Results (kPa) G1 (kPa) G2 (kPa) Source
Free vibration Rhesus monkey 31 E* = E1 + i E2 E1 = 91.2 30.3 18.0 [6]
(15 min) E2 + 53.9
Free vibration Human (6–12 hr) 34 E* = E1 + i E2 E1 = 66.8 22.3 8.7 [6]
E2 = 26.3
Harmonic shear Human white matter 9–10 G* = G1+ i G2 Min: G* = 0.75 + i 0.3 0.75–1.41 0.3–0.6 [7]
(10–62 hr) Max G* = 1.41 + i 0.6
Driving point Rhesus (in vivo) 80 Theoretical G1 = 19.6 19.6 11.2 [7, 8]
impedance approx of data G2 = 11.2
G* = G1 + i G2
Quasi-static Rhesus monkey live 0 Static elastic E1 = 10–60 live and dead 10–60 [9]
expansion (in vivo) dead Modulus, E1 E1 = 40–120 fixed
of balloon (5–45 min) fixed
within tissue (formaldehyde)
Sudden Human (acceleration compare tissue Shear modulus G = l.l7–2.19 1.7 [10]
acceleration duration ≈ displacement kPa (average = 1.7) Kinematic (Tissue
of a cylinder 17.5 ms.) with that of a viscosity ν = 14–124 cm2/s temp
filled with tissue Voigt solid (average = 89) unknown)
cylinder
S.S. Margulies and D.F. Meaney
Harmonic shear Human (<3 hr) 5–350 G* = G1+ i G2 G1 G2 7.6–33.9 2.76–81.4 [11]
5 7.6 2.8
15 8.4 3.5
35 11.7 5.2
85 19.3 13.4
105 21.4 18.0
B5 Brain Tissues
Table B5.3 Creep Modulus of Normal Brain at 37°C (typical values, not averages)
Testing
technique* Species Model Results Source
Compression Rhesus monkey J(t) = C1 + C2 In(t) C1 = 2.97 kPa-1 [6]
C -1
(15 min) 2 = 0.18 kPa
t >0.1s
Compression Human (6–12 hr) J(t) = C1 – C2 In(t) C1 = 2.45 kPa-1 [6]
C2 = 0.18 kPa-1
t >0.1s
Compression Human (6–12 hr) Nonlinear solid K1 = 25.77 kPa [12]
1/ 2 K2 = 20.46 kPa,
æ k22 + 4s 0 k3 ö
s e- m t ç s 0 ( ) k K3 = 104.04 kPa
ϵ( t ) = 0 + + - 2 ÷ . 1 - e- m t
( )
k2 çç k1 2 k3 2 k3 ÷÷ C1 = 651.8 kPa s
è ø σo = 3.44, 4.82, 6.89 kPa
K2
Where m =
C1
Compression Human (6–12 hr) Nonlinear fluid Kl = 74.41 kPa [12]
1/ 2
s 0 C1C2 C1 ( C22 + 4s 0C3 ) 1 é 2 1/ 2
K2 = 20.67 kPa
Î (t ) = + - + (C2 + 4s 0C3 ) - C2 ùúû C1 = 1266.38 kPa
K 2 2 K1C3 2 K1C3 2C3 êë C2 = 36599.7 kPa s
æ C1 - mt ö s 0 é C1 ù - mt
C3 = 1.38 kPa S2
ç t + e ÷ + . ê1 + ú (1 - e ) σo = 3.44, 4.82, 6.89 kPa
è K1 ø K1 ë C2 û
S.S. Margulies and D.F. Meaney
Compression Human (6–12 hr) Hyperelastic with material dissipation C1 = 6.89 kPa [12]
. C2 = 17.23 kPa
s 11 é 2 1 ù é C2 ù B1 l
= l - ú ê1 + ú+ . B1 = 0.55 kPa S2
2C1 êë l û ë C1l û C1 l 3 λ = stretch ratio
[l 2 - 6l 3 + 1] κ = stretch rate
B5 Brain Tissues
B5.7 Comments
The list of primate brain tissue properties presented in this chapter highlights the
numerous areas where data are either unavailable or largely incomplete. With a
renewal of interest in modeling the normal and pathological response of the brain to
various stimuli, it is possible that some of the missing tissue property data will be
generated in the near future. Investigators should be cautioned, however, that
because brain tissue properties vary with environmental factors, measurements
made under unphysiologic conditions may differ from those of living tissue. As an
example, the brain is highly vascularized, and the role of blood flow, volume, and
pressure on tissue behavior remains to be determined.
The caution exercised by the experimentalist in generating new data should be
matched by sound skepticism on the part of the investigators who are developing ana-
lytical or computational models of brain functional and structural response. Although
the ability to calculate detailed responses has improved greatly in the past decade, these
sophisticated models are limited by the available experimental data used to develop
and validate the models. To create a realistic representation of normal or pathological
response of the brain, it is essential that the model parameters be based on measured
tissue properties and that any conclusions drawn from the models be validated with
measured response data.Therefore, it is clear that future experimental studies are
needed to determine the properties and response of living primate brain tissue.
Additional Reading
References
X. Deng and R. Guidoin
B6.1 Introduction
Blood and lymphatic vessels are soft tissues with densities which exhibit nonlinear
stress-strain relationships [1]. The walls of blood and lymphatic vessels show not
only elastic [2, 3] or pseudoelastic [4] behavior, but also possess distinctive inelas-
tic character [5, 6] as well, including viscosity, creep, stress relaxation and pres-
sure diameter hysteresis. The mechanical properties of these vessels depend largely
on the constituents of their walls, especially the collagen, elastin, and vascular
smooth muscle content. In general, the walls of blood and lymphatic vessels are
anisotropic. Moreover, their properties are affected by age and disease state. This
section presents the data concerning the characteristic dimensions of arterial tree
and venous system; the constituents and mechanical properties of the vessel walls.
Water permeability or hydraulic conductivity of blood vessel walls have been also
included, because this transport property of blood vessel wall is believed to be
important both in nourishing the vessel walls and in affecting development of ath-
erosclerosis [7–9].
The data are collected primarily from human tissue but animal results are also
included in places for completeness. Among the three kinds of vessels, the arte-
rial wall has been extensively investigated while studies of lymphatic vessels are
very rare.
Detailed measurements of the number and size of blood vessels in the living
body are very difficult to perform, so reliable information is scarce. Moreover,
data collected on vessels in one tissue or organ are not applicable to another.
Thus one should be cautious in using morphometric data; only the data for large
vessels are reliable.
The aorta is tapered, but most other arteries can be considered to have a constant
diameter between branches. The rate of taper varies from individual to individual,
presumably, between species. However, in the dog, the change of aortic cross
sectional area can be described by the exponential equation:
A = Ao e( - b x Ro ) (B6.1)
Table B6.1 Morphometric and Related Properties of the Human Systemic Circulation*
Diameter Wall thickness Length Blood velocity Reynolds
Vessel (mm) (mm) (cm) (cm/sec) number
Ascending aorta 32 1.6 5–5.5 63 3600–5800
Arch of aorta 25–30 4–5
Thoracic aorta 20 1.2 16 27 1200–1500
Abdominal 17–20 0.9 15
aorta
Femoral artery 8 0.5 32
Carotid artery 9 0.75 18
Radial artery 4 0.35 23
Large artery 2–6 20–50 110–850
Capillaries 0.005–0.01 0.05–0.1 0.0007–0.003
Large veins 5–10 15–20 210–570
Vena cava 20 11–16 630–900
* Source [11–17]. Note: The Reynolds numbers were calculated assuming a value for the viscosity
of the blood of 0.035 poise.
B6 Arteries, Veins and Lymphatic Vessels 79
where A is the cross sectional area of the aorta, Ao and Ro are the respective cross
sectional area and radius at the upstream site, x is the distance from the upstream
site, and β a taper factor, which varies between 0.02 and 0.05 [10]. In man, the taper
is found not to be as smooth as implied by the above equation; thus values of β are
unavailable.
Morphometric and related data are given in Tables B6.1, B6.2 and B6.3.
The main constituents of normal human arterial tissues from young adult subjects
(20–39 years) are listed in Table B6.4 [20]. The major part of the dry matter in the
arterial wall consists of proteins such as elastin and collagen. Because the impor-
tance of elastin and collagen in the mechanical properties of arterial wall, the com-
position of media and adventitial layers in terms of collagen, elastin, smooth muscle
and ground substance is listed in Table B6.5 for three different arterial tissues [1].
Table B6.3 Morphometric and Related Properties of the Canine Systemic Circulation*†
Ascending Descending Abdominal Femoral Vena cava, Pulmonary
Site aorta aorta aorta artery Carotid artery Arteriole Capillary Venule inferior artery, main
Internal 1.5 1.3 0.9 0.4 0.5 5 x 10-3 6 x 10-4 4 x 10-3 1.0 1.7
diameter, (1.0–2.4) (0.8–1.8) (0.5–1.2) (0.2–0.8) (0.2–0.8) (1–8 x 10-3) (4–8 x 10-4) (1–7.5 x 10-3) (0.6–1.5) (1.0–2.0)
di (cm)
Wall thickness, 0.065 – 0.05 0.04 0.03 2 x 10-3 1 x 10-4 2 X 10-4 0.015 0.02
h (cm) (0.05–0.08) (0.04–0.06) (0.02–0.06) (0.02–0.04) (0.01–0.02) (0.01–0.03)
h/di 0.07 – 0.06 0.07 0.08 0.4 0.17 0.05 0.015 0.01
(0.055–0.084) (0.04–0.09) (0.055–0.11) (0.055–0.095)
In vivo length 5 20 15 10 15 0.15 0.06 0.15 30 3.5
(cm) (10–20) (0.1–0.2) (0.02–0.1) (0.1–0.2) (20–40) (3–4)
Cross-section 2 1.3 0.6 0.2 0.2 2x10-5 3x10-7 2x10-5 0.8 2.3
area (cm2)
Total vascular 2 2 2 3 3 125 600 570 3.0 2.3
cross-section at
each level (cm2)
Blood velocity 0.2 1.05 0.55 1.0 – 0.75 0.07 0.35 0.25 0.7
(peak) (ms-1) (0.4–2.9) (0.25–2.5) (0.5–0.6) (1.0–1.2) (0.15–0.4)
Blood velocity 0.2 0.2 0.15 0.1 – (5–10 x 10-3 ) 2–17 x10-4 2–5 x10-3 – 0.15
(mean) (ms-1) (0.1–0.4) (0.1–0.4) (0.08–0.2) (0.1–0.15) (0.06–0.28)
Peak Reynolds 4500 3400 1250 1000 – 0.09 0.001 0.035 700 3000
number, R̂ e
* Source: Adapted from [19].
† Normal values for canine cardiovascular parameters. An approximate average value, and then the range, is given whe re possible. All values are for the dog except those for
arteriole, capillary, and venule, which have only been measured in smaller mammals.
B6 Arteries, Veins and Lymphatic Vessels 81
The collagen in adventitia and media is mostly Type III, some Type I, and a trace of
Type V while the collagen of the basal lamina is Type IV [1].
Table B6.6 lists the constituents of additional arteries (canine), and the the ratio
of collagen to elastin [21].
Composition changes of arterial tissues with age
The composition of normal human arterial tissues is altered with age in many
aspects. Table B6.7 lists the observed changes in human aorta, pulmonary and fem-
oral arteries [20]. There is a tendency that both the dry matter and nitrogen content
of arterial tissues decreases with age. However, the relative quantity of collagen [22]
and elastin [23, 24] in the arterial wall remains almost unchanged with age. Below
the age of 39, the wall of human thoracic aorta has 32.1 ± 5.5% elastin, between the
age 40–69, the wall contains 34.4 ± 9.3%, and from 70–89, the elastin content is
36.5 ± 10.1 [24].
82
Table B6.7 Variation of Normal Human Arterial Tissue Composition with Age*
Aorta
Age Acid
group Dry Nitro- Total Choles- Total Cal- Total soluble Potas
(years) matter gen lipids terol ash cium PO4 PO4 sium
10–19 29.5† 4.38 1.23 0.15 – 0.03 0.25 0.16 0.055
30–39 28.5 4.03 1.75 0.28 0.81 0.09 0.40 0.29 0.040
50–59 28.0 3.67 1.90 0.48 1.55 0.21 0.54 0.41 0.039
70–79 28.0 3.38 – 0.71 2.80 0.39 – – 0.033
Pulmonary artery Femoral artery
Age
group Dry Nitro- Choles- Potas Dry Nitro Choles- Total Cal-
(years) matter gen terol Calcium ium Matter gen terol ash cium
10–19 26.5 3.91 0.12 0.025 0.033 26.9 3.84 0.11 0.59 0.14
30–39 25.7 3.71 0.17 0.028 0.031 24.4 3.30 0.15 0.71 0.18
50–59 24.9 3.45 0.22 0.027 0.026 22.8 2.83 0.23 1.15 0.40
70–79 23.0 3.25 – 0.060 0.025 25.3 2.90 0.55 3.17 1.07
* Source: Adapted from [20]; see for additional age group values.
† Values expressed in percentage of wet tissue weight.
The main constituents of normal human venous tissue are listed in Table B6.10.
B6 Arteries, Veins and Lymphatic Vessels 85
B6.4.2 C
hanges with age in composition of normal venous
tissues
The changes with age in the composition of normal venous vessels are listed in
Table B6.11.
86 X. Deng and R. Guidoin
The blood vessel wall consists of three layers: the intima, media, and adventitia. The
intima contains mainly the endothelial cells that contribute little to the strength of
the blood vessels. The media and the adventitia contain smooth muscle cells, elastin
and collagen. Elastin is the most ‘linearly’ elastic biosolid material known. Unlike
elastin, collagen does not obey Hooke’s Law. However, collagen is the main load
carrying element of blood vessels. Table B6.12 lists the mechanical properties of
tissues composing the blood vessel wall.
Studies of arterial wall mechanics have clearly established the anisotropic nature of
arteries [1, 29, 30]. In vivo pressurized arteries are deformed simultaneously in all
directions. But, experimental studies [31] have demonstrated that arteries deform
orthotropically. Therefore, arterial deformations may be examined in three orthogo-
nal directions, namely, the longitudinal, circumferential, and radial directions.
There are nine elastic parameters: the three elastic moduli, Eθ, Ez and Er; and six
Poisson’s ratios, σrθ,σθr,σzθ,σθz,σrz andσzr. As far as hemodynamics is concerned, how-
ever, among the three elastic moduli the circumferential one is most important.
The cicumferential elastic modulus is termed the incremental modulus and deter-
mined by the following equation (B6.2):
Dp i 2 (1- s 2 ) R 0 R i 2
Eq = (B6.2)
DR 0 ( R 0 2 - R 2i )
where Δpi is the transmural pressure increment, Ro and Ri are the respective external
and internal diameter of the vessel, ΔRo is the change in the external diameter due
to Δpi, and σ the Poisson’s ratio (the ratio of transverse strain to longitudinal strain).
The detailed technique for measuring the circumferential incremental elastic modu-
lus of the arteries was described by Bergel [32].
Table B6.13 presents the circumferential incremental elastic modulus of human
arterial walls from young adults (≤ 35 years) at a transmural pressure of 100 mmHg.
Experimental data by Bader [34] demonstrated that the circumferential elastic
modulus of human thoracic aorta increased almost linearly with age. Table B6.14
gives the variation with age in the elastic moduli for human thoracic and abdominal
aorta at a pressure of 100 mmHg.
It should be mentioned that all the circumferential elastic moduli given in Table
B6.14 are based on the assumption of a Poisson’s ratio of 0.5. This is not strictly
true when large strains are considered [36]. Patel et al. [37] measured Poisson’s
ratios for the aorta in living dogs at a transmural pressure of about 110 mmHg, as
well as the circumferential, longitudinal and radial incremental elastic moduli and
determined that individual values vary between 0.29 and 0.71.
88 X. Deng and R. Guidoin
B6.5.2 C
ompliance, pressure cross-sectional area relationship,
and retraction
By measuring the static elastic properties of human thoracic and abdominal aortas
in vitro, Langewouters et al. [35] proposed the following empirical relationship
between the cross-sectional area of the lumen (A) and the pressure in the vessel (p):
ïì 1 1 æ p - p0 ö üï
A( p) = Am í + tan -1 ç ÷ý (B6.3)
îï 2 p è p1 ø þï
in which Am, Po and p1 are three independent parameters that are defined in Equation
(B6.4) below.
Another important mechanical property of blood vessels in the compliance.
Langewouters et al. [35] defined the ‘static’ compliance as the derivative of equa-
tion (B6.3) with respect to pressure
Cm Am
C ( p) = ; Cm = (B6.4)
æ p - p0 ö
2
p p1
1+ ç ÷
è p1 ø
where Cm and Am are the maximum compliance and the maximum crosssectional
area of the vessel, respectively; Po is the pressure at which aortic compliance reaches
its maximum; and p1 is the half-width pressure, i.e. at p0 ± p1 , aortic compliance is
equal to Cm/2. According to Langewouters et al. [35], the ‘static’ compliance values
of human thoracic aorta at 100 mmHg range from 1.9 to 17 x 10·3 cm2/mmHg; and
0.6 to 4.4 x 10-3 cm2/mmHg for abdominal aorta.
Table B6.15 lists the relative wall thickness and the retraction on excision for a
variety of blood vessels. The retraction of a vessel is the amount by which a segment
of vessel shortens on removal from the body, expressed as a percentage of the length
of the segment in situ. The relative wall thickness is the ratio of wall thickness to
mean diameter of the vessel.
Table B6.16 presents typical data for the tensile properties of arterial tissues from
Yamada [43]. The test specimens of tissues were strips each with a reduced middle
region 10 mm length, 2–3 mm in width, and a length to width ratio of 3:1. In the
tables:
1. Tensile breaking load per unit width (g/mm) = ultimate tensile strength
(g/mm2) x thickness (mm)
tensile breaking load ( g )
2. Ultimate tensile strength ( g mm 2 ) =
cross - section area of the test section
B6 Arteries, Veins and Lymphatic Vessels 89
Table B6.15 Relative Wall Thickness and Retraction on Excision of Various blood Vessels*
Vessel Species γx% Retraction % Source
Thoracic aorta Dog 10.5 32 [2]
Abdominal aorta Dog 10.5 34 [2]
Femoral artery Dog 11.5 42 [2]
Carotid artery Dog 13.2 35 [2]
Iliac artery Dog – 40 [2]
Carotid artery Dog 13.6 – [38]
Carotid artery Cat 14.5 – [38]
Carotid artery Rabbit 11.2 – [38]
Carotid sinus Dog 20.0 – [38]
Carotid sinus Cat 16.0 – [38]
Carotid sinus Rabbit 12.0 – [38]
Thoracic aorta Dog 14.0 – [39]
Abdominal aorta Dog 12.0 – [39]
Femoral artery Dog 13.0 – [39]
Pulmonary artery Man 2.0 – [40]
Abdominal vena cava Dog 2.3 30 [41]
Thoracic aorta Man 6–9 25–15 [33]†
Abdominal aorta Man 8–13 30–17 [33]†
Femoral artery Man 12–19 40–25 [33]†
Carotid artery Man 2–15 25–18 [33]†
Source: Adapted from [42].
† Measurement from [33] of young (<35) and old (>35) subjects respectively.
Table B6.16 Tensile Properties of Human Coronary Arterial Tissue* (Longitudinal Direction)
Tensile Breaking Ultimate Tensile Ultimate
Age (yrs) Load/Unit width (g/mm) Strength (g/mm2) Elongation (%)
10–19 85±3.1 140±3.0 99±2.4
20–39 82±1.8 114±9.3 78±1.6
40–59 82±1.8 104±4.7 68±3.5
60–79 79±2.9 104±4.7 4.5±3.8
Adult 81 107 64
(average)
* Adapted from [43].
3. Ultimate percentage elongation ( % ) = breaking elongation ( mm )
´100
original length of the specimen ( mm )
Table B6.16 lists the tensile data for human coronary arterial tissue in the longi-
tudinal direction.
90 X. Deng and R. Guidoin
Other arteries have similar properties [43]. Yamada [43] provides the tensile
properties of animal tissues in various tables.
The most direct way to study arterial viscoelasticity is to determine the response of
the test tissue to oscillatory stresses. If the arterial wall is conceived to be repre-
sented by a simple Kelvin-Voigt model consisting of a spring and a dashpot in paral-
lel, the dynamic elastic component and the viscous component of a vessel can be
expressed as
ED = E cos f (B6.5)
hw = E sin f
where E is the complex dynamic elastic modulus that is identical to the incremental
elastic modulus under static stresses; ϕ is the phase lag of the strain behind the
stress (in the case of circumferential direction, it is the phase lag of diameter behind
the pressure); ED is the dynamic elastic modulus of the vessel; and ηω is the viscous
retarding modulus (η is the viscous constant and w the angular frequency). For mea-
suring the circumferential dynamic mechanical properties, the test vessel is usually
subjected to an oscillatory pressure. The pressure oscillations are in a sinusoidal
form. In circumferential direction, E can be calculated from equation (B6.2) with
the recorded diameter of the blood vessel and the oscillatory pressure [44].
Table B6.17 lists the dynamic mechanical properties of different arteries at a
frequency of 2.0 Hz at a mean pressure of 100 mmHg. In this table, EP is the circum-
ferential pressure-strain modulus defined as
V p2 = Eh 2 Rp (B6.7)
where E = elastic modulus of the wall, h = wall thickness, R = mean radius of the
vessel, ρ = density of blood.
There is general understanding that the dynamic elastic modulus is not strongly
frequency dependent above 2–4 Hz and that it increases from the static value at
quite low frequencies. Bergel [42] provides additional values for canine vessels.
When a subject is suddenly strained and then the strain is maintained constant after-
ward, the corresponding stresses induced in the subject decrease with time, this
phenomenon is called stress relaxation. If the subject is suddenly stressed and then
the stress is maintained afterwards, the subject continues to deform, this phenome-
non is called creep. Creep and stress relaxation are another two important phenom-
ena in the arterial viscoelasiticity. Langewouters et al. [46] studied the creep
responses of human thoracic and abdominal aortic segments. The pressure in the
segments was changed in steps of 20 mmHg between 20 and 180 mmHg. Aortic
( 10 -3 cm 2
1.5–12.3 5.1 2.5 0.5–3.3 1.5 0.78
C
mmHg )
Age 30–78 63 14 30–78 58 15
* Adapted from [46] by permission.
α1, α2, creep fractions; t1, t2, time constants; C, compliance; S.D., standard deviation.
92 X. Deng and R. Guidoin
creep curves at each pressure level were described individually by a constant plus
biexponential creep model (C-model, [47])
é æ -t ö æ -t öù
Aic = DA ê1 - a1 exp ç i ÷ - a 2 exp ç i ÷ú (B6.8)
êë è t1 ø è t2 ø úû
where A = aortic cross-sectional area; Aci = sample value of aortic creep response at
time ti; δA = change in aortic area upon a 20 mmHg pressure step; t = time; i =
sample number; a1,a2 = creep fraction; τ1,τ2 = time constant. Table B6.18 lists the
creep fractions and time constants for all aortic segments [46].
Stress relaxation relations for human arteries are not available; however, Tanaka
and Fung [48] studied the stress relaxation spectrum of the canine aorta. They
expressed the stress history with respect to a step change in strain in the form:
where G (t) is the normalized relaxation function of time; T(e) (l) is a function of
strain l, called elastic response. This is the tensile stress instantaneously generated
in the aortic tissue when a step elongation, l, is imposed on the specimen.
If the relaxation function is written as:
1é
1 + ò s (t ) e - t t dt ù
µ
G (t ) = (B6.10)
A ê
ë 0 ûú
in which
A = é1 + ò s (t ) dt ù
µ
ëê 0 ûú
S (t ) = c t for t1 £ t £ t2
(B6.11)
= 0 for t1 < t1 , t > t2
Values for segments of the canine aorta are given in [48].
B6 Arteries, Veins and Lymphatic Vessels 93
Table B6.19 Comparison of elastic Moduli between Venous and arterial Segments*
Pressure Extension ratio Incremental venous elastic Carotid artery Incremental
(cm H2O) modulus modulus
λθ λz Eθ (dynes/cm2 Ez (dynes/cm2 Eθ (dynes/cm2 Ez (dynes/cm2
x 106) x 106) x 106) x 106)
Canine Jugular Vein
10 1.457 1.481 15 ± 3* 1.2 ± 0.18† 7.62 5.16
25 1.463 1.530 47 ± 6† 4.4 ± 0.35† 8.39 7.15
50 1.472 1.597 88 ± 7† 11.8 ± 2.1 9.51 10.69
75 1.478 1.646 98 ± 7† 46 ± 13* 10.37 13.97
100 1.482 1.675 117 ± l0† 67 ± 25 10.92 16.24
125 1.484 1.686 134 ± 24† 89 ± 32 11.16 17.17
150 1.484 1.686 171 ± 9† 113 ± 13† 11.16 17.17
Human Saphenous Vein
10 1.357 1.169 0.27 ± 0.12† 1.61 ± 0.32* 5.30 0.017
25 1.417 1.206 0.65 ± 0.13† 2.03 ± 0.39* 5.82 0.328
50 1.500 1.266 1.89 ± 0.41† 2.75 ± 0.78 6.00 0.735
75 1.561 1.325 9.85 ± 1.6 3.18 ± 0.76 9.66 1.80
100 1.602 1.381 15.0 ± 2.6 3.56 ± 0.58 12.77 3.15
125 1.621 1.430 20.4 ± 1.6* 3.98 ± 0.96 14.79 4.59
150 1.621 1.470 25.1 ± 7.5 4.75 ± 1.2 15.51 5.93
* Adapted from [50] by permission.
* p < 0.05 for the comparison between the venous and carotid moduli.
† P < 0.01 for the comparison between the venous and carotid moduli.
The structure of the venous walls is basically similar to that of the arterial walls. The
main difference is that they contain less muscle and elastic tissue than the arterial
walls, which raises the static elastic modulus two to fourfold [49]. Because the
venous walls are much thinner than the arterial wall, they are easily collapsible
when they are subject to external compressions.
Table B6.19 lists the static incremental elastic moduli of the canine jugular vein
and human saphenous vein. For the purpose of comparison, the static increment
elastic modulus of the canine carotid artery segments are also presented in the table.
This comparison is of interest because in some arterial reconstructive surgeries, a
vein is used as a substitute for an artery.
Sobin [51] obtained data on the mechanical properties of human vena cava from
autopsy material. The data may be expressed by the following equation:
94 X. Deng and R. Guidoin
T = a ECl[ a( E 2 - E *2 )]
E = (l 2 - 1) 2 (B6.12)
T * = a CE * l *
where λ. is the ratio of the changed length of the specimen divided by the reference
length of the specimen, C and a are material constants, and E* is the strain that cor-
responds to a selected value of stress S*. The product of the constant αC is similar
to the elastic modulus, provided that the modulus is defined as the ratio, S*/E*,
where S* = T*/λ*. Fung [1] provides typical values of these constants obtained
experimentally; however, no universal constants have been discovered.
The tensile properties of human venous tissues are presented in Table B6.20. For the
testing method and definitions of the terms in the table, please refer to Section B6.5
on Mechanical properties of arteries.
B6 Arteries, Veins and Lymphatic Vessels 95
The problem concerning the ontogenesis of the lymphatic vessels is still not com-
pletely solved. However, most of the evidence indicates that the large lymphatic
vessels are derived from the veins [52]. Therefore, lymphatics can be considered
modified veins. According to the reports by Ohhashi et al. [53, 54], the circumfer-
ential elastic modulus of the bovine mesenteric lymphatics ranged from 4.2 x 104 to
2.7 x 105 dynes/cm2 at a pressure range from 0 to about 20 mmHg, and the elastic
modulus of canine thoracic duct is about 2.0 x 105 dynes/cm2. These values are less
than those of veins obtained by Bergel [55]. Therefore, the lymphatics are more
distensible than the veins.
Under transmural pressures, fluid or plasma will flow across the walls of blood ves-
sels. On one hand, convective fluid motion through the blood vessel wall plays a
very important role in nourishing the vessel walls, on the other hand, it is involved
in atherogenesis by promoting the transport of macromolecules such as lipoproteins
into the arterial wall [55–58], possibly through leaky endothelial cell junctions in
regions of high endothelial cell turnover [59]. Table B6.21 lists the filtration proper-
ties of excised, presumably normal, human iliac blood vessels.
96 X. Deng and R. Guidoin
B6.9.1 Age
Two well known changes accompany aging of the cardiovascular system are dila-
tion of thoracic aorta [60] and increased thickness of arterial wall [61]. Arterial
walls become less distensible with aging [34, 62, 63]. Both dry matter and nitrogen
content of artery tissue show a tendency to decrease with age in large and medium
sized arteries [20]. But the relative quantity of collagen [22] and elastin [23, 24] in
the arterial wall remains essentially unchanged. With aging the arterial wall becomes
progressively stiffer. Bader [34] found that the circumferential elastic modulus of
human thoracic aorta increased linearly with age. At 100 mmHg, the static circum-
ferential modulus of ‘young’ (<35 years) human thoracic aorta averaged 7.5 x 106
dynes/cm2, and for the ‘old’ (>35 years) the average was 16.6 x 106 dynes/cm2 [33].
Young peripheral arteries tend to have a greater viscosity (ƞω) than the older ones
[33]. Despite the overall increase in stiffness, the arterial wall itself is considerably
weaker than the younger/older one [43].
B6.9.2 Hypertension
Several studies have shown that the water content of human, rat and dog arteries is
increased in hypertension, and this increased water content may be associated with
an increased wall thickness [64, 65]. Due to the limitations in studying samples
from human subjects, animal models (mainly rats) have been employed. Mallov
[66] found that the aorta from hypertensive rats had more smooth muscle than nor-
mal aorta. Greenwald and Berry [67] reported increased elastin and decreased col-
lagen content in the aorta from spontaneous hypertensive rats when compared with
the normal aorta. Wolinsky [25] observed an increase in the absolute amounts of
both medial elastin and collagen contents in hypertensive rats. However, the relative
percentage of these elements remained essentially constant. Experimental studies
[67–69] showed an increase in vessel stiffness with the development of hyperten-
sion. This increase in vessel stiffness results in a smaller vessel diameter for a given
distending pressure, i.e. a decrease in the distensibility [70].
B6.9.3 Atherosclerosis
It is generally accepted that substantial changes in the arterial wall occur with ath-
erosclerosis in man. In human atherosclerotic arteries, it appears that there may be
an absolute increase in collagen and a decrease in muscle fibers when compared
B6 Arteries, Veins and Lymphatic Vessels 97
with normal arteries [28]. In canine iliac artery the ratio of collagen to elastin was
found to be increased, while the ratio in carotid was decreased [26]. The elastic
moduli of the diseased aorta and common iliac arteries are several times higher than
those reported [33] for normal arteries. The most popular model used to study the
effect of atherosclerosis on arterial wall properties is the rabbit subject to a high
cholesterol diet. Cox and Detweiler [26] have shown that in the iliac arteries from
high cholesterol fed greyhounds, collagen and elastin contents are decreased, but
the ratio of collagen to elastin is increased. In carotid arteries from the treated ani-
mals, the elastin content is increased and the collagen to elastin ratio is decreased.
Their results also show that the elastic modulus of the iliacs from the cholesterol fed
animals is higher than that of the normal iliacs while the treated carotids are
unchanged. By using a rabbit model, Pynadath and Mukherjee [71] found that cho-
lesterol feeding had no effect on the longitudinal dynamic elastic modulus of the
aorta, but the circumferential one was affected significantly. After six weeks of
feeding, the circumferential dynamic elastic modulus increased from the normal
value of 2.7 x 106 dynes/cm2 to 4.0 x 10 6 dynes/cm2. This increase showed a remark-
able correlation with the cholesterol content in the aortas. Although animal models
shed some light on the effects of cholesterol feeding, since atherosclerotic changes
in the arterial wall are so closely related to aging, it is difficult to separate the effect
of atherosclerosis from those of aging. The effects of atherosclerosis on the mechan-
ical properties of the arterial wall remain unclear. Confusion with the effects of
aging and other forms of arteriosclerosis such as medial calcification make interpre-
tation of the results difficult.
Mechanical properties of the arteries from human and various animals have been
extensively studied. However, literature on lymphatic vessels is very scarce. The
data on the circumferential elastic modulus of the lymphatic vessels obtained by
Ohhashi et al. [53, 54] seem to be too low considering that the lymphatics are origi-
nating from the veins.
The overall viscoelastic properties of a large blood vessel such as the aorta are
known to be nonlinear [2, 44]and anisotropic [37]. But due to the fact that the blood
vessel wall is incompressible [72] and deforms orthotropically [31], the mechanical
properties of blood vessels can be described mainly by six coefficients: an elastic
and a viscous moduli in the longitudinal, circumferential and radial directions.
Among them, only the moduli in the circumferential and longitudinal directions
have been studied widely. Much fewer data in the radial direction can be found in
literature. In calculation of the circumferential elastic moduli, it was usually
assumed that the Poisson’s ratio was 0.5 that is not strictly true when large strains
are considered [36]. In fact, the measured data [37, 73] show that the Poisson’s ratio
is about 0.3, not 0.5 as would be predicted for an isotropic material. But the Poisson’s
98 X. Deng and R. Guidoin
ratio was almost constant with respect to circumferential strain and pressure in both
relaxed and constricted canine carotid arteries [73].
To our best knowledge, water diffusion properties of blood vessels have been
studied extensively, but their electrical and thermal properties are still unknown.
Acknowledgement This work was supported by the Medical Research Council of Canada (Grant
MT-7879). The assistance of Y. Marois, M. King and Y. Douville in preparation of this material is
gratefully acknowledged.
Additional Reading
Canfield, T.R. and Dorbin, P.B. (1987) Static properties of blood vessels, in
Handbook of Bioengineering (eds R. Skalak and S. Chien), McGraw-Hill Book
Company, New York, pp. 16.1–16.28.
The authors discuss the mechanical behavior of arteries and the mathematical
method required for quantification of such data. The discussion is entirely con-
cerned with the elastic or pseudoelastic behavior of blood vessels. It should be
emphasized that the arterial wall also exhibits inelastic properties, such as viscosity,
creep, stress & relaxation and pressure-diameter hysteresis. Very few data on
mechanical properties of blood vessels are presented.
Dorbin, P.B. (1983) Vascular mechanics, in Handbook of Physiology, Vol. 3 The
Cardiovascular System, (eds J.T. Shepherd and F. Abboud), Amer. Physiol. Soc.,
Washington, DC, Section 2, pp. 65–102.
This chapter reviews the essential concepts of vascular mechanics and its meth-
ods of quantification. Some of the important controversies are discussed, and further
research areas are pointed out. Detailed information is provided on the structural
and mechanical changes of arteries with age. The effect of vascular disorders such
as arterosclerosis and hypertension on the mechanical behavior of blood vessels is
discussed as well. Extensive addition literature sources are provided.
Schneck, D.J.(1995) An outline of cardiovascular structure and function, in The
Biomedical Engineering Handbook,(ed. J.D. Bronzino), CRC Press, Boca Raton,
pp. 3–14.
The cardiovascular system is described as a highway network, which includes a
pumping station (the heart), a working fluid (blood), a complex branching configu-
ration of distributing and collecting pipes and channels (the blood vessels), and a
sophisticated means for both intrinsic (inherent) and extrinsic (antonomic and endo-
crine) control. Data on both the arterial and venous systems are tabulated. However,
no detailed sources are provided for the data listed. This is a very suitable reference
for biomedical engineers.
Hargen, A.R. and Villavicenco, J.L. (1995) Mechanics of tissue/lymphatic sup-
port, in The Biomedical Engineering Handbook, (ed J.D. Bronzino), CRC Press,
Boca Raton, pp. 493–504.
B6 Arteries, Veins and Lymphatic Vessels 99
From an engineering point of view, the authors discuss the lymphatic system as
a drainage system for fluids and waste products from tissues. Basic concepts of
lymphatic transport along with clinical disorders are discused, although briefly.
Extensive additional sources are cited.
References
1. Fung, Y.C. (1993) Biomechanics, Mechanical Properties of Living Tissues, 2nd edn, Springer-
Verlag, New York, pp. 321–391.
2. Bergel, D.H. (1961) The static elastic properties of the arterial wall. J. Physiol., 156,
445–457.
3. Patel, D.J. and Vaishnav, R.N. (1972) The rheology of large blood vessels, Cardiovascular
Fluid Dynamics, D.H. Bergel (ed.), Academic Press, New York, Vol. 2, pp. 1–64.
4. Fung, Y.C., Fronek, K. and Patitucci, P. (1979) Pseudoelasticity of arteries and the choice of its
mathematical expression. Am. J. Physiol., 237, H620–H631.
5. Remington, J.W. (1955) Hysteresis loop behavior of the aorta and other extensible tissues. Am.
J. Physiol., 180, 83–95.
6. Alexander, R.S. (1971) Contribution of plastoelasticity to the tone of the cat portal vein. Circ.
Res., 28, 461–469.
7. Anitschkov, N. (1933) Experimental arteriosclerosis in animals, in Arteriosclerosis,
E.V. Cowdry, Editor, Macmillan, New York, pp. 298–299.
8. Wilens, S.L. and McCluskey, R.T. (1954) The permeability of excised arteries and other tis-
sues to serum lipid. Circ. Res., 2, 175–182.
9. Baldwin, A.L., Wilson, L.M. and Simon, B.R. (1992) Effect of pressure on aortic hydraulic
conductance. Arteria. Thromb., 12, 163–171.
10. Fung, Y.C. (1984) Biodynamics, Circulation, Springer-Verlag, New York, p. 77.
11. Brecher, G.A. (1956) Venous Return, Grune & Stratton, New York, pp. 30–67.
12. Helps, E.P.W. and McDonald, D.A. (1954) Observation on laminar flow in veins. J. Physiol.,
124, 631–639.
13. Spencer, M.P. and Denison, A.B. (1963) Pulsatile blood flow in the vascular system, in
Handbook of Physiology, Section 2, Circulation, W.F. Hamilton and P. Dow, Editors, Am.
Physiol. Soc., Washington, pp. 839–864.
14. Fry, D.L. and Greefield, J.C. Jr. (1964) The mathematical approach to hemodynamics, with
particular reference to Womersley’s theory, in Pulsatile blood Flow, Attinger, E.O. (ed),
McGraw-Hill, New York, pp. 85–99.
15. Maggio, E. (1965) Microhemocirculation, Observable Variables and Their Biological Control,
C.C. Thomas, Springfield.
16. Whitmore, R.L. (1968) Rheology of the Circulation, Pergamon Press, Oxford, pp. 99–108.
17. Morse, D.E. (1979) The normal aorta: embryology, anatomy, and histology of the aorta, in The
Aorta, J. Lindsay, Jr. and J.W. Hurst, Editors, Grune & Stratton, New York, pp. 15–37.
18. Singhal, S., Henderson, R., Horsfield, K. et al. (1973) Morphometry of human pulmonary arte-
rial tree. Circ. Res., 33, 190–197.
19. Fung, Y.C. (1984) Biodynamics, Circulation, Springer-Verlag, New York, pp. 77–165.
20. Kirk, J.E. (1962) Arterial and arteriolar system, biochemistry, in Blood Vessels and Lymphatics,
D.l. Abramson, Editor, Academic Press Inc., London, pp. 82–95.
21. Fischer, G.M. and Llaurado, J.G. (1966) Collagen and elastin content in canine arteries
selected from functionally different vascular beds. Circ. Res., 19, 394–399.
22. Kanabrocki, E.L., Fels, I.G. and Kaplan, E. (1960) Calcium, cholesterol and collagen levels in
human aorta. J. Gerontal., 15, 383–387.
23. Lansing, A., Alex, M. and Rosenthal, T.B. (1950) Calcium and elastin in human arteriosclero-
sis. J. Gerontal., 5, 112–119.
100 X. Deng and R. Guidoin
24. Kraemer, D.M. and Miller, H. (1953) Elastin content of the abdominal fraction of human aor-
tae. Arch. Pathol., 55, 70–72.
25. Wolinsky, H. (1970) Response of the rat aortic media to hypertension. Morphological and
chemical studies. Circ. Res., 26, 507–523.
26. Cox, R.H. and Detweiler, D.K. (1979) Arterial wall properties and dietary atherosclerosis in
the racing greyhound. Am. J. Physiol., 5, H790–H797.
27. Kirk, J.E. (1962) Venous system, biochemistry, in Blood Vessels and Lymphatics,
D.I. Abramson, Editor, Academic Press Inc., London, pp. 211–213.
28. Strandness, D.E. and Sumner, D.S. (eds) (1975) Hemodynamics for Surgeons, Grune &
Stratton, New York, New York, pp. 161–205.
29. Patel, D.J., Greenfield, J.C. and Fry, D.L. (1964) In vivo pressure-length-radius relationships
of certain vessels in man and dogs, in Pulsatile blood Flow, E.O. Attinger, Editor, McGraw-
Hill, New York, pp. 293–302.
30. Cox, R.H. (1975) Anisotropic properties of the canine carotid artery in vitro. J. Biomech., 8,
293–300.
31. Patel, D.J. and Fry, D.L. (1969) The elastic symmetry of arterial segments in dogs. Circ. Res.,
24, 1–8.
32. Bergel, D.H. (1958) A photo-electric method for the determination of the elasto-viscous
behavior of the arterial wall. J. Physiol., 141, 22–23.
33. Learoyd, B.M. and Taylor, M.G. (1966) Alteration with age in the viscoelastic properties of
human arterial walls. Circ. Res., 18, 278–292.
34. Bader, H. (1967) Dependence of wall stress in the human thoracic aorta on age and pressure.
Circ. Res., 20, 354–361.
35. Langewouters, G.J., Wesseling, K.H. and Goedhard, W.J.A. (1984) The static elastic proper-
ties of 45 human thoracic and 20 abdominal aortas in vitro and the parameters of a new model.
J. Biomech., 17, 425–435.
36. Bergel, D.H. (1960) The Visco-elastic Properties of the Arterial Wall. PhD thesis, University
of London.
37. Patel, D.J., Janicki, J.S. and Carew, T.E. (1969) Static anisotropic elastic properties of the aorta
in living dogs. Circ. Res., 25, 765–779.
38. Rees, P.M. and Jepson, P. (1970) Measurement of arterial geometry and wall composition in
the carotid sinus baroreceptor area. Circ. Res., 26, 461–467.
39. Gow, B.S. and Taylor, M.G. (1968) Measurement of viscoelastic properties of arteries in the
living dog. Circ. Res., 23, 111–122.
40. Reid, L. (1968) Structural and functional reappraisal of the pulmonary artery system, in
Scientific Basis of Medicine, Annual Reviews, Vol. 8, pp. 289–307.
41. Yates, W.G. (1969) Experimental studies of the variations in the mechanical properties of the
canine abdominal vena cava, in SUDAAR Report 393, Stanford University, California.
42. Bergel, D.H. (1972) The properties of blood vessels, in Biomechanics, Its Functions and
Objectives, Y.C. Fung, Editor, Prentice-Hall, Englewood Cliffs, N.J., p. 110 and p. 131.
43. Yamada, H. (1970) Strength of Biological Materials, F.G. Evans, Editor, Williams and Wilkins,
Baltimore, pp. 106–277.
44. Bergel, D.H. (1961) The dynamic elastic properties of arterial wall. J. Physiol., 156,
458–469.
45. Arndt, J.O., Klauske, J. and Mersch, F. (1968) The diameter of the intact carotid artery in man
and its change with pulse pressure. Pfülger’s Arch,, 30, 230–240.
46. Langewouters, G.J., Wesseling, K.H. and Goedhard, W.J.A. (1985) The pressure dependent
dynamic elasticity of 35 thoracic and 16 abdominal human aortas in vitro described by a five
component model. J. Biomech., 18, 613–620.
47. Langewouters, G.J. (1982) Visco-Elasticity of the Human Aorta in vitro in Relation to Pressure
and Age. PhD thesis, Free University, Amsterdam, The Netherland.
48. Tanaka, T.T. and Fung, Y.C. (1974) Elastic and inelastic properties of the canine aorta and their
variation along the aortic trees. J. Biomech., 7, 357–370.
B6 Arteries, Veins and Lymphatic Vessels 101
49. Attinger, E.O. (1967) Modelling of pressure-flow relations in arteries and veins (Abstract).
Biorheology, 4, 84.
50. Wesly, R.L.R., Vaishnav, R.N., Fuchs, J.C.A. et al. (1975) Static linear and nonlinear elastic
properties of normal and arterialized venous tissue in dog and man. Circ. Res., 37, 509–520.
51. Sobin, P. (1977) Mechanical Properties of Human Veins. M.S. thesis, University of California,
San Diego, California.
52. Susznyák, I., Földi, M. and Szabó, G. (1967) Lymphatics and Lymph Circulation: Physiology
and Pathology, Pergamon Press, London, pp. 33–50.
53. Ohhashi, T., Azuma, T. and Sakaguchi, M. (1980) Active and passive mechanical characteris-
tics of bovine mesenteric lymphatics. Am. J. Physiol., 239, H88–H95.
54. Ohhashi, T. (1987) Comparison of viscoelastic properties of walls and functional characteris-
tics of valves in lymphatic and venous vessels. Lymphatic, 20, 219–223.
55. Bergel, D.H. (1964) Arterial viscoelasticity, in Pulsatile blood Flow, E.O. Attinger, Editor,
McGraw-Hill, New York, pp. 275–292.
56. Wilens, S.L. and McCluskey, R.T. (1952) The comparative filtration filtration properties of
excised arteries and veins. Am. J. Med. Sci., 224, 540–547.
57. Tedgui, A. and Lever, M.J. (1985) The interaction of convection and diffusion in the transport
of 131I-albumin within the media of rabbit thoracic aorta. Circ. Res., 57, 856–863.
58. Tedgui, A. and Lever, M.J. (1987) Effect of pressure and intimal damage on 131I-albumin and
14
C-sucrose spaces in aorta. Am. J. Physiol., 253, H1530–H1539.
59. Weinbaum, S., Tzeghai, G., Ganatos, P. et al. (1985) Effect of cell turnover and leaky junctions
on arterial macromolecular transport. Am. J. Physiol., 248, H945–H960.
60. Bazett, H.C., Cotton, F.S., Laplace, L.B. et al. (1935) The calculation of cardiac output and
effective peripheral resistance from blood pressure measurements with an appendix on the size
of the aorta in man. Am. J. Physiol., 113, 312–334.
61. Anderson, J.R. (1980), Muir’s Textbook of Pathology, Arnold, London, pp. 360–395.
62. Nakashima, T. and Tanikawa, J. (1971) A study of human aortic distensibility with relation to
atherosclerosis and aging. Angiology, 22, 477–490.
63. Mozersky, D.J., Summer, D.S., Hokanson, D.E. et al. (1972) Transcutaneous measurement of
the elastic properties of the human femoral artery. Circulation, 46, 948–955.
64. Tobian, L. (1960) Interrelationship of electrolytes, juxtaglomerular cells and hypertension.
Physiol. Rev., 40, 280–312.
65. Peterson, L.H. (1963) Systems behavior, feed-back loops and high blood pressure research.
Circ. Res., 12, 585–596.
66. Mallov, S. (1959) Comparative reactivities of aortic strips for hypertensive and normotensive
rats to epinephrine and levarterenol. Circ. Res., 7, 196–201
67. Greenwald, S.E. and Berry, C.L. (1978) Static mechanical properties and chemical composi-
tion of the aorta of spontaneously hypertensive rats, a comparison with the effects of induced
hypertension. Cardiovas. Res., 12, 364–372.
68. Feigl, E.O., Peterson, L.H. and Jones, A.W. (1963} Mechanical and chemical properties of
arteries in experimental hypertension. J. Clin. Invest., 42, 1640–1647.
69. Bandick, N. and Sparks, H. (1970) Viscoelastic properties of the aortas of hypertensive rats.
Proc. Soc. Exp. Bioi. Med., 134, 56–60.
70. Greene, M.A., Friedlander, R., Boltax, A.J. et al. (1966) Distensibility of arteries in human
hypertension. Proc. Roy. Soc. Exp. Biol., 121, 580–585.
71. Pynadath, T.I. and Mukherjee, D.P. (1977) Dynamic mechanical properties of atherosclerotic
aorta: A correlation between the cholesterol ester content and the viscoelastic properties of
atherosclerotic aorta. Atherosclerosis, 26, 311–318.
72. Carew, T.E.R., Vaishnav, R.N. and Patel, D.J. (1968) Compressibility of the arterial wall. Circ.
Res., 23, 61–68.
73. Dobrin, P.B. and Doyle, J.M. (1970) Vascular smooth muscle and the anisotropy of dog carotid
artery. Circ. Res., 27, 105–119.
Chapter B7
The Intraocular Lens
B7.1 Introduction
Although the existence of the intraocular crystalline lens in the eye was recognized
by the scholars of the Hellenistic period (about 2000 years ago), the actual role of
the lens in vision was properly understood much later. Truly scientific approaches
to the lens measurements and properties began to be applied only in the nineteenth
century [1]. For instance, a simple property—the weight of the human intraocular
lens—was first reported in 1883 [2].
The intraocular crystalline lens is positioned between the aqueous humor and
vitreous body of the eye. It is usually referred to as “the lens,” which is rather con-
venient considering that the term “intraocular lens” (with the acronym IOL) is now-
adays used exclusively to designate an artificial device implanted after surgical
removal of a cataractous (opaque) natural lens.
The lens refracts the light which enters the eye through the pupil and focuses it
on the retina. Its main functions are the following: (1) provides refractive power to
the optical system of the eye; (2) provides the accommodation necessary for normal
vision; (3) maintains its own transparency; and (4) absorbs UV radiation and blue
light, both deleterious to the subsequent ocular segments. These functions are all
important, but the lens’ contribution to the process of accommodation is crucial for
normal vision. Ocular accommodation is the ability of the eye to adjust the focal
length from far to near through changes in the shape of the lens. The consequence
of losing this ability with age is known as presbyopia, and is one of the major causes
of the need for visual correction in the middle-aged humans. Although the existing
theories attempting to explain presbyopia are still debated [3, 4], many investigators
believe that this condition is related to changes within the lens, especially its age-
induced stiffening. As a result, the stiffness of the human lens has been extensively
investigated, and serves as an important parameter in the modeling of the
accommodation and presbyopia. The topic has been critically appraised and supple-
mented with improved experimental techniques in a recent study [5].
The lens is a biconvex body similar to a flattened globe. For descriptive pur-
poses, it has two poles (anterior and posterior), an equator, and therefore two diam-
eters (polar diameter, or thickness of the lens, and equatorial diameter).
The lens is composed of epithelial cells which become anuclear and elongated as
they are displaced further toward the center. Because of the enormous length finally
attained by these cells, they are referred to as lens fibers. The lens is surrounded by
a transparent acellular capsule of variable thickness. A proper epithelium underlies
the capsule along the anterior side and equator, but not under the posterior capsule.
The superficial layers of cells and fibers constitute the lens cortex, while the lens
nucleus is situated in the center. The fibers are continuously formed throughout life
and the new fibers cover the old ones which are displaced toward the nucleus.
Currently, there are significantly more compositional and physical property data
available on animal lenses than on the human lens. The spread of the data measured
for the human lens, compiled here from various sources, results from large varia-
tions in methodologies from one research group to another, and in the individual
characteristics of the tissue from one human donor to another. Even the determina-
tion of a straightforward property like the water content has led to variable results
(Tables B7.1 and B7.2), likely due both to nonuniform distribution of the water in
the lens and to the variability of methods employed for measurements. While data
on the inorganic content of the human lens (Table B7.3) are generally in agreement,
there was a larger variation in reporting the organic content. This is presumably due
to the greater sensitivity of the organic metabolism to age and disease. The most
reliable results are included in Table B7.4.
Tables B7.5, B7.6, B7.7, and B7.8 provide some key dimensional properties of
the lens, while Tables B7.9 and B7.10 are focused on perhaps its most important
feature, the optical properties.
Mechanical characteristics of the lens required sophisticated procedures for their
measurement and the data are likely difficult to reproduce. The currently available
mechanical properties of the human lens are summarized in Tables B7.11, B7.12,
B7.13, B7.14, B7.15, B7.16, B7.17, and B7.18. Although the lens was traditionally
regarded as an elastic material, most investigators agree now that it is in fact a vis-
coelastic solid. The first rheological measurements on human lenses were reported
B7 The Intraocular Lens 105
by Itoi et al. [30], who found an apparent elastic modulus of 10–100 kPa and a
loss tangent (phase shift) of 0.3–0.4. More recent studies using rheometry [31]
confirmed the increase of stiffness with age and its variation with the distance from
the center of the lens. We have to mention, however, that the magnitude of the
values reported for the viscoelastic parameters of the human lens varies from one
laboratory to another. Table B7.19 comprises the most recent results determined by
rheometry.
106 T.V. Chirila and S. Suzuki
Table B7.3 Inorganic ion content of the normal adult human lens
Ion Representative value Source
Sodium 91 mg/100 g wet wt [6]
Potassium 170 mg/100 g wet wt [6]
Calcium 1.4 mg/100 g wet wt [6]
Magnesium 0.29 mg/100 g wet wt [6]
6.2 μg/g dry wt [14]
Zinc 21 μg/g dry wt [6]
25 μg/g dry wt [14]
Copper <1 μg/g dry wt [6]
0.6 μg/g dry wt [14]
Manganese 0.2 μg/g dry wt [14]
Iron 0.4 μg/g dry wt [14]
Rubidium 6.8 μ/g dry wt [14]
Chloride 35.3 mg/100 g wet wt [6]
Phosphate 25 mg/100 g wet wt [15]
Sulfate 24 mg/100 g wet wt [15]
pH 7.3–7.7 [16]
Table B7.5 Dimensional variation with age of the human lens [14]
Dimension Value (mm)
Polar diameter (lens thickness) 3.5–5
Equatorial diameter 6.5–9
Anterior radius of curvature 8–14
Posterior radius of curvature 4.5–7.5
Table B7.8 Weight, volume, and density of the human lens in adult life [20]
Age interval (years) Weight, mean (mg) Volume, mean (mm3) Density (g/cm3)a
20–30 172.0 162.9 1.034
30–40 190.3 177.3 1.048
40–50 202.4 188.1 1.061
50–60 222.3 205.4 1.072
60–70 230.1 213.0 1.082
70–80 237.1 218.3 1.091
a
Calculated at the beginning of a decade
Table B7.12 Variation with age of the shear modulus of the human lens [4]
Shear modulus, G (Pa)a
Age interval (years) Cortex Nucleus
<30b 98.3 ± 64.5 39.0 ± 13.8
>60c 2040 ± 710 17,400 ± 4900
± Standard deviation
a
Determined from penetration measurements
b
Six samples
c
Twelve samples
Table B7.14 Mechanical properties of the human anterior lens capsule [25]a
Age (years) Property Value
20 Tensile modulus (MPa) 5.6
50 4.0
80 1.5
<20 Ultimate tensile stress (MPa) 2.3
>70 0.7
–b Elongation (%) 29
–b Poisson’s ratio 0.47 ± 0.5
± Standard deviation
a
Determined from the volume–pressure relationship upon distension with iso-
tonic saline
b
Found to be independent of age
Table B7.15 Variation with age of the mechanical properties of the human anterior lens capsule [26]
Ultimate strain Ultimate stress Ultimate elastic
Age (MPa) (MPa) modulus (MPa)
7 months 108 17.5 44.8
98 years 40 1.5 4.4
Table B7.16 Variation with age of the mechanical properties of the human posterior lens capsule [27]
Age Ultimate strain (MPa) Ultimate stress (MPa) Ultimate elastic modulus (MPa)
1 101 16.9 55.7
94 34 1.1 5.4
Large variations in the electrical properties of animal lenses have been reported,
but it seems that the only measurements performed on human lenses are those
shown in Table B7.20.
Additional Reading
Bellows, J.G. (ed.) (1975) Cataract and Abnormalities of the Lens, Grune &
Stratton, New York.
A valuable collection of 42 contributions on the lens, its pathology, and surgery,
written by known experts such as Barraquer, Bellows, Choyce, Girard, Hockwin,
Kaufman, Rosen, and Yanoff. The first five introductory chapters present historical
aspects, development, and characterization of the lens. However, most of the book
is dedicated to cataract and its treatment.
Spector, A. (1982) Aging of the lens and cataract formation, in Aging and Human
Visual Function (eds R. Sekuler, D. Kline and K. Dismukes), Alan R. Liss, Inc.,
New York, pp. 27–43.
A brief but comprehensive account of the changes which take place in the com-
position and metabolism of the lens during aging and cataractogenesis.
Duncan, G. and Jacob, T.J.C. (1984) The lens as a physicochemical system, in
The Eye, vol. lb, 3rd edn (ed H. Davson), Academic Press, Orlando, FL,
pp. 159–206.
This text develops some topics usually neglected in other books, including the
structural order in the lens, optical properties of the lens, role of lens membranes,
and electrolyte transport and distribution in the lens.
Cotlier, E. (1987) The lens, in Adler’s Physiology of the Eye, 8th edn (eds
R.A. Moses and W.M. Hart), C.V. Mosby Co., St. Louis, pp. 268–290.
A systematic presentation of the anatomy, biochemistry, and physiology of the
lens.
Moses, R.A. (1987) Accommodation, in Adler’s Physiology of the Eye, 8th edn,
(eds R.A. Moses and W.M. Hart), C.V. Mosby Co., St. Louis, pp. 291–310.
A thorough exposition of all aspects of the mechanism of accommodation and
the role of the lens in vision. A text, by now classic, on a topic much more complex
than it appears.
Jones, W.L. (1991) Traumatic injury to the lens. Optom. Clin., 1, 125–42.
This review article analyzes the effects of concussive trauma to the eye, empha-
sizing the types of injuries to the lens. The mechanical response of the anterior and
posterior segments of the eye to external forces is also described.
Zampighi, G.A. (2006) The lens, in The Biology of the Eye, (ed J. Fischbarg),
Elsevier, Amsterdam, pp. 149–79.
A concise but informative chapter on relevant aspects of the human lens, includ-
ing its cellular architecture and the mechanisms of the molecular processes and
fluxes.
Beebe, D.C. (2011) The lens, in Adler’s Physiology of the Eye, 11th edn (eds
L.A. Levin et al.), Elsevier, Edinburgh, pp. 131–63.
Written by an expert, and beautifully illustrated, this text covers the current
knowledge on the human lens and is supported by almost 500 references.
Recommended sources for ophthalmic terminology:
112 T.V. Chirila and S. Suzuki
Cassin, B., Solomon, S.A.B. and Rubin, M.L. (1990) Dictionary of Eye
Terminology, 2nd edn, Triad Publishing Co., Gainesville, FL, 286 pp.
Stein, H.A., Slatt, B.J. and Stein, R.M. (1992) Ophthalmic Terminology. Speller
and Vocabulary Builder, 3rd edn, Mosby-Year Book Inc., St. Louis, MO, pp. 3–33,
243–257.
Myles, W.M. (1993) Ophthalmic etymology. Surv. Ophthalmol., 37, 306–9.
References
1. Grom E (1975) History of the crystalline lens. In: Bellows JG (ed) Cataract and abnormalities
of the lens. Grune & Stratton, New York, pp 1–28
2. Smith P (1883) Diseases of crystalline lens and capsule. 1. On the growth of the crystalline
lens. Trans Ophthalmol Soc UK 3:79–99
3. Burd HJ, Judge SJ, Flavell MJ (1999) Mechanics of accommodation of the human eye. Vision
Res 39:1591–1595
4. Heys KR, Cram SL, Truscott JW (2004) Massive increase in stiffness of the human lens
nucleus with age: the basis for presbyopia? Mol Vis 10:956–963
5. Wilde GS (2011) Measurement of human lens stiffness for modelling presbyopic treatments.
Ph.D. Thesis, University of Oxford, 219 pp. http://www.eng.ox.ac.uk/civil/publications/the-
ses/wilde.pdf. Accessed 12 Feb 2014
6. Kuck JFR (1970) Chemical constituents of the lens. In: Graymore CN (ed) Biochemistry of the
eye. Academic, London, pp 183–260
7. Fisher RF, Pettet BE (1973) Presbyopia and the water content of the human crystalline lens.
J Physiol 234:443–447
8. Bours J, Fodisch HJ, Hockwin O (1987) Age-related changes in water and crystalline content
of the fetal and adult human lens, demonstrated by a microsectioning technique. Ophthalmic
Res 19:235–239
9. Huizinga A, Bot ACC, de Mul FFM, Vrensen GFJM, Greve J (1989) Local variation in abso-
lute water content of human and rabbit eye lenses measured by Raman microspectroscopy. Exp
Eye Res 48:487–496
10. Deussen A, Pau H (1989) Regional water content of clear and cataractous human lenses.
Ophthalmic Res 21:374–380
11. Siebinga I, Vrensen GFJM, de Mul FFM, Greve J (1991) Age-related changes in local water
and protein content of human eye lenses measured by Raman microspectroscopy. Exp Eye Res
53:233–239
12. Salit PW (1943) Mineral constituents of sclerosed human lenses. Arch Ophthalmol
30:255–258
13. Tabandeh H, Thompson GM, Heyworth P, Dorey S, Woods AJ, Lynch D (1994) Water content,
lens hardness and cataract appearance. Eye 8:125–129
14. Harding JJ, Crabbe MJC (1984) The lens: development, proteins, metabolism and cataract. In:
Davson H (ed) The eye, vol lb, 3rd edn. Academic, Orlando, FL, pp 207–492
15. Paterson CA (1985) Crystalline lens. In: Duane TD, Jaeger EA (eds) Biomedical foundations
of ophthalmology, vol 2, 2nd edn. Harper & Row, Philadelphia, Chapter 10
16. Kuck JFR (1970) Metabolism of the lens. In: Graymore CN (ed) Biochemistry of the eye.
Academic, London, pp 261–318
17. Davson H (1990) Physiology of the eye, 5th edn. Macmillan, London, Chapter 4
18. Dische Z, Zil H (1951) Studies on the oxidation of cysteine to cystine in lens proteins during
cataract formation. Am J Ophthalmol 34:104–113
B7 The Intraocular Lens 113
19. Strenk SA, Semmlow JL, Strenk LM, Munoz P, Gronlund-Jacob J, DeMarco JK (1999) Age-
related changes in human ciliary muscle and lens: a magnetic resonance imaging study. Invest
Ophthalmol Vis Sci 40:1162–1169
20. Scammon RE, Hesdorfer MB (1937) Growth in mass and volume of the human lens in postna-
tal life. Arch Ophthalmol 17:104–112
21. Moffat BA, Atchison DA, Pope JM (2002) Explanation of the lens paradox. Optom Vis Sci
79:148–150
22. Lerman S (1987) Chemical and physical properties of the normal and ageing lens: spectro-
scopic (UV, fluorescence, phosphorescence, and NMR) analyses. Am J Optom Physiol Optics
64:11–22
23. Fisher RF (1971) The elastic constants of the human lens. J Physiol 212:147–180
24. Heyworth P, Thompson GM, Tabandeh H, McGuigan S (1993) The relationship between clini-
cal classification of cataract and lens hardness. Eye 7:726–730
25. Fisher RF (1969) Elastic constants of the human lens capsule. J Physiol 201:1–19
26. Krag S, Olsen T, Andreassen TT (1997) Biomechanical characteristics of the human anterior
lens capsule in relation to age. Invest Ophthalmol Vis Sci 38:357–363
27. Krag S, Andreassen TT (2003) Mechanical properties of the human posterior lens capsule.
Invest Ophthalmol Vis Sci 44:691–696
28. Fisher RF (1977) The force of contraction of the human ciliary muscle during accommodation.
J Physiol 270:51–74
29. van Alphen GWHM, Graebel WP (1991) Elasticity of tissues involved in accommodation.
Vision Res 31:1417–1438
30. Itoi M, Ito N, Kaneko H (1965) Visco-elastic properties of the lens. Exp Eye Res 4:168–173
31. Weeber HA, Eckert G, Pechhold W, van der Heijde RGL (2007) Stiffness gradient in the crys-
talline lens. Graefe’s Arch Clin Exp Ophthalmol 245:1357–1366
32. Schachar RA, Chan RW, Fu M (2011) Viscoelastic properties of fresh human lenses under 40
years of age: implications for the aetiology of presbyopia. Br J Ophthalmol 95:1010–1013
33. Platsch KD, Wiederholt M (1981) Effect of ion substitution and ouabain on short circuit cur-
rent in the isolated human and rabbit lens. Exp Eye Res 32:615–625
34. Gabriel C (1996) Compilation of the dielectric properties of body tissue at RF and microwave
frequencies. Final Technical Report AL/OE-TR-1996-0004, Prepared for Brooks Air Force
Base, Texas, 16 pp. www.dtic.mil/cgi-bin/GetTRDoc?AD = ADA303903. Accessed 27 Feb
2014
Chapter C1
Blood and Related Fluids
C1.1—Introduction
This section provides data for several human biological fluids including blood,
plasma or serum, cerebrospinal (CS) fluid, lymph, synovial fluid, and tear fluid. The
material presented here was gleaned from a variety of sources, with emphasis placed
on the most recently published work, and includes physicochemical properties
(Table C1.1), cellular compositions (Table C1.2), concentrations of inorganics
(Table C1.3), organics (Table C1.4), and major proteins (Table C1.5). In addition,
various properties of the major proteins are presented in Table C1.7, while Tables
C1.8 and C1.9 contain information regarding the components of the coagulation
and complement cascades, respectively. Because of the variability in values for
many properties of biological fluids, in many cases a normal singular range of such
values is listed. In all cases, the data are those compiled for normal human adults
and, where possible, differences with respect to gender are included. It must be
stressed that fluid properties can readily change as a result of disease, aging, or drug
ingestion.
The following equations can be used to estimate blood volume (BV, mL), eryth-
rocyte volume (EV, mL), and plasma volume (PV, mL) from the known body mass
(b, kg) with a coefficient of variation of approximately 10%:
V. Turitto (*)
Department of Biomedical Engineering, University of Memphis, Memphis, TN 38152, USA
S.M. Slack
Department of Biomedical Engineering, University of Memphis,
Campus Box 526582, Memphis, TN 38152-6502, USA
These equations, relating BY, PY, and EY to body weight, are taken from
Lentner [12].
Additional correlations relating these volumes to body weight and surface area
are available from the same source.
C1 Blood and Related Fluids 117
Additional Reading
Ditmer, D.S. (ed.) (1961) Blood and Other Body Fluids, Federation of American
Societies for Experimental Biology, Washington, D.C.
This text provides a thorough compilation of the physical properties and compo-
sition of numerous biological fluids. Unlike the Geigy Scientific Tables, this book
also reports data for many non-human species. However, citations and some mea-
surement techniques are somewhat outdated.
Kjeldsberg C.R. and Knight J.A. (eds) (1993) Body Fluids: Laboratory
Examination of Amniotic, Cerebrospinal, Seminal, Serous & Synovial Fluids, 3rd
ed., American Society of Clinical Pathologists, Chicago.
An excellent source of information, especially for a clinician or medical tech-
nologist. Includes numerous color photographs of fluids and cells. Discusses abnor-
mal amounts or types of specific proteins and cells in fluids as potentially diagnostic
of disease states.
Lentner, C. (ed.) (1984) Geigy Scientific Tables, Ciba-Geigy, Basle.
This is the most comprehensive source of information available on properties
and composition of body fluids. Volumes 1 and 3 provide extensive data, generally
in tabular form, on fluid content (as well as measurement technique), related to
gender, age and disease state.
124 V. Turitto and S.M. Slack
References
1. Ditmer, D.S. (ed.) (1961) Blood and Other Body Fluids, Federation of American Societies for
Experimental Biology, Washington, D.C.
2. Fullard, R.J. (1988) Current Eye Research, 7, 163–179.
3. Chmiel, H. and Walitza, E. (1980) On the Rheology of Blood and Synovial Fluids, Research
Studies Press, New York.
4. Barbanel, J.C., Lowe, G.D.O. and Forbes, C.D. (1984), The viscosity of blood. in Mathematics
in Medicine and Biomechanics, G.F. Roach (ed.), Shiva Publications, Nantwich, p. 19.
5. Begg, T.B. and Hearns, J.B. (1966) Components in blood viscosity: The relative contribution
of hematocrit, plasma fibrinogen and other proteins. Clinical Science, 31, 87–93.
6. Whitmore, R.L. (1968) Rheology of The Circulation, Pergamon Press, New York.
7. Merrill, E.W. (1969) Rheology of blood. Physiology Reviews, 49, 863–867.
8. Harkness, J. (1971) The viscosity of human blood plasma: Its measurement in health and dis-
ease. Biorheology, 8, 171–193.
9. Lowe, G.D.O., Barbanel, J.C. and Forbes, C.D. (eds) (1981) Clinical Aspects of Blood Viscosity
and Cell Deformability, Springer-Verlag, New York.
10. Lowe, G.D.O and Barbanel, J.C. (1988) Plasma and blood viscosity, in Clinical Blood
Rheology, G.D.O. Lowe (ed.), CRC Press, Boca Raton, pp. 11–44.
11. Schmidt-Schonbein, H. (1988) Fluid dynamics and hemorheology, in Clinical Blood Rheology,
G.D.O. Lowe (ed.), CRC Press, Boca Raton, pp. 129–220.
12. Lentner, C. (ed.) (1984) Geigy Scientific Tables, Ciba-Geigy, Basle.
13. Bicks, R.L. (1993) Hematology: Clinical and Laboratory Practice, Mosby, St Louis.
14. Sokoloff , L. (ed.) (1978) The Joints and Synovial Fluid, Academic Press, New York.
15. Hermens, W.T., Willems, G.M. and Visser, M.P. (1982) Quantification of Circulating Proteins:
Theory and Applications Based on Analysis of Plasma Protein Levels, Martinus Nijhoff, The
Hague.
16. Colman, R.W., Hirsh, J., Marder, V.J., et al. (eds) (1993) Hemostasis and Thrombosis,
Lippincott, Philadelphia.
17. Schultze, H.E. and Heremans, J.F. (1966) Nature and Metabolism of Extracellular Proteins,
Elsevier, Amsterdam.
18. Bing, D.H. (ed.) (1978) The Chemistry and Physiology of the Human Plasma Proteins,
Pergamon Press, Boston.
19. Stamatoyannopoulos, G., Nienhuis, A.W., Majerus, P.W., et al. (eds) (1994) The Molecular
Basis of Blood Diseases, W.B. Saunders, Philadelphia.
Chapter C2
The Vitreous Humor
C2.1 Introduction
The vitreous humor, also termed vitreous body, vitreus, or vitreous, is a clear and
transparent mass (gel or liquid or a mixture of both) that fills the posterior cavity of
the eye in vertebrates, between the lens and the retina.
The mammalian vitreous humor can be defined as an avascular, virtually acellu-
lar, highly hydrated gel located in the vitreous cavity of the eye and consisting of a
dilute network of heterotypic collagen fibrils surrounded by a mixture of glycosami-
noglycans where hyaluronan is the predominant component. In humans, the vitre-
ous humor is perceived as a gel-like body that can provide an adequate support for
the retina, allows the diffusion of metabolic solutes, and allows the light to reach the
retina. Historically, there are two differing concepts on the nature of vitreous humor.
A significant amount of evidence supports the view that the vitreous humor is basi-
cally an extracellular matrix. Another model has been developed in which the vitre-
ous body is considered as a specialized, but simple, connective tissue. The two
concepts are not yet reconciled; therefore the structure and role of the vitreous
humor are usually regarded from both points of view. It is generally accepted that
this gel-like material possesses a unique macromolecular organization in the form
of a double-network system consisting of a scaffold of randomly spaced rodlike
collagen fibrils surrounded by and entangled with a network of very large coiled-up
macromolecules of hyaluronic acid (hyaluronan). The double-network model
explains satisfactorily most of the properties of the vitreous body, as well as its
remarkable mechanical stability, although it probably overestimates the importance
of hyaluronan. The natural vitreous gel displays true viscoelastic properties which
enable it to resist sudden compression shocks, offering much the best protection for
the retina against contusion trauma. It is generally believed that the hyaluronan
network imparts the latter feature, while the collagen network is responsible for the
plasticity and tensile strength of the vitreous humor. The above considerations illus-
trate the understanding of the supramolecular organization of the vitreous humor as
presented it in the first edition of this book [1]. Since that time, there was not much
progress in this understanding [2]. However, our appreciation of the physiological
role and functions of the vitreous humor in the eye has improved significantly [3, 4].
Further research revealed its central role in major diseases of the eye such as dia-
betic retinopathy, retinal vein occlusion, age-related macular degeneration, nuclear
sclerotic cataract, and primary open-angle glaucoma [5].
Biochemically, the vitreous humor consists of collagen types I, V, IX, and XI,
non-collagenous proteins (opticin, fibrillin-1, VIT1, fibronectin, proteoglycans, and
others), and glycosaminoglycans [3, 6]. There has been considerable controversy
regarding the existence of direct interactions between hyaluronan and the collagen
fibrils, and the role of the former in maintaining the long-range spacing in the col-
lagen network. There is no clear evidence for such interactions, but some experi-
ments have suggested [7] that about 6 % of the total hyaluronan may be involved in
maintaining the vitreous gel in a distended state. However, it is rather believed [3]
that, due to indirect interfibrillar interactions, the collagen network has sufficient
mechanical resilience to support itself in solution. It was suggested that the vitreous
humor plays only a minor role in the regulation of intraocular pressure [8].
The role of the vitreous humor and its physiological functions in the human eye
can be summarized as such: (a) contributes to the growth of the eye and to maintain-
ing its volume, elasticity, and resilience, and provides protection during mechanical
trauma; (b) contributes to the total transparency of the ocular pathways; (c) serves
as a support for the intraocular lens and contributes to the accommodation; (d) acts
as a repository and transport conduit for the substances involved in the metabolism
of the surrounding ocular tissues; (e) acts as a barrier to biomolecules, biomacro-
molecules, and cells, and as an inhibitor of inflammation and neovascularization; (f)
prevents the onset of certain types of cataract by protecting the intraocular lens
against oxygen-induced damage; and (g) plays a significant role in the etiology of a
number of major ocular diseases.
As is the case with many other structural elements of the eye, there are presently
much more data on animal vitreous humor than on the human counterpart. The
numerical data tabulated here cover almost everything reported so far on the analy-
sis of the human vitreous. However, in many cases it is not possible to select a most
reliable single value; dependable values representing a range are thus provided in
several of the following tables, illustrating the large variability in the experimental
methodology from one laboratory to another, in collecting and processing the
C2 The Vitreous Humor 127
Table C2.4 Organic content of the human vitreous humor (high-molecular-weight components)
Component Representative value (μg/cm3) Source
Proteinsa 280–1360 [25]
450–1100 [27]
Hyaluronan 100–400 [27]
42–399 [14]
65–210 [28]
Versicanb 60 [29]
Collagen 40–120 [27]
30–532 [14]
Albumin 293 ± 18 [30]
Immunoglobulin (IgG) 33.5 ± 3 [30]
α1-Antitrypsin 141 ± 2.9 [30]
α1-Acid glycoprotein 4 ± 0.7 [30]
(±) represents standard deviation
a
Total protein content
b
A high-molecular-weight proteoglycan based on chondroitin sulfate
Table C2.7 Gel and liquid volume of the human vitreous as a function of age
(adapted from [21])a
Age (years) Gel volume (cm3) Liquid volume (cm3)
Birth 1.6 0
5 3.3 0
10 3.5 0.7
20 3.9 0.9
30 3.9 0.9
40 3.9 0.9
50 3.5 1.3
60 3.2 1.6
70 2.8 2.0
80 2.5 2.3
90 2.2 2.6
a
The liquid vitreous appears first in childhood and by the seventh decade it
occupies half of the vitreous [21, 22]
Additional Reading
Balazs, E.A. (1968) The molecular biology of the vitreous, in New and Controversial
Aspects of Retinal Detachment (ed A. McPherson), Harper & Row, New York,
pp. 3–15.
This is a landmark paper on the nature of the vitreous body, describing the
“mechanochemical” (or “double-network”) model. This model explains satisfacto-
rily the correlations between some properties of the vitreous (composition, rheol-
ogy, volume, cell population, transparency) and the physicochemical principles
governing its stability (frictional interaction, expansion/contraction, the excluded-
volume concept, and the molecular-sieve effect).
Berman, E.R. and Voaden, M. (1970) The vitreous body, in Biochemistry of the
Eye (ed C.N. Graymore), Academic Press, London, pp. 373–471.
A comprehensive summary of knowledge at that time on animal and human vit-
reous humor, including development, chemical composition, metabolism, and aging
effects.
Shields, J. A. (1976) Pathology of the vitreous, in Current Concepts of the
Vitreous Including Vitrectomy (ed K.A. Gitter), C.V. Mosby Co., St. Louis,
pp. 14–42.
This book chapter presents competently the pathologic vitreous, including devel-
opmental abnormalities, inflammation, hemorrhage, effects of trauma, systemic dis-
eases, and degenerative processes.
Gloor, B.P. (1987) The vitreous, in Adler’s Physiology of the Eye, 8th edn (eds
R.A. Moses and W.M. Hart), C.V. Mosby Co., St. Louis, pp. 246–267.
A concise description of all aspects of the vitreous humor, including properties,
development, anatomy, structure, biochemistry, metabolism, and pathology.
Sebag, J. (1989) The Vitreous. Structure, Function, and Pathology, Springer-
Verlag, New York.
This is probably only the second single-authored book in this century to be dedi-
cated entirely to the topic of vitreous humor, written by the leading scholar in the
field. It is a well-structured and updated compendium. The first half of the book is
dedicated to structure, properties, and physiology of the vitreous. Pathology of the
vitreous is analyzed in the other half from a biological angle. Although a clinician,
the author manages to avoid typical clinical descriptions and to provide a text which
integrates the basic scientific knowledge for both clinicians and scientists.
Williams, G.A. and Blumenkranz, M.S. (1992), Vitreous humor, in Duane’s
Foundations of Clinical Ophthalmology, vol. 2 (eds W. Tasman and E.A. Jaeger),
J.B. Lippincott Co., Philadelphia, chapter 11.
This chapter (27 pages) presents the modern concepts in the pathophysiologic
mechanisms of vitreous diseases, and in the clinical conditions involving the vitre-
ous (detachment, macular holes and membranes, diabetes, proliferative vitreoreti-
nopathy, hyalosis, amyloidosis). Aspects such as separation of the vitreous from the
retina and traction of the vitreous by hypocellular gel contraction are explained
according to the most recent findings.
132 T.V. Chirila and Y. Hong
Lund-Andersen, H., Sebag, J., Sander, B. and la Cour, M. (2006) The vitreous, in
The Biology of the Eye (ed J. Fischbarg), Elsevier, Amsterdam, pp. 181–94.
A concise but very informative text covering all aspects of the vitreous humor.
Lund-Andersen, H. and Sander, B. (2011) The vitreous, in Adler’s Physiology of
the Eye, 11th edn (eds L.A. Levin et al.), Elsevier, Edinburgh, pp. 164–81.
The most recent chapter on the vitreous humor to appear in the traditional text-
book, covering the modern views on this part of the eye, and emphasizing biophysi-
cal and physiological aspects.
Sebag, J. and Green, W.R. (2013) Vitreous and vitreoretinal surface, in Retina,
5th edn (ed S.J. Ryan), vol. 1, part 2, Elsevier, Amsterdam, pp. 482–516.
A concise, update, and beautifully written account of the current knowledge of
the vitreous humor, providing a balanced presentation of the structure, anatomy,
physiology, and pathology of the vitreous and of its interface with the retina. This
chapter in the classical monumental treatise displays also remarkable graphics.
Recommended sources for ophthalmic terminology:
Cassin, B., Solomon, S.A.B. and Rubin, M.L. (1990) Dictionary of Eye
Terminology, 2nd edn, Triad Publishing Co., Gainesville, FL, 286 pp.
Stein, H.A., Slatt, B.J. and Stein, R.M. (1992) Ophthalmic Terminology. Speller
and Vocabulary Builder, 3rd edn, Mosby-Year Book Inc., St. Louis, MO, pp. 3–33,
183–198, 275–278.
Myles, W.M. (1993) Ophthalmic etymology. Surv. Ophthalmol., 37, 306–9.
References
1. Chirila TV, Hong Y (1998) The vitreous humor. In: Black J, Hastings G (eds) Handbook of
biomaterial properties. Chapman & Hall, London, pp 125–131
2. Sebag J (2009) Vitreous: the resplendent enigma. Br J Ophthalmol 93:989–991
3. Bishop PN (2000) Structural macromolecules and supramolecular organization of the vitreous
gel. Prog Retin Eye Res 19:232–244
4. Kleinberg TT, Tzekov RT, Stein L, Ravi N, Kaushal S (2011) Vitreous substitutes: a compre-
hensive review. Surv Ophthalmol 56:300–323
5. Holekamp NM (2010) The vitreous gel: more than meets the eye. Am J Ophthalmol
149:32–36
6. Angi M, Kalirai H, Coupland SE, Damato BE, Semeraro F, Romano MR (2012) Proteomic
analyses of the vitreous humour. Mediat Inflamm 2012; Article ID 148039 (7 pages).
doi:10.1155/2012/148039
7. Bishop PN, McLeod D, Reardon A (1999) Effects of hyaluronan lyase, hyaluronidase, and
chondroitin ABC lyase on mammalian vitreous gel. Invest Ophthalmol Vis Sci
40:2173–2178
8. Collins R, van der Werff TJ (1980) Mathematical models of the dynamics of the human eye.
Lect Notes Biomath 34:5–6
9. Redslob E (1932) Le corps vitré. Masson & Cie, Paris, pp 299–305
10. Duke-Elder WS (1929) The physico-chemical properties of the vitreous body. J Physiol
68:155–165
11. Nordmann J (1968) Chimie. In: Brini A, Bronner A, Gerhard JP et al (eds) Biologie et chirurgie
du corps vitré. Masson & Cie, Paris, pp 95–167
C2 The Vitreous Humor 133
12. Mörner CT (1894) Untersuchung der Proteïnsubstanzen in den licht-brechenden Medien des
Auges. Z Physiol Chem 18:233–256
13. Gala A (1925) Observations on the hydrogen ion concentration in the vitreous body of the eye
with reference to glaucoma. Br J Ophthalmol 9:516–519
14. Lee B (1994) Comparative rheological studies of the vitreous body of the eye. Ph.D. Thesis,
University of Pennsylvania, 1992. U.M.I./Bell & Howell Co., Ann Arbor, MI, pp. 102,
138–152
15. Sturner WQ, Dowdey ABC, Putnam RS, Dempsey JL (1972) Osmolality and other chemical
determinations in postmortem human vitreous humor. J Forensic Sci 17:387–393
16. Visser-Heerema J (1936) Über das spezifische Gewicht der bei der Operation von
Netzhautablösungen gewonnenen Flüssigkeit. Arch Augenheilkd 109:543–561
17. Berman ER, Michaelson IC (1964) The chemical composition of the human vitreous body as
related to age and myopia. Exp Eye Res 3:9–15
18. Shafer DM (1965) Intraocular injections as adjuncts to other retinal detachment procedures.
In: Schepens CL, Regan CDJ (eds) Controversial aspects of the management of retinal detach-
ment. Little, Brown & Co., Boston, pp 186–204
19. Guggenheim I, Franceschetti A (1928) Refraktometrische Untersuchungen des Glaskörpers
von Kaninchen und Mensch (unter physiologischen und pathologischen Bedingungen). Arch
Augenheilkd 98:448–482
20. Larsen JS (1971) The sagittal growth of the eye. III. Ultrasonic measurement of the posterior
segment (axial length of the vitreous) from birth to puberty. Acta Ophthalmol 49:441–453
21. Balazs EA (1992) Functional anatomy of the vitreus. In: Tasman W, Jaeger EA (eds) Duane’s
foundations of clinical ophthalmology, vol 1. J.B. Lippincott Co., Philadelphia, Chapter 17
22. Balazs EA, Denlinger JL (1982) Aging changes in the vitreus. In: Sekuler R, Kline D,
Dismukes K (eds) Aging and human visual function. Alan R. Liss, Inc., New York, pp 45–57
23. Naumann HN (1959) Postmortem chemistry of the vitreous body in man. Arch Ophthalmol
62:356–363
24. Coe JI (1969) Postmortem chemistries of human vitreous humor. Am J Clin Pathol
51:741–750
25. Swann DA (1980) Chemistry and biology of the vitreous body. Int Rev Exp Pathol 22:1–64
26. Süllmann H (1951) Chemie des Auges. Tabul Biol 22:1–119
27. Balazs EA, Denlinger JL (1984) The vitreus. In: Davson H (ed) The eye, vol la, 3rd edn.
Academic, Orlando, FL, pp 533–589
28. Grimshaw J, Kane A, Trocha-Grimshaw J, Douglas A, Chakravarthy U, Archer D (1994)
Quantitative analysis of hyaluronan in vitreous humor using capillary electrophoresis.
Electrophoresis 15:936–940
29. Theocharis DA, Skandalis SS, Noulas AV, Papageorgakopoulou N, Theocharis AD, Karamanos
NK (2008) Hyaluronan and chondroitin sulfate proteoglycans in the supramolecular organiza-
tion of the mammalian vitreous body. Connect Tissue Res 49:124–128
30. Clausen R, Weller M, Wiedemann P et al (1991) An immunochemical quantitative analysis of
the protein pattern in physiologic and pathologic vitreous. Graefe’s Arch Clin Exp Ophthalmol
229:186–190
31. Boettner EA, Wolter JR (1962) Transmission of the ocular media. Invest Ophthalmol
1:776–783
32. Chirila TV, Hong Y, Dalton PD, Constable IJ, Refojo MF (1998) The use of hydrophilic poly-
mers as artificial vitreous. Prog Polym Sci 23:475–508
33. Swindle-Reilly KE, Ravi N (2010) Designing hydrogels as vitreous substitutes in ophthalmic
surgery. In: Chirila T (ed) Biomaterials and regenerative medicine in ophthalmology.
Woodhead Publishing Ltd/CRC, Oxford/Boca Raton, pp 339–373
34. Nickerson CS, Karageozian HL, Park J, Kornfield JA (2005) Internal tension: a novel hypoth-
esis concerning the mechanical properties of the vitreous humor. Macromol Symp
227:183–189
134 T.V. Chirila and Y. Hong
35. Nickerson CS, Park J, Kornfield JA, Karageozian H (2008) Rheological properties of the vitre-
ous and the role of hyaluronic acid. J Biomech 41:1840–1846
36. Zimberlin JA, McManus JJ, Crosby AJ (2010) Cavitation rheology of the vitreous: mechanical
properties of biological tissue. Soft Matter 6:3632–3635
37. Sharif-Kashani P, Nishida K, Kavehpour HP, Schwartz SD, Hubschman JP (2013) Effect of cut
rates on fluidic behavior of chopped vitreous. Retina 33:166–169
38. Abdelkawi SA, Abdel-Salam AM, Ghoniem DF, Ghaly SK (2013) Vitreous humor rheology
after Nd:YAG laser photo disruption. Cell Biochem Biophys 68:267–274
39. Watts F, Tan LE, Wilson CG, Girkin JM, Tassieri M, Wright AJ (2014) Investigating the micro-
rheology of the vitreous humor using an optically trapped local probe. J Opt 16; Article ID
#015301 (7 pp)
40. Lee B, Litt M, Buchsbaum G (1992) Rheology of the vitreous body. Part I: Viscoelasticity of
human vitreous. Biorheology 29:521–533
41. Nickerson CS (2006) Engineering the mechanical properties of ocular tissues. Ph.D. Thesis,
California Institute of Technology, 2005, pp. 57–59. http://resolver.caltech.edu./
CaltechETD:etd-03172005-145045. Accessed 11 Dec 2013
Chapter C3
The Cornea
C3.1 Introduction
It is rather surprising that, in spite of the prominence given to its study, the human
cornea has been considerably less investigated than the animal corneas in some
respects, perhaps less than any ocular element, which is rather peculiar considering
that excised tissue rims from postmortem corneas are routinely discarded in the eye
banks. Experimental data concerning certain properties of the human cornea, such
as its composition or electrical properties, are actually available to a little extent,
C3 The Cornea 137
and the investigators must rely upon information abundantly available for the cor-
nea of a variety of other vertebrate species. Table C3.1 shows information available
on the composition of the human cornea, as summarized by Ehlers and Hjortdal
[A6] (likely as rounded-off values). The fact that our literature search did not result
in more data confirms the above statement.
Another peculiarity of the study of the human cornea is the great attention and
amount of research dedicated to its dimensional characteristics (diameter, thickness,
geometry). However, this aspect has a justification. For instance, the corneal diam-
eter is an important tool in identifying the congenital and developmental anomalies
of the cornea, and for the design and fitting of contact lenses, intraocular lenses, and
capsular tension rings. The early literature on the measurement of the corneal diam-
eter has been reviewed by Martin and Holden [9]; more recent results, with an
emphasis on the corneas of non-Caucasian subjects, have been summarized by
Mashige [10]. As such, the concept of “corneal diameter” is poorly defined and still
debated, and consequently the reported estimations differ to some extent between
investigators. Table C3.2 presents a selection of data for the corneal diameter.
Similarly, the measurement of the thickness of the human cornea triggered an
inordinate amount of investigation. Since the corneal thickness is very sensitive to
an abnormal hydration of the stroma, and also related to the intraocular pressure in
the eye, it became a diagnostic tool for certain pathological conditions, such as
corneal degenerations. Exact measurement of the corneal thickness is also of great
significance in the assessment of patients for corneal refractive surgery. The mea-
surement of this quantity is associated with some difficulties; for instance, the thick-
ness of postmortem corneas is much higher than in living persons due to uncontrolled
hydration. Also, the thickness is not uniform throughout the cornea: it is the thinnest
in the central region and becomes thicker toward the periphery. How important is
this dimension is illustrated in the amount of research that has been reviewed in no
less than four major papers [10, 13–15]. Table C3.3 contains representative values
for the central corneal thickness (CCT).
An essential function of the cornea is to allow the formation of an image on the
retina, and one of the prerequisites for this task is the existence of a curvature of the
corneal surface(s) that can refract light sufficiently to focus on the retina. Corneal cur-
vature, expressed commonly as the radius of curvature, has been of interest to vision
scientists for centuries. A system of methodology and instrumentation, covered by the
term “keratometry,” has been developed to measure the curvature of the central cornea.
The refractive power of the cornea can be calculated from its radius of curvature, thick-
ness, and refractive index. Representative values of the corneal radius of curvature and
other derived parameters for the corneal optic system are summarized in Table C3.4.
Another parameter of significance for the light refraction through the cornea,
intrinsic to the tissue as such, is the refractive index. Its measurement has a recorded
history dating back to the nineteenth century, even before the introduction of refrac-
tometers. A refractive index of 1.376, which is commonly used in applications, has
originated in the value of 1.3763 measured in human cornea by Matthiesen in 1891.
However, as seen in Table C3.5, it has been seldom that refractive indices in the
vicinity of 1.376 were reported. There is also substantial variability in the measured
values, likely due to variation in tissues, subjects, and procedures. A similar level of
data spread can be noticed with respect to the measurement of radiation transmis-
sion through the cornea (Table C3.6).
Although the electrical properties of the ocular elements have been extensively
investigated in the rabbit eye [47], there is little amount of such data on the human
eye. Table C3.7 contains the available information regarding the human cornea.
Like any tissue in our body, the cornea undergoes changes with the passage of
time. However, this tissue appears to be remarkably stable and to withstand elevated
physical stress for long intervals, in most of the cases for our natural life duration.
This dimensional stability, which is essential for normal vision, is a result of the
inherent mechanical properties of the corneal tissues, and it is believed that the latter
are ultimately determined by the properties and integrity of the stroma and the under-
lying Descemet’s membrane. The biomechanics of the cornea is a rather well-
established field. However, there are large discrepancies between the results of the
mechanical characteristics measured for the human cornea. For instance, the varia-
tion in the reported experimental values for the elastic modulus covers three orders
C3 The Cornea 141
Table C3.6 Transmission of ultraviolet and visible radiation through the human cornea
Transmittance (%)
Wavelength (nm) Source: [42]a,b [43]b,c [44]b,d [45]b,e [46]f
270 0 0 − <0.01 −
290 0 0 − 0.1 −
300 0 ~7 − ~2 −
310 58 ~8 − ~8 27
325 63 33 − ~10 −
350 72 50 − ~12 −
400 81 58 − ~15 74
500 90 64 89.2 − −
600 92 71 93.7 − −
700 94 78 95.4 − −
a
measured in 9 enucleated eyes, aged between 4 weeks and 75 years, on isolated cornea; b extrap-
olated from the transmission spectra; c measured on a 24-year-old cornea; d measured in 8 postmor-
tem eyes, aged between 22 and 45 years, on the whole eyes; e measured in 20 corneal rims discarded
form the eye bank; f measured in the central region of 4 postmortem corneas (aged 35 to 67 years)
after removal of the epithelia; - not measured
Method Remarks Young’s modulus (MPa) Ultimate strength (MPa) Elongation (%) Source
Uniaxial tensile test Cornea strips 57 ± 4.1b, c 12.7 ± 0.6b, c 35 ± 3b, c [55]
Central cornea strips 0.34d [56]
Central cornea strips 4.1e [56]
Central cornea strip with 0.282 ± 0.159f [57]
Bowman’s layer
Central cornea strip without 0.245 ± 0.088f [57]
Bowman’s layer
Peripheral strips 3.81 ± 0.4g [58]
h
Central strips 1.3 [59]
Inflation Anterior hemisphere of globe 0.37i [60]
Whole globe 1.11j [61]
Descemet’s membrane 5k [62]
Whole globe 2.87 ± 0.38b, l, m [63]
Whole globe 8.55 ± 0.83b, l, n [63]
Whole globe 19.5 ± 0.98b, l, o [63]
Excised cornea 0.79 ± 0.22p [64]
Descemet’s membrane 11.8 ± 1b, q 1.72 ± 0.19b, q 31.2 ± 1.4b, q [65]
Cornea with scleral ring 0.624r, s [66]
Cornea with scleral ring 0.777r, t [66]
Cornea with scleral ring 0.961r, u [66]
Compression Whole globe 0.036v [67]
Ultrasonic Whole globe in saline 5.3 ± 1.1w [68]
Whole globe in 15 % 20 ± 1.0x [68]
dextran solution
T.V. Chirila and S. Suzuki
C3
cut)
Optical methods Living subjects 0.0245 ± 0.0057bb [70]
Living subjects 9.03 ± 0.42cc [71]
Living subjects 2.48 ± 091dd [72]
AFM indentation Anterior basement membrane 0.0075 ± 0.0042ee [73]
Descemet’s membrane 0.050 ± 0.0178ff [73]
Bowman’s layer 0.1098 ± 0.0132gg [74]
Anterior stroma 0.0331 ± 0.0061hh [74]
Anterior stroma 1.14-2.63ii [75]
Anterior stroma 0.25 ± 0.21jj [76]
Posterior stroma 0.1 ± 0.06kk [76]
a
values are mean, or mean ± standard deviation, except in some cases as further specified; b mean ± standard error; c measured in 11 corneas, aged 14 to 76 years;
d
intraocular pressure (IOP) = 30 mm Hg; e IOP = 10 mm Hg; f measured in 5 pairs of eyes, aged 59 to 82 years, IOP = 38 mm Hg; g measured in 10 corneal rings
left from keratoplasty; h measured in 5 enucleated eyes, IOP = 38 mm Hg; i at a strain of ~ 3.3 × 10-2 cm/cm for enucleated corneas; j measured in 5 pairs of eyes,
aged 39 to 67 years, IOP = 19 mm Hg; k aged 45 to 90 years, IOP = 11 mm Hg; l measured in 8 eyes, aged 64 to 88 years; m IOP = 2–10 mm Hg; n IOP = 10–25 mm Hg;
o
IOP = 25–100 mm Hg; p measured in 12 corneas, aged 20 to 69 years, IOP = 15 mm Hg; q measured in 7 Descemet’s membranes, aged 81.3 ± 12.3 years;
r
pressure rate = 37.5 mm Hg/min, IOP = 30 mm Hg; s measured in 4 corneas, aged 50 to 64 years; t measured in 6 corneas, aged 65 to 79 years; u measured in 13
corneas, aged 80 to 95 years; v measured in 2 enucleated eyes, IOP = 35 or 40 mm Hg; w measured in 6 eyes, aged 75 to 82 years; x measured in 6 eyes, aged 78 to
84 years; y measured in 4 corneas; z measured in 3 corneas; aa measured in 7 corneas; bb measured in 28 persons in both eyes, aged 16 to 66 years using photokera-
tometry and topographic pachometry, IOP = 10 mm Hg; cc measured in 29 persons, aged 17 to 66 years, using slit-lamp, photokeratoscopy, topographic pachometry,
and Goldmann applanation tonometry, IOP = 30 mm Hg; dd measured in 5 persons, aged 27 to 80 years, using Reichert ocular response analyzer (ORA); ee measured
in 6 corneas, aged 58 to 67 years; ff measured in 5 corneas, aged 53 to 68 years; gg measured in 6 corneas, aged 19 to 73 years; hh measured in 5 corneas, aged 19 to
73 years; ii measured in 4 corneas; jj measured in 6 pairs of corneas, aged 39 to 89 years; kk measured in 6 pairs of corneas, aged 56 to 88 years
143
144 T.V. Chirila and S. Suzuki
are carried out on living subjects. Table C3.8 presents a compilation, hopefully
exhaustive, of the mechanical property measurements of the human cornea.
Additional Reading
[A1] Maurice, D.M. (1962) The cornea and sclera, in The Eye. Vol. 1: Vegetative
Physiology and Biochemistry (ed H. Davson), Academic Press, New York,
pp. 289-368.
This was the defining text to many generations of corneal researchers, written by
one of the greatest scholars on the topic.
[A2] King, Jr., J.H. and McTigue, J.W. (eds) (1965) The Cornea World Congress,
Butterworths, Washington, 754 pp.
A book containing the proceedings of the First World Congress on the Cornea,
held in Washington, D.C. in 1964, attended by the most illustrious corneal surgeons
and investigators of the period, such as Joachim Barraquer, José Barraquer, Bietti,
Cardona, Castroviejo, DeVoe, Dohlman, Franceschetti, Maumenee, Refojo, Rycroft,
and many others. The book covers the entire knowledge on the cornea up to that
time. It is still a valuable reference book for all cornea scholars.
[A3] Sherrard, E.S. (1977) The cornea: structure and transparency, in Scientific
Foundations of Ophthalmology (eds E.S. Perkins and D.W. Hill), Heinemann
Medical Books, London, pp. 29-35.
A concise yet very informative text, explaining in accessible terms the structure
and function of the cornea.
[A4] Fehér, J. (1996) Pathophysiology of the Eye, Vol. 3: Transparency and
Refraction of the Cornea, Akadémiai Kiadó, Budapest, 230 pp.
This book presents in organized fashion many aspects of the cornea including its
development, anatomy, structure, physiology, and new approaches to certain inter-
relations between the parts of the cornea.
[A5] Foster, C.S, Azar, D.T. and Dohlman, C.H. (eds) (2005) Smolin and Thoft’s
The Cornea: Scientific Foundations and Clinical Practice, 4th edn, Lippincott
Williams & Wilkins, Philadelphia, 1323 pp. + 189 plates.
This monumental treatise does not need any recommendation, as its importance in
the study of the cornea stood the test of time for generations of corneal investigators.
[A6] Ehlers, N. and Hjortdal, J. (2006) The cornea: epithelium and stroma, in
The Biology of the Eye (ed J. Fischbarg), Elsevier, Amsterdam, pp. 83–111.
A well-balanced description of essential aspects of the corneal properties and
function. Highly recommended as an introductory meaningful text for corneal
investigators.
[A7] Fischbarg, J. (2006) The corneal endothelium, in The Biology of the Eye
(ed J. Fischbarg), Elsevier, Amsterdam, pp. 113–25.
A significant contribution to the understanding of the corneal endothelium and
the mechanism of its pump-barrier function, written by the proponent of the latest
theory on the transendothelial fluid transport.
C3 The Cornea 145
[A8] Dawson, D.G., Ubels, J.L. and Edelhauser, H.F. (2011) Cornea and sclera,
in Adler’s Physiology of the Eye, 11th edn (eds L.A. Levin et al.), Elsevier,
Edinburgh, pp. 71–130.
The most recent chapter on the cornea in a traditional textbook, beautifully writ-
ten and illustrated, and covering the latest views on the structure and function of this
part of the eye.
[A9] Krachmer, J.H., Mannis, M.J. and Holland, E.J. (eds) (2011) Cornea, 3rd
edn, vol. I & vol. II, Mosby Elsevier, St. Louis, MO, 1913 pp.
A comprehensive, state-of-the art coverage of the topic throughout 171 chapters
organized in two volumes.
Cassin, B., Solomon, S.A.B. and Rubin, M.L. (1990) Dictionary of Eye Terminology, 2nd edn,
Triad Publishing Co., Gainesville, FL, 286 pp.
Stein, H.A., Slatt, B.J. and Stein, R.M. (1992) Ophthalmic Terminology. Speller and Vocabulary
Builder, 3rd edn, Mosby-Year Book Inc., St. Louis, MO, pp. 3–33, 135-55, 261-5.
Myles, W.M. (1993) Ophthalmic etymology. Surv. Ophthalmol., 37, 306–9.
References
1. Grayson, M. (1979) Diseases of the Cornea, C.V. Mosby Co., St. Louis.
2. Donaldson, D.D. (1980) Atlas of External Diseases of the Eye. Vol. III: Cornea and Sclera, 2nd
edn, C.V. Mosby Co., St. Louis.
3. Coster, D.J. (2002) Cornea, Fundamentals of Clinical Ophthalmology Series, BMJ Books,
London.
4. Holland, E.J. and Mannis, M.J. (2002) Ocular Surface Diseases. Medical and Surgical
Management, Springer-Verlag, New York.
5. Zirm, E. (1906) Eine erfolgreiche totale Keratoplastik. Graefes Arch. Ophthalmol., 64, 580-93.
6. Chirila, T.V., Hicks, C.R., Dalton, P.D., Vijayasekaran, S., Lou, X., Hong, Y., Clayton, A.B.,
Ziegelaar, B.W., Fitton, J.H., Platten, S., Crawford, G.J. and Constable, I.J. (1998) Artificial
cornea. Prog. Polym. Sci., 23, 447-73.
7. Evans, M.D.M. and Sweeney, D.F. (2010) Synthetic corneal implants, in Biomaterials and
Regenerative Medicine in Ophthalmology (ed T. Chirila), Woodhead Publishing Ltd, Oxford &
CRC, Boca Raton, pp. 65-133.
8. Proulx, S., Guillemette, M., Carrier, P., Auger, F.A., Germain, L., Giasson, C.J., Gaudreault,
M. and Guérin, S.L. (2010) Tissue engineering of human cornea, in Biomaterials and
Regenerative Medicine in Ophthalmology (ed T. Chirila), Woodhead Publishing Ltd, Oxford &
CRC, Boca Raton, pp. 150-192.
9. Martin, D.K. and Holden, B.A. (1982) A new method for measuring the diameter of the in vivo
human cornea. Am. J. Optom. Physiol. Opt., 59, 436-41.
10. Mashige, K.P. (2013) A review of corneal diameter, curvature and thickness values and influ-
encing factors. S. Afr. Optom., 72, 185-94.
11. Rüfer, F., Schröder, A. and Erb, C. (2005) White-to-white corneal diameter. Normal values in
healthy humans obtained with the Orbscan II topography system. Cornea, 24, 259-61.
12. Khng, C. and Osher, R.H. (2008) Evaluation of the relationship between corneal diameter and
lens diameter. J. Cataract Refract. Surg., 34, 475-9.
146 T.V. Chirila and S. Suzuki
37. Patel, S., Alió, J.L. and Pérez-Santonja, J. (2004) Refractive index change in bovine and
human corneal stroma before and after LASIK: a study of untreated and re-treated corneas
implicating stromal hydration. Invest. Ophthalmol. Vis. Sci., 45, 3523-30.
38. Lin, R.C., Shure, M.A., Rollins, A.M., Izatt, J.A. and Huang, D. (2004) Group index of the
human cornea at 1.3-μm wavelength obtained in vitro by optical coherence domain refractom-
etry. Opt. Lett., 29, 83-5.
39. Vasudevan, B., Simpson, T.L. and Sivak, J.G. (2008) Regional variation in the refractive-index
of the bovine and human cornea. Optom. Vis. Sci., 85, 977-81.
40. Patel, S., Alió, J.L., Amparo, F. and Rodriguez-Prats, J.L. (2011) The influence of age on the
refractive index of the human corneal stroma resected using a mechanical microkeratome.
Cornea, 30, 1353-7.
41. Amparo, F., Patel, S., Alió, J.L., Rodriguez-Prats, J.L. and Moreno, L.J. (2012) Relationship
between patient age and refractive index of the corneal stroma during refractive surgery
assisted by femtosecond laser flap creation. Cornea, 31, 751-5.
42. Boettner, E.A. and Wolter, J.R. (1962) Transmission of the ocular media. Invest. Ophthalmol.,
1, 776-83.
43. Lerman, S. (1984) Biophysical aspects of corneal and lenticular transparency. Curr. Eye Res.,
3, 3-14.
44. Beems, E.M. and van Best, J.A. (1990) Light transmission of the cornea in whole human eyes.
Exp. Eye Res., 50, 393-5.
45. Kolozsvári, L., Nógrádi, A., Hopp, B. and Bor, Z. (2002) UV absorbance of the human cornea
in the 240- to 400-nm range. Invest. Ophthalmol. Vis. Sci., 43, 2165-8.
46. Doutch, J.J., Quantock, A.J., Joyce, N.C. and Meek, K.M. (2012) Unltraviolet light transmis-
sion through the human corneal stroma is reduced in the periphery. Biophys. J., 102, 1258-64.
47. Gabriel, C., Sheppard, R.J. and Grant, E.H. (1983) Dielectric properties of ocular tissue at
37 °C. Phys. Med. Biol., 28, 43-9.
48. Gabriel, C. (1996) Compilation of the dielectric properties of body tissues at RF and micro-
wave frequencies, Final Technical Report AL/OE-TR-1996-0004, Prepared for Brooks
Air Force Base, Texas, www.dtic.mil/cgi-bin/GetTRDoc?AD = ADA303903 (accessed
02/27/2014), 16 pp.
49. Gabriel, S., Lau, R.W. and Gabriel, C. (1996) The dielectric properties of biological tissues:
III. Parametric models for the dielectric spectrum of tissues. Phys. Med. Biol., 41, 2271-93.
50. Gabriel, C. and Gabriel, S. (1996) Compilation of the dielectric properties of body tissues at
RF and microwave frequencies, Final Report AL/OE-TR-1996-0037, Prepared for AFOSR/NL
Bolling AFB DC 20332-0001, http://niremf.ifac.cnr.it/docs/DIELECTRIC/Report.html
(accessed 03/03/2014), 10 pp.
51. Oksala, A. and Lehtinen, A. (1959) Comparative studies on the electrical conductivity of aque-
ous humour, vitreous body, cornea and sclera. Acta Ophthalmol., 37, 388-94.
52. Wigham, C. and Hodson, S. (1981) Bicarbonate and the trans-endothelial short circuit current
of human cornea. Curr. Eye Res., 1, 285-90.
53. Hodson, S., Wigham, C., Williams, L., Mayes, K.R. and Graham, M.V. (1981) Observations
on the human cornea in vitro. Exp. Eye Res., 32, 353-60.
54. Hodson, S. and Wigham, C. (1983) The permeability of rabbit and human corneal endothe-
lium. J. Physiol., 342, 409-19.
55. Andreassen, T.T., Simonsen, A.H. and Oxlund, H. (1980) Biomechanical properties of kerato-
conus and normal corneas. Exp. Eye Res., 31, 435-41.
56. Hoeltzel, D.A., Altman, P., Buzard, K. and Choe, K. (1992) Strip extensiometry for compari-
son of the mechanical response of bovine, rabbit, and human corneas. J. Biomech. Eng., 114,
202-15.
57. Seiler, T., Matallana, M., Sendler, S. and Bende, T. (1992) Does Bowman’s layer determine the
biomechanical properties of the cornea? Refr. Corneal Surg., 8, 139-42.
58. Zeng, Y., Yang, J., Huang, K., Lee, Z. and Lee, X. (2001) A comparison of biomechanical
properties between human and porcine cornea. J. Biomech., 34, 533-7.
148 T.V. Chirila and S. Suzuki
59. Wollensak, G., Spoerl, E. and Seiler, T. (2003) Stress-strain measurements of human and por-
cine corneas after riboflavin–ultraviolet-A-induced cross-linking. J. Cataract Refract. Surg.,
29, 1780-5.
60. Woo, S.L.-Y., Kobayashi, A.S., Schlegel, W.A. and Lawrence, C. (1972) Nonlinear material
properties of intact cornea and sclera. Exp. Eye Res., 14, 29-39.
61. Kobayashi, A.S., Staberg, L.G. and Schlegel, W.A. (1973) Viscoelastic properties of human
cornea. Exp. Mech., 13, 497-503.
62. Jue, B. and Maurice, D.M. (1986) The mechanical properties of the rabbit and human cornea.
J. Biomech., 19, 847-53.
63. Hjordtal, J.Ø. (1996) Regional elastic performance of the human cornea. J. Biomech., 29,
931-42.
64. Bryant, M.R. and McDonnell, P.J. (1996) Constitutive laws for biomechanical modeling of
refractive surgery. J. Biomech. Eng., 118, 473-81.
65. Danielsen, C.C. (2004) Tensile mechanical and creep properties of Descemet’s membrane and
lens capsule. Exp. Eye Res., 79, 343-50.
66. Elsheikh, A., Wang, D., Brown, M., Rama, P., Campanelli, M. and Pye, D. (2007) Assessment
of corneal biomechanical properties and their variation with age. Curr. Eye Res., 32, 11-9.
67. Schwartz, N.J., Mackay, R.S. and Sackman, J.L. (1966) A theoretical and experimental study
of the mechanical behavior of the cornea with application to the measurement of intraocular
pressure. Bull. Math. Biophys., 28, 585-643.
68. Wang, H., Prendiville, P.L., McDonnell, P.J. and Chang, W.V. (1996) An ultrasonic technique
for the measurement of the elastic moduli of human cornea. J. Biomech., 29, 1633-6.
69. Jayasuriya, A.C., Ghosh, S., Scheinbeim, J.I., Lubkin, V., Bennett, G. and Kramer, P. (2003).
A study of piezoelectric and mechanical anisotropies of the human cornea. Biosens.
Bioelectron., 18, 381-7.
70. Sjøntoft, E. and Edmund, C. (1987) In vivo determination of Young’s modulus for the human
cornea. Bull. Math. Biol., 49, 217-32.
71. Edmund, C. (1988) Corneal elasticity and ocular rigidity in normal and keratoconic eyes. Acta
Ophthalmol., 66, 134-40.
72. Glass, D.H. (2008) Characterization of the biomechanical properties of the in vivo human
cornea, Ph.D. Thesis, Ohio State University, https://etd.ohiolink.edu/ap/10?0::NO:10:P10_
ACCESSION_NUM:osu1211920104 (accessed 02/27/2014), p. 60.
73. Last, J.A., Liliensiek, S.J., Nealey, P.F. and Murphy, C.J. (2009) Determining the mechanical
properties of human corneal basement membranes with Atomic Force Microscopy. J. Struct.
Biol., 167, 19-24.
74. Last, J.A., Thomasy, S.M., Croasdale, C.R., Russell, P. and Murphy, C.J. (2012) Compliance
profile of the human cornea as measured by atomic force microscopy. Micron, 43, 1293-8.
75. Lombardo, M., Lombardo, G., Carbone, G., De Santo, M.P., Barberi, R. and Serrao, S. (2012)
Biomechanics of the anterior human corneal tissue investigated with atomic force microscopy.
Invest. Ophthalmol. Vis. Sci., 53, 1050-7.
76. Dias, J., Diakonis, V.F., Kankariya, V.P., Yoo, S.H. and Zierbath, N.M. (2013) Anterior and
posterior corneal stroma elasticity after corneal collagen crosslinking treatment. Exp. Eye Res.,
116, 58-62.
Part II
Chapter 1a
Metallic Biomaterials: Introduction
1a.1 Introduction
Metallic biomaterials are one of the most commonly used biomaterial groups along
with ceramics, synthetic polymers, and naturally derived products. The utility of
these metallic materials is based largely on the formation of a thin but protective
oxide layer. The oxide layer forms upon exposure to oxygen and re-forms within
milliseconds after damage [1]. This layer reduces corrosion in vivo, one of the
major requirements of a robust biomaterial. The other requirements include biotol-
erability, bioadhesion, biofunctionality (bioactivity), and processability. In practice
each metal or alloy, i.e., titanium alloys, stainless steel, and Co-Cr alloys, has their
own advantages for different applications based on mechanical, chemical, and bio-
functional properties. Generally, metallic biomaterials are used for structural appli-
cations such as implants, pins, and bone scaffolding due to their excellent mechanical
properties such as Young’s modulus, tensile strength, ductility, fatigue, and wear
resistance. However, they can be used for unloaded, purely functional devices such
as cages for pumps, valves, and heart pacemakers. The first generation of metallic
biomaterials was designed for minimal toxicity. The second generation has been
designed for functionality at both at the mechanical and molecular level to enhance
integration of the material into the biological environment and increase longevity of
the implant. The third generation has focused not only on functionality but also on
regeneration of the surrounding tissue in conjunction with the bioactive material [2].
H. Breme • V. Biehl
Lehrstuhl für Metallische Werkstoffe, Universität des Saarlandes, Saarbrücken, Germany
N. Reger • E. Gawalt (*)
Department of Chemistry and Biochemistry, Bayer School of Natural and Environmental
Sciences, Mellon 337, 600 Forbes Avenue, Pittsburgh, PA 15282, USA
e-mail: [email protected]
Most metallic biomaterials in use today are alloys with Ti, Fe, Co, Cr, Ni, V, Al, and
Ta as common components. Each alloy has different mechanical properties detailed
herein making them preferred materials for different applications. Stainless steel
316 L (SS316L), Co-Cr, Ti6Al4V, and Nitinol are among the most common [3].
SS316L is a strong material that is used for small fixation devices such as pins,
screws, and plates. It was originally used for prosthetics such as hip and knee
replacements but the market size for SS316L prosthetics has been consistently
shrinking due to the prevalent use of lighter alloys. The material is composed of
66 % Fe, 19 % Cr, 9 % Ni, 3 % Mo, and 2 % Mn with Si and C. The oxide surface
of SS316L is predominately composed of Fe2O3 and Cr2O3 [4]. This surface is sus-
ceptible to pitting and corrosion, which is the largest cause of failure [5]. Subsequent
to the corrosion, Ni ions may be released into the system. If the Ni ions collect in
organs, toxicity becomes an issue [6]. Therefore, alloys where Ni is replaced with
nitrogen are being introduced into the arena [7].
Cobalt-chromium (Co-Cr) alloyed with W, Mo, or other metals are used for both
weight-bearing and non-weight-bearing applications. These lightweight materials
are common in hip and knee prosthetics. Wear and fatigue are an issue with Co-Cr
alloys. Therefore, metal-on-metal applications are limited and Co-Cr is used in
metal-polymer applications or portions of an implant. This class of alloys is also
common in arterial stent applications [8]. As a result of the wear issues, Cr ion tox-
icity questions have led to a reduction in usage.
Titanium and its alloys are the most common metallic biomaterials. Ti6Al4V is
used in prosthetic devices including knees, hips, and vertebrae. The material is light
with a favorable Young’s modulus [9]. The oxide surface consists mainly of TiO2
and Al2O3 in approximately the same percentage as the bulk. TiO2 is an extremely
stable and inert oxide under physiological and electrochemical conditions coupled
with a high coefficient of friction making the material an ideal prosthetic and for use
in metal-on-metal applications. Even though there are no detectable amounts of V
present on the surface, concerns with V toxicity have been addressed and Ta or Nb
has been used as replacements. Nitinol is a shape memory alloy that is used in stent
and orthodontic applications. The surface of this 50–50 alloy is generally consid-
ered to be TiO2. In these applications, the inert oxide surface is important because
integration into the surrounding tissue in its applications is not critical and in some
cases undesirable. Due to these benefits, Ti alloys dominate the metallic biomateri-
als market [3].
1a Metallic Biomaterials: Introduction 153
The primary corrosion products of metallic implants are mainly responsible for the
biotolerability of the implanted alloy and their sustained interactions. The outer
oxide layer of the metals and alloys provides protection from corrosion [5, 6].
However, the oxide layer does associate with hydroxyl groups and/or bound water
that can facilitate corrosion either through electrochemical reactions or through
reactions with biological fluids (in vivo or simulated). This can result in pitting or
crevice corrosion and changes in surface composition, which can ultimately lead to
metal ion release. This type of corrosion is particularly important in biological sys-
tems where the ions can be transported to various organs and due to an enrichment
of the metal, toxicity may ensue.
The primary corrosion products of the most abundant elements in metallic implant
materials vary due to their thermodynamic stability. While the oxides or hydroxides
of Al, Cr, Nb, Ta, Ti, and V are stable due to a more negative heat of formation than
that of water, the oxides and hydroxides of Co and Ni are unstable because of a less
negative heat of formation than that of water [10]. The interaction between the oxide
or hydroxide and body fluid is increased if the heat of formation for the oxide or
hydroxide is increased. Therefore, the thermodynamically stable corrosion products
have a low solution product and a low solubility in the body fluid. This is directly
demonstrated by the pKa values (negative logarithm) of the solution product of the
primary corrosion products [11]. While Ti-, Ta-, Nb-, and Cr-oxides have pKa values
of >14, i.e., hydrolysis cannot play a role, Co-, Fe-, and Ni-oxides possess negative
pKa values which cause considerable solubility. In spite of a high negative heat of
formation for Fe2O3 and Fe- and Cr-hydroxide negative pKa values and a high solubil-
ity are reported [11]. A remarkable solubility of Cr in serum was observed [12], while
titanium is practically insoluble due to the formation of the thermodynamically stable
oxide TiO2. It is this stability that favors Ti as a biomaterial over other materials.
154 H. Breme et al.
Further, the oxide layers of Ti, Ta, and Nb are semi- or nonconductive oxides.
These oxides are able to prevent to a great extent an exchange of electrons and
therefore a flow of ions through the tissue [13] due to their isolating effect. This
isolating effect may be demonstrated by the dielectric constants of the different
metal oxides. While TiO2 (rutile), Fe2O3, and Nb2O5 have constants even higher than
that of water, A12O3, Cr2O3, and Ta2Os have a lower isolating effect and a higher
conductivity. For Ni- and V-oxides, dielectric constants are not available because of
their high conductivity [14]. As a result of these electrochemical properties, it is
clear why Ti alloys are much less susceptible to pitting and crevice corrosion than
SS316L.
1a.5 Biofunctionality
The new generation of metallic biomaterials must be biofunctional not just biotoler-
able. A biofunctional material will integrate with the existing surrounding tissue or
induce tissue growth de novo. This begins with cell adhesion, spreading, prolifera-
tion, and eventually tissue growth. However, the surface oxide layers are inert
towards the surrounding tissue, protein- and environment-reducing integration, and
potential functionality [15]. Therefore, the chemical bonding of a metallic implant
with the tissue, which is observed between bioactive ceramics like hydroxyapatite
and bone, seems to be limited and improbable, and the adhesion strength between
the bone and the metal will have a primarily mechanical character. While metallic
biomaterials, in general, do not induce de novo tissue formation they support adhe-
sion and spreading of many cell types including fibroblast, endothelial, and osteo-
blast cells. Cell adhesion and growth have been studied in many systems with
varying surface properties on the molecular, nano, micro, or macroscale.
Osseointegration of orthopedic alloys has been studied under many conditions,
surface pretreatments, and modifications. For example, the slow formation of cal-
cium phosphate mineral, the mineral component of bone, occurs on TiO2, Nitinol,
and SS316L both in vivo and in Hanks’ buffered saline solution without surface
modification [1]. In another study, the bone ingrowth behavior of miniplates of cp-Ti
and SS316L was investigated by the implantation of these plates on the legs of
Hanford minipigs. The miniplates were fixed to the legs of the pigs by screws. After
an 8-week exposure time, in the animals where Ti plates had been used, a new forma-
tion of bone could be observed in close contact to the surface of the plates. In contrast
to this result, when SS316L was used, there was less new bone formation and unde-
sired granulated tissue was found between the metallic surface and surrounding
bone. This granulated tissue at the interface bone/implant has the disadvantage that
it is not supplied with blood and unable to transfer or sustain forces, so that an even-
tual loosening of the implant will take place [16]. Successful growth of bone in close
contact with metal and metal alloys has also been reported in similar investigations
in which the contact area tissue/implant has been studied in detail [3, 5–7, 17–23].
1a Metallic Biomaterials: Introduction 155
strength was not increased. A Ti5A12.5Fe implant coated with hydroxyapatite had
a maximum adhesion strength of 1.97 N/mm2 [33].
The shape of non-weight-bearing biomaterials is also important. For example,
those consisting of nanostructures of Al and Ti have shown to enhance cell adhe-
sion and proliferation. The metals were formed into nanofibers and nanospheres.
Both nanoshapes provided a platform for Ca deposition. This step is critical to
long-term osseointegration of the implant material into the body. However, the
nanofibers provided a superior platform for bone formation when compared to the
nanospheres [9, 36, 37].
Traditionally, dental implants have outpaced the orthopedic implants in terms of
longevity, biointegration, and functionality lasting up to 25 years. These materials
include metals and ceramics. For example, implants made from different materials
inserted into the jaws of dogs varied in their behavior. With Ti implants the bone
grew in close contact to the metal surface. In contrast, when SS316L was used as the
implant material, a fibrous encapsulation, which separated the implant from the sur-
rounding tissue, was formed. In Ti implants instead of this fibrous encapsulation, an
intercellular substance appeared. A similar unfavorable behavior was found in the
case of dental implants of a Co-Cr alloy in dogs [38]. As in orthopedic research,
recent dental studies have emphasized the surface properties of the materials. Due
to the oxide nature of the implant surfaces, there are hydroxyl and μ-oxo groups on
the surface. The presence of these hydrophilic groups increases mesenchymal stem
cell adhesion and differentiation leading to enhanced integration. Combining micro-
roughness with these hydrophilic groups provided the best environment for osseo-
integration in vivo [25].
References
1. Minagar S, Wang J, Berndt CC, Ivanova EP, Wen C (2013) Cell response of anodized nano-
tubes on titanium and titanium alloys. J Biomed Mater Res A 101A(9):2726–2739
2. Mantripragada VP, Lecka-Czernik B, Ebraheim NA, Jayasuriya AC (2013) An overview of
recent advances in designing orthopedic and craniofacial implants. J Biomed Mater Res A
101A(11):3349–3364
3. Niinomi M, Nakai M, Hieda J (2012) Development of new metallic alloys for biomedical
applications. Acta Biomater 8:3888–3903
4. Raman A, Dubey M, Gouzman I, Gawalt ES (2006) Formation of Self-assembled monolayers
of alkylphosphonic acid on the native oxide surface of SS316L. Langmuir 22:6469–6472
5. Antunes RA, Lopes de Oliveria MC (2012) Corrosion fatigue of biomedical metallic alloys:
mechanisms and mitigation. Acta Biomater 8:937–962
6. Singh R, Dahotre NB (2007) Corrosion degradation and prevention by surface modification of
biometallic materials. J Mater Sci Mater Med 18:725–751
7. Talha M, Behera CK, Sinha OP (2013) A review on nickel-free nitrogen containing austenitic
stainless steels for biomedical applications. Mater Sci Eng C 33:3563–3575
8. Zhang K, Liu T, Li J, Chen J, Wang J, Huang N (2014) Surface modification of implanted
cardiovascular metal stents: from antithrombosis and antirestenosis to endothelialization.
J Biomed Mater Res A 102A(2):588–609
9. Guduru D, Niepel M, Vogel J, Groth T (2011) Nanostructured material surfaces—preparation,
effect on cellular behavior, and potential biomedical applications: a review. J Artif Organs
34(10):963–985
10. Kubashewski O, Evans EC, Alcock CB (1967) Metallurgical thermochemistry. Pergamon
Press, London
11. Steinemann SG, Perren SM (1984) Titanium alloys as metallic biomaterials. In: Proc. of the
5th world conf. on titanium, vol 2, pp. 1327–1334
12. Zitter H (1976) Schädigung des Gewebes durch metallische Implantate. Unfallheilkunde
79:91
13. Zitter K, Plenk H, Strassl H (1980) Tissue and cell reactions in vivo and in vitro to different
metals for dental implant. In: Heimke G (ed) Dental implants. C. Hanser, Miinchen, p 15
14. Ferguson AB, Akahashi Y, Laing PG, Hodge ES (1962) Characteristics of trace ions released
from embedded metal implants in the rabbit. J Bone Joint Surg 44:323–336
15. Pound BG (2014) Passive films on metallic biomaterials under stimulated physiological condi-
tions. J Biomed Mater Res A 102A:1595–1604
16. Breme J, Steinhäuser E, Paulus G (1988) Commercially pure titanium Steinhäuser plate–screw
system for maxillo facial surgery. Biomaterials 9:310–313
17. Lyndon JA, Boyd BJ, Birbilis N (2014) Metallic implant drug/device combinations for con-
trolled drug release in orthopaedic applications. J Control Release 179:63–75
18. Krekeler G, Schilli W (1984) Das ITI-Implantat Typ H: Technische Entwicklung,
Tierexperiment und klinische Erfahrung. Chirurgische Zahnheilkunde 12:2253–2263
19. Kirsch A (1980) Titan-spritzbeschichtetes Zahnwurzel-implantat unter physiologischer
Belastung beim Menschen. Dt ZahnärztlZ 35:112–114
158 H. Breme et al.
20. Schröder A, van der Zypen E, Sutter F (1981) The reaction of bone, connective tissue and
epithelium to endosteal implants with titanium- sprayed surface. J Maxillofac Surg 1981(9):15
21. Brånemark PI, Adell R, Albrektsson T, Lekholm U, Lundkvist S, Rockier B (1983)
Osseointegrated titanium fixtures in the treatment of edentulousness. Biomaterials 4:25
22. Schröder A, Stich H, Strautmann F, Sutter F (1978) Über die Anlagerung von Osteozement an
einem belasteten Implantatkörper. Schweiz Monatsschr Zahnheilk 4:1051–1058
23. Kydd WL, Daly CH (1976) Bone–titanium implant response to mechanical stress. J Prosthet
Dent 35:567–571
24. Golish SR, Anderson PA (2012) Bearing surfaces for total disc arthroplasty: metal-on-metal
versus metal-on-polyethylene and other biomaterials. Spine J 12:693–701
25. Bose S, Tarafder S (2012) Calcium phosphate ceramic systems in growth factor and drug
delivery for bone tissues engineering: a review. Acta Biomater 8:1401–1421
26. Gittens RA, Scheideler L, Rupp F, Hyzy SL, Geis-Gerstorfer J, Schwartz Z, Boyan BD (2014)
A review on the wettability of dental implant surfaces: II. Biological and clinical aspects. Acta
Biomater 10:2907–2918
27. Purnama A, Hermawan H, Couet J, Mantovani D (2010) Assessing the biocompatibility of
degradable metallic materials: state-of-the-art and focus on the potential of genetic regulation.
Acta Biomater 6:1800–1807
28. Richards RG, Moriarty TF, Miclau T, McClellan RT, Grainger DW (2012) Advances in bioma-
terials and surface technologies. J Orthop Trauma 26(12):703–707
29. Gittens RA, Olivares-Navarrete R, Schwartz Z, Boyan BD (2014) Implant osseointegration
and the role of microroughness and nanostructures: lessons for spine implants. Acta Biomater
10:3363–3371
30. Bjursten LM, Rasmusson L, Oh S, Smith GC, Brammer KS, Jin S (2010) Titanium dioxide
nanotubes enhance bone bonding in vivo. J Biomed Mater Res A 92:1218–1224
31. Tsukimura N, Ueno T, Iwasa F, Minamikawa H, Sugita Y, Ishizaki K et al (2011) Bone integra-
tion capability of alkali- and heat-treated nanobimorphic Ti-15Mo-5Zr-3Al. Acta Biomater
7:4267–4277
32. Thakral GK, Thakral R, Sharma N, Seth J, Vashisht P (2014) Nanosurface—the future of
implants. J Clin Diagn Res 8(5):7–10
33. Schmitz HJ, Gross V, Kinne R, Fuhrmann G, Strunz V (1988) Der Einfluβ unterschiedlicher
Oberflächenstrukturierung plastischer Implantate auf das histologische Zugfestigkeitsverhalten
des Interface. DVM-Vortragsreihe Implantate 7
34. Hayes JS, Richards RG (2010) The use of titanium and stainless steel in fracture fixation.
Expert Rev Med Dev 7:843–853
35. Hayes JS, Richards RG (2010) Surfaces to control tissue adhesion for osteosynthesis with
metal implants: in vitro and in vivo studies to bring solutions to the patient. Expert Rev Med
Dev 7:131–142
36. Webster TJ, Ergun C, Doremus RH, Siegel RW, Bizios R (2000) Enhanced functions of osteo-
blasts on nanophase ceramics. Biomaterials 21(17):1803–1810
37. Webster TJ, Gutwein LG, Tepper F (2008) Increased osteoblast function on nanofibered alu-
mina. In: 26th annual conference on composites, advanced ceramics, materials, and structures
B: ceramic engineering and science proceedings. John Wiley & Sons, Inc., pp. 817–824
38. Strassl H (1978) Experimentelle Studie über das Verhalten von titanbeschichteten Werkstoffen
hinsichtlich der Gewebekompatibilität im Vergleich zu anderen Metallimplanteten. Teil 1.
Osterr Z Stomatol 75(4):134–146
Chapter 1b
Metallic Biomaterials: Cobalt-Chromium
Alloys
Gopinath Mani
1b.1 Introduction
Cobalt-chromium (CoCr) alloys have been extensively used for making various
orthopedic, dental, and cardiovascular implants and devices. These alloys possess
superior mechanical properties with high resistance to corrosion, wear, and fatigue.
The biocompatibility and blood compatibility of these alloys have also been well
demonstrated. In this chapter, the four types of CoCr alloys (ASTM F75 alloy,
ASTM F799 alloy, ASTM F90 alloy, and ASTM F562 alloy) that are currently used
for biomedical implants and devices are discussed. Also, the other types of CoCr
alloy (ASTM F563, ASTM F1058, and Havar) that have been considered and inves-
tigated for biomedical implants and devices are discussed in this chapter. The chap-
ter is divided into subsections based on these individual alloys.
G. Mani (*)
Department of Biomedical Engineering, University of South Dakota,
4800 North Career Avenue, Sioux Falls, SD 57107, USA
e-mail: [email protected]
This alloy is made into the required shape of the final device by a process known
as investment casting [1–4]. During this process, initially, a wax mold is made rep-
licating the shape of the final device. Then, the mold is coated with a ceramic fol-
lowed by melting the wax to leave a ceramic mold. After that, the alloy is melted at
a temperature range of 1350–1450 °C, and poured into the ceramic mold. Once the
metal is solidified, the ceramic mold is broken away, and the metal is further pro-
cessed to obtain the final device.
The way how the casting conditions are carried out can produce different micro-
structural features and can adversely affect the mechanical properties of the alloy
[1–4]. Firstly, as-cast F75 alloy typically consists of a Co-rich matrix (alpha-phase)
with carbides (Co23C6, Cr23C6, and Mo23C6) in interdendritic spaces and grain
boundaries. The relative amounts of alpha and carbide phases in the alloy are
expected to be 85 % and 15 %, respectively. However, if there are any non-
equilibrium conditions happened during cooling, the interdendritic regions become
rich in solutes such as Cr, Mo, and C, along with carbides, and the dendrites will be
rich in Co with lack of Cr. This situation is considered to be unfavorable since the
regions with lack of Cr can act as an anode with respective to the rest of the alloy
microstructure. However, this situation can be improved by solution annealing at
1225 °C for 1 h following casting. Secondly, the solidification step in the casting
process not only creates dendrites but also large-size grains. This is not favorable
since large-size grains can significantly decrease the yield strength of the alloy.
Thirdly, defects may occur in the alloy during casting. For instance, if a ceramic
mold piece is broken and entrapped in the alloy during its solidification, it could
lead to stress concentrations and initiate cracks at the site where the ceramic piece
is present. Such defects can cause fatigue fracture of the device under in vivo condi-
tions. For the similar reasons, it is vital to avoid any porosity (macro- or micro-
pores) in the alloy during casting conditions.
To circumvent the above-mentioned limitations and to improve the microstruc-
tural features and mechanical properties of the alloy, techniques such as hot isotac-
tic pressing (HIP) have been developed [1–4]. In HIP, a fine powder of F75 alloy is
compacted and sintered together under a pressure of 100 MPa at 1100 °C for 1 h,
and then forged to obtain the final shape. The alloy prepared using HIP typically
showed a much smaller grain size (approximately 8 μm), which not only provided
higher yield strength but also improved fatigue properties when compared to those
of as-cast alloy. The mechanical properties of as-cast and HIP F75 alloys are pro-
vided in Table 1b.2. F75 alloy is commonly used to make total hip arthroplasty
components (femoral stems, modular head, and acetabular cup liner) and total knee
arthroplasty components (metal femoral and tibial components). Also, this alloy is
commonly used to make dental implants such as partial dentures and bridgeworks.
1b Metallic Biomaterials: Cobalt-Chromium Alloys 161
Haynes 25 (L605). In this alloy, tungsten (W) and nickel (Ni) are added to improve
the machinability and fabrication properties. The yield and tensile strengths of F90
alloy in annealed state are equivalent to those of F75 alloys. However, when the F90
alloy is cold worked to 44 %, the yield and tensile strengths significantly increased
when compared to those of F75 alloys. The mechanical properties of F90 alloys in
annealed and cold worked states are provided in Table 1b.6. F90 alloy is commonly
used to make cardiovascular implants such as stents and mechanical heart valves.
Also, this alloy is used in orthopedic applications such as cerclage cables and
guide rods.
F562 alloy, the processing involves cold working to 50 %, which greatly supports
the transformation of FCC to HCC crystal structure. The HCP phase exists as plate-
lets within the FCC grains to significantly strengthen the alloy. The alloy can also
be further strengthened by a heat treatment of 430–650 °C. This heat treatment
produces cobalt molybdenum (Co3Mo) precipitates on the HCP platelets. Hence,
this alloy is truly a multiphase material. The higher strength of the alloy is mainly
due to the combination of a variety of processing methods such as cold working,
solid solution strengthening, and precipitation hardening. The mechanical proper-
ties of F562 alloy in hot-forged, and cold-worked and aged forms are provided in
Table 1b.8. As can be seen from this table, the tensile strength of cold-worked and
aged F562 alloy is greater than that of all the other types (F75, F799, and F90) of
CoCr alloy that are used for biomedical applications. F562 alloy is used for cardio-
vascular stents, lead wires for pacemakers, and heart valve components.
Although the above-discussed four types (F75, F799, F90, and F562) of CoCr alloy
are currently used for making various biomedical implants and devices, the few
other types of CoCr alloy such as ASTM F563, ASTM F1058, and UNS R30004
have also been investigated for biomedical applications. Each of these alloys (F563,
F1058, and UNS R30004) has been shown to produce no harmful effects such as
cytotoxicity, systemic toxicity, hemolysis, and pyrogenicity in intracutaneous irrita-
tion, intramuscular implantation, and skin sensitization tests [1].
164 G. Mani
References
1. Davis JR (2003) Handbook of materials for medical devices. ASM International, Materials
Park
2. Park JB, Kim YK (2003) Metallic biomaterials. In: Park JB, Bronzino JD (eds) Biomaterials
principles and applications. CRC Press, Boca Raton, pp 1–20
3. Brunski JB (2004) Metals. In: Ratner BD, Hoffman AS, Schoen FJ, Lemons JE (eds)
Biomaterials science: an introduction to materials in medicine. Elsevier Academic Press,
London, pp 137–153
4. Agrawal CM, Ong JL, Appleford MR, Mani G (2014) Introduction to biomaterials basic the-
ory with engineering applications. Cambridge University Press, New York
5. Breme J, Biehl V (1998) Metallic biomaterials. In: Black J, Hastings G (eds) Handbook of
biomaterial properties. Springer, New York, pp 135–213
6. Klarstrom D, Crook P (2001) Cobalt alloys: alloy designation system. In: Buschow KHJ, Cahn
R, Flemings M, Ilschner B, Kramer E, Mahajan S, Veyssiere P (eds) Encyclopedia of materi-
als: science and technology. Pergamon, Headington Hill Hall, pp 1279–1280
7. Chen Q, Thou G (2015) Biomaterials: a basic introduction. CRC Press, Boca Raton
8. Narushima T, Ueda K, Alfirano (2015) Co-Cr alloys as effective metallic biomaterials. In:
Niinomi M, Narushima T, Nakai M (eds) Advances in metallic biomaterials tissues, materials
and biological reactions. Springer, New York, pp 157–178
9. Robinson M (2005) Havar®, a Co-Cr biocompatible alloy for medical implants, encyclopedia
of materials: science and technology. Elsevier, pp 1–6
10. Robinson M (2004) Havar®, a new, old Co-Cr biocompatible alloy for implants. In: Shrivastava
S (ed) Medical device materials proceedings of the materials & processes for medical devices
conferences. ASM International, Materials Park, pp 324–328
Chapter 1c
Metallic Biomaterials: Titanium
and Titanium Alloys
1c.1 Composition
Table 1c.1 Comparison of international standards for titanium and titanium alloys (Refs. 1, 2a)
International
United Organization for United
Germany Kingdom Standardization States Japan
Alloy DIN BS France AIR ISO ASTM JIS
Ti-1 17850 2TA1 9182 T-35 5832/II F67/Grade 1 H4600
Ti-2 17850 2TA2 9182 T-35 F67/Grade 2
Ti-3 17850 2TA6 9182 T-50 F67/Grade 3
Ti-4 17850 2TA6, 2AT7 9182 T-60 F67/Grade 4
2AT8, 2AT9
Ti6Al4V 17851 2TA10, 9183T-A6V 5832/III F136-13 H4607
2TA11
2TA13,
2TA28
TA56, TA59
Ti5Al2.5Fe 17865 BS7252–10 – 5832/X – –
Ti6Al7Nb 17851 S94–081 5832/XI Fl295 –
a
Ti-l–Ti-4 = commercially pure unalloyed titanium; others: direct chemical composition
H. Breme • V. Biehl
Lehrstuhl für Metallische Werkstoffe, Universität des Saarlandes, Saarbrücken, Germany
N. Reger • E. Gawalt (*)
Department of Chemistry and Biochemistry, Bayer School of Natural and Environmental
Sciences, Mellon 337, 600 Forbes Ave., Pittsburgh, PA 15282, USA
e-mail: [email protected]
Table 1c.2 Chemical composition of commercially pure (cp)-titanium (wt%) (Refs. 1–3)
cp-Ti Fe max. O approx. N max. C max. H max. Ti
Grade 1 0.2 0.18 0.03 0.1 0.015 Balance
Grade 2 0.3 0.25 0.03 0.08 0.015 Balance
Grade 3 0.3 0.35 0.05 0.1 0.015 Balance
Grade 4 0.5 0.40 0.05 0.1 0.015 Balance
Table 1c.4 Chemical composition (wt%) of β and near-β titanium alloys (Refs. 4–7)
Alloy Al Mo Zr Ta Nb Sn Ti
Til5Mo5Zr3Al 3.8 15 5 77
Til2Mo5Zr5Sn 12 5 4.5 Balance
Ti30Nb 30 70
Ti30Ta 29.6 Balance
Table 1c.7 Influence of alloying elements and heat treatment on Young’s modulus (Refs. 5,
15–18)
Alloy Heat treatment Young’s modulus (GPa)
Ti30Ta As rolled 69
1 h/1000 °C/H2O 63
Ti30Nb As rolled 80
Til5Mo5Zr3Al ELIa Solution heat treated at 840 °C 75
Solution heat treated at 740 °C 88
Solution heat treated at 740 °C 113
+ aged at 600 °C
a
ELI = extra-low interstitial
Provided that the following characteristics of titanium are taken into consideration,
almost all processing procedures are possible:
1. High affinity to oxygen, nitrogen, and hydrogen gases
2. High reactivity to all metals to produce intermetallic compounds
3. Relatively low Young’s modulus and therefore backspringing
4. Relatively low thermal conductivity
5. Tendency to stick to tools
Titanium and titanium alloys are fabricated into semifinished products by conven-
tional methods such as forging, rolling, pressing, and drawing. When Ti materials
are heated, care must be taken to avoid an excessive adsorption of oxygen, nitrogen,
and hydrogen. Therefore, heating and annealing should take place in a neutral or
170 H. Breme et al.
At room temperature cp-Ti grades 1 and 2 can be worked very well, and grades 3
and 4 only moderately well. Titanium alloys, because of their high yield/tensile
strength ratio, can be worked only under certain conditions.
Table 1c.9 Recommendation for the heat treatment of cp-Ti and Ti alloys (Refs. 2, 8)
Solution annealing + age
Alloy Stress relief Soft annealing hardening
Ti grade 1 15 min–4 h at 480–595 °C/air 6 min–2 h at −
Ti grade 2 650–750 °C/air
Ti grade 3
Ti grade 4
Ti6A14V 3 min/mm Min. 1 h 15 min–1 h at
1–4 h at 480–650 °C/air max. 4 h at 955–970 °C/H2O
700–850 °C +2 h at 480–595 °C/H2O
air or furnace cool
Ti5A12.5Fe 3 min/mm Min. 10 min 15–1 h at 800–920 °C/H2O
15 min–4 h at 540–650 °C/air max. 4 h at +2–4 h at 480–600 °C/air
720–845 °C
air or furnace cool
1c Metallic Biomaterials: Titanium and Titanium Alloys 171
For deep drawing, special coatings in the form of polymer foils have proved to
be effective. At high temperatures colloidal graphite and common hot press greases
with graphite or molybdenum disulfide additives have been successful. The Fe-,
Ni-, and Cr-contents should be limited to 0.05, 0.1, and 0.33 wt%, respectively, to
allow during a short annealing treatment (1–5 min at 750–850 °C) a grain growth
producing an average grain size of 50–70 μm. Due to this grain growth a deforma-
tion by twinning, with a resulting increased deep drawability, occurs (Ref. 19).
Superplastic forming is a material-saving and cost-reducing process for manufac-
turing parts of a complicated shape because it can be carried out together with diffu-
sion welding in a single operation. Fine-grained alloys like Ti6Al4V and Ti5Al2.5Fe
can be used for superplastic deformation. Other special deformation processes such
as stretch forming, spinning, or explosion forming are also possible.
1c.3.3 Descaling
The average thickness of the oxide layer on the surface of cp-Ti as a function of
temperature can be found in Table 1c.10 and the composition of the oxide layer on
cp-Ti and Ti alloys can be found in Table 1c.11. The oxide layer on the surface of
thick-walled pieces generated during deformation and/or heat treatment is removed
by sandblasting and/or pickling. The workpiece is treated in an aqueous solution of
20 wt% HNO3 and 2 wt% HF.
Thin-walled pieces are merely pickled in an electrolytic solution or salt bath. It
is important that not only the surface layer of oxide but also the underlying oxygen
enriched diffusion zone is removed. Otherwise, the machinability and the service
life of turning and milling tools would be negatively affected.
Table 1c.11 Nature of oxides formed on cp-Ti and titanium alloys (Ref. 22)
Oxide
Alloy TiO2 Al2O3 Nb2O5 MoO3/MoO2 ZrO2
cp-Ti ×
Ti6Al4V × ×
Ti5Al2.5Fe × ×
Ti6Al7Nb × × ×
Ti15Mo5Zr3Al × × × ×
1c.3.4 Machining
Immediately before soldering and brazing, the oxide layer present on the surface of
titanium material must be removed. For direct applications using a torch, alumi-
num–zinc and tin–zinc solders are suitable. The higher temperatures required for
brazing present the difficulty of avoiding the formation of intermetallic phases. As
with almost all metals, titanium forms brittle intermetallic phases in the fusion zone.
The only exception is silver, so that this metal forms one of the main constituents of
brazers. The sources of heat used for brazing and soldering are the acetylene torch,
high-frequency induction coils, infrared radiation, an inert-gas-shielded arc with
graphite or tungsten electrodes, furnaces with an argon atmosphere (min. 99.99 %
1c Metallic Biomaterials: Titanium and Titanium Alloys 173
Table 1c.12 Recommendations for the cutting and milling of cp-Ti and Ti alloys (Ref. 8)
Cutting speed Feed (mm/ Cutting Relief
Alloy (m/min) rev) Depth of cut (mm) angle angle
Rough cutting
cp-Ti TC 30–75 0.2–0.4 0–6° 6–8°
HSS 7.5–4.0 0.1–1.25 6–15° 5–7°
>2.5
Ti alloys TC 15–25 0.2–0.4 0–6° 6–8°
HSS 3–15 0.1–0.4 6–15° 5–7°
Forecutting
cp-Ti TC 60–100 0.1–0.4 6–6° 6–8°
HSS 18–50 0.075–0.2 6–15° 5–7°
0.75–2.5
Ti alloys TC 20–50 0.1–0.4 6–6° 6–8°
HSS 5–15 0.075–0.2 6–15° 5–7°
Finish cutting
cp-Ti TC 60–100 0.075–0.3 0–15° 6–8°
HSS 20–50 0.05–0.1 5–6° 5–7°
0.1–0.75
Ti alloys TC 20–70 0.075–0.3 0–15° 6–8°
HSS 9–15 0.05–0.1 5–6° 5–7°
Milling
cp-Ti TC 25–30 0.07–0.15
HSS 50–60 0.1–0.2 >1.25 face cutter – –
Ti alloys TC 7.5–20 0.07–0.2 >2.5 gear cutter
HSS 15–30 0.1–0.2
TC hard metal (tungsten carbide)
HSS high-speed steel
and/or a dew point below −50 °C), and high-vacuum furnaces. If brazing is not
performed under vacuum or in a controlled atmosphere, fluxes are necessary to
dissolve the oxide layers and prevent a pickup of gases.
1c.3.6 Welding
The inert-gas-shielded arc processes (TIG and MIG) are mainly used for fusion weld-
ing. In special cases resistance, ultrasonic, electron beam, diffusion, and laser welding
are applied. With the cp-Ti grades the weld attains mechanical properties approximat-
ing those of the base metal. A slight decrease in ductility may occur with high tensile
grades. Under passivating conditions, titanium welds have the same corrosion resis-
tance as the base metal. On the contrary, in reducing media the weld may be subjected
to a more severe corrosive attack than the base metal. During the welding operation,
the weld, the heat-affected zone, and the underside of the weld are shielded from the
atmosphere. The filler rod used is an uncoated wire of the same grade or of a grade
174 H. Breme et al.
with a lower hardness than the base metal. Careful preparation of the joint is neces-
sary; that is, surface impurities must be removed by grinding or pickling in order to
avoid porosity. Even fingermarks can produce a hardening of the weld. A single layer
can weld sheets up to 2.5 mm thick. In order to avoid local oxygen concentrations
oxidation products, such as those found at the tip of the electrode, must be cut off. The
effectiveness of the inert gas is responsible for the welding rate. The optimum argon
flow rate has proved to be about 6–8 l/mm. After welding, the appearance of a dark
blue or gray oxide layer indicates an insufficient inert gas shielding and an embrittle-
ment of the weld due to oxygen and/or nitrogen pickup. The hardness of a good weld
may exceed that of the fully recrystallized base metal by a maximum of 50 VHN. If,
after a slight grinding of the surface, a hardness test should give a higher value, the
weld must be completely removed because of embrittlement.
Electron beam welding is particularly suitable for titanium materials. It offers
many advantages such as very narrow seams and small heat-affected zones, weldabil-
ity of thick diameters, high welding speed, and reproducibility of even complex welds.
Titanium materials can be spot welded without any particular preparation under
similar conditions to those of stainless steel. Using flat-tipped electrodes, spot weld-
ing can be performed without inert gas. A hardening of the zone by up to 50 VHN
compared with the base metal is regarded as normal and does not diminish the
strength of the joint. Seam and flashbutt welding are also possible if an argon atmo-
sphere is used.
Diffusion welding is of particular importance for titanium materials because these
materials are more amenable to a homogeneous band in the solid state than other
metals. After welding, the joint zone shows a higher temperature under high vacuum
or, in an inert atmosphere, a microstructure very similar to that of the base metal.
Table 1c.14 Influence of a cold deformation on the mechanical properties of commercially pure
titanium (Ref. 23)
Tensile yield Elongation
Condition strength Ultimate tensile Ratio yield/ at fracture
cp-Titanium (%) (0.2 %) (MPa) strength (MPa) tensile strength (%)
Ti grade 1 30 555 635 0.87 18
40 560 645 0.87 16
55 605 710 0.85 15
60 620 725 0.86 14
65 640 730 0.88 14.5
Ti grade 2 30 605 680 0.89 18
40 645 740 0.87 17
50 680 780 0.87 16
60 685 795 0.86 16
65 692 810 0.85 16.5
Table 1c.15 Mechanical properties of β and near-β-titanium alloys (experimental alloys) (Refs. 6,
15, 24–26)
Tensile yield Ratio yield/ Elongation
strength Ultimate tensile tensile at fracture Reduction
Alloy (0.2 %) (MPa) strength (MPa) strength (%) of area (%)
Til5Mo5Zr3Al 838 852 0.98 25 48
Ti30Nb 500 700 0.71 20 60
Ti30Ta 590 740 0.80 28 58
Table 1c.16 Influence of heat treatment on the mechanical properties of β- and near-β-titanium
alloys (Refs. 6, 15)
Tensile
yield Ultimate Ratio
strength tensile yield/ Elongation
(0.2 %) strength tensile at fracture Reduction
Alloy Condition (MPa) (MPa) strength (%) of area (%)
Til5Mo5Zr3Al SHT at 840 °C 870 882 0.99 20 83.2
SHT at 740 °C 968 975 0.99 16.9 64.5
SHT at 740 °C 1087 1099 0.99 15.3 57.5
+ aged
at 600 °C
45 % cold worked 1284 1312 0.98 11.3 43.8
+ aged
at 600 °C
Ti30Ta Annealed at 1100 °C 650 800 0.81 8 42
Annealed at 1200 °C 660 800 0.83 8 38
Annealed at 1300 °C 665 800 0.83 8 30
Annealed at 1400 °C 680 800 0.83 6 18
SHT solution heat treatment
176
Table 1c.18 Influence of a solution treating and ageing on the mechanical properties of Ti6Al4V
Tensile
yield Ultimate
strength tensile Ratio yield/ Elongation
(0.2 %) strength tensile at fracture Reduction
Condition (MPa) (MPa) strength (%) of area (%)
Sheet ≤ 12.5 mm 15–60 min 1070 1140 0.94 8 20
at 800–920 °C/H2O + 2–4 h
480–600 °C/air
Table 1c.19 Influence of a plasma nitriding (PVD) on the mechanical properties of Ti6Al4V (Ref. 30)
Tensile yield Elongation
strength (0.2 %) Ultimate tensile Ratio yield/ at fracture
Treatment (MPa) strength (MPa) tensile strength (%)
Untreated 809 894 0.90 20
Vacuum annealed 815 924 0.88 21
20 h/850 °C
Plasma nitrideda 805 914 0.88 20
20 h/850 °C/N2
Plasma nitrideda 880 984 0.89 20
20 h/850 °C/NH3
a
Plasma nitriding at 20–40 kW
1c.5 Fatigue
Table 1c.22 Rotating bending fatigue tests of unnotched and notched titanium alloys (Ref. 31)
Stress Fatigue strength for
concentration alternating tensile stresses
Alloy R Condition factor (>107 cycles) (MPa)
Ti6Al4V −1 1.0 725
Ti5Al12.5Fe −1 Wrought + annealed 1.0 725
−1 Wrought + solution 3.6 300
−1 Treated + annealed
−1 Cast + HIP 3.6 300
−1 Cast + HIP 1.0 450
Table 1c.23 High cycle fatigue strength of hip endoprostheses of titanium alloys, measured in
0.9 % NaCl solution at 37 °C. Testing conditions similar to DIN 58840 (simulated loosened shaft,
50 mm) (Refs. 34, 36, 38)
Maximum load in 2 × 107 cycles (kN)
Alloy 0.9 % NaCl (f = 2 Hz)
Hot wrought Ti6Al4V 6.5–8.0
Wrought Ti5Al2.5Fe 8
Ti6Al7Nb 3.5–6.0
Table 1c.24 Influence of the mean stress Sm on the fatigue strength of Ti6Al4V (Ref. 39)
Fatigue strength (MPa)
Sm (MPa) R Notch factor Kf at 107 cycles
0 −1 1 400
2.82 250
250 −0.1 1 300
0.33 2.82 125
500 0.33 1 250
0.66 2.82 100
750 0.7 1 125
0.81 2.82 80
1c Metallic Biomaterials: Titanium and Titanium Alloys 179
Table 1c.26 Influence of interstitial elements on the rotating bending strength of Ti6Al4V (Ref. 44)
Chemical composition of the base alloy (wt%)
Al V Fe C N O H Ti
6.03 3.96 0.10 0.02 0.009 0.1 0.005 balance
Notch factor Fatigue strength
Composition Heat treatment Kf R (MPa)
Base alloy 1 h 815 °C/furnace 1 –1 610
600 °C air 3 –1 210
Base alloy 1 h 855 °C/furnace 1 –1 510
+ 0.07 % N 600 °C air 3 –1 180
Base alloy 1 h 870 °C/furnace 1 –1 550
+ 0.2 % O 600 °C air 3 –1 210
Base alloy 1 h 840 °C/furnace 1 –1 560
+ 9.2 % C 600 °C air 3 –1 230
Base alloy 1 h 930 °C/furnace 1 –1 580
+ 0.07 % N 600 °C air 3 –1 240
+ 0.2 % O
+ 0.2 % C
Table 1c.27 Influence of texture and test directions on the rotating bending fatigue strength of
Ti6Al4V (fine equiaxed microstructure in rolled plates) (Ref. 45)
Fatigue strength
(MPa) test direction
Type of texture and method of production R Text direction at 107 cycles
Basal (0002 tilted out of the rolling plane –1 Rolling direction 625
by about 30°) cross rolling in lower
(α + β)-field
Transverse (0002 aligned parallel to the –1 Rolling direction 630
rolling direction) unidirectional in the Transverse 590
higher (α + β)-field, 950 °C direction
Basal/transverse (both are present) –1 Rolling direction 720
Unidirectional roll at about 930 °C Transverse 690
direction
180 H. Breme et al.
Table 1c.28 Influence of heat treatment (annealing and precipitation hardening, respectively) on
the fatigue strength of Ti6Al4V (Ref. 43, 46)
Condition Yield strength (MPa) Tensile strength (MPa)
As annealed 900 955
As hardened 1100 1195
Fatigue strength (MPa)
Condition sm Kf R at 107 cycles
As annealed 0 1 −1 510
0 3.3 −1 300
As hardened 0 1 −1 600
0 3.3 −1 280
As annealed 200 1 −0.3 425
200 3.3 −0.01 200
As hardened 200 1 −0.5 600
200 3.3 0 200
As annealed 300 1 0.14 400
300 3.3 0.23 165
As hardened 300 1 −0.22 550
300 3.3 −0.2 190
Table 1c.29 Influence of the beta field heat treatment on the fatigue strength of Ti6Al4V (Ref. 47)
Fatigue strength (MPa)
Heat treatment R at 107 cycles
0.5 h 1010 °C/AC + 2 h 700 °C/AC 0 525
5 h 1000 °C/AC + 2 h 700 °C/AC 0 620
0.5 h 1010 °C/H2O + 2 h 700 °C/AC 0 750
5 h 1010 °C/H2O + 2 h 700 °C/AC2 0 650
Table 1c.30 Influence of the surface treatment on the rotating bending fatigue (fine lamellar
microstructure, produced by annealing in 15 min/1050 °C/H2O + 1 h/800 °C/H2O) (Ref. 48)
Fatigue strength (MPa)
Surface treatment R at 107 cycles
Electrically polished −1 680
Mechanically polished (7 μm) −1 750
Mechanically polished (80 μm) −1 605
Mechanically polished (80 μm) + 1 h 500 °C −1 550
Mechanically polished (180 μm) −1 600
Mechanically polished (80 μm) + 1 h 800 °C −1 450
1c Metallic Biomaterials: Titanium and Titanium Alloys 181
Table 1c.31 Influence of the surface treatment on the rotating bending fatigue of Ti6Al4V (fine
equiaxed microstructure produced by rolling at 800 °C/H2O + 1 h/800 °C (HP)) (Ref. 48)
Surface treatment R Fatigue strength (MPa)
Electrically polished −1 610
Shot peened −1 710
Shot peened + 1 h 500 °C −1 390
Shot peened + 1 h 500 °C −1 800
20 μm surface removed
Shot peened + 20 μm surface removed −1 820
Table 1c.32 Influence of surface working on the rotating bending of Ti6Al4V (Ref. 40, 49)
Surface working Notch factor (Kf) R Fatigue strength (MPa)
Mechanically polished 1 −1 620
Mechanically polished 1 −1 660
+ cold roll bent
Ground 1 −1 540
Mechanically polished 2.02 −1 330
Table 1c.33 Influence of plasma nitriding (PVD) on the rotating bending fatigue of Ti6Al4V
(Ref. 30)
Maximum bending stress
Treatment R at 107 cycles (MPa)
Untreated −1 600
Vacuum annealed 20 h/850 °C −1 370
Plasma nitrideda 20 h/850 °C/N2 –1 470
Plasma nitrideda 20 h/700 °C/NH3 –1 550
Solution treated 1 h/940 °C/vac. –1 530
+ Ar cooled
Solution treated 1 h/940 °C/vac. –1 500
+ Ar cooled + plasma nitrideda
at 20 h/770 °C/N2
a
Plasma nitriding at 20–40 kW
182 H. Breme et al.
Table 1c.34 Electrochemical data for titanium and titanium alloys in 0.1 M NaCl under different
conditions (Refs. 24, 50–54)
Corrosion potential Passive current density breakdown potential
Alloy Ecorr (mV) Ip (μA/cm2) Eb (mV)
cp-Ti
pH 7 −628 2.5
pH 2 −459 5–10 >1500
Ti5Al2.5Fe
pH 7 −529 0.68
pH 2 −567 0.71 >1500
Ti6Al4V
pH 7 −510 0.92
pH 2 −699 0.69 >1500
Ti6Al7Nb
pH 7 −368 0.53 >1000
Ti30Ta
pH 7 −419 0.3 >1500
Table 1c.35 Polarization current (i) and polarization resistance (Rc) of titanium and titanium alloys
in pure saline at 37 °C (Ref. 50) and in 0.9 % NaCl with a stable redox system [Fe(CN)64VFe(CN)63]
(Ref. 55)
Rc (kΩ cm2)
2
Material i (μA/cm ) Pure saline Saline + redox
cp-Ti 0.010 1000 714
Ti6Al4V 0.008 1250 455
Table 1c.36 Repassivation time in 0.9 % NaCl and breakdown potential in Hanks’ solution of
cp-Ti and Ti alloys (Refs. 24, 32, 56, 57)
Breakdown potential (mV) Repassivation Time (ms)
Alloy vs. calomel electrode −500 mV +500 mV
cp-Ti 2400 43 44.4
cp-Ta 2250 – –
Ti30Ta >1500 41.7 47.5
Ti40Nb >1500 44.6 43.4
Ti6Al4V >2000 37 41
Ti5Al2.5Fe 110–130 120–160
1c Metallic Biomaterials: Titanium and Titanium Alloys 183
Table 1c.37 Electrochemical data for anodic titanium and Ti6Al4V at 37 °C in different solutions
(de-aired) versus standard calomel electrode (SCE) (Ref. 58)
Corrosion potential Passive current density Breakdown potential
Alloy vs. SCE Ep (mV) Ib (μA/c−2) E (mV) Solutiona
cp-Ti −440 to 490 1.0–3.0 1300 1
−94 to 140 0–1.0 1750 2
−94 5.0–9.0 1950 3
Ti6Al4V −200 to 250 0.9–1.0 1155–1240 1
−240 to 300 0.8–1.5 1900 2
−180 to 250 0.9–2.0 1550 3
a
1 = Ringer’s solution; 2 = Hanks’ solution; 3 = 0.17 M NaCl solution
Table 1c.38 Electrochemical data for cp-Ti and Ti alloys after 7 days in artificial saliva (Ref. 59)
Alloy Corrosion potential (mV) Current densities
Icorr (nA/cm2) Ipass (μA/cm2)
cp-Ti −260 21 9.5
Ti6Al4V −230 31 10
Ti6Al7Nb −220 32 9
Ti5Al2.5Fe −180 30 6
Table 1c.40 Influence of the surface treatment on the fretting corrosion behavior of Ti6Al4V (Ref. 61)
Material combination Total weight loss (μg) Ti (μg) V (μg)
Untreated–untreated 2423 3925 78.5
Untreated–nitrogen ion implanted 1295 1260 31.2
Untreated–PVD coated with TiN 1002 902 15.0
Untreated–plasma ion nitrided 807 716 6.4
PVD-PVD 713 470 8.5
Plasma ion nitrided–plasma ion nitrided 273 87 0.5
Testing conditions: plate screw system (micromotion = 100), 14 days in calf serum solution (1 Hz
for 1,200,000 cycles)
Table 1c.41 Influence of the surface treatment on the wear behavior of Ti6Al7Nb as a result of a
pin-on-disk test (Ref. 62)
PVD coated with Oxygen diffusion hardened (ODH)
Property 3 μm TiN layer (30 μm hardened surface)
Wear factor (10−7 mm3/Nm) 2.111 1.353
Coefficient of friction 0.078 0.051
Surface roughness Rz (μm) 0.159 0.330
Wetting angle (°) 47 49
Table 1c.43 Influence of ion implantation of nitrogen on the wear properties of commercial cp-Ti
and Ti6A14V (Ref. 64)
Material Friction couple Total wear (mg) Friction coefficient
cp-Tia Untreated–untreated 632.3 0.48
Nitrided–nitrided 54.3 0.10
4 h/940 °C/N2:H2 = 2:1
Ti6A14Vb Untreated–untreated 600.0 0.46
Nitrided–nitrided 40.1 0.10
4 h/940 ° C/N2: H2 = 2:1
Nitrided–nitrided 92.3 0.12
6 h/800 °C/N2:H2 = 1:1
a
Friction distance = 1257 m
b
Friction distance = 1885 m
Table 1c.45 Biocompatibility of cp-Ti and Ti alloys, survival rate of L132 cells
incubated with powders (Ref. 66)
Alloy Powder concentration (mg/L) Survival rate of cells
cp-Ti >400
Ti6Al4V >400
Ti5Al2.5Fe >400 >80 %
Ti30Ta >400
Ti30Nb >400
Table 1c.46 Influence of the implantation time (in vivo) on the surface roughness and peak-to-valley
(P–V) height of Ti6A14V femoral heads (Ref. 67)
Implantation time
Before
Position implantation 85 months 110 months 124 months
Ra (nm) P–V Ra (nm) P–V Ra (nm) P–V Ra (nm) P–V
Anterior 43 ± 10 370 ± 72 58 ± 50 746 ± 509 250 ± 147 2044 ± 1178 86 ± 81 812 ± 763
Posterior 41 ± 6 591 ± 333 150 ± 125 2281 ± 1842 114 ± 96 1175 ± 778 142 ± 131 1045 ± 890
Medial 51 ± 14 411 ± 159 44 ± 29 649 ± 259 118 ± 69 1224 ± 731 412 ± 11 401 ± 125
Lateral 52 ± 9 364 ± 68 71 + 55 722 ± 474 117 ± 106 1195 ± 1009 40 ± 16 527 ± 156
186 H. Breme et al.
Table 1c.47 Properties of Nitinol (shape memory, Ni45Ti) alloy (Refs. 50, 68–71)
Below
Above (=austenitic) (= martensitic)
Transition temperature (−200 to 110 °C)
Density (g/cm3) 6.45
Melting point (°C) 1310
Young’s modulus (GPa) 83 28–41
Tensile yield strength (0.2 %) (MPa) 195–690 70–140
Ultimate tensile strength (MPa) 640–1380 103–862
Ratio yield/tensile strength 0.75–0.68 0.33–0.16
Elongation at fracture (%) 1–15 up to 60
Thermal expansion coefficient (°C−1) 1.1 × 10−7 6.6 × 10−6
Specific heat (Cal/g °C) 0.20
Electrical conductivity (S/m) 1–1.5
Corrosion potential (mV)
in 0.1 M NaCl pH 7 −431
pH 2 −518
Passive current density (A/cm2)
in 0.1 M NaCl pH 7 0.44
pH 2 0.61
Breakdown potential (mV)
in 0.1 M NaCl pH 7 890
pH 2 960
Table 1c.49 Ion release from Nitinol incubated with L132 cell culture (Ref. 73)
Ni (ppm) Ti (ppm)
3 days 6 days 3 days 6 days
hp-Ni 6.599 ± 0.037 11.364 ± 0.034 n.m. n.m.
cp-Ti n.m. n.m. 0.001 0.002
NiTi 0.081 ± 0.006 0.176 ± 0.008 0.004 0.006
hp-Ni high-purity nickel, n.m. not measured
1c Metallic Biomaterials: Titanium and Titanium Alloys 187
References
1. Cverna F, Horesh J, Whittle S, Yuko I (2001) Worldwide guide to equivalent nonferrous metals
and alloys, 4th edn. ASM International, Materials Park, OH
2. Donachie MJ (2000) Titanium: a technical guide, 2nd edn. ASM International, Materials Park, OH
3. Standard specification for unalloyed titanium, for surgical implant applications (UNS R50250,
UNS R50400, UNS R50550, UNS R50700) (2006) ASTM International, West Conshohocken, PA
4. Brunette DM, Tengvall P, Textor M, Thomsen P (2001) Titanium in medicine: material sci-
ence, surface science, engineering, biological responses and medical applications, 1st edn.
Springer, New York
5. Zhou YL, Niinomi M, Akahori T (2004) Effects of Ta content on Young’s modulus and tensile
properties of binary Ti-Ta alloys for biomedical applications. Mater Sci Eng A 371:283–290
6. Semlitsch M, Staub F, Weber H (1985) Development of a vital, high-strength titanium
aluminium-niobium alloy for surgical implants. In: Proc. fifth European CONF. ON
BIOMATERIALS, Paris, 4–6 Sept 1985
7. Mausli P-A, Steinemann SG, Simpson JP (1984) Properties of surface oxides on titanium and
some titanium alloys. In: Proc. of the sixth world conference on titanium, vol 3, pp. 1759–1764
8. Material properties of titanium and titanium alloys (DIN 17869) (1990) Beuth Veriag, Germany
9. Microstructure standard for titanium alloy bars, Technical Committee of European Titanium
Producers, Publication ETTC2 (1979)
10. Supra Alloys, Inc. (2013) Aerospace materials specifications for titanium and titanium alloys.
http://www.supraalloys.com/specs.php
11. MatWeb Material Property Data. Titanium, Ti. http://www.matweb.com/search/DataSheet.asp
x?MatGUID=66a15d609a3f4c829cb6ad08f0dafc01
12. Eliaz N (2012) Degradation of implants, 2012th edn. Springer, New York
13. Titanium alloys—Ti6Al7Nb properties and applications. AZO materials. http://www.azom.
com/properties.aspx?ArticleID=2064
14. MatWeb Material Property Data. Titanium IMI 367 (Ti-6Al-7Nb). http://www.matweb.com/
search/datasheet.aspx?matguid=71fff43e6722453c8c9783d017d66977&ckck=1
15. Breme J, Wadewitz V, Burger K (1988) Verbund Titanlegierungen/Al2O3-Keramik fur dentale
Implantate—Entwicklung geeigneter Legierungen. In: Ondracek G (ed) Verbundwerkstoffe—
Stoffverbunde. DGM, pp. 123–130
16. Breme J, Biehl V, Schulte W, d'Hoedt B, Donath K (1993) Development and functionality of
isoelastic dental implants of titanium alloys. Biomaterials 14(12):887–892
17. Akahori T, Niinomi M, Fukui H, Ogawa M, Toda H (2005) Improvement in fatigue character-
istics of newly developed beta type titanium alloy for biomedical applications by thermo-
mechanical treatments. Mater Sci Eng C 25:248–254
18. Kim SE, Jeong HW, Hyun YT, Lee YT, Jung CH, Kim SK, Song JS, Lee JH (2007) Elastic modu-
lus and in vitro biocompatibility of Ti-xNb and Ti-xTa alloys. Met Mater Int 13(2):145–149
19. Zwicker U (1974) Titan und Titanlegierungen. Springer, New York
20. Lee W, Lin C (1998) High-temperature deformation behaviour of Ti6Al4V alloy evaluated at
high strain-rate compression tests. J Mater Process Technol 75:127–136
21. Titanium and titanium wrought alloys forgings (hammer and drop forgings)—technical speci-
fication (DIN 17864) (1990) Beuth Veriag, Germany
22. Long M, Rack HJ (2008) Titanium alloys in total joint replacement—a materials science
perspective. Biomaterials 19:1621–1639
188 H. Breme et al.
23. Ramsdell JD, Hull ED (1960) Characteristics of cold-rolled and annealed Ti. Bureau of Mines
Rep. of Inv. 5656
24. Breme J, Wadewitz V (1989) Comparison of Ti-Ta, Ti-Nb alloys. J Oral Max Implants 4(2):
113–118
25. Niinomi M (1998) Mechanical properties of biomedical titanium alloys. Mater Sci Eng A243:
231–236
26. Shuh A, Bigoney J, Hönle W, Zeiler G, Holzwarth U, Forst R (2007) Second generation (low
modulus) titanium alloys in total hip arthroplasty. Materwiss Werkstofftech 38(12):1003–1007
27. Stemlitsch M, Staub F, Weber H (1985) Development of a vital, high-strength titanium
aluminum-niobium alloy for surgical implants. In: Proc. fifth European conf. on biomaterials,
Paris, 4–6 Sept 1985
28. Breme J, Zhou Y, Groh L (1995) Development of a titanium alloy suitable for an optimized
coating with hydroxyapatite. Biomaterials 16:239–244
29. Titanium and titanium alloy strip, sheet and plate—technical conditions of delivery (DIN
17860) (1990) Beuth Veriag, Germany
30. Lanagan J (1988) Properties of plasma nitrided titanium alloys. In: Proc. of the sixth world
conf. on titanium, vol 4, pp. 1957–1962
31. Borowy KH, Kramer KH (1984) On the properties of a new titanium alloy (TiA15Fe2.5) as
implant material. In: Proc. of the fifth world conf. on Ti, vol 2, pp. 1381–1386
32. Thull R (1979) Eigenschaften von Metallen, fur orthopadische Implantate und deren Priifung.
Orthopädie 7:29
33. Breme J, Schmidt H-J (1990) Criteria for the bioinertness of metals for osseo-integrated
implants. In: ed. Heimke G (ed) Osseo-integrated implants, CRC press, vol 1, pp. 31–80
34. Semlitsch M, Weber H (1992) Titanlegierungen fur zementlose Huftprothesen. In: Hipp E,
Gradinger R, Asherl R (eds) Die zementlose Hiiftprothese. Demeter Verlag, Grafelfingen, pp 18–26
35. Semlitsch M, Panic B (1994) 15 years of experience with test criteria for fracture-proof
anchorage stems of artificial hip joints. In: Buchhorn GH, Willert H-G (eds) Technical princi-
ples, design and safety of joint implants. Hogrefe & Huber Publishers, Seattle, pp 23–36
36. Papakyriacou M, Mayer H, Pypen C, Plenk H Jr, Stanzl-Tschegg S (2000) Effects of surface treat-
ments on high cycle corrosion fatigue of metallic implant materials. Int J Fatigue 22:873–886
37. Narayan R (2009) Biomedical materials, 2009th edn. Springer, New York
38. Kunze E (1988) Vergleichende Untersuchungen zum Langzeit-Ermii-dungsverhalten von
Hiiftgelenkprothesen an Luft und in NaCl-Losung. Metall 2:140–145
39. RMI-Titanium (1969) Reactive Metals Inc., Niles, OH
40. Weinberg IG, Hanna IE (1957) An evaluation of the fatigue properties of Ti and Ti-alloys.
TML Rep. 77
41. Mote WM, Frost RB (1958) The engineering properties of commercial Ti-alloys. TML Rep. 92
42. Weltzin RB, Koves G (1968) Impact fatigue testing of Ti alloys. J Mater 3:469–480
43. Properties and selection of metals. In: ASM metals handbook, vol 1. ASM International (1961)
44. Dittmar CB, Bauer GW, Evers D (1957) The effect of microstructural variables and interstitial
elements on the fatigue behavior of Ti and commercial Ti alloys, Mallory, Sharen Titanium
Corp. AD-110726WADC-TR 56–304, 95
45. Liitjering G, Gysler A (1984) Fatigue. In: Proc. of the fifth conf. on Ti, pp. 2065–2083
46. Hempel M, Hillnhagen E (1966) Dauerschwingfestigkeit der technischen Titanlegierungen
TiA15Sn2.5 und TiA16V4. Arch Eisenhiittenwesen 37:263–277
47. Riidinger K, Fischer D (1984) Relationship between primary alpha content, tensile properties
and high cycle fatigue behaviour of Ti6Al4V. In: Proc. of the fifth conf. on Ti, pp. 2123–2130
48. Wagner L, Gerdes C, Liitjering G (1984) Influence of surface treatment on fatigue strength of
Ti6Al4V. In: Proc. of the fifth conf. on Ti, vol 4, pp. 2147–2154
49. Liu HW (1956) Effect of surface finish on the fatigue strength of TiA16V4- alloy. Dept. of
Theoretical and Applied Mechanics, University of Illinois, T.U.A.M. Rep. 533
50. Geis-Gerstorfer J, Weber H, Breme J (1988) Electrochemische Untersuchung auf die
Korrosionsbestandigkeit von Implantatlegierungen. Z Zahnarztl Implantol 4(1):31–36
1c Metallic Biomaterials: Titanium and Titanium Alloys 189
51. Titanium alloys; chemical composition (DIN 17851) (1990) Beuth Veriag, Germany
52. Luiz de Assis S, Wolynec S, Costa I (2006) Corrosion characterization of titanium alloys by
electrochemical techniques. Electrochim Acta 51:1815–1819
53. Kesteven J, Kannan MB, Walter R, Khakbaz H, Choe H (2015) Low elastic modulus Ti–Ta
alloys for load-bearing permanent implants: enhancing the biodegradation resistance by elec-
trochemical surface engineering. Mater Sci Eng C 46:226–231
54. Cai Z, Nakajima H, Woldu M, Berglud A, Bergman M, Okabe T (1999) In vitro corrosion
resistance of titanium made using different fabrication methods. Biomaterials 20:183–190
55. Breme J (1988) Titanium and titanium alloys, biomaterials of preference. In: Proc. of the sixth
world conf. on Ti, vol 1, pp. 57–68
56. Fraker AC, Ruff AW et al (1983) Surface preparation and corrosion. Behaviour of titanium
alloys for surgical implants. In: Luckey HA, Kubli F (eds) Titanium alloys in surgical implants.
ASTM STP796, pp. 206–219
57. Breme J, Wadewitz V, Ecker D (1986) Untersuchungen zum Einfluf* des Geftigezustandes auf das
Korrosionverhalten der Implantatlegierung TiA15Fe2.5. Z./Zahnarztl. Implantologie, Bd. II 32–37
58. Gagg C (1988) Corrosion characteristics of Til5V3Cr3Sn3Al (Til5-3) alloy in a physiological
environment
59. Mareci D, Ungureanu G, Aelenei DM, Mirza Rosca JC (2007) Electrochemical characteristics
of titanium based biomaterials in artificial saliva. Mater Corros 58(11):848–856
60. Wadewitz V, Breme J (1989) Titan-Legierungen fur dentale Implantate. Z Zahnarztl Implantol
V:116–120
61. Maurer A, Brown SA, Merrit K (1992) Effects of surface treatments on fretting corrosion of
Ti6A14V. In: Proc. of the fourth world biomaterials congress, p. 200
62. Streicher RM, Weber H, Schon R, Semlitsch M (1991) Wet behaviour of different ceramic
surfaces in comparison to TiN and OHD treated Ti6A14V alloy paired with polyethylene. In:
Vincenzini P (ed) Ceramics in substitutive and reconstructive surgery. Elsevier, Amsterdam,
pp 197–205
63. Fellah M, Assala O, Labaïz M, Dekhil L, Iost A (2013) Friction and wear behavior of Ti-6Al-4V
and Ti-6Al-7Nb biomaterial alloy. J Biomater Nanobiotechnol 4:374–384
64. Hu Y-S, Dai D-H, Dong Y-L (1988) The ion nitriding of titanium materials and its applications.
In: Proc. of the sixth world confer, on titanium, vol 4, pp. 1801–1804
65. Mears DC (1977) Metals in medicine and surgery. Int Met Rev, Review 218, 119–155
66. Hildebrand HF (1993) Biologische Aspekte beim Einsatz von Implantatwerkstoffen. DGM
Hochschulseminar, Saarbriicken, Germany
67. Easter TL, Graham RM, Jacobs JJ, Black J, LaBerge M (1994) Clinical performance of
Ti6A14V femoral components: wear mechanisms vs. surface profile. In: Proc. of the 20th
annual meeting of the society for biomaterials, p. 185
68. Ebel B (1990) Zur Biokompatibilitat von Implantatwerkstoffen. KfK 4476
69. Tautzenberger P, Stockel D (1987) Vergleich der Eigenschaften von Thermobimetallen und
Memory-Elementen, 1, 26–32
70. Nitinol Technical Properties. Johnson Matthey Medical Components (2015) http://jmmedical.
com/resources/221/Nitinol-Technical-Properties.html
71. Thompson SA (2000) An overview of nickel-titanium alloys used in dentistry. Int Endod
J 33:297–310
72. Trépanier C, Tabrizian M, Yahia L, Bilodeau L, Piron DL (1997) Effect of modification of oxide
layer on NiTi stent corrosion resistance; Nitinol: Freemont, CA. http://www.nitinol.com/media/
reference-library/033.pdf
73. Rocher P, El Medawar L, Hornez JC, Traisnel M, Breme J, Hildebrand HF (2004) Biocorrosion
and cytocompatibility assessment of NiTi shape memory alloys. Scr Mater 50:255–260
74. El Medawar L, Rocher P, Hornez JC, Traisnel M, Breme J, Hildenbrand HF (2002)
Electrochemical and cytocompatibility assessment of NiTiNOL memory shape alloy for orth-
odontic use. Biomol Eng 19(2–6):153–160
Chapter 1d
Dental Restoration Materials
Jonathan Black and Garth Hastings
1D.1 Amalgams
J. Black (*)
Principal: IMN Biomaterials, King of Prussia, PA, USA
G. Hastings
The Biomaterials Programme Institute of Materials Research and Engineering,
National University of Singapore, Singapore, Singapore
For dental restoration (fillings) the mercury is mixed with the (amalgam) alloy
immediately before application. The alloy is delivered in powder or pellet form. The
amount of mercury after amalgamization must be below 50 wt%. Trituration is
accomplished manually in a mortar with a pestle or more often automatically using
a capsule of mercury with a given weight and an alloy pellet. The amalgam is placed
in the cavity of the tooth in small portions which are pressed with a force of
40–50 N. On the surface of the filling a mercury-rich phase should appear, allowing
a good bonding to the following portion. When the cavity is completely filled, the
mercury-rich phase must be removed, and the filling can be modelled to the desired
shape. 24 hours after the amalgamization the filling must be polished in order to
achieve a smooth surface, resulting in a low corrosion rate.
Table 1d.5 Repassivation rates in artificial saliva versus standard calomel electrode (Ref. 3)
Material Corrosion Potential Potential (mV) after
(mV) 1h 2h 36 h
Non-gamma-2 amalgam -400 -408 -382 -290
194 J. Black and G. Hastings
Corrosion of amalgams
LGC-4-C 74.8–89.8 43.0–55.1 29.0–38.5 0.1–0.2 0–19.5 0–0.3 0–0.5 0–9.0 0–1.5 0–.2 0–0.1 0–0.2 0–2.8
* LGC-4 = low gold-containing, extra hard (60 wt% ≤ Au + Pt metals ≤ 75 wt%).
LGC-4-C = low gold-containing, extra hard, fusible (C = ceramic alloy (a bonding with ceramic is possible).
Table 1d.9 Chemical composition of AgPd and Pd-alloys (wt%) (Ref. 5, 8)
1d Dental Restoration Materials
Au + Pt
Type metals Au Pt Pd Ir Ag Cu Sn Zn In Ru Ga Ge
AgPd-1 29.5 2.0 27.5 70.0 0.2 0.3
AgPd-4 29.5–40.0 ≤2.0 27.4–39.9 0.1 52.0–58.5 0–10.5 ≤2.0 1.5–4.0 ≤2.0
Pd-4-C 52–88 0–17 ≤1.0 25–70 7.2–38.0 0–11.6 1.9–7.5 ≤2.0 0–4.0 ≤0.8 0–7.2 ≤0.5
Key: see Table 1d.7
197
198 J. Black and G. Hastings
Casting
Precious metal alloys are normally cast by means of the lost wax process. The well-
known method of wax modelling is applied. For the commonly used casting proce-
dure, centrifugal and vacuum pressure casting, the alloys can be heated by the
following methods:
• resistence
• propane/oxygen torch (reducing flame zone)
• HF induction
• electrical arc
The alloys are melted in graphite or ceramic crucibles. After removal of the cru-
cible the alloys can if required be hardened. After casting or brazing the alloys are
descaled. Mechanical cleaning can be carried out with rotating tools, ceramic grind-
ing wheels or rubber polishers.
1d Dental Restoration Materials 199
Heat Treatment
Depending on the type of alloy and its application, the dental alloys are heat treated.
After the casting the alloys are quenched. A homogenization at 700°C followed by
rapid cooling should be carried out in order to decompose grain segregations, espe-
cially in alloys containing platinum. After a cold deformation the alloys should be
stress-relieved. Precipitation hardening is performed by:
• slow cooling from 700°C to room temperature
• cooling from 450 to 250°C in 30 minutes, followed by quenching
• heating between 350 and 450°C for 15 minutes, followed by quenching
Table 1d.12 gives the recommended heat treatments for various noble metal
alloys used in dental restoration.
Brazing
Brazing can be carried out with a torch or in a furnace. For larger surfaces furnace
brazing is preferable. The best strength properties are obtained with a solder gap of
0.05 to 0.2 mm between the surfaces. Table 1d.11 shows the chemical composition
and the brazing temperatures of various filler metals.
Bonding with Ceramics
Precious metal alloys are cast by the lost wax process. After removal of the crucible with
carbide tools the castings are sand-blasted with alumina (100–150 μm, pressure 2 bar) to
roughen the surfaces and to provide by increasing the surface an improvement in the adhe-
sion strength. After cleaning with water and hot steam an additional cleaning of the alloys
is performed by annealing for 10 minutes at 980°C. The bonding procedure takes place in
the temperature range of 800–900°C. After bonding the surface must be carefully cleaned
in order to provide good corrosion resistance. The final step of the process is a polishing
operation with rotating cotton or wool buffers and a small amount of polishing paste.
Table 1d.11 Chemical composition (wt%) and brazing temperatures of various filler metals (Ref.
5, 8)
Filler Brazing
Metals for Temp.
Brazing (°C) Au Pt Pd Ir Ag Cu Zn In Re
HGC, 700–840 50–73 ≤19 ≤1.0 ≤0.1 8.0–28 0–9.0 6.0–14 ≤2.0 ≤0.1
LGC
AgPd 760–820 73 0.9 1.0 0.1 13.0 12.0
HGC-C, 700–1120 50–73 ≤1.9 ≤1.0 ≤0.1 10.0 0–5.0 12.0– ≤2.0 ≤0.1
LGC-C –28.0 14.0
Pd-4-C 1030 50–73 ≤1.9 ≤1.0 10.0 3.0– 12.0– ≤2.0
–1120 –28.0 5.0 14.0
Key: see Table 1d.7.
200 J. Black and G. Hastings
Table 1d.12 Recommended heat treatments for various noble metal alloys (Ref. 5, 8)
Oxidizing without
Precipitation Hardening Soft Annealing Vacuum
Time Temperature Time Temperature Time Temperature
Alloy (min) (°C) (min) (°C) (min) (°C)
HGC 15 400 15 700–800
HGC-C 15 500–600 15 950 10 960–980
LGC 15 400–500 15 700–800
LGC-C 15 600 15 950 10 980
AgPd 15 550
Pd-4-C 15 600 15 950 10 980
Key: see Table 1d.7.
Table 1d.13 Mechanical properties of high gold-containing dental cast alloys (Ref. 5, 8)
Ultimate
Tensile Yield Tensile Ratio Yield/ Elongation
Strength (0.2 Strength Tensile at Fracture Hardness
Type %) (MPa) (MPa) Strength (%) HVHN
HGC-1 80 170 0.47 45 55
HGC-2 180–240 370–390 0.49–0.61 35–45 95–110
HGC-3 s 330–350 460 0.72–0.76 35–40 145
h 350–390 550–590 0.64–0.66 20–23 170–190
HGC-4 s 300–420 500–580 0.60–0.72 15–37 155–195
h 540–780 710–870 0.76–0.90 5–18 225–295
HGC-l-C s 90–130 220–280 0.41–0.46 29–38 60–75
h 105–140 230–300 0.46–0.47 27–38 70–90
HGC-2-C s 230 400 0.58 20 105
h 240 410 0.59 18 125
HGC-3-C s 370–420 460–515 0.80–0.82 8–15 150–160
h 470–490 530–590 0.83–0.89 6–9 185–200
HGC-4-C s 380–480 530–580 0.72–0.83 7–14 150–200
h 470–600 550–650 0.85–0.92 3–6 220–230
Key: see Table 1d.7.
s = soft, h = hardened.
1d Dental Restoration Materials 201
Table 1d.14 Mechanical properties of low gold-containing dental cast alloys (Ref.: 5, 8)
Ultimate
Tensile Yield Tensile Ratio Yield/
Strength Strength Tensile Elongation at Hardness
Type (0.2%) (MPa) (MPa) Strength Fracture (%) VHN5
LGC-4 s 310–400 480–510 0.65–0.78 18–43 155–170
h 555–830 640–890 0.87–0.93 3–13 220–275
LGC-4C s 310–590 570–790 0.54–0.75 11–26 180–250
h 550–700 710–900 0.77–0.78 6–18 235–285
Key: see Table 1d.7.
s = soft, h = hardened
Table 1d.16 Polarization current (i) and polarization resistance (Rc) of gold
in 0.9% NaCl and in 0.9% NaCl with a stable redox system [Fe(CN)64- /
Fe(CN63-] at 37 °C corresponding to the potential of the body fluid of 400
mV (Ref. 7, 12)
Material i (μA/cm2 Rc (kΩcm2)
0.9% NaCl saline + redox
Au 0.009 1100 0.28
Table 1d.17 Average passive current density, range of passivity, corrosion and breakdown
potential of a high gold containing alloy (HGC) in artifîcial saliva (Ref. 3)
Range of
Passive Current Passivity Corrosion Potential Breakdown Potential
Alloy Density (μA cm-2) (mV vs. SCE) (mVvs. SCE) (mVvs. SCE)
HGC 1.5 -100 – +400 -137 400
Table 1d.18 Repassivation rates in artifical saliva versus standard calomel electrode (Ref. 3)
Material Corrosion Potential Potential (mV) after
(mV) 1h 2h 36h
HGC -137 -48 -26 -26
202 J. Black and G. Hastings
1D.3 CoCr-Alloys
NiCr-alloys are difficult to classify because of the wide range of the chemical com-
position, as shown in Table 1d.19. The NiCr-alloys in dentistry are generally used
for porcelain veneered and unveneered crowns, fixed and removable partial dentures
and bridgework. As the processing is similar to that of the CoCr-alloys, it is not
described here (see Chapter 1b.3). The requirements for these specific applications
determine the chemical composition. The corrosion resistance of the NiCr-alloys is
provided by the chromium content which produces a passive oxide layer on the
surface. Beryllium is added as a solid solution strengthener and supports the self-
fluxing at the porcelain veneering temperature. It is also responsible for the good
chemical bonding to the porcelain. As it lowers the melting temperature, beryllium
also improves the stability.
Aluminium also produces a passive oxide layer, aids in the bonding to the porce-
lain and strengthens the alloy due to the precipitation of AlNi3. Silicon lowers the
melting temperature and, like manganese, acts as a deoxidizer. Molybdenum and
niobium are added to improve corrosion resistance and, like iron, are used to adapt
the thermal expansion coefficient to the coefficient of the porcelain.
The wide composition range results in an equally wide range of physical (Table
1d.20) and mechanical properties (Table 1d.21). The high rigidity and strength of
these alloys as compared to that of the precious metal alloys make them suitable for
the production of small prosthetic devices.
Table 1d.19 Chemical composition (wt%) of the NiCr-alloys used in dentistry Ref. 11)
Ni Co Fe Cr Mo Nb Ti W Be Ga Si C Others
58 0 0 12 0.5 0 0 0 0 0 0 ≤0.5 Al, Ce, B,
–- –- –- –- –- –- –- –- –- –- –- Mn, Sn, Y, V,
82 2 9 26 16 7 3 4 1.5 7.5 3 Ta, La, Cu
References
1. Cahn, R.W., Haasen, P. and Kramer, E.J. (eds) (1992) Materials Science and Technology, Vol.
14, VCH.
2. Combe, E. (1984) Zahnärztliche Werkstoffe, Hanser.
3. Elagli, K., Traisnel, M. and Hildebrand, H.F. (1993) Electrochemical Behaviour of Titanium
and Dental Alloys in Artificial Saliva. Electrochimica Acta, 38(14), 1769–1774. 881.
4. Berglund, A. (1993) An in vitro and in vivo study of the release of mercury vapor for different
types of amalgam alloys. J. Dent. Res. 72 (5), 939–945.
5. Product information, Degussa, Germany.
6. DIN 13906–1 (=EN 21562, ISO 1562) (1990). Beuth.
7. DIN 13906–2 (invalid, replaced by DIN 28891) (1990). Beuth.
8. Product information, Wieland Edelmetalle, Germany.
9. Fraker, A.C., Corrosion of Metallic Implants and Prosthetic Devices, in Metals Handbook, 9th
Ed., Vol. 13: Corrosion, p. 1351.
10. Encyclopedia of Materials Science and Engineering, Vol. 2, (1986) Pergamon Press,
pp. 1057–1058.
11. NiCo-alloy producing industries.
12. Zitter, H. and Plenk, H. (1987) The Electrochemical Behaviour of Metallic Implant Materials
as an Indicator of their Biocompatibility. J. of Biomedical Materials Research, 21, 881.
Chapter 2
Composite Materials
Table 2.2 (continued)
Grade of class
Property A B E S
At 106 Hz – – 0.002 0.003
Volume resistivity at 72 °F (22 °C) and 500 V – – 1015 1016
DC, ½-cm
Optical properties
Index of refraction – – 1013 104
Acoustic Properties
Velocity of sound, ft/sec (m/sec) – – 17,500 19,200
(5330) (5850)
Table 2.5 Typical Properties of Boron fibers and of other Commercially Available Reinforcement
Filaments (1)
Diameter Manufacturing Average strength Density Modulus
Filament material (μm) technique N/m2 (Ksi) (g/cm3) (GN/m2)
Boron 100–150 CVD 34 (500) 2.6 400
SiC-coated boron 100–150 CVD 31 (450) 2.7 400
SiC 100 CVD 27 (400) 3.5 400
B4C 70–100 CVD 24 (350) 2.7 400
Boron on carbon 100 CVD 24 (350) 2.2 —
Al2O3 250 Melt withdrawal 24 (350) 4.0 250
Beryllium 100–250 Wire drawing 13 (200) 1.8 250
Tungsten 150–250 Wire drawing 27 (400) 19.2 400
‘Rocket Wire’ 50–100 Wire drawing 41(600) 7.9 180
AFC-77
Glass fibers are the most common of all reinforcing fibers for polymeric matrix
composites. Their main advantages are low cost, high tensile strength, high chemi-
cal resistance and good insulating properties. On the other hand they display a low
tensile modulus, a relatively high density comparated to the other fibers, a high
2 Composite Materials 209
The most common aramid fiber available is the Kevlar 49. These fibers are com-
posed of a highly oriented crystalline polymer and present the highest tensile
strength/weight ratio. On the other hand the disadvantages that they present are
the low compressive strength, difficulty of manufacturing and a sensitivity to
ultraviolet light and water. However, Kevlar fibers find applications in sporting
goods.
These fibers are characterized by an extremely high tensile modulus coupled with a
large diameter, offering an excellent resistance to buckling that contributes to a high
compressive strength of the composites. Their high cost is due to the processing
operations. For this reason boron fibers find application only in aerospace and mili-
tary contexts.
Carbon fibers owe their success in high performance composites to their extremely
high tensile modulus/weight and tensile strength/weight ratios, high fatigue strength
and low coefficient of thermal expansion, coupled with a low ratio of cost to perfor-
mance. Carbon fibers are commercially available with a variety of moduli ranging
from 270 GPa to 600 GPa. They are produced following two different processes,
depending on the type of precursors.
210 L. Ambrosio et al.
Polymer matrix resins bind the reinforcing fibers and fabrics together in composite
structures. Resins also act as sizing, load distributors, and vibration dampeners in
the composite structure. A wide variety of thermoset and thermoplastic resins are
used in polymer composites. A summary of the most important materials used as
composite matrix and their properties is presented.
Although many commercial applications for filled thermoplastics exist, use of ther-
moplastics in advanced composites is still in the developmental stage. In fact, the
industry is still split on the use of thermoplastics for advanced composites. According
to some, it is difficult to improve the rigidity, stability and thermal and chemical
resistance of the thermoset resins currently available. Others believe that the major-
ity of the thermoplastic resins available today not only possess the high-service
temperature characteristics required, but also hold the potential of quicker and more
economical processing once the existing problems are resolved. The various ther-
moplastics and their characteristics are described in the following paragraphs:
Polyamides commonly referred to as nylons, are produced by condensation
between diamines and diacids or by self-condensation of an amino acid. Like many
other resins, the exact chemistry of polyamides can vary and, therefore, so can the
final properties. However, all polyamides have low-service temperatures and low
melting points.
Polyamideimides are marketed by Amoco Chemicals under the trade name Torlon. A
polyamideimide is also a good candidate for the production of thermoplastic prepregs.
Polycarbonates are noted for an extremely high-impact strength in an u nreinforced
state. However, compared to such other crystalline polymers as PEEK, a polycar-
bonate does not bond well to reinforcing fibers. It also has poor chemical resistance,
which can be greatly improved by alloying it with thermoplastic polyesters.
Although only a small quantity of polycarbonate with carbon fiber reinforcement is
used today, primarily for interior aircraft structures, polycarbonate is potentially one
of the more promising matrix resins for advanced composites, especially in alloyed
versions such as General Electric’s Xenoy and Bayer’s Makroblend.
Polyetheretherketones (PEEK) have excellent properties for use in advanced poly-
mer composites, including low flammability, low smoke and toxic gas emission, and
broad chemical and solvent resistance. PEEK possesses a continuous service tempera-
ture of 200 °C to 240 °C (392 °F to 464 °F) and has a very high melting point of 334 °C
(633 °F). PEEK also provides excellent abrasion resistance at its service temperature,
radiation resistance, excellent fatigue and wear resistance and a relatively low specific
gravity of 1.32. It is possible to process PEEK on conventional extrusion and molding
equipment, and its highly crystalline nature responds well to fiber reinforcement.
2 Composite Materials 211
The basic difference between thermoset and thermoplastic resins is the reaction
heat. A thermoset resin is cured by the application of heat and often by the addition
chemicals called curing agents. Once cured, the material is infusible, unsoluble and
can softened or reworked with the addition of heat.
A thermoplastic, on the other hand, is capable of being repeatedly softened by
addition of heat and hardened by decreasing temperature. The change occurring
thermoplastic resin with the addition of heat is primarily physical, not chemical.
This different provides one major advantage for thermoplastics: any scrap from
fabrication can be reused.
Thermoset resins can vary greatly with respect to service temperature, solvent
resist and other important characteristics. A description follows of the various ther-
moset resin their basic characteristics.
Vinyl ester resins are thermosetting resins that consist of a polymer backbone with
acrylate or methacrylate termination. The backbone component of vinyl ester resins
can derived from an epoxide resins, polyester resins, urethane resin, and so on, but
those base epoxide resins are of particular commercial significance.
Vinyl ester resins are produced by the addition of ethylenically unsaturated
carbo acids (methacrylic or acrylic acid) to an epoxide resin (usually of the bisphe-
nol epichlorohydrin type). The reaction of acid addition to the epoxide ring (esteri-
fication exothermic and produces a hydroxyl group without the formation of
212 L. Ambrosio et al.
Epoxide are materials which contain two or more glycidyl groups per molecule. The
uncured resins range from free flowing liquids to high melting solids, which can be
cross-linked by reaction with an appropriate curing agent.
Typical curing agents include primary and secondary amines, polyamides and
organic anhydrides. Other curing agents used are the catalytic curing agents, such
as the boron trifluoride complexes. No by-products are evolved during cure. The
resultant cured resins are generally hard thermoset materials with excellent mechan-
ical, chemical and electrical properties.
These materials can be conveniently divided into six classes of resins:
• bisphenol A based (e.g. diglycidyl ether of bisphenol-A (DGEBA))
• glycidyl esters
• glycidyl amines (e.g. tetraglycidyl amine of 4,4-diamino-diphenyl-methane)
• novolacs brominated resins (e.g. diglycidyl ether of tetrabromo-bisphenol A)
• cycloaliphatic resins (e.g. Tetraglycidyl ether of tetraphenylene ethane)
2.8 Diluents
Diluents are added to epoxide resins primarily to lower viscosity and thus to improve
handling characteristics. They also modify the cured properties of the resin. Diluents
can be divided into two classes: (a) the reactive diluents and (b) the non-reactive
diluents.
• Benzyl alcohol
• Furfuryl alcohol
• Dibutyl phthalate (DBP)
These are known variously as curing agents, hardeners, activators or catalysts. They
are required to convert liquid and solid epoxide resins into tough infusible thermo-
set polymers. The curing agents promote this curing reaction by opening the epox-
ide ring and become chemically bound into the resin in the process. Each of the
curing agents for epoxide resins will now be discussed in turn.
• Dimethylaminpropylamine (DMAPA)
• m-Xylylenediamine (mXDA)
• N-Aminoethylpiperazine (AEP)
These consist of organic anhydrides and are used in roughly stoichiometric propor-
tions with epoxide resins. The anhydride commonly used are the following:
• Phthalic anhydride (EPA)
• Tetrahydrophthalic anhydride (THPA)
• Methyltetrahydrophthalic anhydride (MTHPA)
• Endomethylenetetrahydrophthalic anhydride (NA)
• Hexahydrophthalic anhydride (HHPA)
• Methylhexahydrophthalic anhydride (MHPA)
• Trimellitic anhydride (TMA)
• Pyromellitic dianhydride (PDMA)
• Dodecenylsuccinic anhydride (DDSA)
The polyamides used to cure epoxide resins are all reactive compounds with free
amine groups. They may be amidopolyamines, aminopolyamides or imidazolines.
They are mostly used in coating systems and in adhesive formulations. Some sup-
pliers of polyamide curing agents are listed below:
• Ancamide (Anchor Chemical Co, Ltd)
• Araldite (Ciba-Geigy Plastics & Additives Co.)
2 Composite Materials 215
Several other types of curing agent are used with epoxide resins for laminating
applications or in moulding compounds. Examples of these curing agents are:
• Dicyandiamide (Dicy)
• Boron trifluoride complexes
• Boron trifluoride monoethylamine
• 2-Ethyl-4-methylimidazole
• N-n-Butylimidazole
The basic materials used to make a polyester resin are a dibasic organic acid or
anhydride and a dihydric alcohol.
Catalyst or initiators for unsaturated polyester resin system consist of organic per-
oxides. Some commercially available catalysts are:
• Diacyl peroxides (benzoyl peroxide, 2,4-dichlorobenzoyl peroxide, dilauroyl
peroxide, diacetyl peroxide).
• Ketone peroxides (methyl ethyl ketone peroxide, cyclohexanone peroxide, acet-
ylacetone peroxide, methyl isobutyl ketone peroxide).
• Hydroperoxides (t-butyl hydroperoxide, cumene hydroperoxide).
• Dialkyl and diaralkyl peroxides (dicumyl peroxide, di-t-butyl peroxide, t-butyl
cumyl peroxide, 2,5-dimethyl-2,5-bis(t-butylperoxy)hexane).
216 L. Ambrosio et al.
These are materials which when used in conjunction with an organic peroxide cata-
lyst increase the rate at which that peroxide breaks down into free radicals. Some of
the commercially available accelerators are:
• Cobalt accelerators: cobalt siccatolate, naphthenate or octoate
• Manganese accelerators: manganese salts
• Vanadium accelerators
• Tertiary amine accelerators: dimethylaniline (used to accelerate diacyl peroxide
catalysed system); diethylaniline (used to accelerate benzoyl peroxide catalysed
system); dimethyl-p-toluidine (used to accelerate benzoyl peroxide catalysed
system).
orientation, which exhibits a significantly decreased moduli and strength, and (iii)
the response of the material to an in-plane shear load.
By the contrast with isotropic metallic materials, an oriented ply, in the form of
a thin sheet, is anisotropic and requires four elastic (plane stress) constants to spec-
ify its stiffness properties in its natural orientation
σ 1 = Q11ε1 + Q12ε1
σ 2 = Q12ε1 + Q22ε2 (2.1)
σ 6 = Q66ε6
where σ6=τ12 and є6=τ12 or in matrix form
σ 1 Q11 Q11 0 ε1
σ 2 = Q11 Q11 0 . ε2 (2.2)
σ 3 0 0 Q11 ε3
where the plane stress stiffness moduli are
Q11 = E11 / (1 −ν 12 ν 21 )
Q22 = E22 / (1 −ν 12 ν 21 )
(2.3)
Q12 = ν 21 E11 / (1 −ν 12 ν 21 )
Q6 = G12
where vij is the Poisson ratio, defined as -є/єj.
If, however, the ply is rotated with respect to the applied stress or strain direction
additional moduli appear, which results in the direction-indicated shear coupling
rotation simple extension
8 ( 11
U1 = 1 3Q + 3Q22 + 2Q12 + 4Q66 )
U2 = 1 (Q − Q )
2 11 22
U3 = 1 ( Q + Q − 2Q − 4Q ) (2.6)
8 11 22 12 66
U4 = 1 ( Q + Q + 6Q − 4Q )
8 11 22 12 66
U5 = 1 ( Q + Q − 2Q − 4Q )
8 11 22 12 66
In addition, lamination can result in up to 18 elastic coefficients and increased
deformational complexities, but the additional coefficients can all be derived from
the four primary coefficients using the concept of rotation and ply-stacking
sequence. These complications are the result of geometric variables. If the laminate
is properly constructed, the in-plane stretching or stiffness properties can still be
specified by four elastic coefficients. We shall consider laminates of this nature.
Note that both short and continuous fibers are handled in the same manner. These
calculations, while tedious, are analytically simple. The ‘plane stress’, the Qij terms,
are employed because lamination neglects the mechanical properties through the
ply thickness. These stiffnesses are sometimes regrouped into new constants called
‘invariants’, the Ui terms, for analytical simplicity. To compute the properties of the
laminate one then sums the ply (hk) properties through the thickness of the laminate,
weighted by the thickness (hk) of each oriented ply
N
Aij = ∑(Qij )k hk
k=1
For a balanced (same number of ±θ) and symmetrical system (+θ or − θ at same
distance above and below the midplane) the laminate solution is
(
E11 = A11 A22 − A212 / A22 )
E22 = (A11 A22 − A 2
12 )/ A
11 (2.8)
ν 12 / E11 = A12 / ( A11 A22 − A212 ) G12 = A66
2 Composite Materials 219
These calculations have been thoroughly tested and agree closely with
e xperiment. The circles and squares are the experimental points and the lines are the
theoretical predictions for a nylon fiber reinforced rubber. The angle ply laminate is
predicted from the ply properties. The ply properties are in turn correlated with the
transformation equations and the micromechanics. The micromechanics employed
in this demonstration are based upon the ‘self-consistent method’ developed by Hill
(8). Hill rigorously modeled the composite as a single fiber, encased in a cylinder of
matrix, with both embedded in an unbounded homogeneous medium which is mac-
roscopically indistinguishable from the composite. Hermann (9) employed this
model to obtain solution in terms of Hill’s ‘reduced moduli’. Halpin and Tsai (10)
reduced Hermann’s solution to simpler analytical form and extended its use for a
variety of filament geometries
E11 = E f V f + EmVm
ν 12 = ν f V f + ν mVm (2.9)
p / pm = (1 + ηξ V f ) / (1 − ηV f )
where
η = ( t f / t m − 1) / ( p f / pm + ζ )
ξ = ( e / d ) ; ζ G12 = 1ζ G 23 = 1 / ( 3 − 4ν m ) (2.10)
p = E22 , G12 , G23 ; p f = E f , G f ; pm = Em , Gm
These equations are suitable for single calculation and were employed p reviously
for the single ply and angle ply properties. The short fiber composite properties are
also given by the Halpin-Tsai equations where the moduli in the fiber orientation
direction is a sensitive function of aspect ratio (1/d) at small aspect ratios and has
the same properties of a continuous fiber composite at large but finite aspect ratios.
If the ply is used in the construction of a balanced and symmetrical 0/90 laminate
and is mechanically tested, bilinear stress/strain curve is obtained, and the stiffness
is the sum, through the thickness of the plane stress stiffness of each layer. As the
laminate is deformed each ply possesses the same in-plane strain, and when the
strain on the 90 layers in the laminate prevents the 90° layer from carrying its share
of the load, Qij(90°) = 0. This load is transferred to the unbroken layers, the 0° layers
for our illustration, and results in a loss of laminate stiffness or modulus. Continual
loading will ultimately produce a catastrophic failure of the laminate when the
strain capability of the unbroken, 0°, layers is exceeded. For a 0/90 construction,
employing glass/epoxy material, the ratio of the ultimate failure stress to the crazing
stress is 6.1. Note a change in stiffness as the 90° and then the 45° layers fail, and
the correspondence of the theoretical ultimate strength of 356 MPa with the experi-
mental results of 346 MPa. While the strain for transverse ply failure is constant
from laminate to laminate, the stress required to craze the system as well as cause
final failure is a function of laminate geometry, because the construction of the
220 L. Ambrosio et al.
Various fabrications used in the reinforced plastics industry are discussed below:
In one of the simplest and labor intensive procedures, pigmented, catalysed resin is
applied to the surface of the mold. This gel coat in room temperature lay-up tech-
niques will end up on the surface of the finished composite (FRP). Catalysed resin-
impregnated mat is then applied over the gel coat and this and subsequent layers are
brushed or rolled to assure good contact between layers and to remove any entrapped
air. This procedure is continued until the desired thickness of the composite is attained.
The assembled composite may be cured at room temperature or at elevated tem-
peratures for faster cycles. This procedure, which was originally called contact mold-
ing, may be upgraded by the application of a vacuum or pressure bag placed over a
Cellophane film on the final layer to reduce void formation in the composite. The
laminate may also be built up by a spray-up process in which a mixture of chopped
glass strands and catalysed resin is sprayed on the gel coat instead of resin-
impregnated mat. In any case, the inner surface will be less smooth than the first layer
formed by the gel coat. Tanks, boats and pipe may be fabricated by this technique.
Matched die molding of a premix of chopped glass, roving, and catalyzed resin is
used for relatively large scale production of reinforced articles. Uncured dough-like
compositions are called bulk molding compounds (BMC). Uncured resin-
impregnated sheets are called sheet molding compound (SMC). These compounds
are supplemented by thick molding compounds (TMC) and XMC. TMC is pro-
duced continuously on a machine that resembles a 2 roll mill. XMC, in which the
continuous impregnated fiber are arranged in an X-pattern, is produced on a fila-
ment winding machine.
Autoclave Molding Autoclave molding, the process of curing thermoset resins at
elevated temperature and pressure in an inert environment, has an important role in
the fabrication of continuous fiber reinforced thermoplastics. While most compa-
nies view thermoplastics as an alternative to traditional autoclave long-cycle pro-
cessing, they have come to accept the following reasons for the autoclave processing
of thermoplastic matrices:
• Availability
• High temperature and pressure capability
• Reduced tooling requirements
• Uniform pressure distribution.
The dominant reason for the autoclaves’ role in thermoplastics production is its
availability. Aerospace first- and second-tier contractors, who conducted much of
the developmental work in parts fabrication, all have autoclaves on hand.
While it may seem defeatist to use the autoclave for thermoplastic resins which
undergo no chemical reaction and lend themselves to rapid fabrication, autoclave
use offers other advantages.
In the industry, many fabricators own autoclaves capable of processing high tem-
perature thermosets (e.g. polyimides) at operating temperatures up to 800 °F and
pressures of 150 to 200 psi; which can also accommodate high temperature thermo-
plastics such as PEEK, which is normally processed in the 700 to 750 °F range. One
disadvantage, particularly for production-sized autoclaves, is the inability to finely
control cool-down rate – a critical process step for the semicrystalline thermoplas-
tics. This situation can be improved by implementing integrally cooled tooling.
222 L. Ambrosio et al.
Where high consolidation pressures are required, the autoclave can reduce tool-
ing costs by eliminating the need for matched metal tooling, when compared to a
molding process (e.g. compression molding or thermoforming). Less expensive
tooling aids can be used for consolidation via the autoclave’s pressurized environ-
ment. The autoclave also provides uniform pressure over the part’s area and elimi-
nates pressure distribution concerns associated with a matched tool in a press.
Thermoforming Thermoforming offers vast potential for high volume thermoplas-
tic composite parts fabrication. There are many thermoforming variations; but by
basic definition, thermoforming is the heating of a reinforced thermoplastic matrix
sheet or kit above the softening temperature, followed by forcing the material
against a contour by mechanical (e.g. matched tooling, plug) or pneumatic (e.g. dif-
ferential air or hydraulic) means. The material is then held and cooled sufficiently
for shape retention, and removed from the mold. Thermoforming implies only those
processes applicable to thermoplastic resins, and is often used in the same context
as the term ‘compression molding’ which also applies to thermosets. In this section,
the process is defined as the preheating of the lay-up or reconsolidated sheet, fol-
lowed forming via a matched mold.
Phillips Petroleum has applied this technology to fabric reinforced Ryton (PPS)
materials. A conveyorized infrared oven is used to rapidly preheat the lay-up to 600
°F in two to three minutes. The charge is quickly transferred to a preheated mold in
a fast closing press for part forming. Total cycle times of one to three minutes are
feasible with this automated approach.
High production rates can be achieved using thermoforming technology.
However, it is difficult to form high quality continuous fiber reinforced thermoplas-
tic parts with demanding geometries, due to the restricted movement of the fiber.
Du Pont’s long discontinuous fiber sheet products with PEEK and J-2 resins
provide easier fabrication of complex shapes. This product form is particularly
attractive to helicopter manufacturers for the press forming of highly contoured
secondary structure parts.
Transfer molding Transfer molding is used for the manufacture of small compo-
nents and is particularly useful with multi-cavity tools and where small inserts are
to be moulded in. Materials used are polyester and epoxide dough moulding com-
pounds, although a new liquid resin injection technique is reported.
Heated steel molds, preferably hard chrome plated, are used, which may be of
multicavity design. Tooling costs are higher than for compression moulding since
appropriate gates and runners must be included in the mould.
A pre-weighed quantity of DMC is placed in a heated transfer pot by hand. A
punch or ram compresses the material and causes it to flow into the heated tool cav-
ity where it cures. The tool is mounted between the platens of a press.
Factors to be considered with transfer molding are transfer and tool clamping
pressures and transfer time. To reduce transfer time and increase overall efficiency
the molding compound may be pre-heated in an oven or high frequency pre-heater
such as a micro-wave oven.
2 Composite Materials 223
Mold temperatures range from 155 to 170 °C both for polyester and epoxide
resin compounds, with molding pressures ranging from 5 to 100 MPa depending on
the type of compound to be processed, mold design and temperature. Cure time in
the mold (excluding pre-heat time) is usually of the order of 10–30 s per millimetre
of wall thickness for both types of compound.
Injection molding Injection molding, a technique used extensively for the process-
ing of thermoplastic materials, has also been developed to process thermosetting
resin systems. Due to high mould costs it is generally only suitable for the large
scale production of small-to-medium sized components. Materials processed in this
way are polyester and epoxy DMC and also phenolics, ureas, melamines and diallyl
phthalate moulding compounds. These latter materials are generally more difficult
to process than either polyester or epoxy DMC.
Thermoset moldings produced by injection moulding are used widely in the
electrical and automotive fields, thus large production runs are common.
Injection molding has advantages over both compression and transfer molding in
that the process is more automated and far higher production rates can be achieved.
Although mould costs are higher than for compression molding, overall finished
component costs are generally lower. With small weight components, scrap from
runners can be high compared with compression molding but for large mouldings
this becomes relatively insignifiant. Injection molding is also better for thick parts
since, with the pre-heating of the DMC before injection into the mold, shorter mold-
ing cycles are possible.
While injection molding machines designed specifically for processing thermo-
set materials are available, a number of manufacturers offer replacement screw and
barrel assemblies and stuffer hoppers to convert conventional thermoplastic injec-
tion moulding machines to process thermosets.
Molding compound is transferred in the cold state by pressure from the material
hopper into the main injection chamber. Here it can be preheated before injection
into the heated mould tool. Injection, through a special nozzle, can be either by ram
or screw pressure. If screw feed is used, the screw must be of the type designed to
process thermosets as opposed to thermoplastics.
Early machines were designed with vertical clamping pressure on the mold but
today horizontal machines are mostly used. Since thermosetting materials are liq-
uid until gelation occurs, clamping pressure has to be maintained on the mould
until the resin has cured. Unless this is done, excessive flash will form. Heated
matched metal molds are used, which may be of multi-cavity design. These molds
must be designed for use with thermosetting resins, taking into account the fact
that thermoset moldings are harder, more rigid and less easily deformed than
thermoplastics.
A typical temperature sequence for injection molding DMC would be: feed hop-
per and feed zone – ambient temperature; metering section 50–60 °C; nozzle 80–90 °C;
mould temperature 135–185 °C for polyester DMC or 160–22 °C for epoxy DMC;
injection pressure 80–160 MPa. Cure time is generally of the order of 10–20 s
per millimetre of wall thickness. Very little finishing of moldings is necessary.
224 L. Ambrosio et al.
Where fully automatic molding machines are used, hydraulic ejection with
p erhaps a ‘joggle’ facility is necessary, since thermosets have a tendency to stick in
the mold.
Filament winding is a technique used for the manufacture of pipes, tubes, cylinders
and spheres and is frequently used for the construction of large tanks and pipework
for the chemical industry. By suitable design, filament wound structures can be
fabricated to withstand very high pressures in service. In general, products fabri-
cated by filament winding have the highest strength to weight ratios and can have
glass contents of up to 80 % by weight.
The process is suitable for use both with polyester and epoxide resin systems and
a variety of fibres including glass, carbon, aramid and metals, providing that these
materials are available in continuous filament lengths. Glass fibre is by far the most
common reinforcement used and will be used as the example in the description of
the process.
Filament winding is basically a simple process, although numerous modifica-
tions have been developed to improve product quality. Moldings can be produced
by either a wet lay-up process or from prepreg.
In recent years filament winding has been extended to the continuous production
of pipe using a continuous steel band mandrel. In this way continuous lengths of
pipe can be produced, with diameters ranging from about 0.3 to 3.5 m.
With the wet lay-up process glass rovings are drawn through a resin bath to impreg-
nate them with resin. The impregnated rovings are then wound under tension round
a rotating mandrel. Generally the feed head supplying the rovings to the mandrel
traverses backwards and forwards along the mandrel.
The mandrel, which may be segmented for large diameter pipes, is generally
wrapped with a release film, such as Cellophane, prior to wrap- ping with glass and
resin. The mandrel may incorporate some means of heating the resin system, such
as embedded electric heaters, or provision for steam heating. Alternatively, the fully
wrapped mandrel and laminate may be transferred to a curing oven to effect cure.
In order to provide a resin-rich, corrosion resistant inner lining to the pipe, the
mandrel may be wrapped with a surfacing tissue followed by one or two layers of
chopped strand mat or woven tape prior to filament winding. This first layer is usu-
ally allowed to cure partially before winding commences to prevent the resin from
being squeezed out into the main laminate.
2 Composite Materials 225
The winding angle used during construction of pipes or tanks depends on the
strength/performance requirements and may vary from longitudinal through helical
to circumferential. Often a combination of different winding patterns is used to give
optimum performance. Accurate fibre alignment is possible.
For pipe construction, steel mandrels are generally used. However, where cylin-
ders or spheres are to be made, an alternative material has to be used so that it can
be removed once the resin system has cured. In these cases the mandrel can be made
from wax, a low melting metal alloy, or an inert plaster held together with a water
soluble binder. Clearly, in these cases the mandrel can only be used once. Material
choice for the mandrel will depend on the cure cycle needed for the resin system.
In addition to winding with continuous filament rovings, machines have been
developed which permit winding with tapes or slit-width chopped strand mat and
woven rovings. These reinforcements may be used alone or combined with continu-
ous filament rovings. Thus considerable design flexibility exists for the production
of large simple shapes.
Improved chemical resistance can be achieved by the use of a thermoplastic or
synthetic rubber liner. If the liner is sufficiently rigid it can be supported on a light
frame and used as the mandrel, if not then it can be wrapped round the mandrel first.
A grade of polypropylene is available which has a woven glass cloth partially rolled
into one side to improve adhesion of the resin system (Celmar).
Dunlop has recently developed a new process for pipe production to produce
pipes in the range 200–2000 mm diameter. Essentially the process is similar to con-
ventional filament winding. A mandrel, suitably coated with release agent, is
wrapped with an epoxide resin impregnated glass tape over this is wound a 150 mm
high-strength steel strip angled to give 50 % overlap. Epoxide resin system is
applied to the steel strip to ensure that each layer is fully encapsulated. From three
to 13 layers of steel may be applied to satisfy different pressure ratings. The pipe is
finished by wrapping with further resin impregnated glass tape and the resin system
cured. Pipe produced in this way has excellent corrosion resistance coupled with a
high strength/weight ratio. It is said to provide up to 50 % weight saving over con-
ventional steel pipe.
Various processes are available for the ‘on-site’ construction of large filament
wound storage tanks. By manufacturing these on site, transport problems are over-
come and integral structures can be produced. With the various processes either
horizontal or vertical mandrels are first constructed from preformed GRP sheets.
These are then wrapped with resin impregnated glass rovings.
Filament wound vessels can be produced from prepreg tapes and rovings. This
technique is often used with carbon fibre to reduce fibre damage during the winding
operation and to permit the use of resin systems which cannot be handled by wet
lay-up techniques. Here, it is essential to use a heated mandrel to melt the resin and
hence displace air and consolidate and cure the laminate. Resin content of the lami-
nate can be controlled more accurately with prepreg since the prepreg can be made
with exactly the right resin content. The use of prepregs also makes for cleaner
operation.
226 L. Ambrosio et al.
Filament winding has been used to provide a protective laminate on the outside
of steel pressure pipes where external corrosion can take place. An example of this
use is in the protection of the splash zone of steel riser pipes used on sea based oil
and gas production platforms. Here, care has to be taken in the design of such a
composite structure since the coefficient of expansion of the filament wound glass
wrap can be lower than that of the steel core. If such a composite structure is pro-
duced using a heat cured resin system (say 120 °C cure) and then subjected to sub-
zero temperatures in use, the steel pipe can shrink away from the laminate and
permit entry of water by capillary action. Thus the object of wrapping the pipe to
prevent corrosion can be defeated, since corrosion can then still take place under the
laminate. once the bond has broken it can never be remade.
To produce a curved tapered pipe, a flexible mold is used. This is bent to the
required shape once the preform has been placed in position. It is also claimed that
base plates can be simultaneously molded on by this method. These base plates are
first preformed and then inserted into the mold where they become firmly bonded into
the pole. Typical applications include poles for street lighting, flag poles and aerials.
For the purposes of this book, only those processes for which polyester resins and,
to a limited extent, epoxide resins are used will be described. It should be noted,
however, that the decorative laminates used in building and transport applications
are in the main manufactured from melamine faced phenolic resin/paper laminates.
Several patented processes exist for continuous sheet production, all of which are
similar in broad principle.
Resin and glass reinforcement are sandwiched between two sheets of release
film, such as Melinex, Mylar or Cellophane and passed through rollers to consoli-
date the laminate before curing in an oven. Resin is applied to the release film either
by spray or trickle process, care being taken to ensure that application is uniform.
Glass reinforcement is laid in the resin and a second layer of release film applied.
This sandwich is passed through a series of rollers to expel all air bubbles and con-
solidate the laminate to the correct thickness. During the next stage the laminate
sandwich is either passed directly through an oven to produce flat sheet or through
rollers or dies and then an oven to produce corrugated sheet. Once cured the sheet
is trimmed to the required width and cut into suitable lengths. Depending on the
process, corrugations may run longitudinally or transversely. Production speeds of
up to 12 m/min are possible.
To produce clear sheeting the refractive index of the resin system must match
that of the glass reinforcement. For this reason special resins have been developed
which match the refractive index of E-glass. For translucent sheeting A-glass may
be used but, due to its low refractive index, it is unsuitable for use in transparent
sheeting. In any case today A-glass is rarely found.
Generally the glass reinforcement used consists of chopped strand mat with a
soluble polyester powder binder or chopped rovings deposited directly into the resin
film. However, for certain applications woven fabrics may be used. With the latter
and to a much lesser extent with chopped strand mat, the glass cloth may be drawn
through a resin bath and excess resin removed between doctor blades or rollers,
before placing between two layers of release film and processing as before.
With high quality sheeting a surfacing tissue may be used to ensure a resin-rich
finish. Such sheeting must be installed with the resin-rich surface exposed to the
weather.
Resin systems are generally specifically formulated for each machine, since gel
time and viscosity must suit the particular operating conditions of the machine. Resin
systems used include those suitable for producing clear fire retardant sheeting.
2 Composite Materials 229
By the correct choice of resin system, sheeting can be manufactured which will
not yellow to any extent after exposure to tropical weather conditions for several
years. However, to ensure that this is the case the resin system must be chosen with
care and must be fully cured. Also the release film must be removed before installa-
tion and the laminate should contain not less than 75 % by weight of resin. In other
cases the resin content of the laminate may fall between 65 and 75 % by weight.
2.12.8 Pultrusion
A summary of the mechanical properties of the most important matrix materials and
their composites with different reinforcing fibers are presented in Tables 2.8–2.39.
Table 2.26 Properties of the Most Common Resin for High Performance Composites
Coefficient
Tensile Flexural Max service of thermal Water
strength modulus Density temperature expansion absorption
Materials (MPa) (MPa) (gcm-3) (°C) (10-50 °C-1) (24 h%)
Epoxy 35–85 15–35 1.38 25–85 8–12 0.1
Polyimide 120 35 1.46 380 9 0.3
PEEK 92 40 1.30 140 6–9 0.1
Polyamide/imide 95 50 1.38 200 6.3 0.3
Polyether/imide 105 35 1.27 200 5.6 0.25
Polyphenylene/ 70 40 1.32 75 9.9 0.2
sulfide
Phenolics 50–55 10–24 1.30 50–175 4.5–11 0.1–0.2
Table 2.27 Properties of Polyester Composites Reinforced by Continuous and Chopped Fiberglass
Tensile Flexural Transverse Transverse
Continuous Chopped Strength Strength Tensile Strength Flexural Strength
Filament (%) Glass (%) MPa MPa MPa MPa
75 0 690 1200 24 35
65 10 660 1135 27 90
45 20 570 980 60 155
25 20 500 810 95 200
15 60 410 680 125 260
Table 2.28 Typical properties of cured polyester resins
2 Composite Materials
Table 2.30 Shear Properties of Composites of Kevlar 49 Fiber in Epoxy Resins (8)
Shear Strain Secant Shear
Cure Cycle Shear Failure at Failure Modulus at 0.5%
Epoxy System hours/°C Stress MPa Stress % Shear Strain MPa
(weight ratio (hours/°F) (CV) (CV) (CV)
XD 7818/ERL 2.5/80 + 2/160 21.4 (2.6) 1.35 (2.2) 1884 (3.9)
4206/Tonox 60–40 (2.5/176 + 2/320)
DER 332/ Jeffamine 24/60 29.4 (2.0) 173 (2.3) 1923 (4.7)
T-403 (100/39) (24/140)
ERL 2256/Tonox 16/50 + 2/120 23.0 (8.6) 1.49 (2.2) 1775 (0.9)
60–40 (100/25/28.3) (16/122 + 2/248)
Epon 826/RD2/ 3/60 + 2/120 23.4 (6.3) 1.91 (6.5) 1520 (3.9)
Tonox 60–40 (3/140 + 2/248)
(100/25/28.3)
XB 2793/Tonox 2/90 + 2/160 21.9 (0.3) 1.69 (2.9) 1600
60–40 (100/25.6) (2/194 + 2.320)
(continued)
Table 2.30 (continued)
Shear Strain Secant Shear
Cure Cycle Shear Failure at Failure Modulus at 0.5%
Epoxy System hours/°C Stress MPa Stress % Shear Strain MPa
(weight ratio (hours/°F) (CV) (CV) (CV)
XD 7818/XD 5/80 + 3/120 39.7 (0.9) 2.43 (2.5) 1852 (1.7)
7575.02/XD (5/176 + 3/248)
7114/Tonox 60/DAP
(50/50/45/14.1/14.1)
XD 7818/XD 5/60 + 3/120 31.9 (3.4) 1.91 (4.5) 1850 (1.5)
7114/Tonox LC
(100/45/50.3)
Table 2.38 Properties of Epoxy Resin Composites with Different Reinforcing Fibers
Aramid
Properties E-Glass S-Glass Kevlar 49 Graphit Boron
Thermal conductivity W/mK 0.9 1.1 0.9 5 1
Linear expansion cm/cm°C×10-5 1.2 1.1 1 2 1
Tensile strength MPa 450 700 800 700 1600
Elastic modulus MPa 24 000 30 000 33 000 60 000 207 000
Many of the polymeric matrices will require some type of antioxidants to improve
aging properties. The most common primary antioxidants are hindered phenols
such as butylated hydroxytoluene (BHT). Typical low toxic antioxidants are
reported in Tables 2.40–2.42.
The toxicity of commonly used polymer stabilizers and additives are classified in
the following table.
2 Composite Materials 245
Table 2.41 Toxicity of commonly used polymer stabilizers and additives (3)
Stabilizer Toxicity
Fatty acid derivates of Low toxicity. Used in non-toxic medical
calcium, zinc applications
and magnesium
Barium fatty acid Moderately toxic
compounds
Lead & cadmium Highly toxic. Not recommended in the US for use in medical
derivates applications (cadmium pigments considered of low toxicity in
England)
Amines (antioxidants) Generally toxic, with aromatic amines showing carcinogenic
tendencies. Newer types less toxic
Butylated hydroxy toluene Considered non-toxic as also used in foods. Recently investigated
(BHT) and found non-carcinogenic
(antioxidant)
Octyl tin compounds Only class of tin compound classified as of low toxicity and used
in medical applications e.g. di-(n-octyl)tin maleate polymer
Table 2.42 Toxicity data on some common plasticizers used in plastic manufacture (3)
Plasticizer Toxicity
Adipates To date animal experiments indicate possible carcinogenicity
Glycolates Generally of low toxicity levels. However studies underway as
commercial form are phthalyl derivates
Phosphates Generally cause irritation to skin and mucous membranes
Phthalates Although commercially used in medical devices, environmental
effects and toxicological properties continually under investigation
Epoxidized soya bean oil Chelating type of plasticizer with low toxicity
The radiation resistance of common polymeric materials used as matrix for com-
posite are shown in Tables 2.43–2.52. Generally, polystyrene and urethane rubber
have the most resistance.
246 L. Ambrosio et al.
Table 2.43 The radiation resistance of common materials used as matrix of polymeric composites (3)
Material Stability effect
ABS Stable for single dose of 2.5 Mrad
Polyamides Suitable for single doses of 2.5 Mrad level
Polyethylene Stable under ordinary conditions at 2.5 Mrad
Polypropylene Embrittles – newer variations more resistant
Poly(vinyl chloride) Withstands single dose radiation cycle – but discolors – some HCl
liberated
Polystyrene Most radiation – stable of common polymers
Poly(tetrafluoroethylene) Poor resistance to radiation – copolymers less affected
Polysulfone Stable under ordinary conditions at 2.5 Mrad
Polyacetals Embrittles – discolors
Polyurethane Stable under ordinary conditions at 2.5 Mrad
Polymethylmethacrylate Embrittles – discolors
Rubber natural Stable under ordinary conditions when properly compounded
Rubber butyl Poor stability at low radiation levels
Rubber silicones Stable under ordinary conditions at 2.5 Mrad
Urethanes Excellent radiation resistance
Table 2.44 The radiation resistance of common materials used as matris of polymeric composites (3)
Material Thermoplastics Stability
Acrylonitrile/Butadiene/Styrene(ABS) Good
Cellulosics esters more stable Fair Undergoes chain scission,
Fluoinated ethylene propylene (FEP) Fair Copolymers more resistant than
Homopolymer
Polyacetal Poor Embrittles
Polyamides aromatic Excellent
Polyamides aliphatic Fair Hardens as levels increased
Polycarbonates Good Yellow – mechanical
properties unchanged
Polyesters (aromatic) Good
Polyethylene Good lowers melt flow
Polymethylmethacrylate Poor Degrades-turns brown
Polyphenylene sulfide Good
Polyproplyene Fair Improved stability if properly
stabilized
Polysulfone Excellent Yellow natural color
Polystyrene Excellent
Polytetrafluoroethylene Poor Acid evolved
Polyvinylchloride homopolymer Good If properly stabilized
Polyvinylchloride Copolymer Fair HCl evolved – turns brown
Styrene/ Acrylonitrile (SAN) Good More resistant than ABS
2 Composite Materials 247
Table 2.45 The radiation resistance of common materials used as matrix of polymeric composites (3)
Material Thermosetting resin Stability
Epoxies Excellent Very stable with the use of aromatic curing
agen
Phenol or urea formaldehyde Good
Polyimides Excellent
Polyesters Good
Polyurethanes Excellent
Table 2.50 Additives incorporated into natural rubber and as bound antioxidants (6)
Hours to absorb 1 % by wt. of O2
Additive Before extraction After extraction
N,N’-Diethyl-p-nitrosoaniline 39 30
p-Nitrosodiphenylaniline 60 53
p-Nitrosophenol 31 30
2,6-Diter-butyl-p-cresol 47 4
Table 2.51 Summary of effects of moisture and ambient aging on epoxy composites (7)
Flexural strength, MN/m2(ksi)
2 Composite Materials
Table 2.52 Summary of thermal aging of epoxy and polyimide system (7)
Aging Test
Material temperature temperature Aging Retention of
system Orientation K(°F) K(°F) time, h tensile strength, %
B/E [O] 394 (250) 450 (350) 1000 99
[CP] 94
[O] 450 (350) 450 (350) 1000 100
[CP] 100
G/E [O] 394 (250) 450 (350) 1000 94
[CP] 100
[0] 450 (350) 450 (350) 1000 100
[CP] 100
G/PI [O] 505 (450) 505 (450) 1000 98
[CP] 92
[O] 561 (550) 561 (550) 1000 87
[CP] 100
In recent years, carbon fiber has been recognized as a material with many exciting
applications in medicine. Several commercial products utilize carbon fiber as a rein-
forcing material which serves to enhance the mechanical properties of the poly-
meric resin systems in which it is included. The attractive feature of carbon
reinforced polymer for this application is that the orientation and fiber content can
be varied in the implant to provide the mechanical property orientation necessary
for good function. The carbon fiber can be distributed in matrix material to provide
strength in only those locations and directions where it is needed. The implant must
be designed in a way that fatigue failure does not occur and the matrix material is
not attacked by the physiological environment. The matrix materials used are listed
in Table 2.53.
These polymers were combined with unsized carbon fiber, into ±15° laminated,
test specimens approximately 2.5 cm×7.5 cm×0.3 cm. Testing was performed in
three point bending giving the results in Table 2.55.
Because a composite hip will be subjected to the physiological environment in
use, an accelerated test was performed to evaluate changes in properties. In this test,
the samples were immersed in 0.9 % saline solution maintained at 90°C for one
week; the results are shown in Table 2.56.
Blood compatible materials are essential to circulatory support devices.
Numerous materials have been considered for use in prosthetic devices.
2 Composite Materials 251
Table 2.53 Polymeric materials used as matrix for carbon fibers composites (3)
Polymer Polymer type Commercial name and manufacturer
Polysulfone Thermoplastic UDEL MG-11, Union Carbide, Dallas, TX
Poly-methyl methacrylate Thermoplastic PMMA I.V. 0.4, Rohm & Haas, Philadelphia,
PA
Epoxy (low viscosity) Thermoset Stycast 1267
Epoxy (high viscosity) Thermoset C-8W795 & H.R. 795, Hysol Corp.,
Los Angeles, CA
Carbons are widely used in prosthetic heart valves, as a result of their favorable
mechanical and biological properties. Pyrolytic carbons, deposited in a fluidized
bed, have high strength, and high fatigue and wear resistance. Compatibility with
blood and soft tissue is good.
252 L. Ambrosio et al.
Table 2.57 Representative Mechanical Properties of LTI, Glassy, and VaporDeposited Carbons (3)
Glassy Vapor deposited LTI carbon with
Property carbon carbon LTI carbon silicon (5–12%)
Density, g/cm3 1.5 1.9 1.9 2.1
Crystallite size, Å 30 10 35 35
Flexural strength 1000 psi 20 80 70 85
Young’s modulus, 106 psi 4.0 2.5 3.0 4.0
Strain to fracture, % 1.0 5.0 2.0 2.0
Fatigue limit/fracture strength 1.0 1.0 1.0 1.0
Strain energy to fracture, 1 12 7 9
100 psi
Virtually every operation requires the use of materials to close the wound for subse-
quent successful healing. The material must retain adequate strength during the
critical period of healing; it should also induce minimal tissue reaction that might
interfere with the healing process. The complexity involved in wound healing calls
for different types of wound closure materials.
The most common matrices are the low-density metals, such as aluminum and alu-
minum alloys, and magnesium and its alloys. Some work has been carried out on
lead alloys, mainly for bearing applications, and there is interest in the reinforce-
ment, for example, of titanium-, nickel- and iron- base alloys for higher-temperature
performance. However, the problems encountered in achieving the thermodynamic
stability of fibers in intimate contact with metals become more severe as the poten-
tial service temperature is raised, and the bulk of development work at present rests
with the light alloys.
2.18.2 Reinforcements
The principal reinforcements for metal matrices include continuous fibers of car-
bon, boron, aluminum oxide, silica, aluminosilicate compositions and silicon car-
bide. Some ceramic fibers are also available in short staple form, and whiskers of
carbon, silicon carbide and silicon nitride can be obtained commercially in limited
quantities. There is also interest in the use of refractory particles to modify alloy
properties such as wear and abrasion resistance. In this case, particle sizes and vol-
ume fractions are greater than those developed metallurgically in conventional
alloys, and incorporation of the particles into the metal is achieved mechanically
rather than by precipitation as a consequence of heat treatment. Most metal-matrix
composites consist of a dispersed reinforcing phase of fibers, whiskers or particles,
with each reinforcing element ideally separated from the next by a region of metal.
A summary of properties of the most important metal matrix composites is pre-
sented in Tables 2.59–2.69.
Table 2.60 Summary of Transverse Tensile Strengths of Various Aluminum- Graphite Composite
Systems (5)
Composite Average
Number
Fiber Matrix (MN/m2) (psi) High (psi) Low (psi) of tests
Thornel-50 Al-12Si 26 3777 6500 433 9
Courtaulds 220 Al 42 6117 8690 3760 20
Courtaulds 356 Al 70 10,008 14,600 5500 26
Courtaulds HM Al -10Mg 29.5 4280 4500 3600 5
Whittaker-Morgan 356 Al 50 7300 11,300 4100 5
Whittaker-Morgan 7075 Al 21 3040 5100 400 5
Table 2.61 Uniaxial Tensile Data for Aluminum-Silicon Alloy-Thornel-50 Composite Thermally
Cycled 20 Times from -193 to +500°C (5)
Sample number Ultimate tensile strength (psi) Rule-of-mixture strength (%)
C7 103 000 103
C8 100 000 99
C9 100 000 00
C10 99 000 99
Average 101 000 100
Table 2.69 Summary of Mechanical Properties of Zinc and Zinc-Graphite Composite (5)
Modulus of
Strength elasticity Density Strength density Modulus/density
System (psi) (10-6 psi) lb/in.3 10-6in. (10-6 in.)
Z-35 v/o 110 900 16.9 0.191 0.58 88.5
Thornel/75
Alloy AG40A 41 000 10.0 0.240 0.17 41.7
Composite structures in ceramics have been developed for two major reasons. First,
they provide a means to enhance dramatically the performance of the so-called
functional ceramics; these are systems where electrical, dielectrical, piezoelectric or
sensitizing properties are greatly amplified by appropriate composite design.
Secondly, they are used to avoid or diminish the brittle behaviour of structural
ceramic systems.
A summary of properties of the most important ceramic matrix composites is
presented in Tables 2.70–2.74.
2 Composite Materials 257
Table 2.74 Room temperature of some unreinforced ceramics and ceramic matrix composite (4)
Material Flexural strength MPa Fracture toughness MPa m1/2
Ai2O3 550 4–5.0
SiC whiskers/Al2O3 800 8.7
SiC 500 4.0
SiC fibers/SiC 750 25.0
ZrO2 200 5.0
SiC/ZrO2 450 22.0
Borosilicate glass 60 0.6
SiC fibers/borosilicate glass 830 18.9
Glass ceramic 200 2.0
SiC fibers/glass ceramics 830 17.0
Reaction bonded Si3N4 260 2–3.0
SiC whiskers/reaction bonded 900 20
Si3N4
Hot pressed Si3N4 470 3.7–4.5
SiC whiskers/hot pressed Si3N4 800 56
References
4. Islam, M.U., Wallance, W. and Kandeil, A.Y. (1985) Artificial Composites for High
Temperature Applications, Noyes Data Corporation, Park Ridge, New Jersey, USA.
5. Kreider, Kenneth G. (1974) Metallic Matrix Composites, Vol. 4, Academic Press, New York
and London.
6. Hawkins, W.L. (1984) Polymer Degradation and Stabilization, Springer-Verlag.
7. Environmental Effects on Advanced Composite Materials, symposium presented at the
seventy-eighth Annual Meeting ASTM, Philadelphia, 1975.
8. Hill, R. (1964) J. Mech. Phys. Solids, 12, 199.
9. Hermann, J.J. (1970) Proc. Konigl. Nederl. Akad. Weteschappen Amsterdam, B70, 1.
10. Halpin, J.C. and Pagano, N.J. (1964) J. Compos. Mater., 3, 720.
Chapter 3
Thermoplastic Polymers In Biomedical
Applications: Structures, Properties
and Processing
3.1 Introduction
conventional polymers. They may come from a batch intended for general purposes,
but are selected on the basis of clean condition or trace element analysis or
mechanical properties. Subsequent processing requires clean room conditions and
care to avoid any contamination. There is still some inherent uncertainty about con-
stituents unless there has been complete disclosure and/or only a ‘pure’ polymer is
used. With new developments in polymeric biomaterials, the situation should
improve.
It is hoped that the following sections will be of value to researchers in science
and engineering and to clinical practitioners who are engaged in the development
and material selection of new thermoplastic polymers for biomedical applications.
3.2 Polyethylene
3.3 Polypropylene
3.4 Polyurethane
3.5 Polytetrafluoroethylene
3.6 Polyvinylchloride
a. high mechanical strength, rigidity and hardness, b. low impact strength in unmod-
ified form, c. translucent to transparent (depending on method of manufacture), d.
good electrical properties in the low voltage and low frequency range, and e. high
chemical resistance.
3.7 Polyamides
Chemically, the polyamides may be divided into two types: a. those based on
diamines and dibasic acids, and b. those based on amino acids or lactams (Chapman
and Chruma, 1985). Commercial use of nylons is dominated by two products, one
from each type, nylon 66 and nylon 6 from Є-caprolactam.
Aliphatic polyamide is linear and easy to crystallize but crystallinity varies
widely with conditions. Crystalline content may be 50–60% by slow cooling and
10% by fast cooling. High interchain attraction in the crystalline zones and flexibil-
ity in the amorphous zones leads to polymers which are tough above their apparent
glass transition temperatures (Brydson, 1982).
Polyamides have excellent fibre-forming capability due to interchain H-bonding
and a high degree of crystallinity which increases strength in the fibre direction
(Park and Lakes, 1992). Polyamides are hygroscopic and lose strength in vivo. The
amorphous region of polyamide chains is sensitive to the attack of water. The greater
the degree of crystallinity, the less the water absorption and hence the less the effect
of humidity on the properties of the polymer. The reversible absorption of water is
associated with a change in volume and thus of dimensions.
The mechanical properties of moulded polyamide materials depend on molecu-
lar weight, crystallinity and moisture content. In the dry, freshly molded state, all
polyamide grades are hard and brittle. When conditioned they are tough and wear
resistant. High melting points result in good mechanical properties up to tempera-
tures in the region of 120–150 °C.
They are only soluble in a few solvents (formic acid, glacial acetic acid, phenols
and cresols), of similar high solubility parameter. Nylons are of exceptionally good
resistance to hydrocarbons. Esters, alkyl halides, and glycols have little effect on
them. Alcohols can swell the polymers and sometime dissolve some copolymers.
Mineral acids attack the nylons but the rate of attack depends on the type of nylon
and the nature and concentration of the acid. Nitric acid is generally active at all
concentrations. The nylons have very good resistance to alkalis at room temperature.
Resistance to all chemicals is more limited at elevated temperature (Brydson, 1989).
Generally, polyamides are characterized by: a. high strength, stiffness and
hardness, b. high heat distortion temperature, c. high wear resistance, good slip and
dry-running properties, d. good damping capacity, e. good resistance to solvents,
fuels and lubricants, f. non-toxicity, g. good processability, h. aliphatic polyamides
are partially crystalline and thus opaque, and i. moisture content impairs mechanical
properties and affects dimensions of moldings.
3.8 Polyacrylates
Polyacrylates are based on acrylic acid, methacrylic acid, and their esters.
Among them, polymethylmethacrylate (PMMA) and polyhydroxy ethylmethacry-
late (PHEMA) have found wide applications as biomedical materials. The clinical
3 Thermoplastic Polymers In Biomedical Applications: Structures, Properties… 269
history of polyacrylates began when it was unexpectedly discovered that the frag-
ments of PMMA plastic aircraft canopies stayed in the body of the wounded without
any adverse chronic reactions (Jones and Denning, 1988; Park and Lakes, 1992).
In normal conditions, PMMA is a hard transparent material. Molecular weight is
the main property determinant. High molecular weight PMMA can be manufactured
by free radical polymerization (bulk, emulsion, and suspension polymerisation).
Bulk polymerization is used for cast semi-finish products (sheet, profiles and even
tubes), and the cast polymer is distinguished by superior mechanical properties and
high surface finish (Brydson, 1982 and Domininghaus, 1993). Cast material has a
number average molecular weight of about 106 whilst the Tg is about 106°C. The
extensive molecular entanglement prevents melting below its decomposition tem-
perature (approx. 170°C).
An amorphous polymer, PMMA has a solubility parameter of about 18.8 MPa½
and is soluble in a number of solvents with similar solubility parameters. Solvents
include ethyl acetate (δ: 18.6 MPa½), ethylene dichloride (δ: 20.0 MPa½), trichloro-
ethylene (δ: 19 MPa½), chloroform (δ: 19 MPa½), and toluene (δ: 20 MPa½). The
polymer is attacked by mineral acids but is resistant to alkalis, water and most aque-
ous inorganic salt solutions. A number of organic materials although not solvents
may cause crazing and cracking (e.g. aliphatic alcohols).
The characteristic properties of PMMA are, a. high hardness, stiffness and
strength, b. homopolymers are brittle, copolymers are tough, c. scratch-resistant, high
gloss surface capable of being polished, d. water-white transparency, copolymers
exhibit inherent yellowish color, e. high heat distortion temperature, f. good electrical
and dielectric properties, g. resistant to weak acids and alkaline solution as well as to
non-polar solvents, grease, oils and water, h. susceptible to stress cracking, i. flam-
mable, j. good processability and machinability, k. rather low resistance to creep at
temperature only slightly above room temperature, and l. high melt viscosity due to
the high chain stiffness caused by restricted rotations about the C-bonds in the back-
bone chains
3.9 Polyacetal
Polyacetal can be divided into two basic types, acetal homoploymer and acetal
copolymer. Both homopolymer and copolymer are available in a range of molecular
weights (Mn = 20 000–100 000). The homopolymer is a polymer of formaldehyde
with a molecular structure of repeated oxymethylene units (Staudinger, 1932).
Large-scale production of polyformaldehyde, i.e. polyacetal, commenced in 1958 in
the USA (US Patent 2 768 994, 1956) (British patent 770 717, 1957). Delrin (1959)
was the first trade mark for this polymer by Du Pont Company. The copolymers
were introduced by the Celanese Corporation of America, and the first commercial
product named Celcon (1960). One of the major advantages of copolymerization is
to stabilize polyacetal because the homopolymer tends to depolymerize and elimi-
nate formaldehyde. The most important stabilization method is structural modifica-
tion of the polymer by, for example, copolymerization with cyclic ether.
270 S.H. Teoh et al.
As can been seen polyacetal has a very simple structure of a polyether. Unlike
polyethylene, polyacetal has no branching, and its molecules can pack more closely
together than those of polyethylene. The resultant polymer is thus harder and has a
higher melting point than polyethylene (175°C for homopolymer), exhibiting a high
crystallinity (77–85%).
No effective solvents have been found for temperatures below 70°C. Swelling
occurs with solvents of similar solubility parameter (δ: 22.4 MPa½). However, poly-
acetal should be kept away from strong acids, strong alkalis, and oxidizing agents.
Water can not degrade it but may swell it or permeate through it and affect the
dimensions of its products. Prolonged exposure to ultra-violet light will induce sur-
face chalking and reduce the molecular weight of the polymers. Polyacetals, both
homopolymer and copolymer are also radiation sensitive. The radiation damage
threshold is estimated at 0.5 Mrad with 25% damage at 1.1 Mrad (Szycher, 1991).
Generally, the polyacetals have the following characteristics, a. high tensile
strength, shear strength, stiffness, and toughness, b. predictable stress/ strain rela-
tionships, c. predictable dimensional behavior, d. chemical and corrosion resistance,
e. abrasion resistance, f. light weight and good appearance, g. acceptability for food
contact application (most grades), h. ease of processing, and i. competitive costs.
The enormous commercial success of the polyacetals is owed to their very high resis-
tance to creep and fatigue. The acetal resins show superior creep resistance to the nylons.
3.10 Polycarbonate
eral acids and dilute alkaline solutions other than caustic soda and caustic potash.
Where the resin comes into contact with organophilic hydrolysing agents such as
ammonia and the amines, the benzene rings give little protection and reaction is
quite rapid. The absence of both secondary and tertiary C-H bonds leads to a high
measure of oxidative stability. Oxidation takes place only when thin films are heated
in air to temperatures above 300°C.
Typical properties include: a. low density, b. high strength, stiffness, hardness
and toughness over the range from -150 to + 135°C unreinforced and from -150 to
+145°C when reinforced, c. crystal clear transparency, high refractive index, high
surface gloss, d. can be colored in all important shades, transparent, translucent or
opaque with great depth of color, e. good electrical insulation properties which are
not impaired by moisture, f. high resistance to weathering for wall thicknesses
greater than 0.75 mm, g. high resistance to high energy radiation, and h. self-
extinguishing after removal of the ignition source.
The main disadvantages are: a. processing requires care, b. limited chemical
resistance, c. notch sensitivity and susceptibility to stress cracking.
3.12 Polyetheretherketone
3.13 Polysulfone
Polysulfone is a polymer which has properties matching those of light metals (Park
and Lakes, 1992). The first commercial polysulfone was introduced in 1965 by
Union Carbide as Bakelite Polysulfone, now called Udel®. In 1967 3M offered
Astrel 360 referred to as a polyarylsulfone. In 1972 ICI introduced a polyethersul-
fone Victrex®. A high toughness polysulfone was released in the late 1970s by
3 Thermoplastic Polymers In Biomedical Applications: Structures, Properties… 273
Union Carbide. Although the commercial polymers are linear and most have regular
structures they are all principally amorphous. The backbone aromatic structure
leads to high values of the glass transition temperature between 190 and 230°C. The
Union Carbide materials have a secondary transition at -100°C and the ICI polymer
at -70°C. Typical Mn values are ca. 23 000. Commercial materials are described
variously as polysulfones (Udel), polyarylsulfones (Astrel), polyether sulfones or
polyarylethersulfones (Victrex) (Brydson, 1982).
The polymer is manufactured from bisphenol A and 4, 4-dichlorosulphonyl
sulfone by multi-step condensation. The most distinctive feature of the backbone
chain of those polymers is the diphenylene sulfone group. The sulphur atom in
each group is in its highest state of oxidation and tends to draw electrons from
the adjacent benzene rings, hence resisting any tendency to lose electrons to an
oxidizing agent. Polysulfones thus show outstanding oxidation resistance. The
aromatic nature of the diphenylene sulphone can absorb considerable energy
applied as heat or radiation and so resists thermal degradation. The diphenylene
sulfone group thus confers on the entire polymer molecule the inherent charac-
teristics of thermal stability, oxidation resistance, and rigidity at elevated
temperatures.
The potential for energy dissipation confers good impact strength and ductility
down to -100°C with high elongation to break and tensile strength. Under most
conditions, impact properties rival those of bisphenol A polycarbonate. Unlike
polycarbonate, however, polysulfone can exhibit excellent resistance to hydrolysis
or reduction of molecular weight even at elevated temperatures. Tests on the hydro-
lysis stability of polysulfones have been carried out up to 10 000 hours without
observed loss of molecular weight.
The polymers are stable in aqueous inorganic acids, alkalis, salt solutions,
aliphatic hydrocarbons, and paraffin oils, are transparent, capable of steam steril-
ization, and free from taste and smell. They should not come in contact with
ketones, aromatic solvents, chlorinated hydrocarbons, and polar organic solvents.
They may show stress crazing on exposure to steam or water. A polyethersulfone,
however, exhibited no crazing even after 300 hours and retained 90% of initial
tensile impact strength. For a thermoplastic material, creep is low at moderate
temperatures but is significant at temperatures approaching the glass transition.
However, the wear properties of this material are not as good as PE and POM
(Teoh et al., 1994).
Generally polysulfone has the following characteristic properties, a. high strength,
stiffness and hardness between -100 and +150°C short-term to 180°C, b. high ther-
mal stability and heat distortion temperature, c. crystal clear (slightly yellowish)
transparency, d. high processing temperature, e. high melt viscosity, f. high chemical
resistance, g. susceptibility to stress cracking with certain solvents, h. high resistance
to β-, γ-, X- and IR-radiation, i. high transmittance for microwaves, and j. high flame
resistance and low smoke development.
Table 3.1 Chemical structures of thermoplastic polymers in biomedical applications
274
S.H. Teoh et al.
(continued)
3 Thermoplastic Polymers In Biomedical Applications: Structures, Properties… 275
b. Mechanical properties
Mechanical property Unit PE PP PU PTFE PVC PA
Bulk modulus GPa 0.8–2.2 1.6–2.5 1.5–2 1–2 3–4 2.4–3.3
Tensile strength MPa 30–40 21–40 28–40 15–40 10–75 44–90
Elongation at break % 130–500 100–300 600–720 250–550 10–400 40–250
Young’s modulus GPa 0.45–1.3 1–1.6 0.0018–0.009 0.3–0.7 1.0–3.8 1.4–2.8
Elastic limit MPa 20–30 20–33 28–40 15–30 23–52 40–58
Endurance limit MPa 13–19.6 11–18.2 21–30 9–18 13.8–31.2 22–31.9
Fracture toughness MPa m½ 2.2–4 1.7–2.1 0.1–0.4 2.5–3 1–4 1.8–2.6
Hardness MPa 60–90 60–100 50–120 27–90 70–155 100–160
Thermoplastic Polymers In Biomedical Applications: Structures, Properties…
b. Mechanical properties
Mechanical property Unit PMMA POM PC PET PEEK PS
Bulk modulus GPa 3–4.8 4–5.6 2.8–4.6 3–4.9 4–4.5 3.8–4.6
Tensile strength MPa 38–80 70–75 56–75 42–80 70–208 50–100
Elongation at break % 2.5–6 15–80 80–130 50–300 1.3–50 25–80
Young’s modulus CPa 1.8–3.3 2.55–3.5 2–2.9 2.2–3.5 3.6–13 2.4–2.9
Elastic limit MPa 35–70 65–72 53–75 50–72 12–60 58–70
Endurance limit MPa 19.3–38.5 28–42 29.2–41.3 30–43.2 33–36 34.8–42
Fracture toughness MPa m½ 0.8–1.3 1–1.5 2.5–3.2 1.2–2 2.3–2.5 1.3–2
Hardness MPa 100–220 110–220 110–180 97–210 100–120 180–240
Compressive strength MPa 45–107 70–80 100–120 65–90 80–120 72–100
Poisson’s ratio 0.4–0.43 0.38–0.43 0.39–0.44 0.38–0.43 0.38–0.43 0.38–0.42
Shear modulus GPa 0.6–1.2 0.79–1 0.95–1.05 0.83–1.1 1.2–1.4 0.8–1
c. Thermal properties
Thermal property Unit PE PP PU PTFE PVC PA
Service temperature in °C 90–130 140 80–130 300 55–100 130–200
air without mechanical
loading (short-term)
Service temperature in air °C 70–100 100 60–80 250 50–85 70–120
without mechanical
loading (long-term)
Minimum service °c −63 to −53 −123 to −23 −123 to −23 −263 to −253 −43 to −28 −60 to −50
temperature
Melting(Tm)/decomposing °c 125–135 160–180 180–250* 322–327 150* 220–267
(Td*) ranges
Glass transition °c −113 to −103 −30 to −3 −73 to −23 20 to 22 −23 to 90 20 to 92
S.H. Teoh et al.
temperature T
3
c. Thermal properties
Thermal property Unit PE PP PU PTFE PVC PA
Softening temperature °C 40–50 70–100 100 40–110 80–200
Specific heat J/g.K 1.95–2.20 1.70–2.35 0.4–1.76 1.00–1.01 0.85–1.80 1.26–1.8
Thermal expansion 106/K 100–200 80–200 150–210 100–150 60–210 80–150
Thermal conductivity W/m K 0.42–0.52 0.12–0.24 0.29–1.80 0.19–0.25 0.13–0.26 0.23–0.29
Lupolen (BASF, DE) Surnikathene (Surnitorno Chern., JP) Norchern (USllExxon Chern., US)
Lotrex (Orkern Norsolor SA, FR) Ultzex (Mitsui Petrochemical, JP) Rexene (EI Paso Chern., US)
News (Neste Oy Chern., FI) Starnylex (DSM, NL) Tenite (Eastman Chern., US)
Visqueen (lCI, GB) Ladene (SABlC, Saudi Arabia)
285
2. Polypropylene (PP)
Asota, Ecofelt (Chemic Linz, AT) Eltex (Solvay, FR) Starnylan (DSM, NL)
Marlex (Phillips Petroleum Co., BE) Carlona (Shell, GB) Bicor (Mobil Chern., US)
Istrono (Chern., Werke J. Dirnitrow, CS) Propathene (ICI, GB) Extrel, Twistlock, Vistalon(Exxon
Chern, US)
Tatren (Petrirnex, CS) Biofol (Chern. Kombinat. Tisza, HU) Fortilene (Soltex Polymer Corp. US)
Hostalen PP (Hoechst, DE) Afax, Moplen, Valtec (Hirnont, IT) Liteplate S (Hercules, US)
Platilon (Deutsche A TOchern., DE) Bifax (Showa denko. JP) Profax (Hirnont, US)
Trovidur (HulsTroisdorf, DE) Eperon (Kanegafuchi Chem., JP) Rexene (EI Paso Chem., US)
Ultralen (Lonza Werke, DE) Noblen (Mitsubishi Petrochemical, JP) Tenite (Eastman Chemical, US)
Vestolen P (Huls, DE) Novatec (Mitsubishi Chem., JP)
Apryl, Lacqtene P (Atochem., FR) PolyPro, Sunlet (Mitsui Petrochem., JP)
3. Polyurethane (PU)
General
Ucefix, Ucellex (UCB, BE) Europolymers (Avalon Chemical Co., GB) Urafil (Akzo-Wilson-Fiberfil, US)
Fabeltan (Tubize Plastics, BE) Jectothane (Dunlop Holdings, GB) Hi-Tuff (J.P. Stevens & Co., US)
Caprolan, Elastolen, Elastolan (Elastogran Pemullex (Pemu Chemolimpex, HU) Esteloc, Estane, Roylar
Polyurethane, DE) (B.F. Goodrich Chemical Co., US)
Desmopan (Bayer, DE) Uthane (Urethanes India, IN) Irogran, Plastothane (Morton
Thiokol, US)
Lurollex (Lehmann u. Voss Co., DE) Pelprene (Toyobo Co., Resins Div., JP) Proplastic (Pro Lam, US)
Oldopren (Busing u. Fasch & Co., DE) Durane (Swanson, US) Q-Thane (Quinn & Co., US)
Durelast (B & T polymers, GB)
Medical special
Angiollex (Abiomed Danvers, MA) Mitrathane (PolyMedica Burlington, MA, US) Corplex (Corvita Miami, FL, US)
Biomer (Ethicon Somerville, NJ, US) Pellethane (Dow Chemical La Porte, TX, US) Unithane 80F (NICPBP Beijing,
S.H. Teoh et al.
China)
3
3. Polyurethane (PU)
Cardiothane (Kontron Everett, MA, US) Surethane (Cardiac Control Palm Coast, FL, US) Corethane (Corvita Miami, FL, US)
Chronollex (PolyMedica Burlington, MA, US) Tecollex (Thermedics Inc Woburn, MA, US) PU 10 (Univ. NSW, Australia)
Hemothane (Sams, Div 3M Ann Arbor, MI, US) Toyobo TM5 (Toyobo Co. Osaka, Japan)
4. Polytetralluoroethylene (PTFE)
Hostallon TF (Hoechst, DE) Neollon (Daikin Ind., JP) RT duroid (Rogers Corp., US)
Pamllon (Norton Pampus, DE) Polyllon (Daikin Kogyo Co., JP) Rulon (Dixon Ind. Corp., US)
Forallon (A TOCHEM, FR) Ferrotron, F1uorosint (Polymer Corp., US) Salox (Allegheny Plastics, US)
Sorellon (Pechiney U.K., FR) F1uoroloy (F1urocatbon, US) Tellon (Du Pont de Numours, US)
Gallon (Plastic Omnium, FR) FluoroMet (Raymark, US) Turcite (W.S. Shamban & Co., US)
Fluon (ICI, GB) Goretex (W.L. Gore Assoc., US) Tygallor (American Cyanamid
Aerospace, US)
Algollon (Montedison, IT) Halon (Ansimont, US) Xylon (Whitford Corp., US)
5. Polyvinylchloride (PVC)
Benvic (Solvay, BE) Corvic, Vynide, Welvic (ICI, GB) Rosevil (Chern. Kombinat Borzesti, RO)
Plastilit, Selchim, Solvic (Solvay & Cie., BE) Ongrovil (Barsodi Veggi Komb., HU) Ensolite (Uniroyal Chern., US)
Vipopham (Lonza, CH) Ravinil, Sicron, Vipla, Viplast (EniChem, IT) Ethyl (Ethyl Corp., Polymer Div., US)
Astralon, Trocal, Trosiplast, Trovidur, Vestolit Vixir (S.I.R. (Montedison), IT) Fiberloc, Geon (B.F. Goodrich, US)
(Huls-Troisdorf, DE)
Decelith (VEB Ellenburger Chemiewerk, DE) Hishiplate (Mitsubishi Plastics Ind., JP) Pliovic, Vinacel, Vycell (Goodyear, US)
Vinnol (Wacker Chemie, DE) Kureha, Viclon (Kureha Chern. Ind., JP) Vygen (General Tire & Rubber Co., US)
Genopak, Genotherm, Hostalit (Hoechst, DE) Vinychlon (Mitsui Toatsu Chern., JP) Vynaloy (B.F. Goodrich Chern. Co., US)
Thermoplastic Polymers In Biomedical Applications: Structures, Properties…
Armodur (Rhone-Poulenc, FR) Vinylfoil (Mitsubishi Gas Ind., JP) Hipnil (Hemijska Industria, YU)
Bipeau, Orgavyl, Polycal (ATOCHEM., FR) Varian (DSM, NL) Jugotherm, Juvinil (Jugovinil, YU)
Ekavyl (PCUK, FR) Norvinyl, Pevikon (Norsk Hydro, NO) Zadrovil (Polikem, YU)
Carina, Duraftex (Shell Intern. Chern Co., GB) Oltivil (Chern. Komb. Pitesti, RO)
287
(continued)
Table 3.4 (continued)
288
6. Polyamide (PA)
PA 6
Fabenyl (Tubize Plastics, Maranyl (ICI, GB) Akulon (Engineering Plastics of AKZO
Plastics, NL)
Grilon (Ems Chemie, CH) Latamid (L.A.T.I., IT) Capron (Allied Signal Engn. Plastics, US)
Durethan B (Bayer, DE) Nivionplast (EniChem, IT) Plaskon (Plaskon Moldings Div., US)
Ultramid B (BASF, DE) Renyl (Snia Technopolimeri, IT) Zytel (Du Pont de Nemours, US)
Orgamide (ATOCHEM, FR) Amilan, Amilon (Toray Ind., JP)
Technyl C (Rhone-Poulenc Specialites Chimiques, FR) Torayca (Toray Ind., JP)
PA 66
Durethan A (Bayer, DE) Leona (Asahi Chemical Ind., JP) Minion (Du Pont de Nemours, US)
Technyl A (Rhone-Poulenc Specialites Chimiques, FR) Torayca (Toray Ind., JP) Zytel (Du Pont de Nemours, US)
Maranyl A (ICI, GB) Akulon (Engineering Plastics of
AKZO Plastics, NL)
7. Polyacrylates
Umaplex (Synthesia, CS) Asterite (ICI, GB) Sumipex (Sumitomo Chern. Co., JP)
Acrifix (Rohm, DE) Diakon (ICI, GB) Casoglas (Casolith, NL)
Lucryl (BASF, DE) Perspex (ICI, GB) Acrylite (Cy/Ro Industries, US)
Degaplast (Degussa, DE) Unilok (British Vita Co., GB) Lucite (Du Pont de Nemours, US)
Deglas, Dewglas (Degussa, DE) Vetredil (Vetril, IT) Corian (Du Pont de Nemours, US)
Dewoglas (Degussa, DE) Vedril (Montedison, IT) Crofon (Du Pont de Nemours, US)
Paraglas (Degussa, DE) Acrypanel (Mitsubishi Rayon Co., JP) Electroglas (Glasftex Corp., US)
Plexidur, Plexiglas (Rohm, DE) Delmer, Depet (Asahi Chem. Ind., JP) Exolite (Cyro Industries, US)
Resarit (Resart, DE) Eska (Mitsubishi Rayon, JP) Gardlite (Southern Plastics Co., US)
Vestiform (Huls, DE) Palapet (Kyowa Gas Chem., JP) Oroglas (Rohm & Haas Co., US)
Altuglas (Altulor, Orekem, FR) Shinkolite (Mitsubishi Rayon Co.,JP) Swedcast (Swedlow, US)
S.H. Teoh et al.
3
9. Polycarbonate (PC)
SparJux (Solvay & Cie., BE) Sinvet (EniChem, IT) Lexan (General Electric Plastics, US)
Durolon (Policarbonateos do Brazil, BR) Novarex (Mitsubishi Chem. Ind., JP) MerIon (Mobay Chemical Corp .. US)
Makrolon (Bayer, DE) Panlite (Teijin Chemicals, JP) Polycarbafil (Akzo-WilsonFiberfil, US)
Orgalan (A TOCHEM, FR) Xantar (DSM, NL) Polygard (Poly tech, US)
Star-C (Ferro Eurostar, FR) Calibre (Dow Chemical Corp., US) Stat-Kon (LNP Corp., US)
Royalite (British Vita Co., GB) Ekonol (Carborundum Co., US)
10. Polyethyleneterephthalate
(PET)
Crastin (Ciba Geigy, CH) Melinar, Melinite (lCI, GB) Pelion (Mobay Chem. Corp., US)
Grilpet (Ems Chemie, CH) Amite (Akzo Engng. Plastics, NL) Petra (Allied Signal, US)
Impet (Hoechst, DE) Etar (Eastman Chem. Intern., US) Ropet (Rohm & Hass Co., US)
Ultradur (BASF, DE) Mindel (Amoco Performance Products, US) Rynite (Du Pont de Nemours, US)
11. Polyetheretherketone
(PEEK)
Thermoplastic Polymers In Biomedical Applications: Structures, Properties…
12. Polysulphone
U. Polysulphone Stabar (ICI, UK) Udel (Amoco Performance Products, US)
289
References
4.1 Introduction
for every application. Some have been found to perform successfully under static
conditions but fail or perform undesirably under dynamic situations. Often, the
design of a device and the demands upon it will determine if the elastomer chosen
is the proper selection. Therefore the material and its use are inseparable. They must
be evaluated together. Merely passing an array of physical and biological tests do
not confer success. Biocompatibility is an essential element of medical grade elas-
tomers. A set of compatibility tests determine the general physiological acceptabil-
ity of an elastomer. These consist of passing USP Class VI tests. Additional testing
may be required depending upon the device, its area of application and the time it is
in contact with tissues. A Master File is often registered by the manufacturer of the
basic elastomer with the FDA to attest to its properties, composition and response
to biological testing. Demands on medical device manufacturers have never been
more stringent. Regulatory pressures, more indepth testing, the threat of litigation
plus the constraints of health care cost containment are affecting all aspects of the
design and development process and the availability of some biomedical elasto-
mers. A variety of elastomeric materials are available to meet the design challenges
presented by medical devices. However, there is still a need for even better
materials.
The elastomers that are listed here should be considered in light of their suitabil-
ity for a specific application. The properties tables should serve as a guide to design
options for those in the early stages of the development process. Keep in mind that
these properties listed in the tables and the compatibility standings are only indica-
tive of the performance characteristics that an elastomer may exhibit.
Thermoplastic elastomers (TPEs) are a special class of materials that process simi-
larly to other thermoplastic polymers, yet possess many of the desirable properties
of thermoset elastomers. Some TPEs are elastomeric alloys consisting of cross-
linked particles of rubber surrounded by a thermoplastic matrix. Others consist of
block copolymers and are typified by polyurethanes and styrene polymers.
4 Biomedical elastomers 293
Thermoplastic vulcanizates
Thermoplastic vulcanizates are a separate class of thermoplastic elastomers (TPEs)
with Santoprene® as the representative biomedical elastomer.
Santoprene®
This thermoplastic vulcanizate is an olefin based elastomer; an elastomeric alloy. It
is totally synthetic and does not contain any natural rubber thereby avoiding many
of the allergic reaction problems associated with natural rubber latex. It exhibits
outstanding flex-fatigue resistance, low temperature flexibility (-40 °C) and resis-
tance to tearing and abrasion. Its resistance to plastic deformation under tensile and
compression stress is another of its features. Santoprene® is reported to be superior
to natural rubber in some situations and replaces silicone elastomers in others. It has
found use in peristaltic pump tubing, syringe plungers, seals, and caps, tracheal and
enteral tubing, vial closures and pump seals, disposable anesthetic hoses, intrave-
nous delivery systems, and other hospital devices. Santoprene® has met USP Class
IV requirements for in vivo biological reactivity and conforms to the Tripartite
Biocompatibility Guidance standards. However, the manufacturer does not recom-
mend Santoprene® for use as part of human implants. The material may be injection
molded, extruded, blow molded and thermoformed. For details on physical proper-
ties, processing and biocompatibility see Tables 4.1, 4.2, 4.13 and 4.14.
Polyurethane-based elastomers
Polyurethanes are another class of TPEs. They are a large family of chemical com-
pounds that can consist of ether-based, ester-based, polycarbonate-based or poly-
propylene-based varieties. A number of copolymers are also included;. polyurethanes
are combinations of macroglycols and diisocyanates that have been polymerized
into tough and elastic materials. TPE polyurethanes have been used for peristaltic
pump tubing, parenteral solution tubing and catheters. The tables list the majority of
those that are commercially available. Among others are those either of limited sup-
ply, available for proprietary use only or that have been successful, but recently
discontinued such as:
• Hemothane Sarns Div. of 3M. Restricted to proprietary use.
• Biomer Ethicon, Inc. No longer available through this source.
• Surethane Cardiac Control Systems, Inc. Redissolved Lycra® thread. Some for-
mulations may have a few percent PDMS blended with it. Limited availability.
• Pellethane™ 2360 Dow Chemical, USA. This material is no longer available for
medical implant use (see also Pellethane™).
• Angioflex ABIOMED, Danvers, Mass. Restricted to proprietary use.
• Cardiothane Kontrol, Inc. A silicone-urethane interpenetrating polymer network.
Limited availability.
Internationally, polyurethanes for medical use have been developed by Italy,
China and Japan.
Biospan®
This TPE is a segmented polyurethane and is reported to be not significantly differ-
ent from biomer in chemistry and molecular weight. It is a polytetra-methyleneoxide-
based aromatic polyurethane urea with mixed aliphatic and cycloaliphatic diamine
chain extenders. A copolymer of diisopropylamino-ethyl methacrylate and decyl
methacrylate are added as a stabilizer. The material is supplied as 25% solids in
dimethyl acetamide solvent (Tables 4.3, 4.12, 4.13, and 4.14).
Biospan®-S
This is a silicone modified analog of Biospan® with a different stabilizer. It pos-
sesses a silicone-rich surface to enhance thromboresistance while maintaining the
bulk properties of Biospan® (Tables 4.3, 4.12, 4.13, and 4.14).
Biospan®-D
This is another version of Biospan® with surface modification by an oligomeric
hydrocarbon covalently bonded to the base polymer during synthesis. The additive
has a pronounced effect on the polymer surface chemistry but little effect on the
bulk properties of the base polymer according to the manufacturer (Tables 4.3, 4.12,
4.13, and 4.14).
4 Biomedical elastomers
(continued)
Table 4.3 (continued)
298
Property
Tensile Modulus
Specific Durometer strength, Elongation, ASTM Tear strength Compression set,
gravity hardness shore psi percent D-412 pli,die C percent
Dow Chemical 2363-80AE 1.12 85 A 4200 650 890 100 420 30
Co. 2363-80A 1.30 81 A 6860 670 970 100
R0120
2363-90A 1.14 90 A 5850 500 1700 100 570 25
2363-90AE 1.14 90 A 6000 550 1475 100 540
PolyBlend™ PB1000-650 65 D to 75 D6500 350 5300 100 - -
1000 and PB1100-55 1.02 55 A 2150 800 135 100 140 55-66
PolyBlend™ PB1100-60 1.02 60 A 2400 210 100 150 50-60
1100,
Poly Medica PB1100-75 1.02 75 A 3250 575 420 100 240 45-50
Biomaterials, Inc. PB1100-80 1.02 80 A 4600 590 555 100 330 25-30
EG60D 1.09 51 D 7829 363 2000 100
Tecoflex®, EG60D-B20 1.32 55 D 7484 370
EG65D 1.10 60 D 8074 335 2500 100
EG65D-B20 - 63 D 6986 321
Thermedics, EG68D 1.10 8686 332
Inc. EG72D 1.11 67 D 7739 307 3400 100
EG80A 1.04 72 A 5640 709 400 100
Thermedics, EG80A-B20 1.24 73 A 5571 715
Inc. EG85A 1.05 77 A 6935 565 700 100
EG85A-B20 1.25 83 A 5282 632
EG85A-B40 1.51 84 A 5093 559
EG93A 1.08 87 A 7127 423 1100 100
J.W. Boretos and S.J. Boretos
EG100A 1.09 94 A 8282 370 1800 100
EG100A-B20 1.29 93 A 7104 369
EG100A-B40 1.54 96 A 5607 360
1055D 1.16 54 D 9600 350 2500 100
Tecothane® 1065D 1.18 64 D 10 000 300 3200 100
1074A 1.10 75 A 6000 530 530 100
1075D 1.19 75 D 8300 240 3600 100
4 Biomedical elastomers
Property
Tensile Modulus
Specific Durometer strength, Elongation, ASTM Tear strength pli,
gravity hardness shore psi percent D-412 die C
Product and ASTM ASTM ASTM ASTM ASTM
Manufacturer Product no. D-792 D-2240 D-412 D-412 psi % D-624
PC-3555D 1.15 60 D 7000 350 1500 100
Carbothane™ PC-3555D-B20 1.36 57 D 8300 380 1600 100
PC-3572D 1.15 71 D 8500 300 4100 100
Thermedics, PC-3572D-B20 1.35 71 D 8400 310 3400 100
Inc. PC-3575A 1.15 73 A 4400 500 380 100
PC-3575A-B20 73 A 3500 600 410 100
PC-3585A 1.15 84 A 6500 390 640 100
PC-3585A-B40 1.68 89 A 3800 521 700 100
PC-3595A 1.15 95 A 6500 520 900 100
PC-3595A-B20 1.36 96 A 8300 390 1100 100
Chronoflex™ Chronoflex™ 70 A 7500 500 700 100
AR, Poly AR
Medica
Bio-materials, Inc.
Corethane® TPE 55D 1.211 55 D 7000- 365- 1850- 100
8500 440 2200
and TPE 75D 1.216 75 D 7000- 255- 5300- 100
9100 320 5700
TPE 80A 1.179 80 A 400- 770-
490 1250
Corhesive™, Corhesive™ 1.179 80 A 6500- 400- 770- 100
Corvita Corp. (cured) 7500 900 1250
J.W. Boretos and S.J. Boretos
Hydrothane™
Hydrothane™ is a TPE hydrogel belonging to the polyurethane family of polymers.
Hydrothane™ is an aliphatic material with water absorption capabilities ranging
from 5 to 25% by weight while still maintaining high tensile strength and elonga-
tion. Because of its water absorption capacity, Hydrothane™ is reported to be bac-
teria-resistant and lubricious. The polymer can be processed by conventional
extrusion and injection molding techniques. It can also be dissolved in dimethyl
acetamide solvent to produce a 25% solids solution suitable for dip-coating and
other solution processing techniques (Tables 4.3, 4.12, and 4.13).
Medicaflex™
The Lambda series of Medicaftex is a polyurethane-based TPE polymer that exhib-
its low modulus characteristics with high tear strength and abrasion resistance.
Those listed in the tables have passed USP Class VI compatibility tests and have
been used as replacements in some natural rubber latex and silicone rubber applica-
tions. The polymer has been applied to uses such as catheters, tubing and films
where softness, low durometer hardness, low modulus or high elongation are needed
(Tables 4.3, 4.12, and 4.13).
PolyBlend™ polyurethane
This TPE has been described as an aromatic elastoplastic polyurethane alloy. It pos-
sesses a low coefficient of friction, low extractables, and dimensional stability.
Hardness ranges from 65 to 75 Shore D. The material is classified for short-term (29
days or less) implantation. Clear and radiopaque formulations are available. Tubing
should be annealed at 80°C for four hours to reduce crystallinity (Tables 4.3, 4.4,
4.12, and 4.14).
Tecoflex® polyurethane
Tecoflex is an aliphatic polyether-based polyurethane that is available in clear and
radiopaque grades. They are reaction products of methylene bis (cyclohexyl) diiso-
cyanate (HMDI), poly (tetramethylene ether glycol) (PTMEG), and 1,4 butane diol
chain extender. The manufacturer claims that the aliphatic composition of Tecoflex®
4 Biomedical elastomers 305
Tecothane®
Tecothane® is an aromatic polyether-based TPE polyurethane polymer. It has pro-
cessibility and biocompatibility characteristics similar to Tecoflex® except that it is
an aromatic rather than an aliphatic polyurethane. Tecothane® is synthesized from
methylene diisocyanate (MDI), polytetramethylene ether glycol and 1,4 butanediol
chain extender. By varying the ratios of the reactants, polymers have been prepared
ranging from soft elastomers to rigid plastics. The manufacturer of Tecoflex® and
Tecothane® point out that there is not much difference between medical-grade, ali-
phatic and aromatic polyether-based polyurethanes with regard to chemical,
mechanical and biological properties. However, they caution that with improper pro-
cessing of Tecothane® (e.g., high moisture content or steam sterilization) it is pos-
sible to form measurable amounts of methylene dianiline (MDA), a listed carcinogen.
The use of ethylene oxide or gamma radiation are suitable sterilizing agents that do
not affect the chemical or physical properties (Tables 4.3, 4.12, and 4.13).
Texin™
There are four basic polymer formulations of Texin polyurethane TPE that may be
suitable for medical applications. They range in hardness and flexural modulus.
Texin elastomers are produced by the reaction of diisocyanate with a high molecular
weight polyester or polyether polymer and a low molecular weight diol. The poly-
ethers (products 5286 and 5265) offer greater hydrolytic stability and stress crack
resistance. The polyesterbased polyurethane (product 5187) and the polyester poly-
urethane/ polycarbonate blend (product 5370) possess high impact strength and
high stiffness along with useful low-temperature properties. Texin is not recom-
mended for implants of greater than 30 days duration. Texin should not be sterilized
by autoclave or use of boiling water. Other advantages offered by Texin TPUs are
that plasticizers are not necessary to achieve flexibility, the amount of extractables
are low, and they possess high tensile strength, high tear strength, and high abrasion
resistance. Texin polyurethanes are hydroscopic and will absorb ambient moisture.
They can be processed by extrusion and injection molding if thoroughly dried
beforehand. As with all chemical systems, the proper use and handling of these
materials can not be over-emphasized (Tables 4.3, 4.12, and 4.13).
Texin™ 5370 is a blend of polyester-based polyurethane and polycarbonate. It
offers high impact strength and high stiffness. Steam sterilization or boiling should
be avoided (Tables 4.3, 4.12, and 4.13).
Table 4.7 Typical Properties of Styrene-based Thermoplastic Elastomers
306
Property
Durometer Tensile Modulus
Specific hardness strength, Elongation ASTM Tear strength Compression set,
gravity shore psi percent D-412 pli,die B percent
Product and ASTM ASTM ASTM ASTM ASTM ASTM
Manufacturer Product no. D-792 D-2240 D-412 D-412 psi % D-624 D-395
R70-001 0.90 50 A 1200 900 150 100 16
C-Flex®, R70-003 0.90 70 A 1280 760 340 100 25
R70-005 0.90 30 A 1400 950 100 100 11
R70-026 0.90 90 A 1830 650 1,010 100
Consolidated R70-028 0.90 35 A 990 800 120 100 13
Polymer R70-046 0.90 34 A 1320 940 110 100 135 12
Technologies, Inc. R70-050 0.90 48 A 1250 880 170 100 100 18
R70-051 0.90 74 A 1140 680 370 100 150 28
R70-058 0.94 70 A 2080 660 300 100 120 55
R70-057 0.92 40 A 1220 890 100 100 90 33
R70-068 0.93 50 A 1630 850 140 100 110 38
R70-072 0.90 60 A 1270 780 240 100 20
R70-081 0.90 45 A 1440 920 120 100 17
R70-082 0.90 61 A 1270 860 230 100 130 19
R70-085 0.90 50 A 1380 750 200 100 17
R70-089 0.90 45 A 1640 700
R70-091 0.90 50 A 1280 780 130 100
R70-116 0.90 30 A 1105 810 100 100 84 24
R70-190 0.90 5 A 270 1010 20 100
R70-214 0.90 18 A 450 780
D-2103 0.94 70 A 4300 880 400 300 205
J.W. Boretos and S.J. Boretos
Kraton®, D-2104 0.93 27 A 1700 1350 200 300 180
D-2109 0.94 44 A 950 800 300 300 160
Shell G-2701 0.90 67 A 1600 800 480 300 260
Chemical G-2703 0.90 63 A 1200 670 470 300 230
Co. G-2705 0.90 55 A 850 700 400 300 140 38
G-2706 0.90 28 A 850 950 130 300 140
G-2712 0.88 42 A 840 820 250 300 140
4 Biomedical elastomers
307
308 J.W. Boretos and S.J. Boretos
Polycarbonate-based polyurethanes
Carbothane™
This medical grade TPE polyurethane is the reaction product of an aliphatic diiso-
cyanate, a polycarbonate-based macrodiol, and a chain terminating low molecular
weight diol (Tables 4.4, 4.12, and 4.13).
ChronoFlex™ AR.
Available as a dimethyl acetamide solution, this segmented, aromatic, polycarbonate-
based TPE polyurethane was designed to mimic Ethicon Corporation’s Biomer. The
polymer is made from the addition of diphenylmethane 4,4’-diisocyanate to a poly-
carbonate diol followed by addition of a mixture of chain extenders and a molecular
weight regulator. The polymer is believed to be resistant to environmental stress
cracking such as that experienced by other polyurethanes coated onto pacemaker
leads (Tables 4.4, 4.12, and 4.13).
Coremer™
Specifically designed as an 80 Shore A durometer TPE, this is a diamine chain
extended version of Corethane®. Coremer™ solution cast films have a low initial
modulus and high flex fatigue life. Information as to long-term biostability is not
available at this time (Tables 4.4 and 4.13).
Corethane®
A polycarbonate TPE polyurethane that claims biostability is achieved through its
replacement of virtually all ether or ester linkages with carbonate groups. The soft
segment is composed of a polycarbonate diol formed by the condensation reaction
of 1,6-hexanediol with ethylene carbonate. The polycarbonate diol is converted to a
high molecular weight polyurethane by the reaction with 1,4-methylene bisphenyl
diisocyanate (MDI) and 1,4-butanediol. It is reported to be resistant to environmen-
tal stress cracking as experienced with insulation on pacemaker lead wires. The
polymer can be extruded, injection molded or compression-molded, and can be
bonded with conventional urethane adhesives and solvents (Tables 4.4, 4.12, 4.13,
and 4.14).
Corhesive™
Corhesive™ is a solvent-free, two-component reaction adhesive system for use
with polyurethanes, plasma treated silicones and certain metals (Tables 4.4, 4.12,
4.13, and 4.14).
flexible grade tubing. The number of stoppers produced from Sarlink annually num-
ber in the billions. The material can be injection molded, blow molded, extruded,
calendered, and thermoformed on standard processing equipment. It can be thermal
bonded or adhesive bonded (Tables 4.5, 4.12, and 4.13).
Elastichem™ PVC.
This polyvinyl chloride compound family is highly elastomeric and exhibits a dry
non-tacky surface even at hardnesses as low as 40 Shore A durometer. Their rubber-
like resilience, high elongation and low permanent set and fatigue resistance offer
advantages over conventional formulations (Tables 4.6, 4.12, and 4.13).
Ellay™ PVC.
Compounds from Ellay Corp. are available with Shore hardness ranges from 55 A
to 100 A. The polymers have been applied to medication delivery systems, blood
collection, processing and storage, gastro-urological devices and collection sys-
tems. Product numbers ending in ‘R’ are special radiation resistant grades (Tables
4.6, 4.12 and 4.13).
Geon® PVC.
Geon® PVC is associated with vinyl examination gloves. For this use, Geon® recom-
mends a combination of Geon® 121 AR and 213. For a more ‘latex type’ feeling,
Goodtouch 250x100 is recommended. Typical film samples have passed patch insult
tests when worn against the skin for extended periods (Tables 4.6, 4.12 and 4.13).
Multichem™ PVC
This line of PVC polymers consist of alloys of PVC in combination with other poly-
mers. They display notable dynamic properties and resistance to migration and
extraction. These non-toxic PVC compounds (includes Multichem™ and
Elastichem™) have over 25 years of experience in the medical field (Tables 4.6,
4.12 and 4.13).
310 J.W. Boretos and S.J. Boretos
Styrene-based elastomers
C-Flex®TPE.
C-Flex® thermoplastic elastomers are based on styrene/ethylene-butylene/styrene
block copolymers. C-Flex® polymers designated as ‘medical grade’ are clear and
can be processed using conventional extrusion and injection molding equipment.
They have been tested using Good Laboratory Practices and have successfully
passed USP Class VI, biocompatibility tests. Translucent versions have high
rebound values at ultimate elongation. Medical tubing, ureteral stents, blood pumps,
feeding tubes and nephrostomy catheters are successful uses of this material (Tables
4.7, 4.12 and 4.13).
Kraton®
Kraton® elastomer consists of block segments of styrene and rubber monomers and
are available as Kraton® D and G series. The D series is based on unsaturated mid-
block styrene-butadiene-styrene copolymers whereas the G series is based on sty-
rene-ethylene/butylene-styrene copolymers with a stable saturated midblock. Listed
among the attributes of both series are such features as low extractables, dimen-
sional stability, vapor and gas transmission properties, ease of sterilization, softness
and clarity. They exhibit elastomeric flexibility coupled with thermoplastic proces-
sibility (Tables 4.7, 4.12, 4.13).
Natural rubber
Natural rubber (cis-polyisoprene) is strong and one of the most flexible of the elas-
tomers. The material has been used for surgeon's gloves, catheters, urinary drains
and vial stoppers. However, because it has the potential to cause allergic reactions
thought to be due to the elution of entrapped natural protein, this elastomer is being
used less now than in the past. Safer substitutes are being selected.
Silicone elastomers
Silicone elastomers have a long history of use in the medical field. They have been
applied to cannulas, catheters, drainage tubes, balloon catheters, finger and toe
joints, pacemaker lead wire insulation, components of artificial heart valves, breast
implants, intraocular lenses, contraceptive devices, burn dressings and a variety of
4 Biomedical elastomers 311
associated medical devices. A silicone reference material has been made available
by the National Institutes of Health to equate the blood compatibility of different
surfaces for vascular applications. This material is available as a silica-free sheet.
Contact the Artificial Heart Program, NHBLI, NIH, Bethesda, Md. for further
information.
The silicone elastomers most commonly used for medical applications are the
high consistency (HC) and liquid injection molding (LIM) types. The former is
most often peroxide cured and the latter platinum cured although there are varia-
tions. Both materials are similar in properties. LIM offers greater advantages to the
medical device molder and is gaining in popularity. This form of silicone may
become the molder’s material of choice within the next few years.
Other silicones
Silicones and polyurethanes have been used to produce denture liner materials and
maxillofacial prostheses. Most of these materials are silicone based, e.g., Flexibase,
312 J.W. Boretos and S.J. Boretos
Table 4.9 Typical Properties of Liquid Injection Molding (LIM) Silicone Elastomers
Property
Tear
Specific Durometer Tensile Elongation, strength
Product gravity hardness, strength, percent pli, die B
and ASTM shore ASTM psi ASTM ASTM ASTM
Manufacturer Product no. D-792 D-2240 D-412 D-412 D-624
40023 1.11 10 A 500 750 80
Applied 40024 1.11 20 A 800 600 140
Silicone
Medical 40025 1.12 30 A 950 600 150
Implant
Grade, 40026 1.12 40 A 980 450 170
Applied 40027 1.13 50 A 1000 400 190
Silicone
Corp. 40028 1.13 60 A 1100 350 220
40029 1.10 30 A 900 300 80
40071 1.14 70 A 1200 350 220
40072 1.10 25 A 650 400 60
40082 1.10 40 A 900 250 110
NuSil MED-6210 1.04 50 A 1000 100 35
Silicone,
MED-6233 1.03 50 A 1200 300 75
NuSil MED-6382 1.13 45 A 400 200
Technology
MED-6820 1.05 40 A 750 125 25
Silastic® Q7-4840 1.12 40 A 950 425 150
Medical Q7-4850 1.14 50 A 1350 550 225
Materials,
Dow Corning Q7-6860 1.16 60 A 1300 450 250
Corp.
Dispersions
Solvent solutions of polyurethane elastomers and silicone elastomers are given in
Table 4.10. These materials are helpful in casting thin films and odd or complex
shapes.
314
Specific polymeric materials traditionally used for medical applications have been
recently withdrawn from the medical market. Silicone elastomers are among those
withdrawn. To maintain continued supply of vital implants, methods of determining
equivalence for withdrawn elastomers with new or existing ones has been adopted
by the FDA in the form of an FDA Guidance Document.
The FDA will allow manufacturers to change sources of silicone elastomers (and
others) if they can show that the replacement material is ‘not substantially differ-
ent’ from materials described in existing approved applications. The device manu-
facturer is still required to certify that the processes of fabrication, cure and
sterilization it uses in the manufacture of its device are appropriate for the new
material and that the device will perform as intended. Premarket notification sub-
mission under section 510(k) of the Federal Food, Drug, and Cosmetic Act (21
USC 360(k) and 21 CFR 807.81(a)(3)(i), or a supplemental premarket approval
application under 21 USC 360(k) section 515 and 21 CFR 814.39 is necessary
when change could significantly affect the safety or effectiveness of the device.
These submissions are required to be submitted and approved before the device
may be marketed with the change.
There are a number of tests necessary for comparison of silicone elastomers as
indicated by ‘Guidance for Manufacturers of Silicone Devices Affected by
Withdrawal of Dow Corning Silastic® Materials’ (Federal Register, Vol. 58, No.
127, Tuesday, July 6, 1993/ Notices, 36207). They compare the physical, chemical
and biological properties of the bulk polymers as they are received from the supplier
and also compare the molded elastomer as it exists in the final medical device.
Table 4.12 Biocompatibility of Various Elastomers
318
Biocompatibility Status*
Product and Intracutaneous Systemic Skin
Classification Manufacturer Product no. Hemolysis Pyrogenicity Injection Injection Sensitization
Santoprene® 281-45 passed passed passed passed passed
Thermoplastic Rubber, 281-55 passed passed passed passed passed
elastomer 281-64 passed passed passed passed passed
Advanced 281-73 passed passed passed passed passed
Elastomer 281-87 passed passed passed passed passed
Systems 283-40 passed passed passed passed passed
PCCE Ecdel™ Elastomer, 9965 passed passed passed
copolyester Eastman Chemical 9966
Co.
elastomer Biospan® Segmented,
Polyurethane- Polyurethane, Biospan® passed passed passed passed passed
based The Polymer
elastomers Technology Group,
Inc.
Hydrothane™,
Poly Medica Hydrothane™
Biomaterials, Inc.
MP-5000
Medicaflex™, MF-5001 passed passed passed
MF-5040
Advanced Resin MF-5041
Technology MF-5056
MF-5057
J.W. Boretos and S.J. Boretos
MF-5062
Pellethane™ Pellethane™
2363 series, 2363 series passed passed passed
Dow Chemical Co.
PolyBlend™ 1000 , PolyBlend™
and PolyBlend™ 1000 and
1100 PolyBlend™
4 Biomedical elastomers
Biocompatibility Status*
Intramuscular Tissue Cell
Product no. 10 days 30 days 90 days Culture Comments
EG60A
EG80D
1055D
1065D
1074A
1075D
1085A
1095A
Texin™ passed passed Passed USP
Class VI
testing.
See text for
status.
PC-3555D
PC-3572D
PC-3575A
PC-3585A
PC-3595A
ChronoFlex™ passed passed Passed USP
AR Class VI
testing.
Coremer™ See text for
status.
TPE 55D passed passed passed passed
TPE 75D passed
TPE 80A passed passed passed passed
J.W. Boretos and S.J. Boretos
Corhesive™ passed passed passed passed
Sarlink® See text for
medical grade status.
Biocompatibility Status*
Classification Product and Product no. Hemolysis Intracutaneous Systemic Skin
Manufacturer Pyrogenicity Injection Injection Sensitization
Polyvinyl chloride Elastichem™PVC, Elastichem™PVC,
elastomers Geon® PVC, B. F. PVC
4 Biomedical elastomers
Goodrich Co.
Plastics Co.
Ellay™ PVC Ellay™ PVC Geon® PVC passed passed passed
Ellay, Inc. Multichem™ PVC
Geon® PVC, B. F.
Goodrich Co.
Multichem™ PVC,
Colorite
Plastics Co.
3300-45 NT
Teknor™ PVC, 3300-50 NT
3300-55 NT
Teknor Apex Co. 3300-60 NT passed passed passed passed
3300-68 NT
3300-75 NT
3300-80 NT
Teknor™ PVC, 3300-85 NT passed passed passed passed
3300-90 NT
(continued)
323
Table 4.12 (continued)
324
PVC
3300-45 NT passed
3300-50 NT passed
3300-55 NT passed
3300-60 NT passed
3300-68 NT passed
3300-75 NT passed
3300-80 NT passed passed Passed USP
4 Biomedical elastomers
Class VI
testing.
3300-85 NT passed Passed USP
Class VI
testing.
3300-90 NT passed
3310-50 NT passed
3310-55 NT passed
3310-60 NT passed
3310-65 NT passed
3310-70 NT passed
3310-75 NT passed
3310-80 NT passed
3310-85 NT passed
3310-90 NT passed
90A471R-60NT passed Passed USP
Class VI
testing.
90A471R-65NT passed Passed USP
Class VI
testing.
(continued)
325
Table 4.12 (continued)
326
(continued)
Table 4.12 (continued)
328
testing.
G-2712 passed passed Passed USP
Class VI
testing.
Biocompatibility Status*
Classification Product and Product no. Hemolysis Intracutaneous Systemic Skin
Manufacturer Pyrogenicity Injection Injection Sensitization
Polydimethylsiloxane Applied Silicone Applied
Medical Implant Silicone
Grade, Applied Medical
Silicone Corp. Implant
Grade
NuSii Silicone, NuSil NuSii Silicone
Technology
MDX4-4210 passed passed passed passed
Silastic ® Q7-4535 passed passed passed passed
Medical Materials, Q7-4550 passed passed passed passed
Q7-4565 passed passed passed passed
Dow Corning Corp. Q7-4720 passed passed passed passed
Q7-4735 passed passed passed passed
Polydimethyl Silastic ® Q7-4750 passed passed passed passed
329
(continued)
Table 4.12 (continued)
330
Two manufacturers, NuSil Technology and Applied Silicone Corp., are providing
equivalent silicone materials for the Dow Corning products that have been with-
drawn. Tables 4.15 and 4.16 gives reported comparisons.
Table 4.15 Equivalent Silicone Elastomers for Existing Dow Corning Silicones
Dow Corning NuSilt‡ Silicone Applied ∆ Silicone Medical Grade Silicone
Silicone* Equivalent Equivalent Description
Medical MED-1137 40064 Medical RTV Adhesive, Acetoxy
Adhesive A System (see also Rehau, Table
4.16)
Q7-4535 MED-4535 40044 High Consistency, 35 Durometer,
Peroxide Cure
Q7-4550 MED-4550 40045 High Consistency, 50 Durometer,
Peroxide Cure
Q7-4565 MED-4565 40046 High Consistency, 65 Durometer,
Peroxide Cure
Q7-4720 MED-4720 40043 High Consistency, 20 Durometer,
Platinum Cure
Q7-4735 MED-4735 40039 High Consistency, 35 Durometer,
Platinum Cure
Q7-4750 MED-4750 40040 High Consistency, 50 Durometer,
Platinum Cure
Q7-4780 MED-4780 40042 High Consistency, 80 Durometer,
Platinum Cure
MDX4-4210 MED-42111 40072 Liquid Silicone, 25 Durometer,
Platinum Cure
40029 Liquid Silicone, 30 Durometer,
Platinum Cure
Q7-4840 MED-4840 40026 Liquid Silicone, 40 Durometer
Platinum Cure
Q7-4850 MED-4850 40027 Liquid Silicone, 50 Durometer,
Platinum Cure
Q7-4865 MED-4865 Liquid Silicone, 65 Durometer,
Platinum Cure
DC-360 MED-360 40047 Medical Grade Silicone Fluid,
1000 cps.
Specify 40073 Medical Grade Silicone Fluid,
350 cps.
viscosity 40074 Medical Grade Silicone Fluid,
20 cps.
* Dow Corning Corp., Midland, Ml. ‡ NuSil Silicone Technology, Carpinteria, CA
∆ Applied Silicone Corp., Ventura, CA Note: It is the user's responsibility to adequately test or
determine that these materials are suitable or safe for any application.
4 Biomedical elastomers 335
Table 4.16 Equivalent Silicone Elastomers for Withdrawn Dow Corning silicones
Dow Corning NuSilt‡ Silicone Applied ∆ Silicone Medical Grade Silicone
Silicone* Equivalent Equivalent Description
MDX4-4515 MED-4515 40045 50 Durometer, peroxide cure
MDX4-4516 MED-4516 40046 60 Durometer, peroxide cure
Q7-2245 MED-2245 40009 40 Durometer, platinum cure
Q7-2213 MED-2213 40076 Dispersion in 1, 1, 1
Rehau 1511¥ 40076 trichloroethane
Medical RTV adhesive,
acetoxy system
*Dow Corning Corp., Midland, Ml. ‡ NuSil Silicone Technology, Carpinteria, CA.
△ Applied Silicone Corp., Ventura, CA. ¥ Rehau AG and Co., Rehau, Germany.
Note: It is the user’s responsibility to adequately test or determine that these materials are suitable
or safe for any application.
Not all materials respond alike when subjected to various means of sterilization.
Some are heat sensitive, some will absorb sterilization fluids, some will be affected
by molecular changes when subjected to radiation sterilization and others will
absorb and hold irritating gases for extended periods of time. Table 4.13 gives ster-
ilization methods that have been judged most appropriate for each elastomer. The
consequences of using an inappropriate method can be loss in physical properties
and an adverse biological response.
Standard methods of testing elastomers used for medical applications are given by
specific ASTM test methods. Physical and biological tests are provided here to
serve as references for the data cited in the tables and listed in Table 4.17. They are
also designated in the FDA Guidance Document.
4.6 Biocompatibility
4.7 Sources
• AlphaGaryAlphaGary, Leominster, MA
• Applied SiliconeApplied Silicone Corp., Ventura, CA
• Biospan®Polymer Technology Group, Inc., Emeryville, CA
• C-Flex®Consolidated Polymer Technologies, Inc., Largo, FL
• Carbothane™Thermedics, Inc., Woburn, MA
• ChronoFlex™PolyMedica Industries, Inc., Woburn, MA
• Coremer™Corvita Corp., Miami, FL
• Corethane®Corvita Corp., Miami, FL
• Corhesive™Corvita Corp., Miami, FL
• Ecdel™Eastman Chemical Co., Kingsport, TN
• Elastichem™Colorite Plastics Co., Ridgefield, NJ
• Ellay™Ellay, Inc., City of Commerce, CA
• Geon®B.F. Goodrich Co., Chemical Group, Cleveland, OH
• Hydrothane™PolyMedica Industries, Inc., Woburn, MA
• Kraton®Shell Chemical Co., Oak Brook, IL
• Medicaflex™Advanced Resin Technology, Manchester, NH
• Multichem™Colorite Plastics Co., Ridgefield, NJ
• Natural rubberExxon Chem. Co., Buffalo Grove, IL Goodyear Tire and Rubber
Co., Akron, OH
• NuSil SiliconeNuSil Technology, Carpinteria, CA
• Pellethane™Dow Chemical Co., Midland, MI
• PolyBlend™PolyMedica Industries, Inc., Woburn, MA
• Santoprene®Advanced Elastomer Systems, St Louis MO
• Sarlink®DSM Thermoplastic Elastomers, Inc., Leominster, MA
• SilasticDow Corning Corp., Midland, MI
• Tecoflex®Thermedics, Inc., Woburn, MA
• Tecothane®Thermedics, Inc., Woburn, MA
• Teknor™Teknor Apex Co., Pawtucket, RI
• Texin™Miles, Inc., Pittsburgh, PA
Chapter 5
Oxide Bioceramics: Inert Ceramic Materials
in Medicine and Dentistry
5.1 Introduction
Single oxide ceramics, e.g. aluminium oxide (A12O3, alumina) and zirconium
dioxide (ZrO2, zirconia), are bioceramics of an inert nature. An inert ceramic does
not form a bonding to bone similar to those bioceramics of bioactive nature. Alumina
bioceramics are in the pure aluminium oxide form, whereas zirconia bioceramics
are partially stabilized by additional oxides, e.g. yttrium oxide, calcium oxide or
magnesium oxide.
Oxide ceramics exhibit superior mechanical properties, corrosion and wear
resistance. Since the oxides are the highest oxidation state of the metal, they are
stable even in the most invasive industrial and biomedical environments. Alumina
and zirconia are utilized as load-bearing hard tissue replacements and fixation
implants in dentistry and surgery.
Although the use of alumina as implants can be traced back to the 1930s as described
by Hulbert et al. (1) (Table 5.1), the extensive use of alumina since the 1980s has
depended on new powder processing technology enabling grain size reduction of
the sintered ceramics from 10 micrometers down to 2 micrometers (Figure 5.1,
microstructure of alumina). This significantly improves the performance of the
J. Li (*)
Centre for Oral Biology, Karolinska Institute, Huddinge S 141–4, Sweden
G.W. Hastings
Institute of Materials Research and Engineering, Singapore, Singapore
Figure 5.1 SEM micrograph of dense alumina, etched in boiling H3PO4 for 6 minutes to show the
microstructure.
alumina ceramic hip balls. Alumina and partially stabilized zirconia are currently in
extensive use as implants in consequence of their high strength, excellent corrosion
and wear resistance and stability, non-toxicity and biocompatibility in vivo. A summary
of alumina- and zirconia-based implants is presented in Table 5.2. The most estab-
lished example is in the total hip endoprosthesis with a combination of metallic
stem, ceramic ball and ultra high molecule weight polyethylene (UHMWPE)
5 Oxide Bioceramics: Inert Ceramic Materials in Medicine and Dentistry 341
acetabular cup. A ten year clinical success rate better than 90% is reported for the
cemented total hip endoprosthesis.
Dental implants of polycrystalline alumina were suggested by Sandhaus in
Germany (4). Type Tübingen was produced by Frialit in the 1970s. These devices
have not been generally accepted, due to the fracture failure of the implants, particu-
larly for those of polycrystalline type produced in the early 1970s. The single crystal
sapphire type, introduced in Japan by Kawahara in the 1970s (18) is, however, still
being used and a recent 10-year clinical follow-up report from Sweden showed a
92% success rate (19) for the single crystal dental implants.
Alumina and zirconia ceramics are also being used for alveolar ridge reconstruc-
tion (20), maxillofacial reconstruction, as ossicular bone substitutes (21), and in
ophthalmology (22), knee prosthesis (8), bone screws as well as other applications
as dental biomaterials, such as dental crown core, post, bracket and inlay (23, 24).
Table 5.3 and 5.4. Resulting from a strong chemical bond between the Al and O
ions, as expected from the value of heat of formation (-400K cal/mol), Al2O3 has a
high melting point, the highest hardness among known oxides, and high mechanical
strength (26).
Table 5.4 Properties of medical-grade ceramic materials according to the standards to the
standards and the manufacturer’s technical date – alumina and zirconia
Alumina
according to
ISO-6474 Zirconia
ASTM Frialit according to
F 603–83 bioceramic ISO/DIS Prozyr®
Property DIN 58 8353 alumina 13356 zirconia
Purity (%) >99.5 >99.5 >99.5* >95
Density (g/cm3) >3.9 >3.98 >6.0 6
Porosity (%) 0 ** 0 0
Grain size (μm) <4.5 >2.5 <0.6 <1
Microhardness (GPa) 23 23 — 13
Young’s modulus (GPa) 380 380 — 220
Flexural strength (MPa) >400 >450 >900 >920
Biaxial flexural strength (MPa) 250 — >550 —
Impact strength (cm MPa) >40 >40 124
Fracture toughness (MPa m1/2) 10
Wear resistance (mm3/h) 0.01 0.001 __
Corrosion resistance (mg/m2d) <0.1 <0.1 — —
* ZrO2+HfO2+Y2O3
** Not available.
Chemical properties
Wear resistance
Arising from the chemical stability and high surface finish and accurate dimensions,
there is a very low friction torque between the alumina femoral heads and the ace-
tabular cup, leading to a low wear rate. Combinations of ceramic head/UHMWPE
cup and ceramic head/ceramic cup were tested and compared to the metal head/
UHMWPE cup. The wear resistance of the ceramic head/UHMWPE cup combina-
tion over metal/UHMWPE has improved from 1.3 to 34 times in the laboratory and
from three to four times clinically (27, 28). No alumina wear particles from retrieved
ceramic/UHMWPE were found, whereas UHMWPE wear particles from microns
to millimetres in size were found in the retrieved surrounding tissues. However,
from the ceramic/ceramic combination, ceramic particles resembling ‘fine grains
and great fragments in the ranges from 0.5 to 10 micrometers diameter, with the
predominant size of about 1 micrometer’ were found in the surrounding tissue (29).
The advantage of ceramic/ceramic combination over ceramic/UHMWPE is, there-
fore, doubtful. For wear tests, we refer to ISO-6474 ASTM F-603.
344 J. Li and G.W. Hastings
Clinical performance
The fracture of ceramic balls in ceramic: UHMWPE combination has been virtually
zero. Fritsch and Gleitz (30) published a failure analysis on 4341 alumina ceramic
heads articulating with 2693 alumina ceramic and 1464 polymer sockets implanted
over 20 years (1974 to 1994), and concluded that the use of ball type neckless heads
brought the fracture rate close to zero. The success rate of 10 years foliow-up is
normally above 90% for the ‘elderly’ patient population. Stem and cup loosening
are the causes of failure, where the consistent wear debris from UHMWPE and bone
cement remain the problems.
Zirconia ceramics are termed polymorphic because they undergo several transfor-
mations on cooling from a molten state to room temperature. It exhibits three well-
defined polymorphs, the monoclinic, tetragonal and cubic phases and a high pressure
orthorhombic form also exists. The monoclinic phase is stable up to about 1170°C
where it transforms to the tetragonal phase, stable up to 2370°C, while the cubic
phase exists up to the melting point 2680°C. A large volume change of 3 to 5%
occurs when zirconia is cooled down and transforms from the tetragonal to the
monoclinic phase.
The volume change due to phase transformation is sufficient to exceed elastic and
fracture limits and causes cracking of the zirconia ceramics. Therefore, additives
such as calcia (CaO), magnesia (MgO) and/or yttria (Y2O3) must be mixed with
zirconia to stabilize the material in either the tetragonal or the cubic phase. PSZ is a
mixture of cubic and tetragonal and /or monoclinic phases, whereas TZP is 100%
tetragonal (phase diagram Figure 5.2). Both PSZ and TZP are suggested for medical
implant applications. Yttria-TZP ceramics have a strength and fracture toughness
approximately twice that of alumina ceramics used in the biomedical field. This
makes zirconia heads less sensitive to stress concentrations at the points of contact
with metal cones.
Zirconia ceramics have a high density because of heavy zirconium ions, and a low
microhardness and elastic modulus, together with high strength and fracture tough-
ness compared to other ceramics including alumina. The superior mechanical strength
provides the possibilities for producing ceramic ball heads of size below 32 mm.
5 Oxide Bioceramics: Inert Ceramic Materials in Medicine and Dentistry 345
Garvie et al. were the first to realize the transformation toughening mechanism for
zirconia ceramics. Increase of both strength and fracture toughness can be obtained
by utilizing the tetragonal-monoclinic phase transformation of metastable tetrago-
nal grains induced by the presence of the stress field ahead of a crack (31). The
volume change and the shear strain developed in the martensitic reaction were rec-
ognized as opposing the opening of the crack and therefore acting to increase the
resistance to crack propagation.
The published results of in vitro wear tests demonstrated that zirconia has a supe-
rior wear resistance. Saikko (32) showed no wear of zirconia femoral heads on
his hip simulator wear test against 10.9 mm UHMWPE cup, and Praveen Kumar
et al. (33) demonstrated the high wear resistance of zirconia against UHMWPE
and the superiority of zirconia ceramics even over alumina ceramics in terms of
low wear and low friction. A significant reduction in the wear rate of zirconia
ball heads compared to the metal ball heads was reported on a pin-on-disc wear
test and on a hip simulator (34). However, there are two potential limitations for
the use of zirconia as bioceramics: degradation and radiation. It is known that the
phase transformation is accelerated in aqueous environment, but little is known
about how this phase transformation will occur in biological environment,
particularly under dynamic loadings. A warning against steam resterilization has
been issued in the UK. Radioactive U-235 impurity was detected in some ‘pure
zirconia’, both alpha- and gamma-irradiation were measured from zirconia fem-
oral balls. Although the radioactivity was low, more work is required to verify
this matter (13).
Clinical performance
The surface degradation of the zirconia balls due to the phase transformation under
loading seems to be a problem, although no significant change in mechanical
strength was reported in some long-term in vivo and in vitro studies (35, 36).
Seriously, catastrophic failure of modular zirconia ceramics femoral head compo-
nents after total hip arthroplasty was reported (37). Since zirconia femoral heads
have a short clinical history and few clinical results are available, more investigation
is required to eliminate the factors which impair the clinical stability of zirconia
ceramics under loading.
5 Oxide Bioceramics: Inert Ceramic Materials in Medicine and Dentistry 347
An advanced ceramic is processed in such a way that the structure of the materials
on different levels, including atomic, electronic, grain boundary, microstructural
and macrostructural, is under strict control. In the manufacturing processes, empha-
sis is placed on producing dense ceramics with a fine microstructure. However,
other factors such as chemical composition, the nature and distribution of the
impurities, crystal structure, grain size, and defects are also of importance to the
performance of the ceramic materials. Three basic processes are involved in the
production of fine ceramic components, namely: 1. powder technology, 2. densifi-
cation or sintering and 3. machining. Both alumina and zirconia hip balls are pro-
duced by compacting fined-grained powder (green bodies), and sintering at
1500–1700 °C and finally grinding or lapping to obtain a high surface finish and
sphericity (Ra<0.02 μm).
Table 5.5 Guidance for Biologic Evaluation Tests of the Implant Device in Contact to
bone Tissue (According to ISO 10993–1:1992 (E))
Contact duration
A-limited B-prolonged C-permanent
Biological tests (>24 h) (<24 h to 30 days) (<30 days)
Cytotoxity x x x
Sensitisation x x x
Irritation/Intracutaneous x x x
Reactivity
Irritation/Intracutaneous x
Genotoxicity x x
Implantation x x
Chronic toxicity x
Carcinogenicity x
The related tests see ISO standards from No. 10993–1 to 10993–6
5.5 Applications
The dominating application of alumina and zirconia is as hip balls as well as cups
of total femoral prosthesis. The neckless hip balls are the most popular design. In
1981, Oonishi et al. (8) reported on the use of an alumina ceramic total knee pros-
thesis. High alumina ceramic middle ear implants (Frialit) are used clinically in
Europe since 1979 (21). An opthalmological implant device consisting of a combi-
nation of a single crystal alumina optional cylinder and a polycrystalline alumina
holding ring was introduced clinically in 1977 (22). Kawahara (12) has reported
extensively on single crystal alumina bone screws.
5 Oxide Bioceramics: Inert Ceramic Materials in Medicine and Dentistry 349
Figure 5.3(b) Zirconia and bone interface 1 month after implantation. Arrows are pointing to the
interfaces.
350 J. Li and G.W. Hastings
Alumina and zirconia ceramics have been utilized for root analogue, endosteal
screws, blades and pin-type dental implants. The root and blade form dental implants
used during the 1970s tended to fracture after a few years in function (41, 42) (Brose
et al., 1987, Driskell, 1987). Although initial testing of these polycrystalline alu-
mina materials showed adequate mechanical strength, the long-term clinical results
demonstrated functional limitations related to material properties and implant
design. However, single crystalline alumina showed mechanical strength superior to
that of polycrystalline alumina. It allows a much higher load. One-stage dental
implants of single crystalline alumina are used clinically with a high success rate.
McKinnery (43) had also reported on single crystal alumina blade and screw dental
implants. Dental implants of zirconia have not been widely used clinically although
zirconia has a similar mechanical strength and a much higher fracture toughness in
addition to lower cost of production compared to single crystalline alumina. The
term dental implant is used only for materials in contact with bone and soft tissue
(14). Alumina and zirconia are also used in other dental applications, alumina
ceramic crowns, Procera® (23), zirconia dental post, (10) and recently a dental
inlay of zirconia was introduced (11). Orthodontic brackets made of oxide ceramics
were also produced, tested and used clinically. Unfortunately, tooth surface damage
was observed when the brackets were taken away (15). Modification of the debond-
ing technique is under developing.
Clarke and Willmann (13) make a comprehensive summary about the bioceramic
manufacturers (Table 5.6). Some dental companies are included.
5 Oxide Bioceramics: Inert Ceramic Materials in Medicine and Dentistry 351
References
1. Hulbert, S.F., Bokros, J.C., Hench, L.L., Wilson, J. and Heimke, G. In Ceramics in Clinical
Applications, ed. by Vincenzini, P. Elsevier, Amsterdam, 1987, pp. 3–27.
2. Rock, M. German Patent 583 589, 1933.
3. Smith, L. Arch. Surg. 1963; 87: 653–661.
4. Sandhaus, S. British Patent 1083769, 1967.
5. Boutin, P. Presse Med. 1971; 79: 639.
6. Mittelmeier, H. Z. Orthop. Ihre Grenzgeb 1974; 112 : 27.
7. Shikita, T. Paper presented at the XIV World Congress of SICOT, Kyoto, Japan, October
15–20, 1978.
8. Oonishi, H., Okabe, N., Hamaguchi, T. and Nabeshima, T. Orthopaedic Ceramic Implants I,
1981; 11–18.
9. Lord, G. et al. Paper presented at the Harrington Arthritis Research Centre Symposium,
November 18–21, 1990.
10. Akagawa, Y. et al. J. Prosthet. Dent. 1993; 69: 599–604.
11. Meyenberg, K.H., Luthy, H. and Scharer, P. J. Esthet. Dent., 1995; 7(2): 73–80.
12. Johansson, B. Tandläkartidningen 1996; 14746–749.
13. Clarke, I.C and Willmann, G. In Bone implant Interface ed. H.V. Cameron,Hugh U., Mosby,
1994, pp. 222.
14. Kawahara, H. In Encyclopedic handbook of biomaterials and bioengineering ed. by Wise,
Donald L. et al. Marcel Dekker, Inc., New York, 1995, pp. 1469–1524.
15. Sinha, P.K., Rohrer, M.D., Nanda, R.S. and Brickman, C.D. American J. Orthodont &
Dentofacial Orthop, 1995; 108: 455–63.
16. Christel, P.S. In Condse Encyclopedia of Medical & Dental Materials, ed. by Williams, D.F.,
Pergamon Press, Oxford, 1990, pp. 375–379.
17. Keith, O., Kusy, R.P. and Whitley, J.Q. American J. Orthodont. & Dentofacial Orthop., 1994;
106(6): 605–614.
18. Kawahara, H. Orthopaedic Ceramic Implants 1,1981; 1–10.
19. Fartash, B. Single Crystal Sapphire Dental Implants: Experimental and Clinical Studies. PhD
thesis, Karolinska Institute, Sweden, 1996.
20. Hammer, N.B., Topazian, R.G., McKinney, P.V. and Hulbert, S.F. J. Dent. Res. 1973; 52:
356–361.
352 J. Li and G.W. Hastings
D. Daily
6.1 Introduction
D. Daily
(deceased)
6.1.2 Microstructure
-1
s g = s ¥ + s 1G 2
σg = strength of the ceramic, G = mean grain size. σ1 and σ∞ are material constants.
This is the Orowan–Petch relation first observed in steels in 1953 and holds for all
crystalline materials. It implies that the finer the grain size, the stronger the material,
although it fails for small grain sizes.
Carniglia [33] has identified a two-stage relationship in which the strength depen-
dence changed at a critical grain-size Go. Davidge and Evans enlarged on this [34]
and related the grain-size/strength dependence to the fracture mechanism. In the
larger grain size region, to which the above relations refer, fracture occurs as a con-
sequence of the extension of inherent flaws while for the lower grain sizes, plastic
flow is the dominant cause. The changeover in behaviour is a property of the material
considered, related to its purity, and environmental factors such as temperature. In
ceramics where there are large and small grains, an initial flaw in a large grain region
propagates until it reaches a finer region where the apparent fracture surface energy
is higher. Here it is held up until the stress becomes critical for the new region.
Producers of structural bio ceramic components devote much effort towards the
choice and quality of raw materials used in the manufacture of these devices. This is
of particular importance to load-bearing orthopaedic implant applications where
strength and fracture toughness properties are critical to the long-term performance
of the device. The need therefore is to produce high purity, dense ceramic with fine
grain crystalline microstructures, free from inclusions and/or impurities often intro-
duced through raw materials and production processes. The introduction inhomoge-
neity within the microstructure of the ceramic does have a detrimental effect on the
strength and fracture toughness properties of the device. Inclusion or defects as
small as ~10 μm in a sintered ceramic microstructure have contributed in vivo fail-
ure. It must also be mentioned that other factors such as device design (see Chap.
7—Ceramic hip joint endoprosthesis) and processing surface treatment, for example
grinding, honing and polishing, do influence the strength of the ceramic component
and ultimately the long-term success of a device.
The majority of ceramics currently used in the manufacture of orthopaedic
devices are high-purity zirconia toughened alumina (ZTA), alumina (>99.9 %), zir-
conia, and all contain grain size substantially sub-micrometre. In recent years vari-
ous ISO, BSI, ASTM and CEN Standards have been developed and published,
specifying the material performance requirements and testing methodologies for
implantable grade ceramics. Microstructural specifications including grain size are
stated in the ISO 6474: part 1 for alumina, ISO 6474: part 2 for ZTA and the ISO
13356 for zirconia. The ISO 6474 part 1 grain size specification for load bearing
orthopaedic implantable alumina is ≤2.5 μm although the grain size in commercial
grades has significant proportion of sub-micrometre microstructure. Zirconia (ISO
13356) specifies a grain size ≤0.4 μm; commercial grades are typically 0.3–0.4 μm.
6 Ceramic Materials Testing and Fracture Mechanics 355
ZTA forms a matrix of alumina and zirconia have two size specifications ≤1.5 μm
for the crystalline alumina phase and ≤0.6 μm for the zirconia, the typical grain size
0.5–0.8 μm for alumina phase and 0.1–0.2 μm for zirconia.
Sapphire single crystals are stronger than polycrystalline alumina, so a crack would
be expected to follow the grain boundaries, so that the smaller the grains, the greater
the surface free energy of sapphire,—about 6–7 J/m, while for polycrystalline alu-
mina, this is about 30 J/m. It should therefore be harder to create new surfaces in
ceramic alumina than in sapphire. The only explanation for this is that in the
ceramic, the cracks are there already, as surface and volume flaws, while in sapphire
they need to be nucleated.
Pure cubic zirconia is a relatively weak material and normal sintering invariably
generates very large grains. One concludes that it is easier for a crack to go across a
grain rather than travel through the grain boundaries.
New considerations have to be introduced for these materials such as (1) the coher-
ency of the boundaries between different crystalline components; (2) the stress
imparted at the interfaces between components (and this can be severe enough to
crack the material) as a consequence of cooling components or phases of different
thermal expansion; (3) different elastic moduli of non-equilibrium of solutes within
phases once these can create internal stresses which oppose crack propagation.
6.1.3 Porosity
s = s 0 exp - bP
where σ = strength of the porous material, σ0 = strength of the fully dense material,
P = porosity, b is a constant.
Coble and Kingery measured the flexural strength of polycrystalline aluminas
for porosities between 5 % and 50 % porosity, and using the symbols above,
proposed the following empirical equation:
-s
0.6 P = exp
8000 (1 - p )
356 D. Daily
Knudsen [36] and Passmore et al. [37] derived formulae to describe the coincident
effects of grain size and porosity on alumina at both 25 and 1200 °C. These were of
the form:
In general covalent substances are stronger than ionic. Diamond, silicon carbide,
silicon nitride, cubic boron nitride (borazon), boron carbide are typical; they contain
very few flaws and usually have a rigid framework structure.
Alumina and silica are substances with both ionic and covalent character. At very
high stresses (e.g. at the tips of cracks) deformation by slip is possible causing a
dissipation strain energy.
The relationship between failure stress, surface energy and concentration of
flaws is a complicated one, but a number of approaches to this problem have been
made, giving rise to the science of fracture mechanics.
In contrast with metals, stress does not cause deformation by slip in ceramics, and
eventual failure along the planes of maximum shear stress. Ceramics fail by the
propagation of cracks, and in particular when one of them reaches a critical length.
Ceramics also are 8–10 times stronger in compression than in tension, and so the
ceramic member is designed for use in compression, while is specified its strength
as a result of tensile test.
6 Ceramic Materials Testing and Fracture Mechanics 357
Metals are usually tested in tension on a tensile machine where a machined spec-
imen is trained and the strain is measured as the increase in separation of a pair of
marks divided by the original distance separating them (called the “gauge length”),
and the stress is determined from a load cell reading. This method of testing is not
convenient for ceramics for the following reasons.
The test piece will probably be dumbbell shaped and carefully have avoided sharp
corners which would create a stress concentration. It would be supported by a ring,
and another ring would be used to supply the extension force.
The alignment of such a test piece would be critical, requiring a universal joint
between the extending crosshead of the machine and the test piece. Failure to do this
would cause a stress intensification where the ring makes contact.
In the process of bending a material, cross section mm and pp (Fig 2.1) rotate with
respect to each other about axes perpendicular to the plane of bending, so that lon-
gitudinal fibres on the convex side are extended and longitudinal fibres on the con-
cave side are compressed.
358 D. Daily
r O
M m p
M
n n¢
s δ y s¢ x
m p
y
n′
dA
y
n
x
There is also some intermediates surface which is neither extended nor com-
pressed, called the “neutral surface” represented by nn′. The radius of curvature r is
measured from the neutral surface nn′. Let ss′ be any surface at a distance y from the
neutral axis, then, from similar triangles,
Ey
Force acting on dA = s x dA = dA (6.3)
r
6 Ceramic Materials Testing and Fracture Mechanics 359
Sum of all such forces over the cross section represent a couple resultant in the x
direction, therefore = ∅
Ey E
ò r
dA = òydA = f
r
(6.4)
Ey
Moment of force acting on dA, with respect to nn′ is dA × y .
r
Adding all the moments over the cross section, and equating to the moment of
the external forces M,
Ey 2 EI
dA = z = M (6.5)
r r
1 M
This defines the relation = , where I = òy 2 dA (6.6)
r EI
Iz is the “moment of inertia of bending” (analogous to the expression used in describ-
ing a rotating body).
Fig. 2.3 Rectangular
Section Test Piece y
h/
2
dy
C z
h/
2
b z
360 D. Daily
For a rectangle,
h/2
bh 3
I z = 2 ò y 2 bdy = . (6.7)
12
0
For a circular section (diameter d)
p d4
Iz = . (6.8)
64
Eliminating r, we have,
My
sz =
Iz
For a rectangle,
h
2
bh 3
I z = 2 ò y 2 bdy = (6.9)
12
0
For a circular section (diameter d)
TTd 4
Iz = (6.10)
64
Eliminating r, we have,
My
sz =
Iz
Maximum tensile or compressive stress is in the outermost fibre. If nn′ is the middle
of the section, and the thickness = h
Mh
(s x )max = (6.11)
2I z
P ( L - 2)
M=
2 2
6 Ceramic Materials Testing and Fracture Mechanics 361
p/2 p/2
L
`
p ( L - l ) h 12
Stress in outermost fibre = × × × 3
2 2 2 bh
3 p(L - l)
=
2 bh 2
Frequently 1/4 point loading is used, i.e. where L = 4 a and = 2a so that maximum
3Pa
stress = 2 . This is the failure stress (sometimes called modulus of rupture)
bh
3Pa ì 3 PL ü
sF - í= ý
bh 2 î 4 bh 2 þ
æ 16 Pa ö
ç For a cyliner s F =
è p d 3 ÷ø
6 Pa 3PL
sF = = (6.12)
bh 2 2bh 2
æ 32 Pa ö
ç For a cylinder s F = (6.13)
è p d 3 ÷ø
EI d2L
P (L - x) = M = = EI 2 *
R dx
Integrate,
dy P æ x2 ö
= ç Lx - ÷+ A
dx EI è 2 ø
362 D. Daily
L
O x
x,y
P
y
dy
Now A = Æ , because = Æ at the origin
dx
P ì Lx 2 x 3 ü
y= í - ý+ B
EI î 2 6þ
Now B = Æ because y = Æ when x = Æ at origin.
Depression at the free end is therefore
PL3
D max = (6.14)
3EI
3
é1 + ( dy / dx )2 ù 2
1
R=ë 2 2
û =
d y / dx d y / dx 2
2
( P / 2 )( L / 2 )
3
PL3
Dmax = = (6.15)
3EI 48 EI
6 Ceramic Materials Testing and Fracture Mechanics 363
L
2 2
p
p p
2 2
We require to determine:
(a) radius of curvature.
(b) maximum elevation of the centre of the beam above the support points.
(c) depression of ends below support points.
Let h be elevation above supports, then from the similar triangles OPC and CPQ
R - h ( / 2)
=
( / 2 ) ( 2h )
Neglecting h cf. R,
2 2
= 2 Rh, h =
4 8R
The important aspect of the four-point bend test is that there is a circular arc with a
constant bending moment between the loading points (B and C in Fig. 7). Effectively
the specimen selects its own weakest point for failure within this region.
P ( L - ) EI
Over BC, M = × =
2 2 R
2 EI
R=
( P / 2)( L - )
( P / 2 ) ( L - ) 2
h=
16 EI
(6.16)
( / 2 )
Let C represent the origin. Inclination to horizontal at c =
R
( P / 2)( L - )
= arcsin q q =
4 EI
Let right hand part of the beam contain the points (x, y).
Taking moments,
P ìï ( L - ) üï d 2 y
M= í - x ý = 2 EI
2 îï 2 þï dx
364 D. Daily
h Q θ
p B C (x,y) p
2
l p l 2
2 2
p p
2 2
dy ( P / 2 ) ìï ( L - ) x - x üï
2
= í ý+ A
dx EI îï 2 2 þï
æ dy ö
At C, the origin, x = Æ, A = ç ÷ = q
è dx ø
dy ( P / 2 ) ïì ( L - ) üï
= í( L - ) x - x + × ý
2
dx 2 EI îï 2 þï
By integration,
( P / 2 ) ïì ( L - ) x ( L - ) × x ïü
y= í x- + ý+ B
2 EI ïî 2 3 2 ïþ
At C, x = f , y = f , therefore B = f . L -
To find the depression at the ends, let x = .
2
Depression of the ends below the origin,
( P / 2 ) ìï ( L - ) ( L - ) üï
3 3
+ ( L - ) × ý
2
= í -
8 EI ï 2 6
î þï
For ¼ point, symmetrical loading, let L = 4 a , 1 = 2 a
6 Ceramic Materials Testing and Fracture Mechanics 365
( P / 2 ) æ 8a 3
8a 3 ö
= ç - + 8a 3 ÷
8 EI è 2 6 ø (6.17)
2 Pa
=
3EI
These deflection formulae may be used to determine the elastic modulus. At high
stresses the machine deflection becomes significant, and a blank run has to be per-
formed to determine this.
RL is the loaded radius; RO is the outer radius; RS is the supported radius; t is the
plate thickness; P is the loading force; V is Poisson’s ratio.
Principal stresses in the plate are s r = s ¶
3P ìï RS (R32 - RL2 ) üï
sr = s¶ = í(1 + v )1n + (1 + v ) ý (6.18)
2p t 2 RL 2 R02
îï þï
These are effectively the failure stresses for the ceramic.
This test is especially valuable for ceramics in that:
(a) it tests simultaneously in two dimensions at right angles.
(b) there are no sharp edges in the stress field to initiate cracks.
(c) discs are easy to press, fire and polish.
(d) they take up little room in a furnace, and a batch of ten can expect to have iden-
tical firings.
RL
Ro o
RS
366 D. Daily
Vertical guide
rods for ball
Ball bearing
Rs
Rc
Alternatively the hardened support ring may be replaced by three balls at the
apexes of an equilateral triangle. This overcomes the problem of flatness in discs.
RS becomes the radius of the circumcircle of the latter.
(Ref. Ph.D. Thesis, A.D. Sivill, University of Nottingham, 1974) [32].
A disc specimen is compressed between two flat plates. Tensile splitting occurs
along the loaded diameter AB. It was developed originally in Brazil and Japan for
testing concrete about 1943. A biaxial stress distribution is set up in the disc, pro-
vided the thickness is small compared with the diameter.
2P
s max = (6.19)
DtTT
6 Ceramic Materials Testing and Fracture Mechanics 367
6W × ( OD + ID )
= . (6.20)
p × t ( OD - ID )
2
(Tensile stresses are indicated at W, X, Y, Z.)
This configuration is especially useful for high temperature testing in that it
avoids the use of testing jigs (at 1600–1700 °C it would be deforming as fast as the
test piece). The specimen undergoes fracture in four places, often showing separate
maxima on the testing machine. In some respects it resembles two three-point bend
tests superimposed above and below the horizontal diameter. The mathematical
derivation is tedious, but the test is capable of high reproducibility.
This is also used for high temperature strength determinations. Photoelastic studies
indicate that central bar is in pure uniform tension. This pattern was selected out of
60 shapes, but the difficulty of precision shaping has restricted its application. This
I.D
w z
OD
y
t = thickness
D
368 D. Daily
s = kP / Dt (6.21)
Fig. 2.13 Stress/Strain
Graph of Non- Ferrous
Metal
The following expression was given by Weibull for the probability of failure, and it
is especially applicable to brittle materials.
ìïæ -s - s ö m üï
Pf = 1 - exp íç u
÷ ý (6.22)
ïîè s 0 ø ïþ
370 D. Daily
Fig. 2.15 Stress/Strain
graph of Ceramic
(Hookean behaviour)
*
*
*
σ * CERAMIC
σu is a “threshold stress” below which the probability of failure is zero. The thresh-
old stress when compared to the mean stress is relatively small and typically set as
zero rather than given a finite number which may result in over estimation of the
probability of failure. It does have significance for glasses.
σ0 is a normalising factor.
σ is the stress in the material.
m is the “Weibull modulus” and it is a measure of the variability of σ. The higher
the value of m, the more consistent the material.
The function is valid for all stresses in the range s u £ s £ ¥ .
Failure probabilities are in the range 0 £ Pf < 1 .
For ceramic materials it is usual to find m in the range 5 < m < 20.
The curve can be fitted to the experimental points by the manipulation of the
values m, σu and σ0.
Stress/mean stress of batch.
(Ph.D. University of Nottingham 1973)
While this may be the best method of obtaining m, it is time-consuming and
laborious, but with the help of a few assumptions a very rapid procedure is possible,
particularly if a computer is employed.
Assume s u = 0. Let F be the probability of failure corresponding to an applied
stress σ.
ìï æ s ö m üï
Then F = 1 - exp í- ç ÷ ý
ïî è s 0 ø ïþ
m -- -- Takelogs
1 æs ö
Then = exp ç ÷
1- F è s0 ø
m -- -- Takelogs again
1 æs ö
ln =ç 0 ÷ ,
1- F è s ø
6 Ceramic Materials Testing and Fracture Mechanics 371
1.0
Firing batch C
Weibull modulus m = 20.6
Number of Specimens N = 49
0.9
0.8
0.7
0.6
Failure propability p
0.5
0.4
0.3
0.2
0.1
0.0
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3
Stress/mean stress of batch
Fig. 2.16 Weibull distribution of the fracture data of a sintered batch of Reaction bonded silicon
nitride
1
= m ln s - m ln s 0 .
ln. ln. (6.23)
1- F
æ 1 ö
Weibull graph paper is ruled with ln.ln.ç ÷ as the ordinate units and one
è 1- F ø
simply plots F against σ to obtain straight line graph of which m is the slope.
1 1
Note : when s = s 0 ln. ln. = 0 ln =1
1- F 1- F
372 D. Daily
One must remember that the purpose of the Weibull statistics is to be able to take a
limited number of data from the breakage of test-specimens and obtain a reasonable
value for m. This will act as a control for our processing schedule. Twenty is a con-
venient number, but if the quantity of raw material is limited, even ten results can be
useful.
Consider a very large population of N data. If one value has to be chosen at ran-
dom, then each has a 1/N chance of being picked.
If the data is placed in ascending order, the chance of picking a value less than,
or equal to the ith number in the list is i/N.
Each value in the data list is therefore associated with a cumulative probability
of P = i/N.
Let the large circle contain N data (in a bag if the analogy is helpful). Let us
withdraw i′ data at random and arrange it in ascending order. (Think of i′ = 20 for a
convenient number.)
Assemble the data in order on the ladder diagram (Fig. 2.18) and represent it with
the full lines. Return the data to the bag, shake it and withdraw another 20 (i″).
Set these up in ascending order, and indicate them with the short dashed lines.
Return, shake and withdraw another 20 (i‴).
i′′
i′
i′′′′
i′′′
6 Ceramic Materials Testing and Fracture Mechanics 373
i10′′′′
′
i10
′′′
i10 R10
i10
i9′′′ i9′
i9′′′′
i9′′ R9
i8′′′′ i8′′
i8′ R8
i8′′′
i7′′
i7′ R7
i7′′′′
i7′′′
i6′
i6′′′′ i′′′ R6
6
i6′′
i5′′′′ i5′′′
i5′ R5
i5′′
i4′′′ R4
i4′′′′ i4′′ i4′
i3′′
i3′′′′ i3′ R3
i3′′′
i2′′′′
i2′′′
i2′ R2
i2′′
i1′′′′ i1′′′
R1
i1′′ i1′
Set these up in ascending order, and indicate with the long dashed lines. Return,
shake and withdraw another 2(i″″) (dotted lines).
Repeat until i′ is obtained.
Each sequence of first values i′, i″, i‴, .. .. .. .. i′
second values i′, i″, i‴, .. .. .. .. i′
third values i′, i″, i‴, .. .. .. .. i′
twentieth values i′, i″, i‴, .. .. .. .. i′
is called a “rank”.
Each value is a binomial distribution of 20 values (or other number if chosen)
which will have a most probable value represented by the mean or median of the
distribution. There are formulae for calculating these or there are tabulated values
for convenient numbers. Procedure:
When an appropriate number of tests have been made,
1 . the results are arranged in sequence.
2. the table of “F” values for the number of tests is obtained (Table 6.1).
3. using Weibull paper, we plot F is ordinate and σ as abscissa the lowest value of
F is made to correspond with the lowest failure stress and so on.
4. using the least squares method, the best line is obtained through the points.
5. gradient of the line is m, the Weibull modulus.
374 D. Daily
To compare three point bend specimens with those tested in pure tension, the fol-
lowing expression relates
1/ m
s B3 é v ù
= ê( m + 1) × T ú
2
(6.24)
sT ë vB 3 û
For a Weibull modulus of 10 and specimens of equal volume,
s B3
= 1.614 (6.25)
sT
6 Ceramic Materials Testing and Fracture Mechanics 375
0.632
pf
0.100
0.010
0.001
1 10 100 1000
σf / MP a
Fig. 2.19 95 % alumina 1600 °C 120 min. 50 mm × 3 mm bars, ¼ point loading
pf
0.100
0.010
0.001
1 10 100 1000
σf / Kg
0.632
pf
0.100
0.010
0.001
1 10 100 1000
σf / Kg
0.632
pf
0.100
0.010
0.001
1 10 100 1000
σf / MPa
1/ m
s B4 é V ù
= ê( m + 1) × T ú (6.26)
sT ë VB4 û
and for a Weibull modulus of 10 and equal volumes of test pieces
s B4
= 1.27 (6.27)
sT
Relating four point to three point under the same conditions gives
s B4
= 0.785 (6.28)
s B3
Because brittle fracture depends on the largest crack in the stress field growing until
it reaches a critical length, the larger the volume of the test specimen the higher will
be the probability of a crack being present which is large enough to cause failure. It
is for this reason that one uses large test pieces, and chooses four-point loading
instead of three-point. The relationship is simple, i.e.
1
s v1 æ V2 ö m
ç ÷ (6.29)
s v2 è V1 ø
In a rearranged from, s V1 V1 = s V2 V2 . This expression will be used in the following
m m
chapter.
The shear stress at which deformation should begin in an ideally perfect crystal was
calculated by Frenkel in 1926. Referring to Fig. 2.1 representing two adjacent
planes of atoms in a simple crystal structure, we see that if the upper plane were
displaced relative to the lower in the slip direction b, causing atoms A′, B′, C′, etc.
to lie over B, C, D, etc. respectively, slip would have taken place. The crystal would
remain perfect as the atoms had moved to equivalent lattice sites. The shear stress
378 D. Daily
necessary to produce this permanent plastic deformation is the critical shear stress
τc of the ideally perfect crystal.
To obtain its value we first consider the form of the stress displacement curve
(Fig. 2.2) measuring displacements indicated in Fig. 2.1.
Now τ = 0, when x = 0 and x = b;
The crystal is in equilibrium in both cases.
1
When x = b , the atoms in the upper layer can be seen to be in a position of
2
unstable equilibrium; the force in the slip direction is zero, for each atom is equally
attracted to the right and to the left.
1
Consequently τ = 0, also at x = b .
2
Further, we know from the elastic behaviour of crystals that the shear stress
increases linearly with strain within the elastic range. The slope of the displacement
curve at x = 0 and x = b must therefore comply with Hook’s law.
t =G (6.30)
g
where
Gx
g = x / a,t = (6.31)
a
Thus
dx
= G / a ( x a)
dx
Now τ(x) must be a periodic function with wavelength b, and we therefore taken
the relation as typical,
æ 2p x ö
t = K sin ç ÷ (6.32)
è b ø
1
Other harmonic functions satisfying the boundary conditions at x = 0, b and b
2
could have been chosen, without affecting the result. As Eq. (6.32) must reduce to
Eq. (6.30) for small strains, one finds that
x ( 2p x )
t =G =K
a b
Gb
K= (6.33)
2p a
6 Ceramic Materials Testing and Fracture Mechanics 379
x 1
2p = p
b 2
1
i.e. at x = × b , and is numerically equal to K. In practice a and b will not differ
4
from one another appreciably and one therefore obtains the result that the critical
shear stress of the ideally perfect crystal is
t c » G / 10 (6.34)
1
The maximum elastic shear strain occurs when x = b and is therefore equal to
4
1 b which, with a ≈ b, yields
,
4 a
g c » 25 % (6.35)
Perfect crystals necessary to check this theory can only be obtained in the form
of microscopically thin whiskers. They have, nevertheless, established the correct-
ness of the foregoing results. However, the critical shear stress of pure copper is
only about 20 kg/cm2, while G/10, Eq. (6.34) yields about 40,000 kg/cm2. The
strength of the ideal crystal is therefore about 2000 greater than the real one. In met-
als this discrepancy is accounted for by the presence of dislocations, while in non-
metals, and in particular in ceramic substances, the discrepancy is caused by
cracking, although both processes can be present to some degree in either class of
material.
æ du ö 1
s » ç ÷× 2 (6.36)
è dr ø r0
380 D. Daily
dr = force
r0 = area of one atom plane
i.e. the force divided over the approximate area over which it operates. Now
dr
ds = cd Î= c (6.37)
r0
and thus
ds 1 æ ¶ 2 u ö
c = r0 ~ ç ÷ (6.38)
dr r0 è ¶r 2 ø
For a shear deformation, the Born repulsive term may be neglected, and thus for
ionic crystals u µ 1 r Substitution in Eq. (6.32) predicts a proportionality between
0
C44 and r0−4.
When we try to shear the lattice we have to overcome the peak of maximum
resistance. Over small displacements the stress-distance curve is approximately a
sine wave so we may write
P
s = s th sin ( r - r0 ) (6.39)
2a
where a is the atomic displacement corresponding to the theoretical strength, and
the material is assumed isotropic. ds
If E is the elastic modulus, then, E = , and
dÎ
ds ds r0s th P P
= r0 = cos ( r - r0 )
dÎ dr 2a 2a
For very small strains, r–r0 → 0, the cosine term → 1, so that
2 Ea
s th = (6.40)
P r0
To estimate the constant a, Orowan’s argument is that the work done in stressing
the material to σth, must be at least equal to the energy required to create two new
fracture faces. If γ0 is the surface energy per unit area,
r0 + 2a
2 Ea P
ò s dr = ò sin ( r - r0 ) dr = 2g 0 (6.41)
r0
P r0 2a
and this reduces to
4 Ea 2
= g0 (6.42)
P 2 r0
6 Ceramic Materials Testing and Fracture Mechanics 381
1
æ Eg ö 2
s th = ç 0 ÷ (6.43)
è r0 ø
High surface energy and stiffness, together with a small lattice spacing, contrib-
ute to high strength, although a high fabricated density is also required. It has been
shown that α ≈ 0.14 r0 for ionic crystals. Note in the previous hypothetical approach
r
a = 0 substituting in Eq. (6.37), gives
4
Er0
g0 » (6.44)
125
-10
Substituting values for alumina where E = 4 ´ 10 N / m , r0 = 4 ´ 10 m gives
11 2
-2
g 0 » 1 × 2 J m (This is for a sapphire and not a polycrystal) and agrees with experi-
mental data.
In general, two lines of approach to the study of fracture have been made.
These are:
1. The Griffith’s approach, where the reduction of strain energy in the test sample
is equated to the increase of surface energy of the crack.
Formulae for fracture energy have been derived and practical evaluation methods
described.
2. The elastic continuum approach, where the intensified stresses at the edge of the
crack-tip are considered. From this the ‘Stress Intensity Factor’ can be defined,
the critical value of which is a materials property can be used in engineering
design. Equations from the two approaches can be related to provide a number
of useful testing methods to evaluate the fracture toughness of a material.
Let a plate contain a thin elliptical crack of length 2c, represented by AOB in
Fig. 3.1, under a uniform stress σ. To find an expression for the relaxation of the
strain energy as the crack is formed, assume that the plate consists of a number of
independently operating linear elements. Behind the crack front there will be a
stress free zone, caused by the shrinkage of the elements. Let it be assumed that the
length of this zone, measured along each element from the line AOB representing
the centre of the crack is y0. This figure will obviously vary from zero at the crack
edges to y0 in the centre.
382 D. Daily
σ
2
0
A B z
σ
2
yo
yo
(max)
uy
A 0 x B
C C
my E
= (6.45)
y0 s /2
E
Therefore m y = y0
s /2
uy represent the y coordinates of the ellipse and y are lengths of the stress-free zone.
For the ellipse,
x2 my
2
b2 2
+
c2 b2
= 1, m 2
y =
c2
c - x2 ( ) (6.46)
2
æ E ö b
y02 × ç ÷ =
ès / 2 ø c
2 (
c2 - x 2 ) (6.47)
6 Ceramic Materials Testing and Fracture Mechanics 383
ès / 2 ø
y02 c 2
y02( max )
(
= c2 - x 2 )
The locus of the lengths of y0 is a circle if
c2
=1
y02( max )
We may confirm this if we let x = c/2
y2
= 3 / 4 (Fig. 3.2)
y02( max )
y 3
=
y0( max ) 2
3
Since 0 x = c 2 , xy = , then 0 y = c . Which is a radius of the same circle.
2
Therefore y0 lies on a circle, and y0 = c - x
2 2
( )
In order to compute these relaxing stresses, we require to know the displacement
of the upper surface of the ellipse, given by the following expression:
æ E ö
my = (c 2
- x2 × ç)
çs ÷
÷ (6.48)
è 2 ø
Work done by the relaxation of the elements (με) is also the strain energy released,
and is given by,
UÎ = 1 sm y dx ( per element )
2
c
UÎ = 1 òsm y dx ( per quadrant )
2
0
c
UÎ = 2 òsm y dx ( for circle ) (6.49)
0
384 D. Daily
That is
s2 c
UÎ = ò (c )
- x 2 dx
2
(6.50)
E
0
c
ò (c )
- x 2 dx = 1 / 2 x (c )
- x 2 + 1 / 2c 2 is sin(x/a) + c
2 2
0
When x = ∅, expr. = ∅; when x = c, expr = Πc2/4.
Leaving the notion of independent elements and returning to the continuous
sheet, the elastic constraints must now be put in, so that the expression becomes
modified to
s2 c
UÎ =
E
(1 + r ) ( k + 1) ò (c 2
- x2 ) (6.51)
0
where k = 3–4r for plane strain,
(3 - r )
and k = for plane stress
(1 + r )
(r is Poisson’s ratio).
On integration, Eq. (6.51) gives
s 2 P c2
U= (1 + r ) ( k + 1) (6.52)
4E
This expression represents the total crack surface and it should be noted that the
expression is independent of the semi-minor axis of the ellipse b
According to the Griffith criterion, the surface energy, Us, needed to create new
surfaces, is supplied by the energy released, Uε, The total free energy, U, for the
cracked body is expressed as,
m = ms - mÎ + m (6.53)
where Ū is the component of the strain energy independent of the crack in the body
æ du ö s2
ç i.e. dc = Æ ÷ and is equal to 2E per unit volume for this system. In Griffith’s
è ø
terminology, U is the total potential energy for the system. In terms of the total
energy of the system
du
< 0 when the crack is unstable
dc
du
= 0 when the crack is in equilibrium
dc
6 Ceramic Materials Testing and Fracture Mechanics 385
du
> 0 when the crack isstable
dc
The surface energy per unit thickness, Us, equals 4cγ0. Substituting for Us and Uε
in Eq. (6.53) we have,
s 2 P c2
u = 4cg 0 - (1 + r ) ( k + 1) + m (6.54)
4E
du
For the equilibrium situation = Æ , and if the applied stress be equal to σf.
dc
du dus duÎ s 2P c
= - = 4g 0 - f (1 + r ) ( k + 1) = 0
dc dc dc 2E
Therefore
8g 0
sf = (6.55)
P c (1 + r ) ( k + 1)
For plane stress and plane strain the expression tales the form,
2 Eg 0
sf = Plane strain (6.56)
(1 + r ) P c
2
2 Eg 0
sf = plane stress (6.57)
Pc
The results of tests carried out by Griffith indicated that the theoretical predic-
tions were justified provided that there was little or no plastic flow in the region of
the crack tip. Because there is always some dissipation of energy due to plastic flow,
the Griffith expression tends to give a higher value than the true thermodynamic one.
The theory does bring out a very important point, namely that the strength of a
solid is dependent on the flaw size in a direction normal to the applied stress, and the
probability of failure will, in turn, depend on the probability of there being a flaw in
excess of the critical size.
We define G as the energy release rate. This is either the strain energy release or the
potential energy release rate, and it is the energy released per unit extension of the
crack front per unit thickness of the body (rate in this instance relates to distance
rather than time).
386 D. Daily
In the opening mode (or mode 1, in Fig. 4.3), if the crack moves at both ends a
distance 2dc, relative to a crack size 2c,
From equation, at equilibrium,
du dus duÎ
= - = 0. (6.58)
dc dc dc
duÎ æ dus ö
G= = dus = 4g 0 dC (6.59)
dc çè dc ÷ø
For the case where the crack operates from one end only,
duÎ æ dus ö
G= = dus = 2ro dC (6.60)
dc çè dc ÷ø
where the crack size is critical (Fig. 3.2) c = c and G is equal to twice the thermody-
namic surface energy.
I.e. G = 4g 0 (for the double-ended crack) [1–5] and 2γ0 (for the single ended
crack).
This defines the strain-energy-release rate per unit width of crack front.
The conclusion of this analysis is that we can relate the failure stress σf with the
size of the crack c, provided we have some knowledge of γ0 the fracture energy, or
G, the strain energy release rate.
What follows is a discussion of the significance of G, followed by practical
evaluation methods for γ0
th K
ow
Stress
gr
a ck
t cr
ou
th wt
h
wi
gro
with
o N M
Strain
δc 2c δc
An energy balance criterion has been developed by Davidge [3] using a simple geo-
metrical argument (Fig. 3.3). Suppose the stress/strain curve of the sheet containing
the crack is represented by OJ. Let the crack now grow in length from 2C to
2(C + dC), i.e. equal amounts at both ends. The sheet will then be less stiff. The
stress/strain curve is now represented by OKL. When deformation is by dead load-
ing (i.e. constant stress), the strain increases by JL When a rigid machine is used the
strain is constant and therefore the reduction of stress is given by JK. Let the frac-
ture energy per unit area be given by γo. As the crack grows, the energy required to
form new fracture faces is 4γodC.
The elastic strain energy is du.
The external work done on the sheet is dW.
The condition for fracture is therefore,
d ( w - u ) ³ 4g o dC (6.61)
We need to estimate W and U.
Note that the energy is proportional to the area under the graph, so that when
fracture occurs at constant strain, dW = ∅ and an amount of elastic strain energy,
proportional to OKJ is released (dU is negative).
When fracture occurs at constant stress, dW is proportional to JLMN and the
elastic strain energy increases proportional to OLM − OJN = OJL. Because JKL
becomes negligible in the limit, d(W − U) is again proportional to OJK.
388 D. Daily
Note that dW = 2dU; half the external work is absorbed as elastic energy, and half
is available to assist crack propagation.
For either type of loading, we can continue the discussion solely in terms of U.
The condition for fracture reduces to
du
³ 4g o (6.62)
dC
d = Pl (6.63)
where δ is the elongation of the test piece and λ is the elastic compliance. The strain
energy is therefore,
d
1
uÎ = òPdd = Pd
2
o
(P is a function of δ and is area under curve)
1 2 1 d2
= P l= (6.64)
2 2 l
δ
6 Ceramic Materials Testing and Fracture Mechanics 389
d w » Pd (d ) = P 2dl (6.65)
1 2
du = P dl (6.66)
2
and the total change in mechanical energy is
-1 2
d ( -w + u ) » P dl (6.67)
2
2. Constant displacement
dW = O
1 (6.68)
d u = - P 2 dl
2
giving
1
d ( -w + u ) = - P 2dl (6.69)
2
It will be seen that the two results are identical, i.e. the mechanical energy
released during incremental crack extension is independent of the loading configu-
ration. We can therefore define a crack extension force, which is the strain energy
release rate with respect to the crack length in a test piece of unit thickness.
The strain energy release rate is a measure of available potential energy, and the
given the symbol G and is equal to
-d ( -w + u )
dC
and is equivalent to
æ ¶u ö
G = -ç ÷
è ¶c ød
Because W = ∅, for the constant strain situation (i.e. where the fixation is rigid).
390 D. Daily
From equations G =
1 2 dl
2
P
dc and 2 (
1 d2
l2 )dl , G can be deduced in two
dc
ways. At constant stress, all that is needed is the load, and a calibration for the com-
pliance. At constant strain, one needs to know the compliance, and then its rate of
change with the crack-length. Because of the negative sign the strain energy actu-
ally decreases under constant strain, but under constant stress it increases. For this
reason constant strain methods are preferable, because the crack configuration is
stable.
Davidge and Tappin [10] describe a practical method for compliance analysis
with a notched bar specimen. The “stiffness” k is determined from the load/deflec-
tion curve, and it is given by:
P = kd (6.70)
PFd F
U=
2
æ ¶u ö
Now g c = - ç ÷ , fracture occurring at a fixed deflection, or
è ¶A ød
æ ¶u ö æ ¶K ö
gc = -ç ÷ ×ç ÷
è ¶k ød è ¶A ød
But,
æ ¶u ö
ç ¶k ÷ = d
2
è ø
and thus
æ ¶K ö
g c = -d 2 ç ÷/2 (6.71)
è ¶A ø
1 2 dl
G= P
2 dc
æ1ö
dç ÷
1 k 1 1 dk
= P2 è ø = - P2 2
2 dc 2 k dc
d 2 æ dk ö bd 2 æç ¶k ö÷
=- ç ÷ =
2 è dc ød 2 ç¶A ÷
è 2 ød
æ ö
d 2 ç ¶k ÷
=- ç ÷
2 ç A÷
ç¶ ÷
è 2 ød
(both sides of a single-ended crack)
Since
G = 2g o (at equilibrium)
G -d 2 æ ¶k ö
gc = = (6.72)
2 2 çè ¶A ÷ød
6.3.8 E
xperimental Technique for Compliance:
Calculating of γ [10]
Five test bars of ∅.5 cm square section and 4 cm length are prepared as accurately
as possible (allowing for sintering shrinkage) with progressively increasing notch-
depths, cut accurately with a diamond saw, so that the crack area increases conve-
niently in units of ∅. 1 cm of cross-sectional area. The bar is then subject to a
three-point bend test, and the stiffness is obtained from dividing the applied load by
the deflection obtained. It is advisable to find the average value of a number of tests,
particularly at low crack areas. This deflection may be measured directly by a trans-
ducer or in terms of movement of the chart, making corrections for the deformation
of the load cell of the testing-machine at comparable loads. Note that the first objec-
tive is not to extend the crack or break the specimen, but merely to obtain plots from
æ ¶k ö
which k and ç ÷ can be broken and δF obtained (at the point of fracture).
è ¶A ød
The tabulation is completed for each specimen, the stiffness against crack area
æ ¶k ö
curve drawn, and ç ÷ obtained for each specimen area, by measuring the gradi-
è ¶A ød
ents of the curve.
392 D. Daily
δκ
δA δ
STIFFNESS κ
CRACK AREA A - 2bc
General form of κ versus A curve
8
STIFFNESS κ
0
0 0.1 0.2 0.3 0.4 0.5
CRACK AREA A - 2bc (cm2)
κ versus A data for PMMA.
This requires a crack to propagate in a specimen while its growth is under control.
The stronger the specimen the more difficult this becomes. It is essential to have:
(a) a very hard machine.
(b) the facility for very slow rates of strain.
6 Ceramic Materials Testing and Fracture Mechanics 393
10
0
0 0.1 0.2 0.3 0.4 0.5
CRACK AREA A - 2bc (cm2)
2.5
STIFFNESS κ (109 dyn/cm)
2.0
1.5
1.0
0.5
0
0 0.1 0.2 0.3 0.4 0.5
æ ¶k ö
-ç ÷
P δ k Ave k b c A = zbc è ¶A ø δF
Fig. 3.10 Specimen L
geometry. I = span;
b = breadth; d = crack depth
b
c
d T
Fig. 3.11 Load/deflection
curve. Pf = fracture load; PF
δ1 = fracture deflection; σF = 3 PF L
k = specimen stiffness; 2 bd
2
σf = fracture stress
γF U
=
2b (d-c)
LOAD P
σF
κ
DEFLECTION δ
Controlled cracking can only occur in brittle material when the notch depth is so
large that the test specimen is substantially weakened. One then arrives at a situation
where the residual strain energy in a specimen is less than that of the increase in
surface energy required when fracture occurs. Increasing the external strain then
goes directly into providing the energy for overcoming the surface and propagating
the crack.
Measure the area under the curve to obtain U Record P Calculate γF and δF
6 Ceramic Materials Testing and Fracture Mechanics 395
LOAD (κ8)
6
0
1 2 3 4
−2
DEFLECTION (10 cm)
Fig. 3.13 Load/deflection 20
curves for
PMMA. Numbers on graph (PMMA)
refer to notch depth ratio
c/d. X indicates
catastrophic failure
15
0
LOAD P (κ8)
0.1
10 x
0.2
x
0.3
x
5
0.4
x
0.5
0.6
0.7
0.8
0.9
0
-1
10 cm
-1
DEFLECTION δ (10 cm)
396 D. Daily
Davidge [3] also draws attention to the fact that the γo in the Griffith equation is an
ideal value, requiring perfectly sharp cracks to propagate through a material and
creating planer fracture surfaces. Fracture in ceramics is not far from ideal, but there
are energy-dissipating processes in what is a relatively complex process. For exam-
ple fracture may be due to:
(a) a combination of transcrystalline and intercrystalline microstructure, manifest-
ing as steps in the fracture surface.
(b) Because of the high stresses near the crack-tip ( approaching…) there is a pos-
sibility that plastic flow may occur.
(c) Cracking branching in the microstructure causing subsidiary cracking.
All of these are energy-consuming processes which make the apparent fracture
energy γ1, somewhat higher than the thermodynamic surface energy γ0. The effective
surface energy γ1. is only a valid concept when plastic flow is localised near the crack
tip and would not apply to metals where plastic flow is the normal deformation mode.
Attach the test piece show above, to the Instron in the tensile mode, i.e. P (up) to the
load-cell and P (down) to the cross-head via a universal coupling. Follow the prog-
ress of the crack front (L) and measure 2d with a transducer system.
P up
y
2b
2δ
x w
L
Pdown
Bending moment is distributed along the beam, and is caused by the applied
force P, so that
P 2 L2 æ wb3 ö
L
1
( )
2 EI ò0
U- M 2
x dx = ç where I = ÷ (6.73)
6E × I è 12 ø
(E is the elastic modulus and I the moment of inertia of bending)
Deflection of beam at point of application of force is obtained using Castigliano’s
theorem.
æ ¶U ö PL3
d =ç ÷ = (6.74)
è ¶P ø x = L 3EI
3EId 2
U= (6.75)
2 L3
dL dd
æ ¶U ö æ ¶U ö 9 EId dL 2
dU = ç ÷ +ç ÷ dd = - + Pdd
è ¶L ød è ¶d ø L 2 L4
9 EId 2 3Eb3d 2
Therefore g w = or g d = ( measuring d ) (6.77)
2 L3 8 L4
6 P 2 L2
( 3.24 ) and ( 3.27 ) give g P = ( measuringP ) (6.78)
Ew 2 b3
Both of these formulas can be used. A problem is to make the crack grow straight,
and to clearly see the tip of it. This will be a problem with opaque bodies, where one
cannot use total internal reflection to see the crack tip. The crack may be guided by
the use a rectangular grooved. Specimen
398 D. Daily
w2 w1
a
b
dS = g w1 dL
æ ¶U ö -9 EId
2
æ w2 b3 ö
ç ¶L ÷ = 4 ç where I =
12
and where a <<< b ÷ (6.79)
è ø 2L è ø
3Ew2 b d
3 2
6P L2 2
Thus g d = and g P =
8w1 L4 w2 w1b3 E
An important historical precursor to the Griffith study was the street analysis of
Inglis (1913) of an elliptical hole in a uniformly stressed plate. This showed that the
local stresses about a sharp notch or corner could rise to several times that of the
applied stress. It thus became apparent that even submicroscopic flaws might be
potential sources of weakness in solids. In the limiting case a crack could be
regarded as an infinitesimally narrow ellipse.
Let us therefore examine the modifying effect of the hole on the distribution of
stresses in the solid plate. Let it be assumed that Hooke’s law applies everywhere in
the plate, and the dimensions b, c (Fig. 4.1) are small in comparison with those of
the plate. The problem then reduces to one of linear elasticity.
Beginning with the equation of an ellipse
x 2 y2
+ =1 (6.80)
c2 b2
One can show that at C the radius of curvature is:
b2
P= (6.81)
c
It is at the point C where the greatest concentration of stress occurs. The appar-
ently simple equation (6.81) is obtained as a result of a complex analysis [1–3, 5]
æ 2C ö é æ C ö 12 ù
s yy ( C ,q ) = s A ç 1 + = s ê
A 1+ ç ÷ ú. (6.82)
è b ÷ø êë è P ø úû
6 Ceramic Materials Testing and Fracture Mechanics 399
−2b O
C X
2c
σA
4
Y
3
2 σ
yy
B
1
b σxx
o
O C X
It appears that stress concentration depends on the shape of the hole rather that
its size. The variation of local stress along OX is also of interest. Figure 4.2 illus-
trates the case where C = 3b. σyy drops from its maximum at C approach the value of
σA asymptotically for large values of x, while σxx rises to a sharp peak within a small
distance from the stress-free surface and subsequently drops to zero at high x. (It can
be appreciated that when C / b = 5, s yy = s max = 11 .)
400 D. Daily
1
s A p c = s max p P . (6.84)
2
in which we can call the L.H.S. K1 = s p C , and is known as the “stress intensity
factor”.
It is important to realise that K1 contains only the macroscopic quantities (the
external tensile stress) and the half crack-length, C, both of which are measurable.
It should be appreciated that the Eq. (6.95) implies a physical limit restricted by the
properties P and σmax without breaking at the tip of the crack, for P is given by the
microstructure of the material, and σmax cannot be larger than the internal molecular
strength, σM. There should exist therefore a specific material limit given by
1
K1C = s M p P (6.85)
2
the so-called “critical stress intensity factor”, for “fracture toughness”.
In Davidge’s treatment [3] the expression (6.85) is related to the Orowan value
for the theoretical strengths of solids, i.e. the latter value for σth is made equivalent
to the stress at the edge of the crack tip. For the failure this gives
E P
sf = × . (6.86)
4C ro
Now if the tip of the crack were infinitesimally small the stresses would be infi-
nitely large. It is reasonable to assign the value half an atomic spacing to P if the
crack passes between adjacent planes of atoms. Substitution in (6.38) gives a second
value for i.e.
1
æ Eg ö 2
sf = ç o ÷ (6.87)
è SC ø
The stress is somewhat lower than the Griffith stress at which therefore, fracture
should certainly occur and Orowan’s value for could only be accurate to within a
factor of two. Hence we conclude the Griffith criterion is adequate and sufficient.
A criticism of this solution is that the elastic stresses in the direction of the
applied force rapidly reach infinity as the crack becomes sharper, and so become
insensitive to the differences in the mode of cracking and the stress configurations.
6 Ceramic Materials Testing and Fracture Mechanics 401
K1 = s p a . (6.88)
é æ pa ö 2 ù
1
K1 = s ê2 w tan ç ÷ ú (6.89)
êë è 2W ø úû
Expanding the above expression we have,
é p 22 ù
K1 = s p a ê1 + a ú ± (6.90)
ë 24W û
as a/w → θ, becomes
3. Surface crack in a semi infinite body.
K1 = 1.12s p a (6.91)
2a
σ
402 D. Daily
2σ
2w
1
é æpa ö æ p a öù 2
K1 = s ê2 w tan ç ÷ + q .2 w sin ç ÷ú . (6.92)
ë è 2w ø è w øû
K1 = 2s a / p (6.93)
6 Ceramic Materials Testing and Fracture Mechanics 403
Fig. 4.13a Two σ
Symmetrical edge cracks
in plate of finite width
2w
2a
A B
A B
Plan
3 P1
a= . 2 (6.94)
2 bd
3P1 3 é ù
1 3 5
æaö 2 æaö 2 æaö 2
K1 = ê1.93 ç ÷ - 3.07 ç ÷ + 14.53 ç ÷ +ú (6.95)
2bd 2 ê èdø èdø èdø úû
ë
for ∅.2 < a/d < ∅.6.
In terms of the flexural strength, the expression (6.95) becomes
é æaö ù
K1s a ê1.93 - 3.07 ç ÷ + ú
ë èdø û
A crack may propagate in three different modes as represented in Fig. 4.3. While the
opening mode is likely to predominate, the other modes do exist, and combinations
of mode 1 with others are possible. The other modes will only operate if the crack
or the test piece is not free to move into a favourable orientation for mode 1. The
modes are referred to as
1 . the opening mode
2. the shearing or sliding mode
3. the tearing mode
6 Ceramic Materials Testing and Fracture Mechanics 405
P
P P
P P
Fig. 4.3 Open Mode, shearing or Sliding Mode, Tearing Mode fracture of a Brittle Material
τxy σyy
y
σxx
τyx
r
The stress distributions near to the end of a Griffith’s crack has been analysed as
follows:
406 D. Daily
s xx ü =
ìcos q
ï 2 ( ) éêë1 - sin (q 2 ) sin (3q 2 )ùúû ( tensile )
ï K ïï
s yy ý =
ï 2p r
ícos q 2
ï
( ) éêë1 + sin (q 2 ) sin (3q 2 )ùúû ( tensile ) (6.96)
t xy þ =
ïsin q
ïî 2 ( ) cos (q 2 ) cos (3q 2 ) ( shear )
s zz = n (s xx + s yy ) ( P. strain ) ( tensile )
= q ( P. stress )
t zz = t yz = q ( shear )
c
sx =s × cos (q / 2 ) (1 - sin (q / 2 ) sin ( 3q / 2 ) (6.98)
2r
σ represents the general stress in the material, and the trigonometrical terms are the
resolved components of it.
s xx = 2 êë ( )( ( ) ( )
ì- sin q é 2 + cos q cos 3q ù
ï 2 2 úû
ïï
K
s yy = p
2p r
í
ï
2 ( ) ( ) ( )
sin q coss q cos 3q
2 2
t xy =
ïî 2 êë ( )( ( ) ( )
ï cos q é 1 - sin q sin 3q ù
2 2 úû
s zz = t xz = t yz = q for plane stress, (6.99)
= n (s xx + s yy ) and t xz = t yz = q for plane strain
6 Ceramic Materials Testing and Fracture Mechanics 407
t xz =
K III
2p r
( 2 ) , andt
- sin q yz =
K III
2p r
( 2)
cos q
(6.100)
s xx = s yy = s zz = t xy = q
The K values depend on the applied loading and crack geometry, and determine
the intensity of the local field.
It is also possible to express the above equations in polar coordinates.
The value of r has certain limitations, i.e. the equations are not valid if r → θ or
becomes large in relation to the length of the crack.
Where two crack modes operate simultaneously, the resolved components of
each in a specific direction are additive.
6.4.5 T
he Application of K1 in terms of is a Materials
Property [2]
y
c δc
O
x
1
2
(s yy m y + t xy mz + t xy mz ) (6.101)
If crack extends from ∅ to δc, strain energy increase =
dc
(d U E ) = 2 òU E dx
o
dc
(6.102)
= ò (s yy m x + t zy m Lz ) dx
o
æ dm E ö dm E
In the limit ç ÷=
è d c ø dc
Let X be measured from the ∅ (closed position) where f < X < d c and let fail-
ure be in mode 1.
Work done by BOTH crack surfaces
dc
dU = 2 ò 1 (6.103)
2
o
Now σ is the stress normal to the crack surface before it opens or extends (i.e.
crack is in the closed position).
U is the normal displacement of the crack surfaces after the crack has opened.
dc
d UE 1
G= = ( in the limit ) òs udx. (6.104)
dc dc o
6 Ceramic Materials Testing and Fracture Mechanics 409
K1
s yy = cos(q / 2 ) × [1 + sin(q / 2 × sin ( 3q / 2 )] (6.105)
2p r
K1 r
m= (1 + u ) éë( 2k + 1) sin(q / 2 - sin ( 3q / 2 )ùû (6.106)
2E 2p
so that for σyy (σ directed towards the surface) q = f and r = x .
And for μ (μ displaced away from the surface) q = p and r = d c - x .
Substituting,
K1 K1 (d c - x )
s yy = m= (1 + v )( 2k + 2 ) (6.107)
2p x 2E 2
1
d Uc 1 K12
dc
(d c - x ) 2 dx
G= ( mode 1) = (1 + y )( k + 1) ò (6.108)
dc d c 2p E x
o
d c ×p
The value of the integral (Hahn, p. 149) is [5].
2
Giving
K12
G= (1 + Y )( k + 1) (6.109)
4E
At fracture, for plane strain and stress
K12c K2
G=
E E
( )
( plane stress ) G = 1c 1 - Y 2 ( plane stress ) (6.110)
By these simple expressions therefore, it is possible to relate fracture energy and
fracture toughness.
σ θθ
y τ rθ σ rr
r σ rr
σθθ
For mode 1
K1
s qq = cos3 (q / 2 ) = K1 fqq (q )
2p r
K1
s rr = cos (q / 2 ) éë1 + sin 2 (q / 2 ) ûù = K1 frr (q ) (6.111)
2p r
K1
t rq = sin (q / 2 ) cos2 (q / 2 ) = K1 frq (q )
2p r
For mode 2
K II
s rr = sin (q / 2 ) éë1 - 3 sin 2 (q / 2 ) ùû = K II frr ( e )
2p r
K II
s qq = - 3 sinn (q / 2 ) cos2 (q / 2 ) = K II fqq ( e ) (6.112)
2p r
K II
t rc = cos (q / 2 ) [1 - 3 sin 2 (/ 2] = K II frq (q )
2p r
Fig. 4.7, is from Lawn and Wilshaw’s “Fracture of Brittle solids and is a graphi-
cal illustration of Figs. 4.32 and 4.33, with θ extended over the range −π to +π [1].
Fig. 4.8, from the same source, demonstrates the effect of adding these curves.
Assuming K I = K II , it is possible to see how an asymmetric function develops for
K when the f(θθ) are added.
An inclined crack in a loaded plate will be subject to two stresses and the stress
intensity factors are transformed accordingly as
Total normal component
f rr
1
fθθ
frθ
-1
-π 0 π
θ
frr Mode II
1
frθ
-1
fθθ
-2
-π 0 π
θ
4.0
Modes I + II
1 1.0
0.25
KII /KI = 0
0
–π 0 π
q
412 D. Daily
I
S
II
SA
2 Eg K1
sf = = (6.114)
pc pc
where c = crack length
K1 = s f p c (6.115)
6 Ceramic Materials Testing and Fracture Mechanics 413
b
P
K1c 3P1
sf = 1
= (6.116)
2bd 2
Y ×c 2
1
3P1Y c 2
K= where
2bd
Y = Ao + A1 ( c / d ) + A2 ( c / d ) + A3 ( c / d ) + A4 ( 4 / d )
2 3 4
(6.117)
The curves (Fig. 5.2) relate Y with c/d. It clearly shows that a c/d value between
0.1 and 0.2 is most favourable and contrasts with the work of fracture experiment.
Values of A coefficients
Loading A0 A1 A2 A3 A4
Type
1/w = 8 +1.96 −2.75 +13.66 −23.98 +25.22
1/w = 4 +1.93 −3.07 +14.53 −25.11 +25.80
Four point +1.99 −2.47 +12.97 −23.17 +24.80
3.0
2.8
2.6
y
2.4
2.2
2.0
6.5.2 The Double Torsion Method [3, 16, 21, 24, 30]
Where a greater degree of control over the crack progress is required, then the dou-
ble torsion or double cantilever methods are preferred. They also lend themselves to
the study of stress corrosion effects.
If y is the displacement of the loading points, P is the applied load and a is the
crack length, from the compliance relationship,
y = P ( Ba + c ) (6.118)
Differentiating w.r.t. time,
dy dp PBda
= ( Ba + c ) + (6.119)
dt dt dt
da
Note that is the crack velocity, which at constant displacement
dt
dp
da (
Ba + c )
dy = f dt
is given by, = (6.120)
dt dt BP
The values for P and a are interrelated, when the specimen is subject to constant
displacement. When the crack has finished growing, we have Pf and af, hence
6 Ceramic Materials Testing and Fracture Mechanics 415
K1 = Pb′ 3 (1 + √−)
b P/2
P/2
bd3dn dn
P/2
d P/2
P ( Ba + c ) = Pf ( Baf + c ) (6.121)
Substituting for (Ba + c) in eq. (6.120), we obtain,
da P dp
= 2f ( Baf + c ) (6.122)
dt p B dt
For a relatively large crack length, where c <<< B, the equation simplifies to
da p dp
= - 2f af (6.123)
dt p dt
The crack velocity can be found if the initial and final loads, the rate of load relax-
ation with time, and the final crack length can be measured. The specimen is loaded
to a suitable level (below K1) and for a fixed displacement, the load decay with time
is followed on the pen recorder coupled to the load-cell of the testing machine. The
final crack length is measured, probably with the aid of a dye penetrant in the crack.
Also we note that
1
é 3 (1 + u ) ù 2
K1 ( orK1c ) = P1 ( or P1c ) wm ê 3 ú (6.124)
ë wt t n û
1
é 3 (1 + u ) ù 2
K I = PWm ê 3 ú
ë wt t n û
416 D. Daily
W
P
tn
P/2 wm
III
crack velocity da
KIt
dt
KId
LOG V
Vt
II
I
KO
VO
da
Figures 5.5 and 5.6 show the characteristic shapes of the curve when is plot-
dt
ted against K1. In the results for polycrystalline alumina, K1c is equal to 5.2 MN m−3/2.
The slope in region 1 is often referred to by the symbol “n” and called the “crack
susceptibility.” One can contrast the behaviour of alumina in moist air with that in
toluene (where air is excluded). In air the onset of stress corrosion is at a lower
stress level than in toluene and in vacuum it is depressed still further.
Region 2 is depressed in high corrosion situations [24] and the effect of Ca and
Mg ions in the alumina depresses the value of n. Increasing the purity of the alumina
increases n to the extent that measurement becomes very difficult because the initial
reading approaches K1c, although the early detection of the movement of the crack
by acoustic emission should simplify this measurement [25].
An important consideration in the double torsion technique is that the crack front
has a curved profile [16]. Allowance can be made for this in the calculation, but a
method which almost eliminates this to use very thin specimens.
6 Ceramic Materials Testing and Fracture Mechanics 417
Alumina of various categories has been used for the manufacture of microcir-
cuits and these can be obtained in 50 mm squares which are very suitable for the
double torsion test. Some difficulty has been encountered in making an initial crack
in the edge to begin the process. The following technique has been successful. A
small cut (<1 mm) is made in the edge, and this is dipped symmetrically into a bath
of Silicone oil at 180–200 °C. When the edge with the crack has attained the tem-
perature, it is then plunged into cold water. With practice the thermal shock can be
used to make a short crack in the side of the specimen which can then be placed in
the equipment and tested.
Results of Double Torsion Experiments (Davidge [3])
This technique was the original method for the investigation of fracture energy and
K1. The formula for K1 is
3.45 Pa é æ t öù
K1 = 3 ê1 + 0.7 ç a ÷ ú (6.125)
bt 2 ë è øû
418 D. Daily
The technique suffers in that it is dependent on a, the crack length. Glass and epoxy
resins etc. create no problems, because the crack can be seen due to total reflection
on the air surface within the crack. K1 for sapphire was also measured in this way.
But for opaque ceramics it is not very successful because the exact end of the crack
cannot be seen, unless the specimen is dismounted and a dye penetrant used to
reveal the end of the crack. This will alter the surface conditions and destroy the
value of the experiment.
Another difficulty is caused by the asymmetric loading of the specimen on a
conventional hard testing machine. The weight of the test piece will influence the
result. It can be used successfully in certain purpose built machines which put the
test piece in tension horizontally [15]. This technique also allows stress-corrosion
measurements to be made.
It has been said [28] that this technique measures the pure K1 fracture mode.
While this is probably true, the results do not differ significantly from those obtained
by the double torsion technique.
Because the DCB method depends on a knowledge of the crack length, an alter-
native method has been proposed which eliminates this, carefully removed and the
specimen dried before use. The moment arms were attached with high strength
epoxy resin which was cured at 70 °C. Small lugs were also incorporated in the
moment arms to meet with a transducer for measuring the separation of the crack
surfaces. Alumina proved to be a very difficult material to measure in this way.
When one attempted to work at constant strain, the crack did not extend even though
one approached within 90 % of K1c. Any further increase simply shattered the speci-
men which was clearly too stiff.
2t
a
Fig. 5.9 Specimen
Fracturing across Width
due to Surface
Imperfections
The crack velocity can be obtained from the measurement of the crack opening,
or from the depression of the loading points by a simple relation.
If the beam deflects through an angle θ, then tan θ = d/L = a/r = aM/EI. Hence
a = (EI/ML) ∙ d.
Like the DT, or DCB this method is easy to operate if there is significant stress
corrosion, but with increased purity of materials the stress intensity factor K1
approaches K1c as the regions I and II of the curve of Fig. 7.4 diminish and if one
420 D. Daily
attempts to allow the stress to relax from an initial constant strain it is more than
likely the specimen will fracture prematurely.
An alternative to this is to perform the experiment under constant stress. This can
be generated by means of a long lever arm on the end of which is a container of
zircon sand (chosen for its high density). This system, which admittedly appears
crude, is capable of very fine adjustment, and the stress actually on the moment
arms is read off from the testing machine chart as before. It proved quite possible to
control the growth of the crack by this method, but the values of K1 in those speci-
mens which fractured were twice as high as expected. This was because in the
constant loading method an equal quantity of energy is used to strain the specimen
as that used in extending the crack (p. 51). Before it is possible to accept these
results as reliable, a check for the influence of crack kinetic energy should be made
[1]. Attempts to use the constant moment method in this manner for pure alumina
gave results which were no better than the double torsion. A large number of plots
were made with K1 ranging from 2.2 to 2.5 MN m−3/2 over a velocity range from 10−8
to 10−5. The constant velocity region (II) was assumed to be about 10 m/s and this
had a particularly wide scatter of results. K1c was in the range 4.5–5 MN m–3/2.
A possible advantage of this configuration is that by adjustment of the position
of the test-specimen one can perform fracture experiments in the K3 mode.
Precautions have to be taken to prevent
(a) the specimen rotating back into the normal K1 mode, and
( b) the specimen fracturing across its width from a surface imperfection near the
point of attachment of the moment arms. Cladding with a thin piece of metal
attached with epoxy resin accomplishes this, and it does not significantly alter
the moment of inertia of bending of the specimen.
An elegant method for the measurement of fracture toughness in the tensile mode
has been reported by Kapp et al. [26]. It is intended for use with hollow cylinder
geometries, and one can appreciate that it can be used in reverse, i.e. from a knowl-
edge of K it is capable of detecting critical sizes of flaw in a hollow cylinder geom-
etry. Much of the mathematical theory is available in Refs. [17, 18, 26], but the
paper of Kapp et al., presents the formula for K in a readily applicable form.
æ ö
ç P ÷é
K =ç 3 X / W + 1.9 + 1.1 a ù
1 ÷ë wû
ç BW ÷
è 2ø
é
( ) ( )
æ 1 - r1 ö ù F a
2
X ê1 + 0.25 1 - a ç
ë w è r2 ÷ø úû w (6.126)
( w ) = éêë( a w ) / (1 - a w ) ùé
( ) ( )
1 3
where F a
2 2
a a
2
a
3
ù
ú ê3.74 - 6.30 w + 6.32 w 2.45 w ú
ûë û
6 Ceramic Materials Testing and Fracture Mechanics 421
Limits:
For 0.45 ≤ a/w ≤ 0.55
For all r1/r2 values, Accuracy within ±1 %
For X/W either 0.5 or 0
For 0.2 ≤ a/w ≤ 1.0
For all r1/r2 values, Accuracy within ±1.5 %
For X/W either 0.5 or 0
For 0.2 ≤ a/w ≤ 1.0
For all r1/r2 values, Accuracy within ±3 %
For 0 ≤ X/W ≤ 1
An application of this method is found in Tan and Davidge’s paper [18]. This
paper also describes an alternative mode of gripping the specimen, and making it
more applicable for evaluating the quality of tube materials.
The basis of this test is a short cylindrical or block specimen which has a notch
machined into it of such a shape that it leaves a “V” configuration behind in the
remaining ceramic. A basic requirement for this must be a very precise fabrication
Fig. 5.10 Short
Cylindrical or Block
Specimen
2θ
d
422 D. Daily
method for the initial cylinder or block, and a means of sawing in the notch to very
precise specifications after firing. Testing is carried out by opening up the notch and
this may be achieved in two ways.
1 . By pulling the specimen apart using a specially shaped tensile jig.
2. Inserting a thin inflatable bag into the notch and pressuring it until the test-piece
breaks.
The latter method is clearly quite attractive if expensive test machines are not
available.
The crack growth from the point of the V-shaped slot in the specimen is initially
stable, even for the most brittle materials. Thus a real crack is obtained during the
initial loading of the specimen and an ever increasing load is required for continued
crack growth until the crack reaches a critical length ac. Thereafter, the load to pro-
duce further crack growth decreases. ac depends upon the specimen geometry, but is
independent of the specimen material, provided only that it behaves in accordance
with linear elastic fracture mechanics. The equation for K1c is as follows:
K1c = AF Pc B (6.127)
Pc is the peak pressure on the notch face during the test.
B is the specimen diameter.
AF is a dimensionless constant dependent on the specimen geometry and loading
configuration.
6 Ceramic Materials Testing and Fracture Mechanics 423
W / B = 1.5000 + 0.006
C / B = 0.03 + 0.002
a0 / B = 0.531 + 0.006
a p / B = 0.482 + 0.002
8 = 55.2 + 0.5°
For block specimens L/B = 0.870.
Figures 5.13 and 5.14 are the alternative configurations. Figure 5.15 is a section
through the notch.
For the ration given above, A = 7.51.
Typical dimensions are going to be determined by the width of the notch. For a
1 mm notch, one requires a 33.3 mm diameter.
Main reference is L.M. Barker in ASTM/STP 678 (p. 73 [6]). (The author dis-
cusses the use of a curved profile on the “V”. This might be achievable with com-
puter machining.)
One of the difficulties of obtaining fracture toughness data is that for the most part
specially prepared specimens with characteristic shapes have to be produced. It
would be very desirable to have a technique which would eliminate these difficulties
even at the expense of accuracy if it could be used to sort bodies into categories, so
that when a promising material was forthcoming, fewer refined test specimens
would have to be prepared. It is possible to do this with a hardness indenter which
is a standard piece of equipment in metallurgical and engineering laboratories.
Figure 5.17 represents the impression left by a Vicker’s diamond pyramid indenter
and showing cracks emanating from each corner.
Figure 5.16 is a longitudinal section through the diagonal of the pyramid base.
The advancing crack front is the semicircle of diameter 2D′ (surface extremities of
the crack) and of sub-surface depth D.
2a is the length of the diagonal indentation, and to be
Effective, D/a ≥ 2.
An expression K1c, given by Lawn and Swain [3] is
(1 - 2g )(aro )
1/ 2 1/ 2
æPö
K1c = çD÷ (6.128)
21/ 2 p 2 b è ø
where
P = indenter load.
D = crack depth.
424 D. Daily
γ = Poisson’s ratio.
ρo = mean contact pressure.
α = dimensionless constant determined by indenter geometry (for Vicker’s a = 2 / p ).
β = dimensionless constant determined by the deformation zone geometry (for
Vicker’s or Knoop indenter β ~ 2).
6 Ceramic Materials Testing and Fracture Mechanics 425
1 æ P ö
K1c = (6.129)
p 3 / 2 tany çè D1/ 2 ÷
ø
where ψ is the indenter cone half angle (Strictly it is only valid for conical indenters)
but it gives a good fit to crack growth obtained with a Vicker’s indenter. In this case
ψ is assumed to be 68°, the half angle between the faces. Analysis predicts that
3
P / D 2 will be a straight line and the hardness is not needed.
Other points about the methods:
1 . Crack length should be ten times the max-grain size.
2. Crack depth should be less than 1/3 specimen thickness.
3. For ceramics loading rate was 5.0 mm/min with as Vickers’s indenter.
4. Measurement of crack length was by travelling microscope with dye-penetrant
to render crack.
5. Data collected at four indenter loads with a min. Of three indents at each load.
For materials with high stress corrosion rates, the indenter may be covered with
paraffin oil otherwise tests are carried out in air. For transparent materials photogra-
phy is probably a better method to measure crack lengths.
2a
2D'
426 D. Daily
Fig. 5.17 Indentation
cracks introduced by a
Vickers indenter. (a)
Schematic showing the
parameters used in the test.
(b) A typical crack pattern
obtained in Si3N4
(polarised light reflected
micrograph)
Fig. 5.18 Indentation cracks introduced by a Vickers indenter (a) schematic showing the param-
eters used in the test
When a crack in a ceramic is exposed to the air, or other environment, the freshly
exposed, very reactive surfaces will undergo a chemical reaction with whatever is
nearest to it. If you consider that the bonds which hold the crystal together have
been disrupted, and there are the free unbounded electrons of the valency forces
available to react, then this effect is inevitable. If it were possible to compress a
6 Ceramic Materials Testing and Fracture Mechanics 427
crack before the atmosphere was able to attack it, the crack would heal up again, and
this can be demonstrated in glass. The net effect of the atmosphere is to stabilise the
crack in the open position. The most deadly corrosion agent is simply water vapour,
and the amount present in normal moist air is quite enough to cause damage.
Obviously more active agents, e.g. saline solution, present in animal tissues will be
even more dangerous. In effect the (OH)− groups from the water (or Cl− ions) satisfy
the released valency forces. The combined effect of stress in the material and the
activation energy of the chemical attack causes a crack to steadily grow until it
eventually reaches the critical Griffith flaw size, and if the stress is sustained, the
material will break.
Referring to the double torsion graph (Fig. 5.5), the linear region (I) is that for
sub-critical crack growth (or stress corrosion) and for a particular stress (on the
abscissa) the appropriate crack velocity can be read off on the ordinate. As the crack
grows the stress intensity factor increases, the velocity increases, and a catastrophic
situation ensues (i.e. when the stress intensity factor becomes critical).
Under constant stress σ time to failure is given by:
Cc
dc
t= ò
Ci
v
(6.130)
1
Ci is the initial crack size and Cc is the critical size, and K1c = g s Cc 2 .
1
Because generally K1 = g s C 2 , differentiation given
2 K1
dC = dK1 (6.131)
s 2g 2
Substituting for dc in Eq. (6.130).
K1 c
2 K3
t=
s g
2 2 ò
K1i
v
dK1 (6.132)
Referring to the double torsion graph (Fig. 5.5), we have,
Region I, log v = const.x log K1 i.e v = α1K1.
Region II, Constant velocity region v = α2.
Region III, A third term is possible, but is usually not considered, because crack is
reaching sonic velocities, and the specimen is in effect broken at the end of
region II.
Let K1* be the limiting value of K1 separating regions I and II
2 æ 1 K1 (1- n ) 1
K1 c
ö
then t( minimum ) = ç ò K1 dK1 + ò K dK ÷ (6.133)
s g
2 2 ç a1 k a2
1
÷
è 1 K1 ø
428 D. Daily
For a real situation, the limit of tolerable stress corrosion should be the end of
region I, so if we simply integrate this part of the equation we get,
2
t( min ) =
s 2g 2a1 ( n - 2 )
( K12i- n - K12*- n ) (6.134)
2-n 2-n
Furthermore if n is large (typically) 10, K 1i >> K 1* so that
2 K12i- n
t( min ) = (6.135)
s 2g 2a1 ( n - 2 )
1
Let a particular specimen have an initial crack length c1, then K1i = sg ci2 ; com-
bining with Eq. (6.135), (6.136) we see that tσn is constant (for this particular speci-
men). Thus the ration of the lifetimes tσi, tσj, at two subcritical stresses σi and σj, is
given by
n
æ si ö ts j
çç ÷÷ = (6.136)
ès j ø ts i
This relation allows the strength/probability/time diagram to be made, combining
the statistical and time-dependent properties of ceramics.
The failure times under constant stress conditions are likely to be inconveniently
short compared with those measured a constant rate of strain. These times are
related by the expression
te = ( n - 1) ts (6.137)
(where n is the slope of the K1/v curve in region I).
Suppose a specimen fails under stress σεi, in testing time tεi then s e i = Ee i te i .
For another specimen failing under stress σεj, in time tεj, then s e j = Ee j te j .
for which
æ s ei ö e i te i
çç ÷÷ = (6.138)
è se j ø e j te j
0.99
0.98
0.95
0.90
Survivel probability
0.80
0.50
0.10
0.05 1055 1045 1035 1025 105 15 0.15
0.01
0.001
6.6.2.1 Procedure
(1) About 100 test specimens (bars or discs) are prepared and the side which
becomes under tension is polished.
(2) Tests on five batches of 20 are conducted over a wide range of strain rates (each
batch having a constant rate). The Weibull modulus for each is obtained.
(3) A representative stress is chosen for each batch (usually the median) and n is
calculated between pairs of batches from the median stress and the straining
rate.
(4) Select a standard failure time (usually 1 s) and normalise the most convenient
Weibull spread to this time (using Eq. 6.137)?.
(5) On the basis of this, draw parallel Weibull slopes in decades of seconds until the
desired life of the component is reached. For convenience remember that
3–107 s is equivalent to 1 year.
For n = 10 each decade represents a reduction of strength of 21 %.
6.6.2.2 Application
This procedure is necessary when ceramic components are designed for use in cor-
rosive environments. Typical applications are hip-joints and grinding wheels. For
the hip prosthesis the corrosive environment is the body fluid which is approxi-
mately normal saline (e.g. tears). A projected life is 50 years, which makes a revi-
sion of the operation unlikely!, and the stress chosen as a basis for calculation is 10
times body weight (i.e. if one jumped down four stairs and landed on one leg.) The
calculation is pessimistic because this is an exceptional stress condition and not a
constantly applied one.
Apart from the knee these strictures do not apply to any other ceramic joint
replacement because they are not under stress.
In the case of grinding wheels, the normal moist air we live in is corrosive agent,
and the centrifugal force of its rotation creates the stress. Failure of the wheel cata-
strophic and fatal accident. Manufacturers of wheel should therefore specify the
maximum number of hours for safe usage.
One should be able to apply this reasoning to any kind of ceramic component
which will be under stress during its useful life. How do you regard ceramic engines
in cars, in lorries, or in aircraft!? Perhaps they may be regarded as disposable items
since they should become cheaper than conventional materials and thereby create
and environmental disaster with all the indestructible ceramic litter.
The inherently brittle nature of ceramics makes their use in high stress situations
attended with some risk, and before use car be contemplated one has to make a very
careful assessment of this risk.
6 Ceramic Materials Testing and Fracture Mechanics 431
K1a = Ys a c p (6.141)
Thus
sa
K1a = K1c (6.142)
sp
2 K12c- n (s a / s p )
2-n
K ( min ) = (6.143)
s a2Ya21 ( n - 2 )
432 D. Daily
Fig. 6.3 S.P.T diagram for Co-WC alloy at different temperatures (Braiden et al. [23])
Choosing different ratios of σa/σp the stress, σa may be plotted against the time for
which it is applied (Eq. 6.143), on a log-log plot this must have a slope of −2. We
can superimpose on this diagram curves which relate the time in seconds to the
stress for a specific probability of failure. These can be obtained from the S.P.D.
diagram. It is possible to predict therefore, for a specific ratio of failures to survi-
vors, what the service time would be under a given stress, and also the proof testing
ratio σp/σa which would guarantee this performance.
Figure 6.3 shows that if we allow a one per thousand (10– 3) proof-testing failure
ratio, a service life of 5 ´ 10 7 s is guaranteed for a service stress of 30 MPa, if the
proof testing ratio had been 2.6/1.
(Check that if we allow one per hundred failures during proof-testing at a ratio of
3.6/1, the service life under the same stress is 1016 s. Since this is about 108 years
they would have to have been made before mankind appeared.)
For hip joint balls designed to operate with a conical fixation, σa is 9000 N
applied over the effective conical surface, and σp is 25,000 N. This should last about
50 years. Because there is some slight degree of failure when the load is removed,
the specimens are reloaded to the working stress level to ascertain they are still
sound.
6 Ceramic Materials Testing and Fracture Mechanics 433
The design of the ceramic head for the hip joint prosthesis requires the solution of a
number of problems, important among which are as follows:
1. Using a conical fixation to the metal stem creates a hoop stress which will be
severe if the cone is forced or impacted too violently into ball.
2. It is unlikely that the conical surfaces will be machined so precisely that stress
will not be concentrated at some point on the cone.1
3. A working stress limit on the ball is considered to be about ten times weight,
taking account of sudden impacts on one limb.
4. The effect of stress corrosion after 30–40 years may weaken the ceramic.
The testing of every combination of a hip device design with multiple modular
heads and neck arrangements and sizes are not practical in terms of time or cost for
the pre-market approval process of a device. To overcome this issue often “the worst
case” prosthesis design/combination is identified and used as the bases for the
approval of the range of similar devices. The identification of the “worse case”
design may not always be obvious; however, it is generally accepted that femoral
heads with the largest off set of +12 mm [29] on the smallest stem in terms of diam-
eter produce the most extreme stress conditions in the neck of the femoral stem. In
particular with new device designs were there is no historical data to draw from then
Finite Element Analysis (FEA) can be a valuable tool to determine the worst case
design combination to test. When this chapter was written in the 1990s FEA was a
relatively recent computational technique; however, it is now widely used in medi-
cal devices to verify a design is “fit for purpose”. FEA is used to simulate the load-
ing conditions within a human body and therefore assisted in the development of
more realistic testing protocols.
In the original investigation, the aim was to answer questions raised (Sect. 6.7)
questions 1, 2 and 3. In the original computational process the optimum mesh of
152 elements was drawn and eight nodes per element were defined at which the
stresses were measured. In modern computational studies it is common to use 4000
[30] mesh units to represent the possible contact areas using a new range of analyti-
cal models; however, the original work describes the basic theory behind the analy-
sis therefore it was decided to retain much of the original texts.
Three load cases and three restraints were defined, as follows (Fig. 7.2).
Accurately machined cone and socket sections from epoxy resin were analysed photoelastically.
1
However, careful the machining, a “high spot” which concentrated the stress was always revealed.
Much depends on the confidence one has in the machining operation.
434 D. Daily
Case (A) was chosen to represent the very small area of contact of the prosthesis
ball with a ceramic acetabular cup.
Case (B) was chosen to represent the wider distribution of the load when a less
rigid UHMWPE (ultra high molecular weight polyethylene) acetabular cup was
used.
Case (C) was chosen to represent the situation in the body, where the load was
not applied axially but spread over an arc between 20° and 55° to the cone axis. This
case is somewhat exaggerated and chosen to accommodate the computational sys-
tem. It serves to indicate that in a realistic situation where the load is applied at 25°
to the axis there will be more distortion than is represented by cases A and B.
6.7.1.2 Restraints
1. The load was transmitted to the shaft via the ball across the lower third of the
conical interface, i.e. at the end of the cone remote from the centre of the ball.
2. Load transmitted via the middle third, and
3. Load transmitted via the upper third, i.e. at the end of the cone nearest the
centre.
The finite element system requires information about the physical properties of the
material e.g. modulus of elasticity, Poisson’s ratio etc. and the load applied, which
was 9000 N (i.e. about ten times body weight. The body in this instance is rather
large, but not untypical of the type of elderly patient needing a hip replacement) [8].
Because the ceramic heads were symmetrical, the problem was considerably
simplified, and the analysis was able to provide plane strain diagrams of the
following:
1. The highly magnified “displaced shape” of the ball under load (Figs. 7.3, 4, 5).
The diagrams were chosen to represent (“worst cases”). The diagrams appear to
display very large distortions, but the scale of these is very large, i.e. 1 cm on the
diagrams represents the order of micrometres.
2. Diagrams representing the magnitude and direction of the two principle stresses
at the nodal points in the xy plane.
3. Diagrams showing equal stress contour lines through the nodal points for the
maximum and minimum principle stresses and for the hoop stress which is in the
z direction.
6 Ceramic Materials Testing and Fracture Mechanics 435
A stress printout for all the nodes which surround each element was also pro-
vided, from which it was possible to derive the average stress at the centre of each
element.
The finite element analysis was able to point to several factors about the design
of the prosthesis shape, the manner of application of the load and to enable specifi-
cations for the alumina to be drawn.
Examination of the diagrams for the principle stress (2) above and the equal stress
contour lines (3) show large tensile stresses at the base of the cone and at the cone
opening. This can be remedied by undercutting the base of the cone to turn it into a
dome, and by relieving the corner at the cone opening.
The displaced shape diagrams (1) and the stress contours (3) clearly show that the
application of the load over a larger area from the polyethylene cup produces a
higher concentration of large tensile stresses than is the case with the alumina ace-
tabular cup.
This involved the application of the four function Weibull equation [9, 32] as
follows:
The probability of failure of a test specimen can be represented by the standard
Weibull equation.
ìï é (s - s ) ù m üï
Pf = 1 - exp í- ê u ú ý (6.144)
ïî ë s 0 û ïþ
Another form of this equation is given by
ïì æ 1 ö æ s ö ïü
m m
Pf = 1 - exp í- ç ÷! ç ÷ ý (6.145)
ïî è m ø è s f ø ïþ
436 D. Daily
which employs the relation s f = s 0 (1 / m )! , where (1/m)! are the “gamma func-
tion” and is available in tables, and s f is the mean stress of the batch.
The failure stress of a specimen is also related to its volume, given by the expres-
sion (based on 6.144)
s f mV = s fv m v (6.146)
s fv is the tensile breaking stress of a test piece of reference volume v. The purpose
of the exercise is to evaluate this.
Combining the two we arrive at the four function Weibull equation
ì ö V m üï
m
ï æ1ö æ 1
Pf = 1 - exp í- ç ÷!m × ç ÷ × ×s ý (6.147)
ç ÷ v
ïî è m ø è s fv ø ïþ
The stress in each volume element dV is made up from the three principal
stresses so σmdV has to be applied three times over in each element, i.e.
( )
s 1m + s 2m + s 3m dV . Account has also to be taken of whether the stress is tensile or
compressive. Use is made of the Heaviside function H(σ). Its value is unity for posi-
tive tensile stresses, and −∝ for negative compressive stresses, ∝ is the modulus of
the ratio of the mean failure stresses per unit volume in uniaxial tension and com-
pression. In this exercise ∝ was conveniently given the value of −10 for the alumina
under compression.
The final expression for the failure probability of the specimen is given by
ì m ü éæ s öm æ s öm æ s ök ù
ï æ1ö æ 1 ö 1ï
Ptotal = 1 - exp í -ç ÷ !m ×ç
m
îï è ø è fr
s
÷ × ý ò êêçè H (s11 ) ÷ø + çè H (s22 ) ÷ø + çè H (s33 ) ÷ø úú dv ( Ref. 10 )
ø v ïþ r
ë û
(6.148)
where the “stress-volume integral” is taken over all the elements in the body.
From the finite element diagrams, the area of each element was calculated in
square millimetres by approximating each to a trapezoid. The volume of revolution
of each element about the centre line of the section was then calculated.
To construct Tables 6.4 and 6.5 the 152 elements were divided into eight separate
computer programmes to estimate the stress-volume-integral, the sum of which is
shown in the first column. For convenience of calculation the stress unit chosen was
10 MPa.
6 Ceramic Materials Testing and Fracture Mechanics 437
3.5
3.0
1012
1010 2.5
100 yr
Time in service s
108 10 yr
1yr
2.0
1min
106
1d
104
1h
R = 1.5
102
1min
1 2 3 4 5 6 8 10 20 30 40 50 60 80 100
Fig. 7.3 Displaced shapes of prosthesis head. Load case A. Constraints 1 and 3
Fig. 7.4 (continued)
Str/vol. Int Ref., vol. Weibull mod. Gamma for “m” value Failure prob. reqd.
Sv(-----)dv (v) (m) (1/m)! (Pr)
1.050E14
8.890E10 mm3 10.000 0.610 0.100
6.890E13 1000.000 ln(1 − Pr)
9.280E15 −0.105
1.030E12 −Gamma
1.140E10 −0.610
5.330E14
5.800E14
Total Total/Ref. V S/V Int × Gamma “A”/E−14 Req. str/10
1.057E16 1.057E13 −6.446E12 1.634 23.91 MPa
Str/vol. Int Ref. vol. Weibull mod. Gamma for “m” value Failure prob. reqd. (Pr)
Sv()dv (v) (m) (1/m)!
1.05E14
9.890E10 mm3 10.000 0.610 0.100
6.890E13 1000.000 In (1 − Pr)
9.280E15 −0.105
1.030E12 −Gamma
1.140E10 −0.610
5.330E14
5.800E14
Total Total/Ref. V S/V Int × −Gamma “A”/E−14 Req. str/10
1.057E16 1.057E13 −6.446E12 0.016 38.099 MP a
“A” = (S/V Int × Gamma)/log(1 − Pr)
With a Weibull modulus of 10 the required tensile strength of a 1 cc test specimen is 381 MPa for a 0.1 % failure probability
The required tensile strength is 239 MPa for a 10 % failure probability
D. Daily
Table 6.5 Survival Probability Vs. Stress
Str/vol. Int Ref., vol. Weibull mod. Gamma for “m” value
Sv(-----)dv (v) (m) (1/m)! Failure Prob. reqd. (P)
3.500E20
3.210E15 mm3 15.000 0.590 0.100
3.520E19 1000.000 ln(1 − Pr)
9.940E18 −0.105
1.460E17 −Gamma
1.140E10 −0.590
2.400E21
8.820E20
Total Total/Ref. V S/V Int × −Gamma “A”/E−14 Req. str/10
3.677E21 3.677E18 −2.170E18 4.856E−6 19.39 MPa
Str/vol. Int Ref. Vol. Weibull mod. Failure prob. reqd.
3.500E20
3.210E15 mm3 15.000 0.590 0.001
6 Ceramic Materials Testing and Fracture Mechanics
An obvious solution to the problems created by the conical fixation is to avoid this
if possible. A cylindrical fixation would have the following advantages:
Dangerous hoop stresses would not be created by impacting the ball in service or
during fixation.
It is easier and less expensive for the manufacturer to press solid spheres and
afterwards machine out the cylindrical fixation hole, rather than having to machine
out a precise conical hole or include it in the original pressing.
To obtain the degree of bonding required needs a very precise frictional fit
between the stem and the ball. This has been achieved by
(a) Covering the stem with a plastic sleeve and forcing it into the cavity.
(b) Diffusion bonding a biocompatible metallised layer of precise thickness into
the surface of the cylindrical cavity in the ceramic so that it becomes integral
with the ceramic. A titanium shaft is subsequently fitted into the cavity. This is
still under active development.
Fig. 7.6 Femoral head under axial loading, Image supplied by Lucideon
References
1. Lawn BR, Wilshaw TR (1975) Fracture of brittle solids. Cambridge University Press, London
2. Jayatilaka A de S (1979) Fracture of engineering brittle materials. Applied Science Publishers,
London
3. Davidge RW (1979) Mechanical behaviour of ceramics. Cambridge University Press,
Cambridge
4. Kerkhof F (1983) Handbook of ceramics. Schmidt, Freiburg im Breisgau
5. Hahn HG (1976) Bruchmechanik. Teubner, Stuttgart
6. Freiman SW (ed) (1979) Fracture mechanics applied to brittle materials. A.S.T.M., S.T.P. 678
7. Braiden P (1975) An introduction to Weibull statistics. A.E.R.E.R 7165
8. Maier HR, Staerk N, Krauth A (1983) In: Hastings GW, Williams DF (eds) Mechanical prop-
erties of biomaterials. Wiley, New York
9. Stanley P, Fessler H, Sivill AD (1973) An engineer’s approach to the prediction of failure prob-
ability of brittle components. Proc Br Ceram Soc 22:453–487
10. Davidge RW, Tappin G (1968) The effective surface energy of brittle materials. J Mater Sci
3:165–173
11. Feltham P (1966) Deformation and strength of materials. Butterworths, London
12. Timoshenko S (1976) Strength of materials, vol 1. Van Nostrand-Reinhold, New York, p 92
13. Durelli AJ, Morse S, Parks V (1962) The theta specimen for determining tensile strength of
brittle materials. Mater Res Stand 2(2):114–117
14. Daniel IM, Weil NA (1962) Effect of non-uniform stress fields. Studies of the brittle behaviour
of ceramic materials. Report AD 277659
15. Griffiths R, Holloway DG (1970) The fracture energy of some epoxy resin materials. J Mater
Sci 5:302–307
16. Evans AG (1972) A method for evaluating the time-dependent failure characteristics of brittle
materials and its application to polycrystalline alumina. J Mater Sci 7:1137–1146
6 Ceramic Materials Testing and Fracture Mechanics 445
17. Evans JRG, Stevens R (1984) The “C”-ring test for the strength of brittle materials. Trans Br
Ceram Soc 83:14–18
18. Tan SR, Davidge RW et al (1980) Strength-controlling flaws in beta-alumina. Trans Br Ceram
Soc 79:120–127
19. Mordfib L, Kerper MJ (1969) Strength testing of ceramics. Office of Aerospace Research,
Arlington, VA
20. Paris PC, Sih GC (1965) Stress analysis of cracks. ASTM Spec Tech Publ 381
21. Williams P, Evans AG (1973) A simple method for studying slow crack growth. J Test Eval
1(4):264–270
22. Evans AG, Wiederhorn SM (1973) Proof testing of ceramic materials—an analytical basis for
failure prediction. Int J Fract 10(3):379–373
23. Braiden PM, Davidge RW, Airey R (1977) Time dependent strength parameters for tungsten
carbides containing 6 or 16% cobalt at room and elevated temperatures. J Mech Phys Solids
25:257–268
24. Lach S, Dailly DF, Hastings GW (1982) Stress corrosion of debased alumina. Proc Br Ceram
Soc 31:191–200
25. Dalgleish BJ, Pratt PL, Rawlings RD, Fakhr A (1980) The fracture toughness testing of ceram-
ics and acoustic emission. Mater Sci Eng 45:9–20
26. Lawn BR, Swain MV (1975) Microfracture beneath point indentations in brittle solids. J Mater
Sci 10:113–122
27. Champonier RP (1979) A.S.T.M./S.T.P. No. 678, p 60
28. Shaw MC, Braiden PM, de Salvo GJ (1975) The disc test for brittle materials. J Eng Ind
97(1):77–87
29. Khan I, Naylor M, Gupta G (2013) Characterization of Orthopaedic Devices. In: Bandyopadhyay
A, Bose S (eds) Characterization of biomaterials. Elsevier, Amsterdam, pp 323–351
30. Cilingir A (2010) Finite element analysis of the contact mechanics of ceramic-on-ceramic hip
resurfacing prostheses. J Bionic Eng 7(3):244–253
31. Smith R (2015) The application of digital image correlation technology in healthcare. Lucideon
White Paper
32. Lach S (1981) Ph.D. Thesis, C.N.A.A.
33. Carniglia SC (1972) Working model for porosity effects on the uniaxial strength of ceramics.
J Am Ceram Soc 55(12):610–618
34. Davidge RW, Evans AG (1970) The strength of ceramics. Mater Sci Eng 6(5):281–298
35. Duckworth W (1953) Discussion of Ryshkewitch paper compression strength of porous sin-
tered alumina and zirconia. J Am Ceram Soc 36(2):68
36. Knudsen FP (1959) Dependence of mechanical strength of brittle polycrystalline specimens on
porosity and grain size. J Am Ceram Soc 42(8):376–387
37. Passmore EM, Spriggs RM, Vasilos T (1965) Strength-grain size-porosity relations in alu-
mina. J Am Ceram Soc 48(1):1–7
38. Eudier M (1962) The mechanical properties of sintered low-alloy steels. Powder Metall
5(9):278–290
Chapter 7
Properties of Bioactive Glasses
and Glass-ceramics
A bioactive material is one that elicits a specific biological response at the interface
of the material which results in the formation of a bond between the tissues and the
material. A common characteristic of bioactive glasses, bioactive glass-ceramics,
and bioactive ceramics is that their surface develops a biologically active hydroxy
carbonate apatite (HCA) layer which bonds with collagen fibrils. The HCA phase
that forms on bioactive implants is equivalent chemically and structurally to the
mineral phase of bone. It is that equivalence which is responsible for interfacial
bonding1–3.
Bioactive materials develop an adherent interface with tissues that resist substantial
mechanical forces. In many cases the interfacial strength of adhesion is equivalent
to or greater than the cohesive strength of bone. The interfacial strength of a bioac-
tive implant bonded to bone is 15–40 times greater than the interfacial adherence of
non-bioactive materials (such as Al2O3) (Table 7.1 and Figure 7.1), tested in the
same animal model (rabbit tibia) (Figure 7.2)4.
Fig. 7.1 Comparison of interfacial bond strengths of bioactive implants with non-bonding
implants (alumina) using ‘pull off’ detaching test [4, 5]
7 Properties of Bioactive Glasses and Glass-ceramics 449
Fig. 7.2 Schematic of ‘pull off’ detaching test for determining bone-implant bonding (based upon
T. Yamamuro, ref. 4)
Fig. 7.3 Time dependence of interfacial bone formation for various types of bioceramic implants
Table 7.2 summarizes the physical properties of the bioactive glasses, glassceram-
ics, and ceramics in clinical use, with references. The bioactive glasses are single
phase amorphous materials which have high IB values (rapidly form a bone bond)
but have low mechanical strength and toughness. These materials should be used in
particulate form (as powders), as coatings, or in low load bearing applications, as
listed in Table 7.3. Bioactive glass-ceramics are multi-phase materials with a fine,
homogeneous grain size and good mechanical strength and toughness5 and interme-
diate IB values. They can be used in moderate load bearing
1.
L.L. Hench and E.C. Ethridge, Biomaterials, An Interfacial Approach, p. 137, Academic Press,
New York, 1982.
2.
ö.H. Andersson, K.H. Karlsson, K. Kangasniemi, and A. Yli-Urpo, Models for Physical Properties
and Bioactivity of Phosphate Opal Glasses, Glastech. Ber., 61, 300–305 (1988).
Table 7.2 Composition and Mechanical Properties of Bioactive Ceramics Used Clinically
Sintered7,8,9
hydroxypatite Sintered 10,11
Bioglass®1 Glass-ceramic3 Glass-ceramic4 Glass-ceramic5 Glass-ceramic6 Ca10(PO4)6 β-3CaO⋅
Composition 45S5 S53P42 Ceravital® Cerabone® A-W Ilmaplant® Ll Bioverit® (OH)2 > 99.2% P2O5 > 99.7%
Na2O 24.5 wt% 22.6 wt% 5–10 wt% 0 wt% 4.6 wt% 3–8 wt%
K2 O 0 0.5–3.0 0 0.2 3–8 wt%
MgO 0 2.5–5.0 4.6 2.8 2–21
CaO 24.5 21.8 30–35 44.7 31.9 10–34
Al2O3 0 0 0 0 8–15
SiO2 45.0 53.9 40–50 34.0 44.3 19–54
P2O5 6.0 1.7 10–50 16.2 11.2 2–10
CaF2 0 0.5 5.0 F 3–23
B2O3 0
Phase Glass Glass Apatite Apatite Apatite Apatite Apatite Whitlockite
(Ca10(PO4)6 (O,F2)) (Ca10(PO4)6 (β-3CaO • P2O5)
(OH)2)
Glass β-Wollastonite β-Wollastonite Phlogopite
(CaO · Si02) ((Na1K)Mg3
(AlSiO10)F2)
Glass Glass Glass
Density (g/cm3) 2.6572 3.07 2.8 3.16 3.07
Hardness 458 ± 9.4 680 500 600
(Vickers) (HV)
Compressive strength (MPa) 500 1080 500 500–1000 460–687
Bending strength (MPa) 42(Tensile) 215 160 100–160 115–200 140–154
Young modulus (GPa) 35 100–150 118 70–88 80–110 33–90
Fracture toughness, KIC(MPa 2.0 2.5 0.5–1.0 1.0
m1/2)
Slow crack growth, n 33 12–27
Index of bioactivity lB12 12.5 3.8 5.6 7.5 (est) 3.1
452 L.L. Hench and T. Kokubo
Fig. 7.4 The compositional dependence of bone bonding to bioactive glasses (region A) contain-
ing 6 weight % P2O5. Soft tissue bonding occurs for compositions with IB values > 8 (see text).
Region B: non-bioactive compositions. Glasses in Region C are resorbable. (Based upon chapters
1 and 3 in ref. 1.)
Fig. 7.5 Compositional dependence of bioactivity for glasses in the CaO–P2O5–SiO2. (Based
upon T. Kokubo, ref. 5.)
7 Properties of Bioactive Glasses and Glass-ceramics 453
sitions with the highest IB values develop interfacial bonding layers (Figure 7.6)
composed of both hydrated silica gel layers and Ca, P-rich layers. Compositions
with low to moderate IB values form thinner bonding zones composed primarily of
Ca-P-rich compounds. Non-bonding implants have a non adherent fibrous tissue
layer at the implant interface.
References
1. Hench, L.L. and Wilson, J. (eds) (1993) Introduction to Bioceramics, World Scientific
Publishers, London and Singapore, pp. 1–24.
2. Gross, U., Kinne, R., Schmitz, H.J. and Strunz, V. (1988) The response of bone to surface
active glass/glass-ceramics. CRC Critical Reviews in Biocompatibility, 4, 2.
3. Yamamuro, T., Hench, L.L. and Wilson, J. (eds) (1990) Handbook of Bioactive Ceramics,
Vol 1: Bioactive Glasses and Glass-Ceramics, CRC Press, Boca Raton, FL.
4. Yamamuro, T. (1993) A/W glass-ceramic: clinical applications, in Introduction to Bioceramics
(eds L.L. Hench and J. Wilson), World Scientific Publishers, London and Singapore,
pp. 89–104.
5. Kokubo, T. (1993) A/W glass-ceramic: processing and properties, in Introduction to
Bioceramics (eds L.L. Hench and J. Wilson), World Scientific Publishers, London and
Singapore, pp. 75–88.
Chapter 8
Wear
M. LaBerge and J.D. Desjardins
8.1 Introduction
Table 8.2 (continued)
Wear process Model Specifications
Fatigue wear Halling model: • Incorporates the concept of
[15] ηγ fatigue failure as well as
Wfa = K 2 FN
e1 H simple plastic deformation
Wfa = wear rate failure
η = line distribution of asperities
γ = constant defining particle size
e1 = strain to failure in one loading cycle
H = hardness of the softer material
K = wear coefficient
Corrosive Quinn model: • E
xplains wear in steel and
wear dA exp [ −Q / RcTc ] assumes that surface
[16, 17] Wcorr = c FN asperity layers formed
3e2 ρ 2 vH
tribochemically are
Wcorr = wear rate detached at a certain critical
ρ = density of material thickness
Ac = Arrhenius constant
Q = activation energy
Rc = gas constant
Tc = contact temperature
d = asperity contact diameter
v = sliding velocity
e = critical thickness of reaction layer
Ultrahigh- Wang model: • F or ultrahigh-molecular-
molecular- d ( µ − µ0 ) 1 1 sin 2α weight polyethylene in
weight k = k′ − × 1 − lubricated multi-directional
polyethylene 2γ c X
c X 0 2α sliding
(UHMWPE) k′ = constant • Assumes the occurrence of
wear [18] d = diameter of fibrils (mm) preferential molecular
μ = coefficient of friction alignment in the principal
μ0 = coefficient of friction for initiating direction of sliding and
surface failure rupture or splitting of the
Xc = cross-link density (mole/g) oriented molecules in the
X0 = critical cross-link density (mole/g) perpendicular direction
γc = C–C bond energy (Joule) associated with secondary
α = maximum cross-shear angle (radian) sliding
a
Note
• The wear factor (k) is a measure of the rate at which a given combination of materials wears in
the environment of the test. k is widely used for comparative purposes (mm3/Nm)
• According to Dowson [19], if the mean contact stresses are not too high the wear of polymer
against a hard surface (metal or ceramic) is obtained with fair accuracy with the relationship
V = kFL
• K is directly influenced by the roughness average (Ra) of the metallic counterface for the contact
UHMWPE–stainless steel in water under reciprocating pin-on-plate conditions given by the
relationship k = 4.0 × 10−5 Ra1.2 [20]
8 Wear 459
Although wear is a very complex process, apparatuses are available which allow for
the accumulation of data resulting in an estimate of the wear resistance of a combi-
nation of materials or a device. Preliminary material studies will commonly be per-
formed on laboratory wear benches while devices will be evaluated with simulators.
The classic arrangement of articulating surfaces in material wear testing involves an
upper bearing part that is attached to a pin, and a lower bearing material that is
attached to a specimen holding plate, or disk. The primary advantage of this con-
figuration is that any combination of candidate bearing materials can be incorpo-
rated into the testing setup. Typically, this involves either a spherical or a flat
material geometry on the pin, and a flat material geometry at the disk counterface.
The sacrificial material (which experiences the greatest amount of wear) can be
fabricated into either the pin or the disk; the decision regarding location of the sac-
rificial material depends on the particular type of analysis desired. Although it has
an important effect on the resulting wear, primarily because the material on the
surface of the pin is under constant load while the disk is loaded only periodically
in discrete regions, the importance of this decision is often overlooked. Historical
configurations of this pin-on-flat wear testing arrangement are most often referred
to as a “pin-on-disk” (POD) wear test. Many variations of this arrangement reported
in the literature include rotary pin-on-disk, reciprocating pin-on-disk, circularly
translating pin-on-disk, and multi-axis pin-on-disk configurations [22]. For clarity,
the translations and rotations of both the pin and disk have been described using
consistent terminology as follows: (1) translations are given with respect to a fixed
point of observation not on either the pin or the disk; circular translation does not
imply rotation about an internal axis; and (2) rotations are given with respect to the
central axis of either the pin or the disk being described, wherein the axis is normal
to the bearing surface.
One of the most basic wear testing configurations is the rotary pin-on-disk
method. In this method the pin is neither translated nor rotated, and the disk is
rotated without translation, producing a circular wear track on the flat specimen.
This configuration is often used to produce constant-velocity, unidirectional shear-
ing forces at the pin and disk surfaces. The pin surface is under continuous load and
unidirectional shear conditions, while only discrete locations of the flat specimen
are periodically loaded in unidirectional shear during each cycle. In this manner, the
pin specimen experiences the majority of the wear potential. Depending upon mate-
rial fabrication and geometric constraints, the candidate material under wear inves-
tigation can be fabricated into either pins or sheets; typically the softer material (the
polymer in a metal/polymer combination) is used as the upper bearing surface.
460 M. LaBerge and J.D. Desjardins
Table 8.3 Example of biomaterial combinations tribologically characterized with POD and
simulators
Material combinations Test apparatus
Stainless steel-UHMWPE Pin-on-disc; joint simulator [36–38]
Co-Cr-alloys-UHMWPE Pin-on-disc; joint simulator [38–43]
Titanium alloys-UHMWPE Pin-on-disc; joint simulator [40, 44, 45]
Alumina-UHMWPE Pin-on-disc; joint simulator [38, 42, 43, 46]
Zirconia-UHMWPE Pin-on-disc; joint simulator [37, 42, 47]
CoCrMo-CoCrMo Pin-on-disc; joint simulator [46, 48–52]
CoCrMo-Delrin Disc-on-flat [53]
PEEK-CoCrMo Pin-on-plate [54, 55]
CoCrMo-alumina Joint simulator [42]
Ti6A14V-alumina Joint simulator [42]
Alumina-alumina Wear and friction benches; joint simulator [46, 51,
56–59]
Elastomers-metal Reciprocating friction benches; joint simulator [22,
60–68]
Dental resins-enamel Pin-on-flat [69]
Modified UHMWPE-metal Pin-on-disc [70–72] Joint Simulator [73–75]
adapted versions of this standard to assess the wear resistance of bearing surfaces
for orthopaedic applications [48, 60, 76–82]. Other ASTM methods and protocols
from the International Organization for Standardization (ISO) pertinent to the
evaluation of wear performance of engineering materials and devices are listed in
Table 8.5.
Wear resistance is usually reported in terms of wear rate, either linear or
volumetric, with different units such as volume lost per 106 cycles (MC), mass loss
per 106 cycles, or linear displacement per 106 cycles. A complete walking cycle is
represented by two steps. One cycle on a reciprocating pin-on-flat system is obtained
462 M. LaBerge and J.D. Desjardins
Table 8.5 (continued)
ISO Implants for surgery—Wear of total knee-joint prostheses—Part 1:
14243-1:2009 Loading and displacement parameters for wear-testing machines with load
control and corresponding environmental conditions for test
ISO Implants for surgery—Wear of total knee-joint prostheses—Part 2:
14243-2:2009 Methods of measurement
ISO Implants for surgery—Wear of total knee-joint prostheses—Part 3:
14243-3:2014 Loading and displacement parameters for wear-testing machines
with displacement control and corresponding environmental conditions
for test
ISO/TS Dental materials—Guidance on testing of wear—Part 2: Wear by two and/
14569-2:2001 or three body contact
ISO 16428:2005 Implants for surgery—Test solutions and environmental conditions for
static and dynamic corrosion tests on implantable materials and medical
devices
ISO 17853:2011 Wear of implant materials—Polymer and metal wear particles—Isolation
and characterization
ISO Implants for surgery—Wear of total intervertebral spinal disc prostheses—
18192-1:2011 Part 1: Loading and displacement parameters for wear testing and
corresponding environmental conditions for test
ISO Implants for surgery—Wear of total intervertebral spinal disc prostheses—
18192-2:2010 Part 2: Nucleus replacements
a
American Society for Testing and Materials, Philadelphia, PA (www.astm.org)
b
International organization for Standardization, Geneva, Switzerland (www.iso.org)
by two passes (return to starting point), while the other cycle on a rotating pin-on-
disc system corresponds to one revolution. It is assumed that a normal individual
will make two million steps per year while an active subject may make more than
10 million steps [21] at a maximum frequency of 1 Hz. Investigators have also
reported wear rates as cubic millimeters (volume) per millimeter (sliding distance)
(mm3/mm). The volume is calculated by measuring the mass loss and using the
density of the polymer as a conversion factor. Tables 8.6 and 8.7 present a critical
selection of wear data available for biomaterial tribosystems useful to the orthopae-
dic design community. Both friction coefficient and wear rate are used as design
parameters. Unless independently monitored, friction coefficients are usually
acquired during wear tests. The static coefficient of friction is calculated using the
force required to initiate motion. The kinetic coefficient of friction may vary during
a test for a constant velocity and should be calculated from averaged force readings
during the duration of the test. The ASTM method G115–10 (2013) proposes a
guide for “Measuring and Reporting Friction Coefficients” which is designed to
assist investigators in the selection of an appropriate method for measuring the fric-
tional properties of materials.
Table 8.6 Friction coefficients of various implant materials from representative in vitro studies
464
• Angular velocity = 0.6–2.4 rad/s
Stainless steel 0.03–0.09 Twelve-channel • Conforming, flat-on-flat configuration [86]
(316L) + UHMWPE (ASTM F friction and wear • Velocity = 100 cyc./min
648) machine FW-12 • Pressure = 6.90 MPa
• Duration: 3.7 × 106 cycles
• Lub: bovine calf serum w/55 sodium azide
CoCrMo Dry 0.13 Lub: 0.21 Rolling-sliding • Velocity = 25 m/min [46]
(Protasul-2) + UHMWPE apparatus • Pressure = 30 N/cm2
• Duration: 20 h
• Lub: none or distilled water
• Room temperature
Co-Cr + UHMWPE 0.05–0.11 Pin-on-disk • Load = 3.45 M Pa [40]
• Velocity = 50 × 106 mm/year
• Lub: serum
• Duration: 2 years test
• 28–32 °C
Co–Cr + UHMWPE 00.07–0.25 Pin-on-flat • Axial load = 223 N [87]
• Duration 250,000 cycles
• Lub: bovine serum
Co–Cr–Mo 0.08–0.15 Pin-on-flat • Contact pressure = 4.8 MPa [42]
(Vitallium) + UHMWPE • Frequency = 1 Hz
• Sliding dist. = 50 mm
• Lub: distilled, deionized H2O
• 37.1 °C
(continued)
465
Table 8.6 (continued)
466
• Duration: 3.7 × 106 cycles
• Lub: bovine calf serum w/55 sodium azide
Co–Cr–Mo + UHMWPE 0.06–0.07 Hip joint simulator • Load 2.5 kN static load [91]
(a,c) • Lub: (a) serum, (b) serum albumin, (c)
0.10–0.12 synovial fluid, (d) veronate buffer
(b,d)
Co–Cr–Mo + UHMWPE 0.04–0.06 Hip joint simulator • Load 2.5 kN static load [92]
• elocity: 30 cycles/min
V
• Duration: 1.8 million cycles
• Lub: serum
Co–Cr–Mo + UHMWPE 0.03–0.05 Pin-on-disk • Load = 100 N [93]
• Lub: Ringer’s solution
• Velocity = 0.05 m/s
• Duration: 48 h
Ti–6A1–4V + UHMWPE 0.04–0.26 Pin-on-flat • Axial load = 223 N [87]
• Duration: 250,000 cycles
• Lub: bovine serum
Ti–6A1–4V + UHMWPE 0.05–0.121 Reciprocating flat • Load = 445 N [38]
on-flat • Velocity = 100 cycles/min
• Lub: bovine serum
• Duration: 4.1 × 106 cycles
Ti–6A1–4V (Ion 0.058 Single-channel hip • Range of motion = 30° [85]
implanted) + UHMWPE (ASTM joint simulator • Lub: deionized water at 37 °C
F 648) • Load = 1–4 kN
• Angular velocity = 0.6–2.4 rad/s
(continued)
467
Table 8.6 (continued)
468
• Angular velocity = 0.6–2.4 rad/s
Zirconia (Y–PSZ) + UHMWPE 0.049 Pin-on-disk • Load = 3 MPa [37]
• Velocity = 60 mm/s
• Lub: bovine serum, 40–50 ml
• 24–26 °C
Zirconia (Zyranox) + UHMWPE 0.05–0.16 Pin-on-flat • Contact pressure = 4.8 MPa [42]
• Frequency = 1 Hz
• Sliding dist. = 50 mm
• Lub: distilled, deionized H2O
• 37.1 °C
Zirconia + UHMWPE (ASTM F 0.059 Single-channel hip • Range of Motion = 30° [85]
648) joint simulator • Lub: deionized water at 37 °C
• Load = 1–4 kN
• Angular velocity = 0.6–2.4 rad/s
Stainless steel (100CR6- 0.6 (a); 0.26 (b); 0.106 Ball-on-disk • 10 mm diameter ball [83]
German) + Stainless steel (c); 0.1 (d) vibrotribometer • Oscillation = 10 Hz; 1.65 mm amplitude
(100CR6-German) (Optimol • Load = 50 and 300 N
SRV-German) • Lub: (a) none; (b) human synovial fluid; (c)
yellow bone marrow; (d) red bone marrow
• 37 °C
Co–Cr–Mo + Co–Cr–Mo 0.03–0.04 Pin-on-disk • Load = 100 N [93]
• Lub: Ringer’s solution
• Velocity = 0.05 m/s
• Duration: 48 h
(continued)
469
470
Table 8.6 (continued)
Average (or range)
Material contact friction coefficient Testing apparatus Tribological conditions References
CoCrMo (Protasul-2) + CoCrMo Dry: 0.4 Rolling-sliding • Velocity = 25 m/min [46]
(Protasul-2) Lub: 0.35 apparatus • Pressure = 30 N/cm2
• Duration: 20 h
• Lub: none or distilled water
• Room temperature
Co–Cr–Mo + Co–Cr–Mo 0.12–0.13 Hip joint simulator • Load 2.5 kN static load [91]
(a,b,c) • Lub: (a) serum, (b) serum albumin, (c)
0.22 (d) synovial fluid, (d) veronate buffer
Co–Cr–Mo + Co–Cr–Mo 0.13–0.14 Hip joint simulator • Load 2.5 kN static load [92]
• Velocity: 30 cycles/min
• Duration: 1.8 million cycles
• Lub: serum
Co–Cr–Mo (ASTM F75) 0.16 Hip simulator machine • Lub: synovial fluid [90]
• Peak load = 150 kg
CoCrMo + CoCrMo Serum and synovial fluid: Dual-hip simulator • Velocity = 30 cycles/min [84]
(McKee-Farrar) 0.12 Saline: 0.22 • Load = 250 kg
• Duration: 1000 h
• Lub: serum and synovial fluid or saline
• Room temperature
M. LaBerge and J.D. Desjardins
Average (or range)
Material contact friction coefficient Testing apparatus Tribological conditions References
CoCrMo + CoCrMo 0.13 Dual-hip simulator • Velocity = 30 cycles/min [84]
(McKee-Farrar) • Load = 250 kg
• Duration: 1000 h
8 Wear
• Lub: serum
• Room temperature
Alumina + alumina 0.26–0.35 Pin-on-disk • Load = 100 N [93]
• Lub: Ringer’s solution
• Velocity = 0.05 m/s
• Duration: 48 h
Alumina + alumina Dry: 0.71 Lub: 0.09 Rolling-sliding • Velocity = 25 m/min [46]
apparatus • Pressure = 30 N/cm2
• Duration: 20 h
• Lub: none or distilled water
• Room temperature
471
472
Table 8.7 Wear rates of various material combination from in vitro studies
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Stainless steel 0.4 mm/year NA • Pin-on-disk Smooth polymer, [40]
(316LVM) + UHMWPE • Luba: bovine serum metal—scratches
• Duration: 2 years test
• Load = 3.45 MPa
• Velocity = 50 × 106 mm/year
• 28–32 °C
Stainless steel 0.17–0.23 mm3/106 NA • Reciprocating flat-on-flat Surface scratching [38]
(316L) + UHMWPE • Load = 445 N
• Velocity = 100 cycles/min
• Lub: bovine serum
• Duration = 3.7 × 106 cycles
Stainless steel (316L) + NA 27.7 × 10−7 mm3/Nm • Pin-on-disk Original machine [37]
UHMWPE • Lub: bovine serum, marks gone, new
40–50 ml wear marks
• Load = 3 MPa
• Velocity = 60 mm/s
• 24–26 °C
Stainless steel Machined UHMWPE: NA • Annular disk on flat pin UHMWPE transfer [45]
(316L) + Machined 3.23 in./in. × 10−9 Molded • Range of motion = 110° film
UHMWPE (HiFax 1900) UHMWPE: 1.70 in./ • Sliding velocity = 43.3 in./
in. × 10−9 min
• Stress = 500 psi
• Lub: Ringer’s solution
Stainless steel + UHMWPE Max. depth of wear: NA • Dual-hip simulator Evidence of brittle [84]
(Charnley) 0.15 mm • Velocity = 30 cycles/min fracture
• Load = 250 kg
• Duration: 1000 g
• Lub: serum
• Room temperature
M. LaBerge and J.D. Desjardins
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Stainless steel (316L) UHMWPE: NA • Twelve-channel friction and Quantification of [86]
0.20 mm3/106 cycles wear machine FW-12 wear separately from
UHMWPE 0.65 μm/year • Conforming, flat-on-flat creep deformation.
8 Wear
configuration Adhesive/abrasive
• Velocity = 100 cyc./min wear emphasizes over
• Pressure = 6.90 MPa fatigue wear
• Duration: 3.7 × 106 cycles
• Lub: bovine calf serum
w/55 sodium azide
Stainless steel + UHMWPE UHMWPE: NA • Ten station hip simulator Highly loaded region [94]
40 mg/106 cycles • Range of motion = 46° of UHMWPE smooth
• Lub: Bovine blood serum at and shiny, peeling,
37 ° C pitting
• Load = Oscillating
0–2030 N
• Frequency = 1 Hz
• Duration: 1 × 106 cycles
Stainless steel + UHMWPE 1.62 × 10−7 mm3/Nm NA • Ball-and-socket simulator Uniform superficial [44]
• Ra < 0.016 scratches, occasional
• Sterilized with deeper marks. Metal
2.5 Mrad g rad particles, acrylic
• Lub: bovine serum with cement particles
sodium azide
• Load = 2000 N
• Speed = 100 mm/s
(continued)
473
474
Table 8.7 (continued)
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Cast Co-Cr-Mo pins PE thickness change: NA • Reciprocating pin-on-flat Abrasive wear of PE. [95]
(ASTM F-75) + UHMWPE 64 ± 13 μm • Sterilized with 2.5 Mrad Transfer of PE on
(GUR 415 plate) • Lub: deionized water Co–Cr pins.
• 36 MPa Line contact stress Oxidative wear of
• Frequency = 2.1 Hz Co–Cr pins
• Stroke length = 15 mm
• Duration = 2 × 106 cycles
• Final friction = 0.079 ± 0.001
Wrought Co-Cr pins PE thickness change: NA • Reciprocating pin-on-flat Abrasive wear of [95]
(ASTM F-90) + UHMWPE 71 ± 25 μm • Sterilized with 2.5 Mrad PE. Transfer of PE on
GUR 415 plate • Lub: deionized water Co–Cr pins.
• 36 MPa Line contact stress Oxidative wear of
• Frequency = 2.1 Hz Co–Cr pins
• Stroke length = 15 mm
• Duration = 2 × 106 cycles
• Final friction = 0.101 ± 0.019
Co–Cr–Mo CoCrMo: 0.1 mg/20 h NA • Rolling-sliding wear and NA [40]
(Protasul-2) + UHMWPE UHMWPE: 1 mg/20 h • Friction apparatus
• Velocity • 25 m/min
• Pressure = 30 N/cm2
• Duration: 20 h
• Dry condition
• Room temperature
Co-Cr-Mo (hot 0.5 mm/year NA • Pin-on-disk Smooth polymer, [87]
pressed) + UHMWPE • Lub: bovine serum metal—scratches
• Duration: 2 years test
• Load = 3.45 M Pa
• Velocity = 50 × 106 mm/year
• 28–32 °C
M. LaBerge and J.D. Desjardins
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Co–Cr + UHMWPE 1.05 mg/106 cycles NA • Pin-on-flat Transfer of PE to [42]
• Lub: distilled, deionized Co-Cr Surface
water Adhesive wear
8 Wear
Table 8.7 (continued)
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Co-Cr-Mo + UHMWPE Max. depth of wear: NA • Dual hip simulator Evidence of brittle [84]
(Charnley–Muller) 0.08 mm • Velocity = 30 cycles/min fracture
• Load = 250 kg
• Duration: 1000 h
• Lub: serum
• Room temperature
Co-Cr-Mo + UHMWPE 0.15 mm/1.8 × 106 cycles NA • Hip joint simulator Creep, abrasion, [91]
(Charnley) • 2.5 kN static load adhesion. Max cup
• Lub: bovine serum wear depth
• Duration: 1.8 × 106 cycles
• Velocity = 30 cycles/min
Co–Cr–Mo + UHMWPE 0.075 mm/1.8 × 106 cycles NA • Hip joint simulator Creep, abrasion, [91]
(Charnley–Muller) • 2.5 kN static load adhesion. Max cup
• Lub: bovine serum wear depth
• Duration: 1.8 × 106 cycles
• Velocity = 30 cycles/min
Co-Cr-Mo + UHMWPE 1.8 mg/105 cycles NA • Knee joint simulator Creep and fatigue [41]
(Duo-Patella) • 700 lb peak load cracks evident
• Velocity = 33 cycles/min
• Duration: 105 cycles
• Lub: double-spun bovine
serum
Co-Cr-Mo + UHMWPE 1.1 mg/105 cycles NA • Knee joint simulator Creep and fatigue [41]
(Ewald) • 700 lb peak load cracks evident
• Velocity = 33 cycles/min
• Duration: 105 cycles
• Lub: double-spun bovine
serum
M. LaBerge and J.D. Desjardins
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Co-Cr-Mo + UHMWPE 0.3 mg/105 NA • Knee joint simulator Creep and fatigue [41]
(Spherocentric) • 700 lb peak load cracks evident
• Velocity = 33 cycles/min
8 Wear
Table 8.7 (continued)
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Ti–6A1–4V PE thickness change: NA • Reciprocating pin-on-flat Adhesive transfer of 95]
Pins + UHMWPE GUR 59 ± 12 μm • Sterilized with 2.5 Mrad PE on cylinders.
415 plate • Lub: deionized water Oxidative wear of
• 36 MPa Line contact stress Ti-6A1-4V
• Frequency = 2.1 Hz Abrasive wear of PE
• Stroke length = 15 mm
• Duration = 2 × 106 cycles
• Final friction = 0.112 ± 0.007
Ti-6A1-4V 0.9 mm/year NA • Pin-on-disk No scratches present [40]
(nitrided) + UHMWPE • Lub: bovine serum
• Duration: 2 years test
• Load = 3.45 MPa
• Velocity = 50 × 106 mm/year
• 28–32 °C
Ti–6A1–4V Machined UHMWPE: NA • Annular disk-on-flat pin UHMWPE transfer [45]
(passivated) + UHMWPE 0.84–2.07 in/in × 10−9 • Range of motion = 110° film
(HiFax 1900) • Sliding velocity = 43.3 in./
min
• Stress = 500 psi
• Lub: Ringer’s solution
Ti-6A1-4V Machined UHMWPE: NA • Annular disk-on-flat pin UHMWPE transfer [45]
(nitrided) + UHMWPE 1.83 in/in × 10−9 • Range of motion = 110° film
(HiFax 1900) • Sliding velocity = 43.3 in./
min
• Stress = 500 psi
• Lub: Ringer’s solution
M. LaBerge and J.D. Desjardins
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Ti-6A1-4V (not Machined UHMWPE: NA • Annular disk-on-flat pin UHMWPE transfer [45]
passivated) + UHMWPE 1.55 in/in × 10−9 • Range of motion = 110° film
(HiFax 1900) • Sliding velocity = 43.3 in./
8 Wear
min
• Stress: 500 psi
• Lub: Ringer’s solution
Ti-6A1-4V + UHMWPE 0.04–0.11 mm3/106 NA • Reciprocating flat-on-flat Surface scratching [38]
• Load = 445 N
• Velocity = 100 cycles/min
• Lub: bovine serum
• Duration: 4.1 × 106 cycles
Ti-6Al-4V (implanted with NA 1.9 × 10−7 mm3/Nm • Ball and socket simulator Uniform superficial [44]
nitrogen) + UHMWPE • Ra < 0.016 scratches, occasional
• Sterilized w/2.5 Mrad g rad deeper marks
• Lub: bovine serum w/ Metal particles,
sodium azide acrylic cement
• Load = 2000 N particles
• Speed = 100 mm/s
Ti-6A1-4V NA 1.98 × 10−7 mm3/Nm • Ball-and-socket simulator Uniform superficial [44]
(conventional) + UHMWPE • Ra < 0.016 scratches, occasional
• Sterilized w/2.5 Mrad g rad deeper marks
• Lub: bovine serum w/ Metal particles,
sodium azide acrylic cement
• Load = 2000 N particles
• Speed = 100 mm/s
Alumina + UHMWPE NA 18.2 × 10−7 mm3/Nm • Pin-on-disk Original machine [37]
• Lub: bovine serum, marks worn,
40–50 ml smoother
• Load = 3 M Pa
• Velocity = 60 mm/s
• 24–26 °C
(continued)
479
480
Table 8.7 (continued)
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Alumina + UHMWPE 0.26 mg/106 cycles NA • Pin-on-flat NA [38]
• Axial load = 223 N
• Duration: 250,000 cycles
• Lub: bovine serum
Alumina + UHMWPE 0.04 mg/106 cycles NA • Pin-on-flat Minimal wear [42]
• Lub: distilled, deionized
water
• Contact pressure: 4.8 MPa
• Frequency = 1 Hz
• Sliding dist. = 50 mm
• 37.1 °C
Alumina + UHMWPE Alumina: 0.1 mg/20 h NA • Rolling-sliding wear and NA [46]
UHMWPE: 0.1 mg/20 h friction apparatus
• Velocity = 25 m/min
• Pressure = 30 N/cm2
• Duration: 20 h
• Dry condition
• Room temperature
Zirconia + UHMWPE 0.03 mg/106 cycles NA • Pin on flat Minimal wear [42]
• Lub: distilled, deionized
water
• Contact pressure: 4.8 MPa
• Frequency = 1 Hz
• Sliding dist. = 50 mm
• 37.1 °C
M. LaBerge and J.D. Desjardins
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Solid yttria Zr02 PE thickness change: NA • Reciprocating pin-on-flat Abrasive wear from [95]
pins + UHMWPE GUR 33 ± 13 μm • Sterilized with 2.5 Mrad the surface roughness
415 plate • Lub: deionized water characteristic and
8 Wear
(continued)
482
Table 8.7 (continued)
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Co-Cr-Mo + Co-Cr-Mo Max. depth of wear: NA • Dual-hip simulator Adhesive and [84]
(McKee-Farrar) 0.01 mm • Velocity = 30 cycles/min abrasive types of
• Load = 250 kg wear
• Duration: 1000 h
• Lub: serum
• Room temperature
Co-Cr-Mo • initial wear: 10–20 mm NA • Stanmore Mk III hip Equal amount of wear [52]
(Protasul-2) + Co-Cr-Mo • linear wear: simulator on both components
(Protasul-2) 2–4 mm/106 cycles • 37 mm diameter head Preferential wear
• Frequency = 1/2 Hz direction, with
• Load = 300–3500 N pronounced grooving
• Duration: min. 2.5 × 106
movements/test
• Lub: Ringer’s solution
w/30 % calf serum
Co–Cr-Mo + Co-Cr-Mo 0.013 mm/1.8 × 106 cycles NA • Hip joint simulator Abrasion, scratching. [91]
(McKee-Farrar) • 2.5 kN static load Max cup wear depth
• Lub: bovine serum
• Duration: 1.8 × 106 cycles
• Velocity = 30 cycles/min
Co-Cr-Mo (Protasul- • I nitial wear: 10–20 mm NA • Stanmore Mk III hip Equal amount of wear [56]
21WF) + Co-Cr-Mo • Linear wear: simulator on both components
(Protasul-21WF) 2–4 mm/106 cycles • 28 and 32 mm heads Preferential wear
• Frequency = 1/2 Hz direction, with
• Load = 300–3500 N pronounced grooving
• Duration: min. 2.5 × 106
movements/test
• Lub: Ringer’s solution
w/30 % calf serum
M. LaBerge and J.D. Desjardins
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Alumina + alumina 1.2 × 10-7 mm/mm NA • Disc-on-plate NA [46]
• Lub: water
• 37 °C
8 Wear
Stainless steel (316 2.26 mm/year 3.4 × 10−5 mm3/Nm • Pin-on-flat No creep. Particles [96]
S16) + PTFE • Lub: bovine serum present
• Ra = 0.0 mm
• Load = 40 n/pin
• Speed = 2 p rad/s
• Sliding dist. = 0.24 m/s
Stainless steel + Delrin 500 31 mm/year NA • Pin-on-flat NA [53]
• Velocity = 100 mm/s
• Load = 6.9 N/mm2
• Lub: serum
• Room temperature
• 4.8 years’ effective use
Stainless steel + Polyester 521 mm/year NA • Pin-on-flat NA [53]
• Velocity = 100 mm/s
• Load = 6.9 N/mm2
• Lub: serum
• Room temperature
• 4.8 years’ effective use
Stainless steel + alumina (mg/cyc) SS: 176, 146, NA • Five station hip simulator Corrosion of stainless [97]
(Protek) 212 (Helsinki) steel heads
A1203: 0.3, 2.1, 0.2 • 32 mm head
• Frequency = 1.08 sH3
• Load = 35 kN
• Duration: 3 × 106 cycles
• Lub: distilled, deionized
water
• 37 °C
(continued)
483
484
Table 8.7 (continued)
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Stainless steel + alumina (mg/cyc) SS: 50.9, 62.4, NA • Five-station hip simulator Corrosion of stainless [97]
(Thackray) 46.0 (Helsinki) steel heads
A12O3: 0.8, 1.0, 0.7 • 22.2 mm head
• Frequency = 1.08 sH3
• Load = 35 kN
• Duration: 3 × 106 cycles
• Lub: distilled, deionized
water
• 37 °C
Co–Cr-Mo + alumina (mg/cyc) CoCrMo: 39.7, NA • Five-station hip simulator NA [97]
(Link) 48.2. 94.0 (Helsinki)
A12O3: 1.5, 0.5, 0.0 • 32 mm head
• Frequency = 1.08 sH3
• Load = 35 kN
• Duration: 3 × 106 cycles
• Lub: distilled, deionized
water
• 37 °C
Co–Cr–Mo + alumina (mg/cyc) CoCrMo: 2.6, NA • Five-station hip simulator NA [97]
(Howmedica) 4.7, 4.3 • (Helsinski)
A12O3: 0.1, 0.0, 0.0 • 32 mm head
• Frequency = 1.08 sH3
• Load = 35 kN
• Duration: 3 × 106 cycles
• Lub: distilled, deionized
water
• 37 °C
M. LaBerge and J.D. Desjardins
Material contact Wear rate Wear coefficient Tribological conditions Wear mechanisms References
Co–Cr-Mo + Delrin 550 37.5 mm/year NA • Disk-on-flat NA [53]
• Velocity = 106 mm/s
• Load = 3.7 N/mm2
8 Wear
• Lub: water
• 37 °C
Si3N4 + UHMWPE 0.27 mg/106 cycles NA • Pin-on-flat NA [42]
• Lub: distilled, deionized
water
• Contact pressure: 4.8 MPa
• Frequency = 1 Hz
• Sliding dist. = 50 mm
• 37.1 °C
a
Lub lubricant
485
486 M. LaBerge and J.D. Desjardins
Baykal and collaborators have critically and systematically reviewed the litera-
ture of multi-axis pin-on-disc wear testing for the past decades with a focus on
UHMWPE tribological properties and relevant wear mechanisms. Data reviewed
have been extracted from 22 published studies and are presented as UHMWPE
average wear rate (mg/MC), average wear rate (mm3/MC), and average wear factor
(mm3/Nm) as well as contact area for bearing couples [77].
Methods have been proposed by Griffith et al. [98] and Deavane et al. [99] to radio-
graphically measure linear wear in UHMWPE acetabular cups of total hip replace-
ments. The overall penetration of a femoral head (Tables 8.8 and 8.9) into a
UHMWPE acetabular cup is a consequence of both creep and wear, where creep is
Table 8.8 Clinical penetration into UHMWPE cups for metallic femoral heads
Reference UHMWPE penetration rate Clinical
[98] • 0.07 mm/year average 491 acetabular cups 8.3-year follow-up (range
(range 0.06–0.24 mm/year) 7–9 years)
[100] • 0.13 mm/year average (range Follow-up at least 9.5 years after implantation
0–0.39 mm/year) for 22 mm (227 of the 22 mm femoral head size; 98 of the
head size 28 mm size; and 60 of the 32 mm head size)
• 0.08 mm/year average (range
0–0.3 mm/year) for 28 mm
head size
• 0.10 mm/year average (range
0–0.32 mm/year) for 32 mm
head size
[101] • 0.21 mm/year (range 87 acetabular cups 9-year service life (range
0.005–0.6 mm/year) <1–17.5 years)
[102] • 0.022 mm/year (range Four explained acetabular cups with no
0.010–0.034 mm/year) apparent deterioration of the finish of the
femoral head 20 year mean service life (range
17–23 years)
Table 8.9 Clinical penetration into UHMWPE cups for alumina ceramic heads
Reference Femoral head Average penetration rate
[103] Alumina 0.098 mm/year (6-year follow-up; 28 mm cups)
[103] Alumina 0.072 mm/year
(g irradiated (6-year follow-up; 28 mm cups)
UHMWPE)
[104] Alumina 0.084 mm/year
[105] Alumina 0.080 mm/year (38 acetabular cups over a period of 6.7 years)
[106] Alumina 0.100 mm/year
8 Wear 487
predominantly observed for the first million loading cycles [19]. Atkinson et al.
[107] estimated that the total UHMWPE residual compression due to creep in total
hip replacements after x can be estimated by
Conclusion
References
1. Jacobs JJ, Shanbhag A, Giant TT, Black J, Galante JO (1994) Wear debris in total joint
replacements. J Am Acad Orthop Surg 2:212–220
2. Mittlmeier T, Walter A (1987) The influence of prosthesis design on wear and loosening
phenomena. CRC Critical Reviews in Biocompatibility, 3(4): 319–419
3. Dowson D, Wright V (1981) An Introduction to the Bio-mechanics of Joints and Joint
Replacement. Dowson D. and Wright V. Editors. Mechanical Engineering Publications, Ltd,
London, England, pp. 49–60
4. Suh NP (1986) Tribophysics. Prentice-Hall, Inc., Englewood Cliffs, NJ, pp 3–11, 195–222
5. Jahanmir S, Suh NP (1977) Mechanics of subsurface void nucleation in delamination wear.
Wear 44:17–38
6. Rowe CN (1980) Lubricated Wear, in CRC Handbook of Lubrication. Theory and Practice of
Tribology Volume II. Booser, R.E. Editor. Boca Raton, FL:CRC Press; pp. 209–225
7. Waterhouse RB (1992) Fretting wear. Frict Lubr Wear Technol 18:242–256
8. Black J (1988) Orthopaedic biomaterials in research and practice. Churchill Livingstone, NY,
p 255
9. Stachowiak G, Batchelor AW (2014) Engineering tribology, 4th edn. Elsevier, Amsterdam,
852 pp
10. Unsworth A, Dowson D, Wright V (1971) Cracking joints: a bioengineering study of cavita-
tion in the metacarpophalangeal joint. Ann Rheum Dis 30:348–358
11. Williams JA (2005) Wear and wear particles—some fundamentals. Tribol Int 38:863–870
12. Archard JF (1980) Wear theory and mechanisms. In: Peterson MB, Winer WO (eds) Wear
control handbook. ASME, New York, pp 35–80
13. Hornbogen E (1975) The role of fracture toughness in the wear of metals. Wear 33:251–259
14. Zum Gahr K-H (1982) Abrasive Wear of ductile materials. Z Metallkd 73:267–276
15. Hailing JA (1975) A contribution to the theory of mechanical wear. Wear 34:239–249
16. Quinn TFJ (1983) A review of oxidational wear. Parts I and II. Tribol Int 16(257–271):
305–315
17. Quinn TFJ, Rowson D, Sullivan JL (1980) Application of the oxidational theory of mild wear
to the sliding wear of low-alloy steel. Wear 65:1–20
18. Wang A (2001) A unified theory of wear for ultra-high molecular weight polyethylene in
multi-directional sliding. Wear 248:38–47
19. Dowson D (1995) A comparative study of the performance of metallic and ceramic femoral
head components in total replacement hip joints. Wear 90:171–183
20. Dowson D, El-Hady Diab MM, Gillis BJ, Atkinson JR (1985) Influence of counterface
topography on the wear of ultra high molecular weight polyethylene under wet conditions. In:
LeeAm L-H (ed) Polymer wear and its control. Chem. soc. symposium series no. 287,
Chapter 12, pp. 171–187
21. Dumbleton JH (1981) Tribology of natural and artificial joints. Elsevier, New York,
pp 94–103, 110–148, 183–214
22. Gevaert MR, LaBerge M, Gordon JM, DesJardins JD (2005) The quantification of physiolog-
ically relevant cross shear wear phenomena on orthopaedic bearing materials using a novel
wear testing machine. J Tribol 127:740–749
23. Bragdon CR, O'Connor DO, Lowenstein JD, Jasty M, Syniuta WD (1996) The importance of
multi-directional motion on the wear of polyethylene. Proc Inst Mech Eng H 210:157–165
24. Wang A, Stark C, Dumbleton JH (1996) Mechanistic and morphological origins of UHMWPE
wear debris in total joint replacement prostheses. Proc Inst Mech Eng H 210:141–155
25. Sambasivan S, Fischer DA, Shen MC, Hsu SM (2004) Molecular orientation of ultrahigh
molecular weight polyethylene induced by various sliding motions. J Biomed Mater Res B
Appl Biomater 70:278–285
490 M. LaBerge and J.D. Desjardins
26. Bragdon CR, O'Connor DO, Lowenstein JD, Jasty M, Biggs SA, Harris WH (2001) A new
pin-on-disk wear testing method for simulating wear of polyethylene on cobalt-chrome alloy
in total hip arthroplasty. J Arthroplasty 16:658–665
27. Hua Z, Zhang H, Fana Y, Jin Z (2014) Development of a BiotriboPOD testing methodology
for the wear evaluation of orthopaedic biomaterials. RSC Adv 4:19987–19991
28. Joyce TJ, Vandelli C, Cartwright T, Unsworth A (2001) A comparison of the wear of cross-
linked polyethylene against itself under reciprocating and multi-directional motion with dif-
ferent lubricants. Wear 250:206–211
29. Saikko V (1998) A multidirectional motion pin-on-disk wear test method for prosthetic joint
materials. J Biomed Mater Res 41:58–64
30. Saikko V (2014) In vitro wear simulation on the RandomPOD wear testing system as a
screening method for bearing materials intended for total knee arthroplasty. J Biomech
47(11):2774–2778
31. Toohey KS, Blanchet TA, Heckelman DD (2003) Effect of accelerated aging conditions on
resultant depth-dependent oxidation and wear resistance of UHMWPE joint replacement
bearing materials. Wear 255:1076–1084
32. Turell M, Wang A, Bellare A (2003) Quantification of the effect of cross-path motion on the
wear rate of ultra-high molecular weight polyethylene. Wear 255:1034–1039
33. Yao JQ, Blanchet TA, Murphy DJ, Laurent MP (2003) Effect of fluid absorption on the wear
resistance of UHMWPE orthopedic bearing surfaces. Wear 255:1113–1120
34. McGloughlin TM, Kavanagh AG (2000) Wear of ultra-high molecular weight polyethylene
(UHMWPE) in total knee prostheses: a review of key influences. Proc InstMech Eng H
214:349–359
35. Walker PS, Blunn GW, Lilley PA (1996) Wear testing of materials and surfaces for total knee
replacement. J Biomed Mater Res 33:159–175
36. Dowson DB, Jobbins B (1988) Design and development of a versatile hip joint simulator and
a preliminary assessment of wear and creep in Charnley total replacement hip joints. Eng
Med 17:111–117
37. Kumar P, Oka M, Ikeuchi K, Shimizu K, Yamamuro T, Okumura H, Kotoura Y (1991) Low
wear rate of UHMWPE against zirconia ceramic (Y-PSZ) in comparison to alumina ceramic
and SUS 316L alloy. J Biomed Mater Res 25:813–828
38. McKellop H, Clarke I, Markolf K, Amstutz H (1981) Friction and wear properties of poly-
mer, metal, and ceramic prosthetic joint materials evaluated on a multichannel screening
device. J Biomed Mater Res 15:619–653
39. DesJardins I, Aurora A, Tanner SL, Pace TB, Acampora KB, LaBerge M (2006) Increased
total knee arthroplasty ultra-high molecular weight polyethylene wear using a clinically rel-
evant hyaluronic acid simulator lubricant. Proc Inst Mech Eng H 220:609–623
40. McKellop HA, Clarke IC, Markolf KL, Amstutz HC (1978) Wear characteristics of UHMW
polyethylene: a method for accurately measuring extremely low wear rates. J Biomed Mater
Res 12:895–927
41. Rose RM, Ries MD, Paul IL, Crugnola AM, Ellis E (1984) On the true wear rate of ultrahigh
molecular weight polyethylene in the total knee prosthesis. J Biomed Mater Res 18:
207–224
42. Saikko V (1993) Wear and friction properties of prosthetic joint materials evaluated on a pin
on flat apparatus. Wear 166:169–178
43. Saikko V, Ahlroos T, Calonius O, Keränen J (2001) Wear simulation of total hip prostheses
with polyethylene against CoCr, alumina and diamond-like carbon. Biomaterials 22(12):
1507–1514
44. McKellop HA, Rostlund TV (1990) The wear behavior of ion-implanted Ti-6A1-4V against
UHMW polyethylene. J Biomed Mater Res 24:1413–1425
45. Miller DA, Ainsworth RD, Dumbleton JH, Page D, Miller EH, Shen C (1974) A comparative
evaluation of the wear of ultra-high-molecular weight polyethylene abraded by
Ti-6A1-4V. Wear 28:207–216
8 Wear 491
46. Semlitsch M, Lehmann M, Weber H, Doerre E, Willert HG (1977) New prospects for a pro-
longed functional life-span of artificial hip joints by using the material combination polyeth-
ylene/aluminium oxide ceramic/metal. J Biomed Mater Res 11:537–552
47. Desjardins JD, Burnikel B, LaBerge M (2008) UHMWPE wear against roughened oxidized
zirconium and CoCr femoral knee components during force-controlled simulation. Wear
264:245–256
48. Medley JB, Krygier JJ, Bobyn JD, Chan FW, Tanzer M (1995) Metal-metal bearing surfaces
in the hip: investigation of factors influencing wear. Trans Orthop Res Soc 20(2):765
49. Chan FW, Medley JB, Krygier JJ, Bobyn JD, Podgorsak GK, Tanzer M (1996) Wear perfor-
mance of cobalt–chromium metal–metal bearing surfaces for total hip arthroplasty. In: Trans.
42nd annual meeting of the orthopaedic research society, vol 21, no. 2, p. 464
50. Chan FW, Bobyn JD, Medley JB, Krygier JJ, Tanzer M (1999) Wear and lubrication of metal-
on-metal hip implants. Clin Orthop Relat Res 369:10–24
51. Galante J, Rostoker W (1973) Wear in total hip prostheses. Acta Orthop Scand Suppl
145:1–46
52. Streicher RM, Schon R, Semlitsch M (1990) Investigation of the tribological behaviour of
metal-on-metal combinations for artificial hip joints. Biomed Tech 35(5):3–7
53. Clarke IC (1982) Wear-screening and joint simulation studies vs. materials selection and
prosthesis design. CRC Crit Rev Biomed Eng 8:29–91
54. Scholes SC, Unsworth A (2009) Wear studies on the likely performance of CFR-PEEK/
CoCrMo for use as artificial joint bearing materials. J Mater Sci Mater Med 20(1):163–170
55. Scholes SC, Unsworth A (2010) The wear performance of PEEK-OPTIMA based self-mating
couples. Wear 268:380–387
56. Boutin P (1972) Arthroplastie total de la hanche par prothèse en alumina frittée. Rev Chir
Orthop 58:229
57. Nevelos J, Ingham E, Doyle C, Streicher R, Nevelos A, Walter W, Fisher J (2000)
Microseparation of the centers of alumina-alumina artificial hip joints during simulator test-
ing produces clinically relevant wear rates and patterns. J Arthroplasty 15:793–795
58. Oonishi H, Clarke IC, Good V, Amino H, Ueno M (2004) Alumina hip joints characterized
by run-in wear and steady-state wear to 14 million cycles in hip-simulator model. J Biomed
Mater Res A 70A:523–532
59. Wroblewski BM, Siney PD, Dowson D, Collins SN (1996) Prospective clinical and joint
simulator studies of a new total hip arthroplasty using alumina ceramic heads and cross-
linked polyethylene cups. J Bone Joint Surg 78(2):280–285, British volume
60. Jin ZM, Dowson D, Fisher J (1993) Wear and friction of medical grade polyurethane sliding
on smooth metal counterfaces. Wear 162–164:627–630
61 Auger DD, Dowson D, Fisher J, Jin MZ (1993) Friction and lubrication in cushion form bear-
ings for artificial hip joints. Proc Inst Mech Eng 207:25–33
62. Auger DD, Dowson D, Fisher J (1995) Cushion form bearings for total knee joint replace-
ment. Part 1: Design, friction, and lubrication. Proc Inst Mech Eng 209:73–81
63. Caravia L, Dowson D, Fisher J (1993) Start up and steady state friction of thin polyurethane
layers. Wear 160:191–197
64. Chow AB, Medley JB, LaBerge M (1994) Mechanical and tribological analyses of elasto-
meric surface layers in load bearing implants. In: Trans of 20th annual meeting of the society
for biomaterials, Boston, MA, p. 434
65. Graham RM, Joseph PF, Dooley RL, LaBerge M (1995) Effect of lubricant viscosity and
bearing surface stiffness on the lubrication mechanism of a point contact as a total hip arthro-
plasty model. Trans Soc Biomater 23:124
66. LaBerge M, Kirkley S, Chow A, Medley JB, Brox WT (1991) Biotribological study of the
contact elastomer-cartilage. In: Proceedings of the combined meeting of orthopaedic research
societies (USA, Japan, Canada), Banff, Alberta, Canada, June 1991
67. Pylios T, Shepherd DET (2008) Wear of medical grade silicone rubber against titanium and
ultrahigh molecular weight polyethylene. J Biomed Mater Res Part B Appl Biomater
84B:520–523
492 M. LaBerge and J.D. Desjardins
68. Teoh SH, Tang ZG, Ramakrishna S (1999) Development of thin elastomeric composite mem-
branes for biomedical applications. J Mater Sci Mater Med 10:343–352
69. Suzuki S, Suzuki SH, Cox CF (1996) Evaluating the antagonistic wear of restorative
materials. J Am Dent Assoc 127:74–80
70. Deng M, Shalaby SW (1997) Properties of self-reinforced ultra-high-molecular-weight poly-
ethylene composites. Biomaterials 18:645–655
71. Roeder RK, Sproul MM, Turner CH (2003) Hydroxyapatite whiskers provide improved
mechanical properties in reinforced polymer composites. J Biomed Mater Res A
67A:801–812
72. Wang M, Chandrasekaran M, Bonfield W (2002) Friction and wear of hydroxyapatite rein-
forced high density polyethylene against the stainless steel counterface. J Mater Sci Mater
Med 13:607–611
73. Harsha AP, Joyce TJ (2013) Comparative wear tests of ultra-high molecular weight polyeth-
ylene and cross-linked polyethylene. Proc IMechE Part H J Eng Med 227:600–608
74. Kyomoto M, Moro T, Miyaji F, Hashimoto M, Kawaguchi H, Takatori Y, Nakamura K,
Ishihara K (2009) Effects of mobility/immobility of surface modification by
2-methacryloyloxyethyl phosphorylcholine polymer on the durability of polyethylene for
artificial joints. J Biomed Mater Res A 90A:362–371
75. Oral E, Christensen SD, Malhi AS, Wannomae KK, Muratoglu OK (2006) Wear resistance
and mechanical properties of highly cross-linked, ultrahigh-molecular weight polyethylene
doped with vitamin E. J Arthroplasty 21:580–591
76. Agrawal CM, Micallef DM, Wirth MA, Lankford J, Dearnaley G, McCabe AR (1993) The
effects of diamond-like-carbon coatings on the friction and wear of enhanced UHMWPE-
metal couples. In: Trans 21st annual meet of the society for biomaterials, Birmingham, AL,
vol 26, p. 10
77. Baykal D, Siskey RS, Haider H, Saikko V, Ahlroos T, Kurtz SM (2014) Advances in tribo-
logical testing of artificial joint biomaterials using multidirectional pin-on-disk testers.
J Mech Behav Biomed Mater 31:117–134
78. Brown KJ, Atkinson JR, Dowson D, Wright V (1976) The wear of ultra high molecular
weight polyethylene and a preliminary study of its relation to the in vivo behaviour of replace-
ment hip joints. Wear 40:255–264
79. Fisher J, Dowson D (1991) Tribology of total artificial joints. Proc Inst Mech Eng
205:73–79
80. Rose RM, Radin EL (1982) Wear of polyethylene in total hip prostheses. Clin Orthop Relat
Res 170:107–115
81. Schmidt MB, Lin M, Greer KW (1995) Wear performance of UHWMPE articulated against
ion implanted CoCr. In: Trans 21st annual meeting of the society for biomaterials, San
Francisco, CA, vol 28, p. 230
82. Streicher RM (1991) Ceramic surfaces as wear partners for polyethylene. In: Bonfiled W,
Hastings GW, Tanner KE (eds) Bioceramics, vol 4. Butterworth-Heinemann Ltd, London,
England, p 9
83. Gavrjushenko NS (1993) Recommendations with respect to the improvement of lubrication
quality of synovial fluid in artificial joints. Proc Inst Mech Eng 207:111–114
84. Simon SR Radin EL (1973) Lubrication and wear of the Charnley, Charnley-Muller, and
McKee-Farrar prostheses with special regard to their clinical behavior. In: Proc. 1st scient.
meet. hip society, Saint Louis, Chapter 5, pp. 33–45
85. Saikko V (1992) Simulator study of friction in total replacement hip joints. Proc Inst Mech
Eng H 206:201–211
86. McKellop H (1981) Wear of artificial joint materials. II. Twelve-channel wear-screening
device: correlation of experimental and clinical results. Eng Med 10(3):123–136
87. Tateishi T, Terui A, Yunoki H (1990) Friction and wear properties for biomaterials for artifi-
cial joint. Bioceramics 2:145–151
8 Wear 493
88. Ruger LM, Love BJ, Drews MJ, Hutton WC, LaBerge M (1996) Effect of antioxidant on the
tribological properties of gamma sterilized ultra high molecular weight polyethylene. In:
Proc. fifth world biomaterials congress, Toronto, Canada, 29 May–2 June
89. Unsworth A, Pearcy MJ, White EFT, White G (1988) Frictional properties of artificial hip
joints. Eng Med 17:101–104
90. Walker PS, Bullough PG (1973) The effects of friction and wear in artificial joints. Orthop
Clin North Am 4:275–293
91. Weightman B, Simon S, Paul LI, Rose R, Radin EL (1972) Lubrication mechanisms of hip
joint replacement prostheses. J Lubr Technol (Trans Am Soc Mech Eng) 94:131
92. Weightman B, Paul IL, Rose RM, Simon SR, Radin EL (1973) A comparative study of total
hip replacement prostheses. J Biomech 6:299
93. Ungethum M, Refior HJ (1974) 1st Aluminiumoxidkeramik als Gleirlagerwekstoff fur
Totalendoprothesen geeignet? Arch Orthop Unfallchir 79:97
94. Eyerer P, Kurth M, McKellop HA, Mittlmeier T (1987) Characterization of UHMWPE Hip
cups run on joint simulators. J Biomed Mater Res 21:275–291
95. Poggie RA, Wert JJ, Mishra AK, Davidson JA (1992) Friction and wear characterization of
UHMWPE in reciprocating sliding contact with Co–Cr, Ti-6AMV, and zirconia implant
bearing surfaces. In: Denton R, Keshavan MK (eds) Wear and friction of elastomers. ASTM
STP 1145. American Society For Testing and Materials, Philadelphia, pp 65–81
96. Dowson D, Wallbridge NC (1985) Laboratory wear tests and clinical observations of the
penetration of femoral heads into acetabular cups in total replacement hip joints. I: Charnley
prosthesis with polytetrafluoroethylene acetabular cups. Wear 104:203–215
97. Saikko V, Paavolainen P, Kleimola M, Slatis P (1992) A five-station hip joint simulator for
wear rate studies. Proc Inst Mech Eng H 206:195–200
98. Griffith MJ, Seidenstein MK, Williams D, Charnley J (1978) Socket wear in Charnley low
friction arthroplasty of the hip. Clin Orthop Relat Res 137:36–47
99. Deavane PA, Bourne RB, Rorabeck CH, Hardie RM, Home JG (1995) Measurement of poly-
ethylene wear in metal-backed acetabular cups. Clin Orthop Relat Res 319:303–316
100. Livermore J, Ilstrup D, Morrey B (1990) Effect of femoral head size on wear of the polyeth-
ylene acetabular component. J Bone Joint Surg 72A:518–528
101. Isaac GI, Wroblewski PJ, Atkinson JR, Dowson D (1992) Tribological study of retrieved hip
prostheses. Clin Orthop Relat Res 276:115–125
102. Wroblewski BM, McCullagh PJ, Siney PD (1992) Quality of the surface finish of the head of
the femoral component and the wear of the socket in long-term results of the Charnley low-
friction arthroplasty. Eng Med 20:181–183
103. Oonishi H, Takayama Y (1989) Comparisons of wear of UHMW polyethylene sliding against
metal and alumina in total hip prostheses. Bioceramics 1:272–277
104. Okumara H, Yamamuro P, Kumar T, Nakamura T, Oka M (1989) Socket wear in total hip
prosthesis with alumina ceramic head. Bioceramics 1:284–189
105. Egli A, Weber BG, Sieber H, Semlitsch M, Dörre E (1990) Experience with the pairing of
polyethylene/ceramic materials in hip endoprostheses. In: Willert H-G, Bucchorn GH, Eyerer
P (eds) Ultra-high molecular weight polyethylene as biomaterial in orthopaedic surgery.
Hoffrefe and Huber, Gottingen, pp 154–158
106. Zichner LP, Willert HG (1992) Comparison of alumina-polyethylene and metal-polyethylene
in clinical trials. Clin Orthop Relat Res 282:86–94
107. Atkinson JR, Dowling JM, Cicek RZ (1980) Materials for internal prostheses: the present
position and possible future developments. Biomaterials 1:89–96
Chapter 9
Degradation/resorption in Bioactive
Ceramics in Orthopaedics
H. Oonishi and H. Oomamiuda
9.1 Introduction
Bioceramics have now been widely used as bone replacement materials in orthopaedic
surgery. In particular, calcium phosphate ceramics have been applied as bioactive
ceramics with bone bonding capacities.
Biological responses such as bone bonding and the biodegradation properties of
these materials are very important in clinical applications. Any convincing conclu-
sion has not yet been reached as to whether these materials are biodegradable or not,
although it has been discussed for a long time.
Degradation is an important characteristic for biomaterials, and it is considered
to have a large influence on the bone bonding properties. This degradation charac-
teristic must be considered from the following two view points. These are the solu-
tion mediated dissolution process and the cellmediated process (phagocytosis).
This chapter overviews the literature regarding the biodegradation processes of
bioactive calcium phosphate ceramics from the viewpoint of in vitro physico-
chemical dissolution processes and in vivo/in vitro biological degradation
processes.
{Ca ( PO ) ( OH ) }
10 4 6 2 solid
10Ca 2 + + 6 PO34− + 2OH −
Therefore, its solubility product constant Ksp is calculated as follows;
10 6 2
K SP ( HA ) = Ca 2 + PO34− OH −
p
where values in brackets represent ionic activities.
There are many reports on solubility product constants of calcium phosphate
compounds obtained by the above equation, and these values are shown in Table 9.1.
Chow [18] calculated, based on these data, the solubility isotherms at 37 °C over
a wide pH range (Figure 9.1). These solubility isotherms were shown as the func-
tion of the concentration of calcium and phosphate ions in a saturated solution of
each calcium phosphate salt. The relative stability of calcium phosphate salts at
various values of pH can be obtained by these solubility isotherms. At a given pH,
a salt with its isotherm lying below that of another salt is less soluble and more
9 Degradation/resorption in Bioactive Ceramics in Orthopaedics 497
stable than the other. Therefore, HAp is the most stable and least soluble salt among
these salts in the range of pH below approximately 4.2 where DCPA is the least
soluble. Similarly, TTCP is the least stable and most soluble salt in the range of pH
below 8.5, above which pH DCPD is the most soluble.
Newesely [19] and Monma and Kanazawa [20] reported that α-TCP was con-
verted to HAp by hydration as follows;
Figure 9.1 Solubility phase diagram for the ternary system Ca(OH)2–H3PO4–H2O at 37°C [18].
9 Degradation/resorption in Bioactive Ceramics in Orthopaedics 499
concluded that the degradation rates of these materials were determined by neck
dissolution rate and neck geometry, the latter factor being dependent on the crystal-
lography and stoichiometry of the material and of the sintering conditions. Ducheyne
et al. [15] compared the dissolution rate and the precipitation rate of the following
six calcium phosphates in calcium and phosphate free solution with pH 7.3. The
dissolution rate increased in the following order:
HAp < CDAp < OHAp < β TCP < α TCP < TTCP
Most of the implanted calcium phosphate materials that were used in animal experi-
ments and clinical applications were TCP, HAp and calcium phosphate glasses.
Table 9.2 shows major reports of these investigations.
Most of the reports on TCP have concluded that TCP is biodegradable although
there are some differences depending on the characteristics of the materials used.
Bhaskar et al. [26] concluded that this biodegradation of TCP was caused by the
Table 9.2 Biodegradation of calcium phosphate compounds in vivo.
500
phagocytosis of the mesenchymal cells. Cameron et al. [28] stated that the ingestion
by giant cells did play a significant role in the degradation of TCP although passive
dissolution occurred. Klein et al [32, 33] stated that the micro-pores played an
important role in the biodegradation rate of TCP. The degradation of TCP started
mostly from the medulla by solution mediated disintegration processes, and fine
particles released were phagocytosed and removed by macrophages in the medulla
to the lymph nodes. Renooij et al. [35] reported that HAp was not affected by bio-
degradation processes, while TCP was subject to extensive bioresorption. Resorption
debris from TCP was found in mononuclear phagocytes and multinuclear osteo-
clastlike cells. Although multinuclear cells were occasionally seen near the surface
of HAP, cells carrying HAP debris were never observed. And it was supposed that
TCP was transformed into HAp in a physiologic environment.
Concerning the biodegradation of HAp, there are reports in the case of no degra-
dation, slow or partial degradation and for the degradable case. The differences in
these results are dependent on the experimental conditions such as the characteris-
tics of the materials, animal species, implanted sites and methods of observation.
Holmes et al. [35–39] carried out investigations using HAp which was derived
from marine coral and reported the results as follows. Significant biodegradation
occurred when implanted in load bearing sites such as mandibles, while minimal
biodegradation was observed in cortical bone of radius and no apparent evidence of
biodegradation was observed in cancellous bone of tibia. In clinical applications,
radiographic observations did not show any irrefutable evidence of biodegradation
and history of biopsies showed no conclusive evidence of biodegradation, while
osteoclasts were occasionally seen along the implant surface. In contrast to these
results, degradation of TCP appeared to occur by passive dissolution and osteoclas-
9 Degradation/resorption in Bioactive Ceramics in Orthopaedics 503
tic resorption, and in many cases it was radiographically observed in clinical trials,
especially where the implant was applied in a diaphyseal onlay fashion. Denissen
et al. [40] reportered no degradation of three different dense HAp varying in its
density. Similarly, Hoogendoorn et al. [41] reported through their long-term study
that porous HAp did not undergo biodegradation during 3.5 years of implantation,
while giant multinucleated cells were occasionally seen in pores near the bone and
ceramic surface.
On the other hand, Kurosawa et al. [42] observed the degradation of highly
porous HAp in their experiments, and concluded that this degradation was caused in
two ways; the mechanical collapse of the material and the ingestion of fine particles
released from the HAp surface by multinuclear giant cells. Similarly, Blitterswijk
et al. [43, 44] observed in their implantation experiments with dense and macropo-
rous HAp that the deposition of calcium, partially in the form of calcium phosphate,
was found on the implant surface, and the resorption of the implant occurred as the
result of phagocytosis by mono- and multi-nuclear cells. Oonishi et al. [45] com-
pared bioactivity for bone formation in several kinds of bioceramics. These materi-
als were divided into three groups; bioinert ceramics (alumina), surface bioactive
ceramics (HAp and Bioglass), and resorbable bioactive ceramics (DCPD, DCPA,
OCP, α-TCP, β-TCP, TTCP and amorphous HAp). In resorbable bioactive ceramics,
bioactivity or bioresorbability might increased in the following order:
Daculsi et al. [46] stated that the bioresorption of macroporous biphasic calcium
phosphate consisting of HAp and β-TCP was conducted by multinucleated cells
(osteoclastlike cells) and was related to the β-TCP content of this material. Bruijin
et al. [48] and Dhert et al. [49] compared the degradation of plasma spray coated
TTCP, MWL (magnesium whitlockite) and three types of HAp with various degrees
of crystallinity. It was revealed that both TTCP and semi-crystalline HAp under-
went distinct bulk degradation and amorphous HAp showed a gradual surface deg-
radation, while the degradation was negligible with the highly crystalline HAp and
MWL. Biodegradation appeared to be related to bone apposition, since more bone
seemed to be present on amorphous HAp and TTCP, as compared to highly crystal-
line HAp and MWL. The degrading surface of TTCP and amorphous HAp coatings
was most likely a dynamic zone in which dissolution and reprecipitation occurs.
This zone was therefore thought to be favourable for rapid bone formation and
bonding. At the interface between bone and MWL, a seam of unmineralized bone-
like tissue was frequently seen, and a substantial amount of aluminum was detected
in the MWL coating and the unmineralized bone-like tissue, which might cause the
impaired mineralization.
Since the discovery of Bioglass by Hench et al. [50], various kinds of bioactive
glasses and glass ceramics have been developed and applied clinically. Hench et al.
[51] summarized their study on Bioglasses which were based on the SiO2–P2O5–
504 H. Oonishi and H. Oomamiuda
interface and might, as a result, have an effect on the bone formation rate and
bonding strength between HAp and bone tissue. Similarly, rat bone marrow cells
were cultured on various plasma sprayed calcium phosphate coatings. Mineralized
extracellular matrix was formed on HAp, TCP and TTCP in 2 weeks, and was
formed on FAp (fluorapatite) in 8 weeks. It was only occasionally observed in some
area on MWL, which phenomenon might have been due to aluminium impurities in
the coating. It was concluded that plasma sprayed calcium phosphates would d isplay
different bone-bonding and biodegradation properties, depending on their chemical
composition and crystal structures.
9.4 Summary
References
47. J.A. Jansen, J.P.C.M. van de Waerden, J.G.C. Wolke and K. de Groot, J. Biomed. Mater. Res.,
25, 973–989, 1991.
48. J.D. de Bruijn, Y.P. Bovell and C.A. van Blitterswijk, Calcium Phosphate Biomaterials: Bone-
bonding and Biodegradation Properties, ed. J.D. de Bruijn, Offsetdrukkerij Haveka B.V.,
Alblasserdam, pp. 79–92, 1993.
49. W.J.A. Dhert, C.P.A.T. Klein, J.A. Jansen, E.A. van der Velde, R.C. Vriesde, P.M. Rozing and
K. de Groot, J. Biomed. Mater. Res., 27, 127–138, 1993.
50. L.L. Hench, R.J. Splinter, W.C. Allen and T.K. Greenlee, J. Biomed. Mater. Res. Symp., 2,
117–141, 1971.
51. L.L. Hench and B. Andersson, Advanced Series in Ceramic – Vol. 1 An Introduction to
Bioceramics eds L.L. Hench and J. Wilson, World Scientific, Singapore, pp. 41–62, 1993.
52. T. Kokubo, S. Ito, S. Sasaki and T. Yamamuro, J. Mater. Sci., 21, 536–540, 1986.
53. T. Kokubo, Advanced Series in Ceramics – Vol. 1 An Introduction to Bioceramics, eds
L.L. Hench and J. Wilson, World Scientific, Singapore, pp. 75–88, 1993.
54. M. Gregoire, I. Orly and J. Menanteau, J. Biomed. Mater. Res., 24, 165–177, 1990.
55. K. Gomi, B. Lowenberg, G. Shapiro and J.E. Davies, Biomaterials, 14, 91–96, 1993.
56. M. Ogura, T. Sakae and J.D. Davies, Bioceramics vol. 4, eds W. Bonfield, G.W. Hastings and
K.E. Tanner, Butterworth-Heinemann, London, UK, pp. 121–126, 1991.
57. J.D. de Bruijn, C.P.A.T. Klein, K. de Groot and C.A. van Blitterswijk, J. Biomed. Mater. Res.,
26, 1365–1382, 1992.
58. J.D. de Bruijn, J.S. Flachl, K. de Groot, C.A. van Blitterswijk and J.E. Davies, Calcium
Phosphate Biomaterials: Bone-bonding and Biodegradation Properties, ed. J.D. de Bruijn,
Offsetdrukkerij Haveka B.V., Alblasserdam, pp. 45–62, 1993.
59. J.D. de Bruijn, C.P.A.T. Klein, K. de Groot and C.A. van Blitterswijk, Calcium Phosphate
Biomaterials: Bone-bonding and Biodegradation Properties, ed. J.D. de Bruijn, Offsetdrukkerij
Haveka B.V., Alblasserdam, pp. 63–78, 1993.
Chapter 10
Corrosion of Metallic Implants
M.A. Barbosa
The surfaces of passive metals are normally attacked at specific points where the
oxide film has been destroyed and massive quantities of metal ions are released.
Depending on the magnification with which surfaces are observed, various degrees
of localized attack can be detected. Sometimes, however, corrosion may not be eas-
ily distinguishable from mechanical imperfections associated with manufacturing
or handling. Even under the scanning electron microscope (SEM) it is often difficult
to distinguish between mechanical indentations and pitting or crevice attack.
After determining the existence of corrosion, the next step is to assess its magni-
tude. This can be done, by corrosion scores, such as those given in Table 10.1
(Thomas et al., 1988). In this table the ‘no surface degradation’ score is obtained
with a magnification of 60x. If a higher magnification was used some ‘non-degraded’
surfaces might fall in one of the ‘surface degradation’ categories. The borderline
between ‘corroded’ and ‘non-corroded’) surfaces is therefore very much dependent
on magnification, as well as on surface preparation, as explained above. With these
limitations in mind, it is useful to have an idea of the incidence of corrosion, i.e. of
the percentage of implants that suffer some degree of attack. Table 10.2 compares
the data obtained by several authors. Vacuum melting (VM) significantly reduces
the susceptibility of 316L stainless steel to attack, while titanium is practically
immune. Corrosion, apart from affecting the mechanical performance of the
implants, also results in contamination of the tissues with metallic ions.
The detection of ions released from metallic implants is dependent on the tech-
nique used. Very minute amounts of ions can be detected by electrothermal atomic
Table 10.1 Grading scale used to evaluate the degree of interface and surface corrosion (Thomas
et al., 1988)
0 = no surface degradation visible at 60x magnification
1 = very mild surface degradation visible at 60x magnification
2 = mild surface degradation visible at 60x magnification
3 = moderate surface degradation visible without magnification
4 = heavy surface degradation visible without magnification
5 = very heavy severe surface degradation visible without magnification
These diagrams indicate the regions of immunity, passivation and corrosion of pure
metallic elements in pure water at 25°C. Fig.10.1 (a, b) gives the potential-pH dia-
grams of Cr and Ti. Cr is the element responsible for the passive behaviour of stain-
less steels and Co-Cr-Mo alloys. In Fig. 10.la passivation by a Cr(OHh film is
assumed. The film is thermodynamically stable over a wide range of pH and poten-
tial values. Below pH 4 the film is unstable and Cr is corroded. Ti is responsible for
the excellent corrosion resistance of Ti-Al-V alloys. In Fig. 10.lb passivation by a
hydrated TiO2 .H2O oxide has been assumed. The passivation region extends to
much higher potentials than in the case of Cr. The passive film is unstable below pH
2.5. In both diagrams the dotted lines give the region of stability of water: below
line a hydrogen is evolved, whereas above line b oxygen is released. Normally, the
corrosion potentials of implant materials do not reach such extreme values.
In spite of their usefulness in predicting the stability of metals and their oxides,
potential-pH diagrams suffer from a number of limitations. They refer to pure metals,
not to alloys, and to pure water, not to environments normally found in practical situa-
tions. For example, the diversity of chemical species, and particularly the presence of
chloride ions in physiological media, is responsible for substantial differences between
practical and predicted behaviour. Localized attack, in the form of crevice, pitting and
corrosion fatigue, is due to the presence of chloride. Furthermore, the kinetics of metal
10 Corrosion of Metallic Implants 511
Fig. 10.1 (a) (b) Theoretical domains of corrosion, immunity and passivation of titanium, at
25°C, considering Tiltz (Pourbaix, 1971).
512 M.A. Barbosa
dissolution or passivation cannot be assessed by these diagrams, which are purely ther-
modynamic. However, if not misused, potential-pH diagrams can give useful informa-
tion which must be complemented by other type of date, namely of kinetic nature.
The evolution of stainless steel composition can be used to illustrate the importance
of materials purity in reducing corrosion susceptibility. Chromium and molybde-
num are the key elements in promoting resistance to pitting and crevice attack of
stainless steels, but high chromium and molybdenum concentrations are not suffi-
cient to ensure an adequate corrosion resistance. Low concentrations of impurities,
like carbon, silicon, phosphorous and sulphur, are required. Type 316L and 316LVM
stainless steels are commonly employed to fabricate a variety of fracture fixation
devices. They both have low carbon concentration, below 0.03 wt%, which is indi-
cated by the letter L. VM stands for vacuum-melted, a technique that enables the
production of metals with very low concentrations of impurities.
A retrieval analysis of Kuntscher intramedullary rods (Cook et al. 1990) has
shown that significant surface corrosion, inclusion content and carbon content
occurred on early materials, which had remained in situ for 10 years or longer (maxi-
mum 23 years). Significant relationships were obtained for surface corrosion score
vs. thin globular oxide inclusion content, and for surface corrosion score vs. sulphide
inclusion content. Fig. 10.2 shows the data obtained for the former correlation.
Due to the presence of a thin oxide film, titanium has a very high corrosion resis-
tance. However, its low resistance under wear conditions may lead to enormously
high titanium concentrations in tissues adjacent to titanium implants (section 9.3.1).
Rapid film formation after surface damage is therefore of critical importance to
guarantee low levels of titanium ions.
The current density (c.d.) required to form a passive film is called the critical c.d.
for passivation, ic. The lower ic the better. Fig. 10.3 shows that Zr, Nb, Ta and Pd
decrease ic, whereas Sn increases it. It has been found (Okazaki et al. 1994) that ic
can be related to the percentages of Pd, Ta, Nb, Zr and Sn by the following
expression
ic(A.m-2 = 10-2{98-89.5[%Pd] - 9.5[%Ta] - 3.4[%Nb]- 0.67[%Zr] + 8[%Sn]}
Fig. 10.2 Relationship between surface corrosion and thin globular oxide inclusion content.
Regression line y= 1.78x+0.52, r= 0.65, n= 18, p <0.05. (Cook et al., 1990.)
Fig. 10.3 Effects of Zr, Nb, Ta, Pd and Sn contents critical current density for passivation in 15%
H2S04 and 5% HCl solutions at 310 K. (Okazaki et al., 1994.)
514 M.A. Barbosa
Fig. 10.4 Current density/potential curves of five different implant materials in Ringer’s solution
bubbled through with nitrogen .■ AISI 316L; ●CoNiCrMo; ▲ Ti6AI-7Nb; ▼ CP-titanium; x
Ti-6AI-7Nb/ODH. (Semlitsch et al., 1992.)
10 Corrosion of Metallic Implants 515
The need to combine different materials may sometimes arise. An example is the
use of hard materials for the head of hip joints in combination with a titanium stem.
Titanium has a very high corrosion resistance, but a very poor wear resistance.
Therefore, either surface hardening treatments (e.g. ion implantation of nitrogen or
surface alloying) or a harder material, e.g. a ceramic, are employed for the femoral
head. Ceramics, like alumina or zirconia, do not cause enhanced electrochemical
dissolution of the titanium stem because of their low electronic conductivity.
However, when another metal (e.g. Co-Cr-Mo alloy) is used instead, the possibility
of a galvanic couple between the stem and the head being formed exists. The situa-
tion illustrated by this example can be extended to other couples, including those
involving carbon. Even in the case of hard coatings galvanic couples between the
coating and the substrate may form.
In a first approximation, the safety of couples involving different materials can
be preditected by a number of experimental techniques. Table 10.3 summarizes the
data obtained by several authors. Notice that the couples between stainless steel and
other materials is unsafe. On the contrary, TiAlV/CoCrMo, CoCrMo/C and
TiAIV/C combinations may be considered safe. However, repeated fracture of the
oxide film at the conical taper region between head and stem of Ti6Al4V/CoCr
combinations has been associated with corrosion. Attack also occurred in CoCr/
CoCr combinations and was proportional to the duration of implantation, as seen in
Fig. 10.5 (Gilbert et al., 1993b). A larger percentage (34.5%) of cases of corrosion
was found with mixed CoCrffiAlV systems than with CoCr/CoCr systems (7%)
(Cook et al., 1994). Corrosion occurred at the interface between head and neck of
modular components. No correlation between the presence or extent of corrosion
with the time in situ was found. In another study the percentage of corroded tapered
516 M.A. Barbosa
Fig. 10.5 Graphs summarizing the percentage of mixed metal components which show signs of
moderate to severe corrosion as a function of duration of implanation. The dotted regression line
was fit to the data for the heads (O) and the solid line was fit to the data for the necks (▲). (Gilbert
et al., 1993b.)
Contaminations of tissue with metals may have two origins. The first is the
release of ionic species resulting from the process of electrochemical dissolution of
the implant. This is normally associated with static corrosion. Under dynamic con-
ditions, and particularly when fretting occurs, small metallic particles detach from
the surface, and become embedded in the soft tissue around the implant. The fate of
these particles may vary, depending on their size and chemical nature. They may,
for instance, undergo a process of corrosion, with the consequent release of metal
ions. This process may take place both in the extracellular matrix or as a result of
macrophage activity. Table 10.4 gives the concentration of Cr, Ni, Fe and Co in
biological samples. It shows that tissues around implants may be orders of magni-
tude richer in these metallic elements than normal blood or normal bone.
Titanium has a tendency to accumulate in tissues. The concentrations can be very
high, as indicated in Table 10.5. Titanium was not excreted in the urine of hamsters
injected with metal salts (Merritt et al., 1992). Small concentrations were found in
the serum, red blood cells and organs. Only 5.5% of the injected titanium was found
in the kidneys, liver, lung and spleen tissues. The authors suggest that titanium accu-
mulates at the injection site due to the high stability of the titanium dioxide that is
formed at physiological conditions. In the same study nearly all the injected vana-
dium was recovered in the urine. This behaviour is similar to that of nickel and
cobalt, and is related to the formation of highly soluble compounds.
High concentrations of metals were found in capsule and fibrous membranes of
loose titanium and Co-Cr stems of total hip prostheses (Dorr et al., 1990). The same
work reports elevated metal ion concentrations in synovial fluid and blood when-
ever cemented and uncemented stems are loose, but no increase when they are fixed.
The average values are given in Table 10.6. The standard deviations (not shown)
were often very large, of the order of magnitude of the averages.
Polyethylene wear debris may artefactually contribute to high ion readings in
periprosthetic tissues, as indicated in Tables 10.7 and 10.8 (Meldrum et al., 1993).
The high concentrations found in UHMWPE are due to the manufacturing pro-
cesses. These tables show that there are statistically significant increases in Co,
Al and Ti in the nonarticulated inserts with respect to bar stock. In retrieved
implants, large increases with respect to bar stock were found for Cr, Mo, Ti and
V. The role of UHMWPE wear debris would be twofold: irritant to tissues and
source of metal ions.
The accumulation of metal ions in periprosthetic tissue is a combination of two
sources: the extracellular matrix and the cells themselves. The ability of fibroblasts
to incorporate metal cations is a linear function of concentration, up to 50% toxicity
concentrations, for Ag+, Au4+, Cd 2+, Cu2+, In3+, Ni2+, Pd 2+ and Zn2+ (Wataha
et al., 1993), as illustrated in Fig. 10.6 for Cu2+, NF+ and Pd2 +. By measuring the
slope of the lines in this figure it is possible to estimate the uptake efficiency
(Table 10.9). The efficiency is highest for In3+ and lowest for Pd2+. Two years after
implantation of femoral components made of Ti-6Al-4V, the titanium and alumin-
ium concentrations measured in the synovial fluid were higher for cemented com-
ponents than for the uncemented (200 μm HA, or porous Ti coatings) components
(Karrholm et al., 1994). Table 10.10 gives the data for the synovial fluid and the
aluminium concentrations in serum and urine. No significant concentrations of
vanadium were found in any of the samples, which was also the case for titanium in
serum and urine. Fast clearance of vanadium from the synovial fluid, due to high
solubility of vanadium complexes, and formation of stable titanium compounds,
10 Corrosion of Metallic Implants 519
Table 10.6 Concentration (μg/l) of metals in tissues and blood retrieved during total arthroplasty
of cementless stems (Dorr et al, 1990)
TiAlV stems CoCr stems
Sample Ti Al V Co Cr Mo Ni
SF 556 654 62 588 385 58 32
SF (control) 13 109 5 5 3 21 5
CAP 1540 2053 288 821 3329 447 5789
CAP (control) 723 951 122 25 133 17 3996
FM 20813 10581 1027 2229 12554 1524 13234
Blood 67 218 23 20 110 1524 29
Blood (control) 17 12.5 5.8 0.1–1.2 2–6 0.5–1.8 2.9–7.0
SF – synovial fluid; CAP – capsule; FM – fibrous membrane.
Table 10.7 Cobalt-chrome alloy ion concentrations in UHMWPE et al. material and manufactured
and retrieved inserts (Meldrum et al., 1993)
Co Cr Mo Ni
Bar stock, n=3 55±5 330±5 5* 650±5
Manufactured inserts, n=9 440±250 520±440 5* 490±600
Retrieved, n=18(all cemented 54±42 1,500±1,400 87±120 1,360±1,300
inserts)
All concentrations are in parts per billion (nanograms/gram).
* This is the minimum detection limit of the spectrometer.
Table 10.8 Titanium alloy ion concentrations in UHMWPE material and manufactured and
retrieved inserts (Meldrum, et al. 1993)
AL Ti Va
Bar Stock n=3 5* 5* 5*
Manufactured inserts, n=9 800±200 2300±980 60±95
Retrieved, n=21 (all metal backed) 5* 6700±4500 220±410
All concentrations are in parts per billion (nanograms/gram).
* This is the minimum detection limit of the spectrophotometer.
e.g. titanium phosphates (Ribeiro et al., 1995), might be reasonable explanations for
these findings.
Experiments with metal salts and with stainless steel and Co–Cr–Mo electrodes
corroded in vivo by applying anodic potentials showed that all the nickel and most
of the cobalt were rapidly excreted (Brown et al., 1988). Acceleration of corrosion
by the use of anodic potentials obeys similar mechanisms both in vivo and in saline
when a potential of 500 mV vs. SCE is applied. This is illustrated by the single
straight line in Fig. 10.7 (weight loss vs. total charge). In particular, this implies that
the valency of the released cations is no different in both media, according to
Faraday’s law.
520 M.A. Barbosa
Fig. 10.6 Plots of the average uptake of metal cation per cell vs. concentration of the metal cation
in the medium for Cu2+, Ni2+, and Pd2+. The least-squaresmethod was used to fit linear curves to the
points. (Wataha et al., 1993.)
Table 10.9 Uptake efficiencies of metal cations by fibroblasts (Wataha et al., 1993)
Uptake efficiency
Metal cation ((fmol/cell )/μM )/h) *
Ag+ 23.8
Au4+ 1.0
Cd2+ 38.0
Cu2+ 0.26
In3+ 45.3
Ni2+ 0.21
Pd2+ 0.11
Zn2+ 0.73
* fmol = femtomoles (10-15 moles).
10 Corrosion of Metallic Implants 521
Table 10.10 Metal concentrations (ng/g) in synovial fluid, serum and urine. Median (range)
(Karrholm et al., 1994)
Cemented HA-coated Porous Controls
Ti/synovial fluid 37 (12-56) 3.5 (0-14) 6.4 (0-7.8) 0 (0-7.5)
Al/synovial fluid 12 (6.7-28) 5.2 (2.6-13) 3.8 (2.9-9.1) 7.3 (1.9-19)
Al/serum 2.1 (0-11) 1.4 (0-5.9) 5.7 (2.1-16) 3.7 (0-17)
Al/urine 6.2 (1.7-17) 4.9 (1.7-7.0) 4.2 (3.7-4.6) 4.6 (2.1-14)
Fig. 10.7 Linear regression analysis of weight loss as a function of total charge for stainless steel
rods at 500 mV (SCE) for 30 min. Symbols: * = in saline, box = in 10% serum, circle = in vivo.
(Brown et al. 1988.)
Acurate analysis of trace elements in tissues is essential to assess the degree of con-
tamination. This is not an easy task, mainly because we are dealing with normal
levels of the order of μg/litre. Sampling and sample preparation are steps prone to
serious contaminations, if the necessary precautions are not taken. As indicated in a
review by Lugowski et al. (1990), reported ‘normal’ levels of Cr in blood span over
four orders of magnitude. Contamination during sampling can be avoided by using
PTFE or polyethylene materials for blood collection and sample storage. For cut-
ting tissues a blade made of a material free from the elements to be analysed should
be used. For example, in our laboratory we have been using pure titanium blades to
522 M.A. Barbosa
cut soft tissues for Cr and Ni analysis. Contamination during sample preparation
can be minimized by: (i) adopting a very strict protocol of labware cleaning; (ii)
chemical treatment with ultrapure reagents, preferably in a microwave oven to
reduce the time necessary for digesting tissues; (iii) use a laminar flow hood to pre-
pare the samples, in order to avoid airborn contamination.
Lugowski has published a number of excellent works where the reader can find
very detailed information on the above and other aspects. The degree of precision and
accuracy to be expected when adequate experimental methods are used is indicated
in Table 10.11. This table refers to an ‘internal’ lab blood standard and to a Standard
Reference Material (SRM) with vanadium concentration certification. The relative
standard deviation (RSD) ranges from ca. 10% for Ni and Co, to ca. 29% for V.
The main constituents of sweat are chlorides (0.3–3.0 g/l), urea (0.12–0.57 g/l) and
lactic acid (0.45–4.5 g/l). When metallic objects come in contact with skin corro-
sion may occur, and if the corrosion products are toxic or irritating they may origi-
nate contact dermatitis. The most common example is dermatitis caused by
nickel-containing jewelry. In North America ca. 10% of men and women have a
history of nickel dermatitis (Randin, 1988). Although the degree of sensitization
may not be directly related to the amount of metal ions released from an object, due
to variability of response from person to person, it is generally considered that a
high corrosion resistance gives rise to fewer allergies.
The corrosion resistance of several materials in artificial sweat is given in
Table 10.12 (Randin, 1988). The composition (g/l) of the medium used was: 20
NaCl, 17.5 NH 4CI, 5 urea, 2.5 acetic acid, 15 lactic acid, pH 4.7. The table gives
the corrosion potential, Ecorr, in 02- and N2-saturated medium, the pitting potential,
Table 10.11 Precision of laboratory standard and precision and accuracy of SRM 909 human
serum (Lugowski et al., 1990)
Element X (μg/litre) n SD RSD (%)
Al 1.88 10 0.35 18.8
Co 2.37 9 0.26 10.9
Cr 0.71 8 0.14 19.6
Ni 2.95 8 0.30 10.1
Ti 4.20 7 0.41 9.8
V 0.28 8 0.08 29.4
V in SRM 909 2.73 13 0.14 5.0
certified value =
2.70±0.56 μg/litre
x – concentration; n – number of measurements; SD – Standard deviation; RSD – Relative SD.
10 Corrosion of Metallic Implants 523
Epit, and the corrosion rate, icorr, measured by the Tafel extrapolation method. icorr, is
given only for those alloys which are in the active state. For the other alloys Epit, is
given. The following materials were found to corrode in the active state: Ni, CuNi25,
NiAl (50:50, 60:40 and 70:30), WC+Ni, white gold, FN42, Nilo Alby K, NiP. Alloys
such as stainless steels, TiC+Mo2C+Ni, NiTi, Hastelloy X, Phydur, PdNi, and SnNi
are in the passive state and may pit under exceptional circumstances. Titanium has
an extremely high Epit and therefore cannot pit under normal use.
524 M.A. Barbosa
Potential
(mV vs S.C.E.)
−100
−200
−300
−400
−500
−600
0 60 120 180 240 300
Time (minutes)
Fig. 10.8 Potential-time curves for pure titanium in 0.9% saline with and without additions of
oxytetracycline: ○(upper line), as received; ○](lower line), abraded; ◊ 0.01 mg ml-1; □0.1 mg
ml-1; ∆ 1.0 mg ml-l. (von Fraunhofer et al., 1989.)
526 M.A. Barbosa
For a detailed description of anodic oxidation of titanium and its alloys the reader
may refer to a review by Aladjem (1973).
The oxide on titanium can grow to thicknesses of the order of 100 nm or more
by applying anodic currents in suitable electrolytes. H3P04 and NaOH baths have
been used for this purpose. The colour of the oxide changes with thickness due to
light interference. A gold colour corresponds to a thickness of the order of
10–25 nm whereas a blue colour is normally associated with thicknesses of
30–60 nm. The corrosion resistance of anodized titanium increases as the oxide
becomes thicker. This is illustrated in Fig. 10.9 (Cigada et al. 1992), which shows
that films formed in H3PO4 are thicker than those formed in air. They are also more
protective, since the passive current density in a buffered physiological solution at
38°C is ca. 10% that measured for specimens oxidized in air. The same figure
shows that anodizing in NaOH is not so effective in reducing the current density as
doing it in H3P04•
Fig. 10.9 Average passivity currents (between 600 and SOO h) and standard deviations in physi-
ological solution of Ti6AI4V specimens, oxidized and anodized in different conditions. (Cigada
et al. 1992.)
10 Corrosion of Metallic Implants 527
The corrosion rate of anodized titanium (solution: 60 ml ethanol, 35 ml water,
10 ml lactic acid, 5 ml phosphoric acid, 5 mg citric acid and 5 mg oxalic acid; 45V,
45s) is much lower than that of passivated titanium (40% volume nitric acid, room
temperature, 30 min.), as indicated in Table 10.15 (Ong et al., 1993). The corrosion
potential of the former is also more noble, as indicated in the same table. The aver-
age thicknesses are given in Table 10.16. The anodized film is ca. 10 times thicker
than the passivated film.
There have recently been reports (Lowenberg et al., 1994l Callen et al.,1995)
indicating that passivation of Ti–6Al–4V in HNO3 increases the release of all three
constituent elements in a culture medium (α-Minimal Essential Medium with 15%
foetal bovine serum and 10% antibiotics). Table 10.17 exemplifies the results
obtained for titanium ions, for three periods of immersion of three days each. The
level of Ti is significantly reduced throughout the 9-day experimental period.
Table 10.16 Titanium oxide thickness (nm), relative to tantalum pentoxide (Ong et al., 1993)
Treatment Mean SD Sample size
Non-passivated 3.1 0.6 1.8
Passivated 4.1 1.8 1.8
Anodized 43.6 4.9 1.5
Table 10.17 Trace Levels of Ti, A, and V in culture medium (Callen et al., 1995)
cpTi Wells Ti6Al4V Wells
Time Control
points Not Passivated Passivated Not Passivated Passivated Values
Ti
1st 23.696±12.892 l5.735±3.354 12.599±3.850 23.338±8.497 4.983±0.977
2nd 12.650±5.275 16.640±4.940 l1.050±1.601 24.645±8.419
3rd 6.444±2.495 8.738±2.983 5.513±1.943 10.486±3.674
Al
1st 4.091±0.677 4. 133±0.523 8.933±1.187 16.878±4.574 3.476±0.392
2nd 4.694±1.039 5.523±2.784 5.703±0.707 9.656±2.750
3rd 5.215±1.096 4.l49±0.397 4.516±0.384 6.614±1.407
V
1st 0.508±0.199 0.366±0.167 6.195±2.191 21.104±8.828 0.246±0.082
2nd 0.255±0.018 0.l71±0.05l 2.789±1.129 10.096±5.697
3rd 0 0.330±0.213 0.588±0.334 4.218±2.003
528 M.A. Barbosa
When metals are used as coatings the possibility of occurrence of galvanic corro-
sion exists, since cracks or pores in the coating enable the corrosive medium to
contact the substrate. Mainly for this reason metallic coatings have not been used in
internal implants. However, surface treatments with inert materials have been
widely applied and are now in clinical practice. The effect of these and other surface
treatments will be addressed in this section.
With the development of ion implantation the plating of practically any element
on any substrate opened new perspectives to surface modification of biomaterials.
Carbon and nitrogen have been the species most widely employed to modify the
corrosion and wear behaviour of stainless steels and titanium alloys. However, the
plating of metallic elements, with a view to modifying either the corrosion perfor-
mance and/or the biological behaviour of metallic implants, is an interesting possi-
bility. This would be particularly valuable in the case of stainless steel substrates.
Very little has been reported in this area. Titanium, niobium and tantalum coatings
on stainless steel act as anodes, therefore indicating that they may retard the transfer
of chromium and nickel into the environment (Gluszek and Masalski, 1992). In the
same medium (Ringer’s solution) the oxide layers formed on titanium, niobium and
tantalum by prolonged (100h) exposure to air are not very stable. Fig. 10.10 shows
this effect (dotted line). The galvanic current first increases, corresponding to modi-
fication/destruction of the original oxide layer, and then decreases, corresponding to
increased stability of the film formed in solution. When freshly ground specimens
are used (solid line) the galvanic current decreases with time, due to film growth,
which follows a logarithmic law [log i ∝ (–log t)].
Laser surface alloying (LSA) of Ti6Al4V with Nb, Mo and Zr, in order to
increase surface hardness, has shown that the latter element is the most promising
(Akgun and Inal, 1994). The hardness increase is almost threefold in comparison to
the substrate and identical to that obtained by laser surface melting (LSM). Since a
nitrogen atmosphere was used in LSA and LSM, TiN formed during melting appears
to be the main reason for the high hardnesses obtained. The hardened zone extends
to a depth of over 0.5 mm. Wear and fretting corrosion could be considerably
reduced with such surface treatments, but no experimental data are yet available.
Radio-frequency (RF) plasma treatments in air (1.0 torr) produced enhanced
ionic release from Co–Cr–Mo and Ti–6Al–4V alloys, without any improvement in
biological behaviour (Kummer et al., 1992). Table 10.18 gives the Cr, Co and Ti
concentrations obtained after 10 days exposure to cell culture fluid (DMEM with
10% FBS). The RF plasma-treated Ti–Al–V alloy shows a 3-fold increase after the
plasma-treatment.
Depassivation of Ti–6Al–4V occurs during planar-planar rubbing against
PMMA in Ringer’s solution (Rabbe et al., 1994). The free corrosion potential drops
to values below -650 mV vs. SCE. This potential is substantially lower than those
obtained for nitrogen ion-implanted and ion-nitrided Ti–6Al–4V, which are of the
10 Corrosion of Metallic Implants 529
Fig. 10.10 Galvanic current density-time relationship for 316L/titanium couple in Ringer’s solu-
tion. (Gluszek and Masalski, 1992.)
Table 10.18 Concentration of Cr, Co and Ti in cell culture fluid after 10 days (Kummer et al.,
1992)
concentration (μ/mL)
Sample Cr Cr Ti
Control 1.1 <2 -
Co–Cr–Mo 96.5 720 –
Co–Cr–Mo/RF 120.5 960 –
Ti–AI–V – – 35.6
Ti–AI–V/RF – – 102.0
RF = Radio-frequency plasma treated.
530 M.A. Barbosa
order of , -100 mV vs. SCE. At high doses (~2xl018 ions/cm2) a TiN layer is formed
on ion-implanted surfaces, whereas TiN and Ti2N form as a result of ion nitriding,
thus increasing the hardness of the alloy surface.
Superalloy MA 956 (Fe–20Cr–4.5Al–0.5Ti–0.5Y2O3, wt%) possesses the inter-
esting ability of developing a fine α-alumina scale on the surface upon isothermal
treatment at 110 °C (Escudero and González-Carrasco, 1994). This layer acts as a
coating, being responsible for an improved corrosion resistance of the alloy, as indi-
cated by the anodic polarization curves given in Fig. 10.11. No pitting corrosion
occurs for potentials up to 700 mV vs. SCE.
Hard ceramic coatings (Al2O3 and SiC) deposited by radio-frequency (RF) sput-
tering on Ti and Co–Cr–Mo alloy resulted in significant corrosion resistance
improvement, as seen in Table 10.19 (Sella et al., 1990). The data in this table were
obtained by applying a constant potential of 1.4 V vs. SCE and measuring the cor-
rosion current density (c.d.) in artificial saliva. SiC coatings deposited on Ti caused
a decrease of c.d. of ca. 300 times. The same coating applied to Co–Cr–Mo was
only effective when an intermediate Ti sublayer was used to avoid cracking. An
Al–Al2O3 cermet sublayer was also very effective in improving the corrosion resis-
tance of Al2O3-coated Co–Cr–Mo alloy; the c.d. decreased 200 times when both
layers were used. The authors indicate that Al2O3 and SiC coatings gave better bio-
compatibility than Ti and that no signs of corrosion were observed on Al2O3-coated
dental implants removed after several years of implantation.
Modification of Ti–6Al–4V alloy surfaces by ion implantation with iridium, at
fluences of 0.74 x 1016 and 1.48x1016 ions/cm2, corresponding to 2.5 and 5.0 at% Ir
peak concentrations, has been reported (Buchanan and Lee, 1990). After pre-
treatment of the implanted surfaces in 1N H2S04 the surfaces become enriched in Ir
Fig. 10.11 Anodic polarization curves for MA956 in the as-received and oxidized conditions after
nine months of immersion in Hank’s solution. ○ Oxidized; □ as-received. (Escudero and
Gonzalez-Carrasco, 1994.)
10 Corrosion of Metallic Implants 531
Table 10.19 Comparison of the corrosion currents of coated and uncoated metals (Sella et al., 1990)
Corrosion current at E=1.4V/SCE (μA/cm2)
Uncoated metal or alloy
Ni-Cr 6000-8000
Co-Cr-Mo 8000
Ti 260
Experimental coatings
SiC (1 μm) on Ti 0.8
SiC (1 μm) on Co-Cr 10000
Ti (1 μm) on Co-Cr 500
SiC (1 μm) + Ti (1 μm) on Co-Cr 28
Al2o3 (0.5 μm) on Co-Cr 1800
Al2o3 (0.5 μm) + Al-Al2o3 cerment on Co-Cr 40
Fig. 10.12 Corrosion potential vs. time in aerated isotonic saline. (Buchanan and Lee, 1990.)
(the concentrations are over 60% and may approach 100%), as a consequence of
alloy dissolution. The result is a corrosion potential in isotonic saline very close to
that of pure Ir, as depicted in Fig. 10.12. Owing to the very high corrosion resistance
of Ir, its implantation onto titanium is of potential interest, particularly if it becomes
significantly enriched on the surface. Galvanic couples formed between Ir and Ti is
a possibility that justifies further research.
532 M.A. Barbosa
Fig. 10.13 Metallic dissolution products released from a polished Co-Cr-Mo alloy after 550 h in
0.17 M NaCl+2.7x10-3 M EDTA solution at 37°C. □ uncoated; ■ TiN coated. (Wisbey et al., 1987.)
10 Corrosion of Metallic Implants 533
Most of the data available on this topic refer to hydroxyapatite deposited by plasma
spraying. Although compounds may form at the metal/hydroxyapatite interface, partic-
ularly in the case of titanium, their existence has not been unequivocally demonstrated.
Titanium phosphates and phosphides, as well as calcium titanates, may exist, but they
probably form very thin layers. The large surface roughness, caused by grit blasting of
the substrate prior to hydroxyapatite deposition, is another factor that renders identifica-
tion of any interfacial compounds by surface analysis techniques difficult.
Table 10.21 shows that the corrosion resistance of stainless steel increases upon
coating with hydroxyapatite. The presence of calcium phosphate in solution, due to
dissolution of hydroxyapatite, seems to be the cause for these changes. The same
table indicates that calcium phosphate is detrimental to the corrosion resistance of
titanium, both in terms of film breakdown potential and corrosion rate under passive
conditions.
Metallic ions may influence the formation of calcium phosphates in different ways.
Some inhibit (nickel, tin, cobalt, manganese, copper, zinc, gallium, thalium, molyb-
denum, cadmium, antimony, magnesium, and mercury), a few stimulate (iron [fer-
ric] and iridium) whereas others have no effect (cerium, titanium, chromium, iron
[ferrous], iridium, palladium, platinum, silver, gold, aluminum, and lead) (Okamoto
and Hidaka, 1994). Fig. 10.14 gives the induction time for calcium phosphate for-
mation vs. concentration for the above metal ions.
Heat treatment of Ti–6Al–4V at 280 °C for 3 h produced a high accumulation of
Ca deposited next to screws implanted in rats (Hazan et al.,1993), as indicated in
Table 10.22. The oxide was twice as thick as that formed on non-treated screws.
534 M.A. Barbosa
Table 10.21 Effect of hydroxyapatite coatings and calcium phosphate solutions on the corrosion
resistance of titanium and stainless steel
Material Solution Effect Ref
316L ss/HA Saline Increase in breakdown potential Hayashi et al.,
1990
Ti-6Al-4V/HA Saline Decrease in breakdown potential Hayashi et al.,
1990
316L ss Saline+ Increase in breakdown potential Sousa and
Ca phosphate Barbosa, 1991
316L ss Saline+ Decrease in corrosion rate (passive Barbosa, 1991 b
Ca phosphate state)
Ti cp Saline+ Decrease in breakdown potential Sousa and
Ca phosphate Barbosa, 1991
Ti cp Saline+ Increase in corrosion rate (passive Barbosa, 1991 b
Ca phosphate state)
Fig. 10.14 The induction time (min) versus concentration of various metal ions (open circle) and
HEBP: 1-hydroxyethylidene-l, 1-biphosphonate (closed circle). (Okamoto and Hidaka, 1994.)
Table 10.22 Calcium deposition (mg) next to control and heat-treated Ti–6Al–4V implants
(Hazan et al. 1993)
Time after
immersion (days) Control Heat treated
4 – –
5 – –
6 2.0±0.2 4.5±6.5
10 3.l±0.5 7.4±1.1
35 4.0±1.0 9.6±1.0
10 Corrosion of Metallic Implants 535
The oxide film on metallic implants is usually very thin (5–10 nm). It is formed as
a result of a spontaneous reaction between the metal and the environment. In spite
of the common use of immersion treatments in nitric acid solutions, usually known
as passivation treatments, they are not necessary to form an oxide. They are often
10 Corrosion of Metallic Implants 537
Table 10.23 Selected physico-chemico properties of metal oxides in water (Tengvall and Lundstrom, 1992)
0.996
Na(FeCN64-
FeCN63-)
Order of Water Dielectric polarization
practical corrosion pKa of Solubility lsoelectric Charge onst for resistance, Rp Corroded in Corrosion
mobility Element product hydrolysis at pH 7 (M) point at pH 7 oxide (KΩcm2) Essential Soft tissue reaction H202 at pH7 product
2 Nb Nb2O5 >20 -10-5 - 280 455 No Inert? No
3 Ta Ta2O5 >20 -10-5 - 12 1430 No Inert? No
4 Au Au2O3 - (pH7) 7x10-2 >10 (calc.) ++ 0.28 No Sequestration Yes AU2O3
Au(OH)3 5x106 E0≃1.04V
7 Ti TiO2 anat. +18 3x106 6.2 -- 48 78 110 714 No No Yes TiO22-
brook. TiO2 TiO2
rutile
14 Ag Ag2O +10 Ag++0.7 104 >1 (Ag+)12 ++ 9 No Sequestrattion Yes AgO-
AgOH
19 Al Al2O3 α 14.6 - 106 10-3 -9 + 5-10 No Sequestrattion ? ?
Al(OH)3
am.
21 Cr Cr2O3 CrO3 -1.8(Cr(OH)2+) 10-11 >10-13 8.4 (Cr3+) + 12 Yes Toxicity Toxicity Yes Yes CrO42-HCrO4-
Cr(OH)3 18.6(CrO2-)
28 Fe Fe2O3 -13.3 (Fe2+) >10-10 10-1 12.4 (Fe2+) + 100 30-38 Stainless steel Yes Sequestration Yes FeO42-?
Fe(OH)2 (pH 9.1) 10-9 8.0 (Fe3+) 316 4.38
Fe(OH)3
29 Ni Ni2+ NiO 12.2 (Ni2+) 10-15 10-11 9.5 + Yes Toxicity Yes NiO42-?
(pH 8.9) NiO2
30 Co C02+ CoO -12.6 10-11 10-12 10.8 + CoCrNi Yes Toxicity Yes CoO2
3.32
- -
40 V V2O5 V2O4 +10.3 (HV2O-) >1 10-4 1-2.5 (V5+) - Yes Toxicity Yes H3V2O7H 2VO4
M.A. Barbosa
10 Corrosion of Metallic Implants 539
Generally, the oxide film grows according to a logarithmic law (log thickness
proportional to log time), reaching a quasi-stationary thickness very rapidly. Under
stationary conditions, film dissolution and film formation rates should be the same.
Normally, film thickness increases slowly with time, after an initial period of rapid
growth. This is illustrated in Fig. 10.18, which depicts film thickening with implan-
tation time (Kasemo and Lausmaa, 1994).
Fig. 10.18 An artist’s attempt to capture some of the complexity involved in the interaction
between a material and living tissue, exemplified here by a titanium implant in bone. Note the wide
range of dimensions and time scales that are relevant. (Kasemo and Lausmaa, 1994.)
540 M.A. Barbosa
Fig. 10.19 Normalized integral passive dissolution kinetics for titanium thin films immersed in
EDTA/SIE (simulated interstitial electrolyte): (a) real time data empirically fitted with two-phase
logarithmic law relationship; (b) a semilogarithmic plot of the data demonstrating the two-phase
logarithm relationship. The correlation coefficient for the least-squares fit of the linear functions
are given. (Healy and Ducheyne, 1992.)
10 Corrosion of Metallic Implants 541
Table 10.24 Type of oxide formed on biomaterials (pure oxygen, 300°C, 30 min.) (Oshida et al.,
1992)
Material Type of oxide
Pure Ti TiO2 (rutile)
Ti-6Al-4V TiO2 with traces of Al2TiO5
Ni-Ti, austenitic and martensitic Mixture of TiO2 and NiTiO3
316L stainless steel spinel-type [(Fe,Ni)O· (Fe,Cr)2O3]a and
corundum-type oxides [(Fe,Cr)2O3]a
a Possible composition.
542 M.A. Barbosa
Table 10.25 Thickness (nm) of titanium oxide films (Pan et al., 1994)
Source Polarized at 0.4 V/SCE
H2O2 in the PBS (mM) Dry-polished Wet-ground 0 1 10
XPS measurements 1.5 4.6 6.3 6.2 5.8
Capacitance measurements 6.7 5.5 6.0
Literature data 1.2-1.6 4-5 6.6
molecules, a property which has been established for other metals, e.g. iron, copper,
cadmium, chromium, lead, mercury, nickel and vanadium (Stohs and Bagchi, 1995).
These metals produce reactive oxygen species, leading to lipid peroxidation, DNA
damage, depletion of sulphydryls, apart from modifying calcium homeostasis. Since
large concentrations of titanium debris may be found around Ti and Ti-alloy implants
(section 9.3.1) the oxidative deterioration of biological molecules induced by the pres-
ence of Ti ions is a process that deserves to be studied.
The oxide formed on titanium upon passivation in HNO3 is composed of regions
of mixed titanium oxides (anatase and rutile), together with areas of amorphous
titanium oxide (Browne and Gregson, 1994). Films formed on anodized titanium
may be one order of magnitude thicker than those formed by passivation (section
9.4.1). The film is predominantly constituted by TiO2, with the presence of carboxyl
groups (Ong, 1993). It appears that upon passivation of cp Ti and Ti-6AI-4V alloy
the film on the former is thinner (3.2 ± 0.8 nm) than that on the latter (8.3 ± 1.2 nm)
(Keller et al., 1994). TiO2 films are generally amorphous, except in the case of thick
films produced by thermal oxidation or anodizing. Table 10.26 summarizes the
characteristics (composition, oxide thickness, surface topography/roughness, and
substrate microstructure/oxide crystallinity) of titanium samples subjected to vari-
ous treatments (Larsson et al., 1994). Electropolished + anodized (1M acetic acid,
room temperature) films are thicker than those formed by electropolishing and on
‘clinical reference’ (machined) surfaces. For 80V the oxide is crystalline.
Table 10.27 summarizes the composition and thickness of oxides formed on a
Co–Cr–Mo alloy exposed to ‘dry air’ and ‘wet steam’ for 1 h (Lewis, 1993a).
Table 10.26 Summary of surface characteristics of the four different types of Ti samples (Larsson
et al., 1994)
Substrate
Oxide Surface topography microstructure Oxide
Preparation Composition thickness Surface roughness crystallinity
Clinical reference TiO2+45-80% 4nm Rough, with Rough, Plastically deformed,
C, traces of Ca, with and amorphous metal
S, Si, P, CI and protrusions, ≤ 10 surface Non-
Na μm Rrms=29±4 nm crystalline oxide
Electropolished TiO2+55-90% 4-5nm Smooth, with Polycrystalline metal
C, traces of Ca, occasional pits, ≤ 1 surface Non-
S, Si, P, Cl and μm crystalline oxide
Na Rrms=2.7±0.9 nm
Electropolished + TiO2+55-70 21 nm Smooth, with pits Polycrystalline metal
anodized, 10 V %C, traces of and porous regions, surface Non-
Ca, S, Si, and ~10 μm crystalline oxide
Cl Rrms=1.5±1 nm
Electropolished + TiO2+~34-40% 180nm Heterogeneous, Polycrystalline metal
anodized, 10 V + C, traces of Ca with smooth or surface. Crystalline
and Cl porous regions, - 10 oxide (anatase)
μm Rrms=16±2 nm
Table 10.28 Initial contact angles and changes in contact angles as function of time of oxidized
surfaces of biomaterials after mechanical and buff polishing (Oshida et al., 1992)
Mechanical polish-oxidizing Buff polish-oxidizing
θ° (deg) δθ/δt θ° (deg) δθ/δt
Pure Ti 54.24 -0.0046
Ti6A14V 32.08 -0.0010 30.85 -0.0015
NiTi (m) 69.88 -0.0055 68.92 -0.0053
NiTi (a) 71.88 -0.0048
316L s.s. 56.46 -0.0024 55.73 -0.0025
Pure Ni 35.72 -0.0016
Co-Cr alloy 62.04 -0.0023 61.85 -0.0021
α-alumina 60.87 -0.0044
m – martensite; a – austenite.
544 M.A. Barbosa
Table 10.29 Critical surface tension of Ti surfaces (Kilpadi and Lemons 1994)
Critical Surface
Specimen Plot Tension, τc (dyn/cm) Comments
I C – No liquids were appropriate
P 46.0±1.08
P 42.5±1.08 Only diiodomethane and bromonaphthalene were
used
II C 31.6±0.48 Water was not used in these analyses, as it did not
fit with Good’s criterion and also did not fall in line
with the other liquids
P 35.4±0.48 Only glycerol and thiodoethanol were used
D 34.9±0.48 Only diiodomethane and bromonaphthalene were
used
III C 40.0±0.41
41.4±0.59
D 40.7±0.59 Only diiodomethane and bromonaphthalene were
used
IV C 41.9±0.79
P 41.5±1.05D
D 42.5±1.05 Only diiodomethane and bromonaphthalene were
used
V C 37.4±0.51
P 31.0±0.30 Water was not included
D 41.8±0.51 Only diiodomethane and bromonaphthalene
were used
I – Non-passivated, TFGD-treated, polished machined flats; II – Non-passivated, unsterilized, pol-
ished machined flats; III – Passivated, dry-heat-sterilized, polished machined flats; IV – Passivated,
dry-heat-sterilized, polished coined flats; V – Passivated, dry-heat-sterilized, unpolished flats; C –
Composite (includes all liquids); P – only polar liquids: D – only dispersive liquids. RFGD – Radio
Frequency Glow Discharge.
10 Corrosion of Metallic Implants 545
discharge (RFGD) treated samples showed the higher CST. Grain size (70 vs. 23
μm) did not affect the CST of polished, passivated, and dry-heat-sterilized titanium
surfaces.
The equilibrium contact angles of cp Ti and Ti–6Al–4V, both passivated in nitric
acid, were 52±2° and 56±4°, respectively (Keller et al., 1994). Wettability was mea-
sured employing water drops. This similarity in contact angles reflets the similarity
in oxide film composition found in the same work. However, the film on the alloy
surface was significantly thicker (8.3±1.2 nm) than that on the cp Ti (3.2±0.8 nm).
Reference
O.V. Akgun and O.T. Inal, 'Laser surface modification of Ti-6A1-4V alloy', Journal of Materials
Science , 29, 1159-1168 (1994).
A. Aladjem, 'Review Anodic oxidation of titanium and its alloys', Journal of Materials Science, 8,
688-704 (1973).
M.A. Barbosa, 'Corrosion mechanisms of metallic biomaterials', Biomaterials Degradation-
Fundamental Aspects and Related Clinical Phenomena, European Materials Research Society
Monographs, Vol. 1 (eds. M.A. Barbosa, F. Burny, J. Cordey, E. Dorre, G. Hastings, D. Muster
and P. Tranquilli-Leali), pp. 227-257, Elsevier Science Publishers, Amsterdam (1991 a).
M.A. Barbosa, 'Electrochemical impedance studies on calcium phosphate-metal interfaces',
Bioceramics, Vol. 4 (eds. W. Bonfield, G.W. Hastings and K.E. Tanner), pp. 326-333,
Butterworth-Heinemann (1991 b).
M.A. Barbosa, 'Surface layers and the reactivity of metallic implants', High-Tech Biomaterials,
European Materials Research Society Monographs, Vol. 3 (eds. D. Muster, M.A. Barbosa,
F. Burny, J. Cordey, E. Dorre, G. Hastings, and P. Tranquilli-Leali), pp. 257-283, Elsevier
Science Publishers B.V., Amsterdam (1992).
J. Black, 'Biological Performance of Tantalum', Clinical Materials, 16, 167-173 (1994)
N.C. Blumenthal and V. Cosma, 'Inhibition of apatite formation by titanium and vanadium ions',
Journal of Biomedical Materials Research: Applied Biomaterials , 23, 13-22 (1989).
M. Browne and P.J. Gregson, 'Surface modification of titanium alloy implants', Biomaterials, 15,
894-898 (1994).
S.A. Brown, L.J. Farnsworth, K. Merritt, and T.D. Crowe, ' In vitro and in vivo metal ion release',
Journal of Biomedical Materials Research, 22, 321-338 (1988).
R.A. Buchanan and I.S. Lee, 'Surface modification of biomaterials through noble metal ion implan-
tation', Journal of Biomedical Materials Research, 24,309-318 (1990).
B.W. Callen, B.F. Lowenberg, S. Lugowski, R.N.S. Sodhi, and J.E. Davies, 'Nitric acid passivation
of Ti6A14V reduces thickness of surface oxide layer and increases trace element release',
Journal of Biomedical Materials Research, 29, 279-290 (1995).
S.K. Chawla, S.A. Brown, K. Merritt, and J.H. Payer, 'Serum protein effects on polarization behav-
ior of 316L stainless steel', Corrosion 46, 147-152 (1990).
A. Cigada, G. Rondelli, B. Vicentini, M. Giacomazzi, and A. Roos, 'Duplex stainless steels for
osteosynthesis devices', Journal of Biomedical Materials Research, 23, 1087-1095 (1989).
A. Cigada, M. Cabrini, and P. Pedeferri, 'Increasing of the corrosion resistance of the Ti6A14V
alloy by high thickness anodic oxidation', Journal of Materials Science: Materials in Medicine,
3, 408-412 (1992).
J.P. Collier, V.A. Surprenant, R.E. Jensen, and M.B. Mayor, 'Corrosion at the interface of cobalt-
alloy heads on titanium-alloy stems', Clinical Orthopaedics, 271, 305-312 (1991).
S.D. Cook, M.R. Brinker, R.C. Anderson, R.J. Tomlinson and J.C. Butler, 'Performance of retrieved
Kuntscher intramedullary rods: improved corrosion resistance with contemporary material
design', Clinical Materials, 5, 53-71 (1990).
546 M.A. Barbosa
S.J. Stohs and D. Bagchi, Oxidative mechanisms in the toxicity of metal ions, Free Radical Biology
and Medicine, 18, 321-336 (1995).
D.S. Sutherland, P.O. Forshaw, G.C. Allen, I.T. Brown and K.R. Williams, Surface analysis of tita-
nium implants, Biomaterials, 14, 893-899 (1993).
P. Tengvall, L. Lundstrom, L. Sjokvist, H. Elwing, and L.M. Bjurstein, Titaniumhydrogen perox-
ide interaction: model studies of the influence of the inflammatory response on titanium
implants, Biomaterials , 10, 166-175 (1989).
P. Tengvall and I. Lundstrom, Physico-chemical considerations of titanium as a biomaterial,
Clinical Materials, 9, 115-134 (1992).
K.A. Thomas, S.D. Cook, A.F. Harding, and R.J. Haddad Jr., Tissue reaction to implant. corrosion
in 38 internal fixation devices, Orthopedics, 11, 441-451 (1988).
H. Tomás, A.P. Freire, and L.M. Abrantes, Cast Co-Cr alloy and pure chromium in proteinaceous
media: an electrochemical characterization, Journal of Materials Science: Materials in
Medicine, 5, 446-451 (1994).
S. Torgersen and N.R. Gjerdet, Retrieval study of stainless steel and titanium mini plates and
screws used in maxillofacial surgery, Journal of Materials Science: Materials in Medicine , 5,
256-262 (1994).
D.L. Tsalev and Z.K. Zaprianov, Atomic absorption spectrometry in occupational and environmen-
tal health practice, Vol. I, CRC Press, Boca Raton, 1983.
J.A. von Fraunhofer, N. Berberich, and D. Seligson, Antibiotic-metal interactions in saline
medium, Biomaterials, 10, 136-138 (1989).
J.C. Wataha, C.T. Hanks, and R.G. Craig, Uptake of metal cations by fibroblasts in vitro, Journal
of Biomedical Materials Research, 27, 227-232 (1993).
A. Wisbey, P.J. Gregson, and M. Tuke Application of PVD TiN coating to CoCr-Mo based surgical
implants, Biomaterials, 8, 477-480 (1987).
R.L. Williams, S.A. Brown, and K. Merritt, Electrochemical studies on the influence of proteins on
the corrosion of implant alloys, Biomaterials , 9, 181-186 (1988).
Chapter 11
Carbons
11.1 Introduction
The biocompatibility of carbon has long been appreciated: ancient man, for exam-
ple, knew that pulverized charcoal could be placed under the skin without any
apparent ill effects (Benson, 1969). The charcoal particles visibly remained indefi-
nitely and thus allowed ancient man the means to decorate himself permanently
with tattoos. However, it was not until the mid-1960s that carbon was first consid-
ered for use as a structural material in implantable prosthetic devices. During this
period, a specific, imperfectly crystalline, man-made, pyrolytic form of carbon was
found to be well suited for application in prosthetic heart valves. Because of the
outstanding clinical success of pyrolytic carbon in long-term structural components
of heart valve prostheses, carbons have assumed a prominent position in our reper-
toire of biomaterials and have sparked investigation of other forms of carbons for
possible in vivo use. A number of these forms are listed in Table 11.1. This chapter
will be devoted to a discussion of the background and historical uses of carbons in
medical devices along with suggestions for future research.
11.1.1 Background
Although only two allotropic crystalline forms of elemental carbon, diamond and
graphite occur in nature, carbon also occurs as a spectrum of imperfect crystalline
forms that range from amorphous through mixed amorphous, graphite-like and
diamond-like to the perfectly crystalline allotropes. Such imperfect crystalline
structures are termed turbostatic and give rise to considerable variability in physical
and mechanical properties (Figure 11.1). Indeed, this ability of carbon to assume
either perfectly crystalline or chaotic, turbostatic structures gives rise to confusion
when considering physical and mechanical properties. For this reason, it is best to
consider carbon as a spectrum of materials and to bear in mind that within this spec-
trum, a number of unique combinations of structure and physical and mechanical
properties occur. This is true with respect to biocompatibility: the fact that one type
of pyrolytic carbon has been used successfully in heart valves does not necessarily
imply that other forms of pyrolytic carbons or indeed other forms of carbon in gen-
eral will also prove useful in this or other prosthetic applications. For example,
pyrolytic and glassy carbon can be finished to have identical appearances, yet the
properties of glassy carbon make it unsuitable for use in prosthetic heart valves
(Haubold et al., 1981).
For successful use in implantable prostheses a material (1) must retain its proper-
ties required for device function in the hostile biological environment and (2) must
not provoke adverse effects either locally or systemically. Most of the pure carbons
are relatively inert and unlikely to provoke severe tissue reactions, however, only
certain pyrolytic carbons have sufficient strength, fatigue resistance, wear resistance
11 Carbons
Figure 11.1 Possible atomic arrangements of crystalline carbon: (a) diamond tetrahydral, cyclohexane chair, crystalline structure, (b) graphite planar hexagonal
layered structure (c) three-dimensional quasicrystalline turbostatic structure. Biomedical carbons have a turbostatic structure.
551
552 A.D. Haubold et al.
11.1.2 Diamond
Diamond, the hardest substance known, has the so called diamond cubic structure
consisting of a network of regular tetrahedral arrays in which each carbon atom is
covalently bonded to four other carbon atoms forming the corners of a regular tetra-
hedron (Figure 11.1). From X-ray diffraction data, there is a single value, 1.54 Å
bond length and a unit cube lattice spacing of 3.56 Å. The entire crystalline array is
a single covalently bonded molecule. Because many covalent bonds must be broken
to break the crystal, a very large amount of energy is required, therefore, the sub-
stance is very hard (Pauling, 1964). There has been an ongoing interest in the use of
diamond or diamond-like coatings (May, 1995). However, suitable manufacturing
processes do not yet exist that allow economical preparation of diamond type mate-
rials in the quantity, quality, shapes and sizes required for durable long term bio-
medical applications. A specific application uniquely requiring diamond or
diamond-like material for clinical success has not been identified as a justification
to compel additional research efforts in preparation techniques.
11.1.3 Graphites
Graphite has a layered hexagonal crystal structure. Each atom forms two single
bonds and one double bond with its three nearest in-plane neighbors to form sheet-
like layers of flat six atom ring arrays (Figure 11.1). Interatomic bond distances are
1.42 Å within the layer and 3.4 Å between each layer. Within each layer, bonding is
covalent and between the layers, bonding is of the much weaker van der Waals type.
Consequently, the layers are easily separated giving rise to the soft, lubricating
properties of graphite.
Naturally occurring graphite is generally found as a isolated small scales or
imperfect single crystal precipitates in granite. These small crystals such as those
11 Carbons 553
Pyrolytic carbons of the type developed at General Atomic for use in bioengineer-
ing were an off-spring of research directed at developing carbon materials that
would be suitable for structural applications in the severe environment of high-
temperature, gas-cooled nuclear reactors. The isotropic carbon forms called high-
temperature-isotropic carbon were derived from the gas phase nucleation and
condensation of droplets formed during the pyrolysis of methane at temperatures in
excess of 2000°C where densification can occur by thermally activated processes.
The carbons called low-temperature-pyrolytic carbons are formed by the pyrolysis
554 A.D. Haubold et al.
The preparation, structure and properties of glassy, or polymeric carbons has been
described in detail by Jenkins and Kawamura (1976). These carbons are derived
from a polymer by a slow pyrolysis process which results in a vitreous residue free
of macroscopic bubbles.
Fabrication of glassy carbon materials is a relatively straightforward, but time
consuming process. A preformed polymeric precursor such as phenol-formaldehyde,
polyfurfuryl alcohol, polyvinyl alcohol or oxidized polystyrene is slowly heated in
an inert atmosphere to a high temperature in excess of 2000 °C. Heating times may
be as short as a day or as long as one month. It is not unusual to encounter exother-
mic temperature regions that must be traversed very slowly (i.e., 1 °c temperature
increase per hour) to avoid the nucleation of bubbles.
There is a volumetric shrinkage of about 50% so the resultant structure formed
in this process is a miniature of the precursor preform. The gases generated within
the preformed structure must have time to diffuse out and not nucleate bubbles so
one dimension of glassy carbon structures is limited to about seven millimeters.
Hence materials or objects are limited to thin flat plates or tubes with thin walls.
Massive equiaxed structures are not possible unless they are small with dimensions
compatible with the diffusional requirements.
Carbon fibers, thought by many to be a relatively new material, actually have a long
history as evidenced by the issuance of the first patent for incandescent electric
lamp filaments (carbon fibers). The patent was issued to Thomas Edison in 1892.
Hiram Maxim (the inventor of the machine gun, among other things) was issued a
process patent for carbon fibers in 1899. Prior to the 1950s, these fibers had mar-
ginal strength and were used primarily for their electrical properties.
High strength carbon fibers were developed in the 1950s for the aerospace indus-
try and military aircraft. The mechanical properties of rayonbased carbon fibers
were enhanced using stress graphitization. Since that time, a variety of other precur-
sors have been used including polyacrylonitrile (PAN), specific fractions of asphalt
or pitch, lignin, lignosulfonates, hetero and nonheterocyclic aromatic polymers,
11 Carbons 555
linear polymers and even coal. The processes for fiber manufacture are as varied as
the precursors themselves. In the patent literature, hundreds of processes and vari-
ants can be found (Sittig, 1980). Nevertheless, generalizations can be made. The
first step in the process is the selection and treatment of a suitable raw material
which can be carbonized to a high yield. The second step generally consists of a low
temperature (250–500 °C heat treatment or preoxidation followed by high tempera-
ture (up to 2800 °C) carbonization and graphitization steps.
The resultant fibers generally are of three types classified according to their
structure and the degree of crystallite orientation. There are the high modulus (50
million psi or above), high strength fibers which, when incorporated into structures,
give the highest stiffness per unit weight. Fibers with a lower modulus (about 30
million psi) but still of high strength are the second class of generally available
fibers. The lowest modulus (less than 20 million psi) do not have structural
applications.
Coatings formed at reduced pressure (<1 atm) are generally termed ‘vapor depos-
ited’. These coatings may be formed by physical vapor deposition, chemical vapor
deposition or combinations thereof. Physical vapor deposition such as evaporation
is probably the oldest technique for depositing thin films and involves generating a
vapor from a source material at reduced pressure. The vapor subsequently con-
denses on the object to be coated. This technique suffers from a number of limita-
tions such as only line-of-sight coating is possible. In the case of a carbon source,
which does not evaporate but rather only sublimes at extreme temperatures, the
object to be coated is exposed to direct thermal radiation from the subliming source.
Few materials can withstand the intense thermal radiation for any great length of
time. Hence coating thickness is limited as is the choice of substrate material to be
coated. Much has been written on physical vapor deposition such as the comprehen-
sive text by Maissel and Glang (1970).
More versatile coating techniques are broadly termed chemical vapor deposition
(CVD). Coatings are formed through chemical reactions in the vapor phase or
through the thermal decomposition or reduction of gases generally at reduced pres-
sures. It is interesting to note that the process used to form pyrolytic carbon is a
CVD process but is generally carried out at ambient pressures. The low temperature
CVD processes are usually assisted by means of catalysts, glow discharge plasmas,
ion beams and the like to activate gas reactions. In fact, production of diamond films
is now almost routine through ion beam assisted disassociation of selected hydro-
carbons in the presence of excess hydrogen. On the other hand, amorphous carbon
films with little or no detectable crystallinity can also be produced by CVD. Thus
chemical vapor deposition techniques are extremely versatile and consequently
films and coatings produced by CVD must be carefully characterized and identified.
An extensive and comprehensive review of thin film deposition technologies can be
found in Bunshah (1982).
556 A.D. Haubold et al.
11.1.8 Composites
Even more complex than the materials described above are a family broadly termed
‘composites’. Three-dimensional structures can be formed by combining a filler
material with an appropriate matrix. Herein lies the difficulty with composites; all
too often, the starting materials are not adequately described and the resultant struc-
ture characterized.
Carbon-carbon composites can be produced with a multitude of structures. The
simplest have two-dimensional order and consist of stacked plies of carbon fabric
held together by a carbon matrix. The fabric fibers may be any of those described
previously, prepared from the pyrolysis of polyacrylonitrile and the like. The matrix
could be derived from petroleum pitch or be infiltrated pyrolytic carbon or even sili-
con carbide. The latter are generally referred to as SiC/C composites. From two-
dimensional, the next progression in structure is three-dimensional on to
n-dimensional. This terminology refers to fiber orientation within the matrix.
Filament wound carbon composites have also been developed. In this case the
desired shape is, as the name implies, wound using carbon fibers onto a suitable
mandrel. The fibers are bonded using an epoxy type or thermoset resin. The bonding
of the matrix to the fibers and direction of fiber orientation in large part determine
the mechanical properties of the composite. The biological properties are generally
governed by the selection of matrix.
Encouraged by the success of the LTI form of pyrolytic carbon in the demanding
mechanical heart valve application, other carbon materials and usages were
explored. Different carbon materials have been evaluated because pyrolytic carbon
was patented on the one hand and processing constraints limited its versatility on
the other hand. Shown in Table 11.2 are examples that demonstrate medical and
engineering ingenuity in attempts to expand the use of carbons in the biological
environment. Although the attempts are numerous and varied, only the use of car-
bon as components for artificial heart valves has achieved widespread usage.
11.2.1 Dental
In another attempt, artificial tooth roots in the form of blades were fabricated
from pyrolytic carbon. Although they were less bulky than the glassy carbon
implants, they were difficult to place. Improper seating of the blade caused micro-
motion after implantation that ultimately caused the prosthesis to fail. The success
rate of 60% after 5 years was judged inadequate. Metal blades coated with carbon
fared a similar fate.
Metallic mandibular reconstruction trays coated with a vapor deposited film of
carbon generally performed well. The application was complicated by the fact that
the trays were custom and many times fashioned directly in the operating room,
making the logistics of coating with carbon unacceptable.
11.2.2 Vascular
Graft prostheses >6mm diameter are generally considered to work well and improve-
ments in performance as a result of modifying the biochemical nature of the graft
lumen with carbon coatings are difficult to quantify and not significant enough to
justify the cost of carbon coating. Improvements in the patency of small diameter
grafts (<4 mm) as a result of carbon coating have been reported but even with the
improvement, these grafts ultimately failed as a result of intimal hyperplasia prolif-
eration at the anastomoses. While the carbon coating may retard clotting of certain
grafts, the coating does little to ameliorate intimal hyperplasia formation. Carbon
vascular attachment prostheses have also been reported to perform poorly, not as a
result of poor biocompatibility, but rather the result of mechanical complications.
11.2.3 Orthopedics
Femoral stems fabricated from carbon composites fitted with a femoral head
fabricated from pyrolytic carbon have been reportedly used successfully by some
for over 10 years (Chen, 1986). Others have experienced disastrous failures through
a lack of attention to engineering and material property details. Such failures natu-
rally lead to questions on the suitability of ‘carbon’ for use in such medical devices.
Attempts are underway to design and fabricate other joint replacements.
Ligaments and tendons have been fabricated from carbon fibers (Béjui and
Drouin, 1988). These fibers in the initial stage perform well as a scaffolding material
that aids in the regeneration of tendons and ligaments in vivo; but in the long term
the fibers fracture and migrate to, for example, the lymph nodes.
Fracture fixation devices that are fabricated from stainless steel are unsuitable for
coating or coupling with carbon because of galvanic effects (Haubold et al., 1986).
Carbon composite devices have been used reportedly with good results but such
usage has not become widespread presumably because of an unaccepted cost/benefit
ratio.
11 Carbons 559
11.2.4 Other
Many applications of devices listed in Table 11.2, while successful, are not in
widespread use because, even in their non-carbon form, usage is limited. For example,
left ventricular apex tubes (Haubold et al., 1979) are used successfully in the
construction of a prosthesis to correct idiopathic hypertrophic subaortic stenosis but
fortunately in man, such a medical condition is rare.
LTI pyrolytic carbons, since their introduction in the late 1960s, have become the
material of choice for use in the fabrication of mechanical prosthetic heart valves.
Over 90% of the mechanical valves implanted worldwide utilize such carbon com-
ponents. To date, more than 2 million valves have been implanted which amounts to
an accumulated experience in excess of 12 million patient years. While this material
has proven to be the most successful carbon biomaterial, it can be improved.
Recently, advances in process control methodologies have allowed refinements in
the pyrolytic carbon coating preparation. These improvements allow the elimination
of the potentially thrombogenic silicon carbide from the biomedical coatings.
Furthermore, this pure pyrolytic carbon can be produced with substantially improved
mechanical properties relative to the silicon alloyed material (Emken et al., 1993;
Ely et al., 1994,). Such improvements in the material open possibilities for improve-
ments in heart valve prosthesis design and performance.
Results from investigations on the suitability of other forms of carbon for in vivo
use yielded mixed and often seemingly contradictory results. Some of the confusion
developed because of misunderstanding of carbon structure and misunderstanding of
the relationship of carbon properties to such structures. The use of the generic label
‘carbon’ compounded the problem. A similar situation exists with ‘polyurethanes’.
There are polyether urethanes, polyester urethanes, polyether urethane ureas and even
polyester urethane ureas – all misappropriately called simply ‘urethanes’. Thus it is
not surprising that the biological responses of ‘carbon’ (Table 11.1) are so varied.
Biocompatibility claims for a particular form of ‘carbon’ based on published
results for a totally different form or structure should be carefully scrutinized. For
example, to claim that carbon fibers have the same biological properties as bulk
pyrolytic carbons or even that all pyrolytic carbons behave similarly is unjustified.
In the case of fibers, geometry plays a significant role. It is well known that bulk
materials may be well tolerated when the same material in particulate form may not.
The lack of characterization and standardization can be devastating.
More and more entrants are anticipated into the field of carbon biomaterials. In the
past, because of technology, patent or cost constraints, there were only several sources
for, for example, pyrolytic carbon. A number of the earlier constraints have now been
removed. Consequently, these materials are being produced in limited but increasing
560 A.D. Haubold et al.
References
Bejui, J. and Drouin, G. (1988). Carbon Fiber Ligaments. In CRC Critical Reviews in
Biocompatibility, 4(2), 79–108.
Benson, J. (1969). Pre-Survey on the Biomedical Applications of Carbon, North American
Rockwell Corporation Report R-7855.
Bokros, J.C. (1969). Deposition, Structure, and Properties of Pyrolytic Carbon. In Chemistry and
Physics of Carbon, Vol. 5 (Walker, P.L., ed.). Dekker, New York, pp. 1–118.
Bokros, J.C. LaGrange, L.D. and Schoen, F.J. (1972). Control of Structure of Carbon For Use in
Bioengineering. In Chemistry and Physics of Carbon, Vol. 9 (Walker, P.L., ed.). Dekker,
New York, 103–171.
Bokros, J.C., Akins, R.J., Shim, H.S., Haubold, A.D. and Agarwal, N.K. (1975). Carbon in
Prosthetic Devices. In Petroleum Derived Carbons (Deviney, M.L., and O’Grady, T.M., eds).
American Chemical Society, Washington, DC, pp. 237–265.
Bunshah, R.F. (ed.) (1982). Deposition Technologies for Films and Coatings., Noyes Publications,
Park Ridge.
Chen, Lan-Tian (1986). Carbon-Titanium Combined Joints. In Chinese Journal of Biomedical
Engineering 3, 55–61.
Ely, J., Haubold, A., Bokros, J. and Emken, M. (1994), New Unalloyed Pyrolytic Carbon with
Improved Properties for Implant Applications, XXI Congress European Society for Artificial
Organs, Oct. 20–22, Barcelona Spain. Also US Patent 5 514 410.
Emken, M., Bokros, J., Accuntis, J. and Wilde, D., (1993) Precise Control of Pyrolytic Carbon
coating, Extended Abstracts & Program Proceedings of the 21st Biennial Conference on
Carbon, Buffalo New York, June 13–18, pp. 531–532. Also US Patent 5 284 676.
Haubold, A.D., Shim, H.S., and Bokros, J.C. (1979). Carbon Cardiovascular Devices. In Assisted
Circulation (Unger, F., ed.) Springer Verlag, Berlin, Heidelberg, New York, pp. 520–532.
Haubold, A.D. (1977). Carbon in Prosthetics. In Annals of the New York Academy of Sciences, Vol.
283, The Behavior of blood and its Components at Interfaces, (Vroman, L. and Leonard E.F.,
eds). New York Academy of Sciences, New York.
Haubold, A.D., Shim, H.S., and Bokros, J.C. (1981). Carbon in Medical Devices. In Biocompatibility
of Clinical Implant Materials, Vol II (Williams, D.F., ed.). CRC Press, Boca Raton, pp. 3–42.
Haubold, A.D., Yapp, R.A., and Bokros, J.C. (1986). Carbons for Biomedical Applications. in
Encyclopedia of Materials Science and Engineering (Bever, M.B., ed.) Pergamon Press,
Oxford, New York, Toronto, Sydney, Frankfurt, pp. 514–520.
Jenkins, G.M. and Kawamura, K. (1976). Polymeric Carbons - Carbon Fibre, Glass and Char.
Cambridge University Press, Cambridge. London, New York, Melbourne.
Lewis, J.C. and Redfern, B., and Cowland, F.B. (1963). Vitreous Carbons as Crucible Materials for
Semiconductors. In Solid State Electronics, 6, 251.
Maissel, L.I. and Giang, R. (1970). Handbook of Thin Film Technology, McGrawHill, New York.
May, P.W. (1995), CVD Diamond - a new Technology for the Future?, Endeavor Magazine 19(3),
101–106.
Pauling, L. (1964), College Chemistry, W.H. Freeman and Co., San Francisco.
Pierson, H.O. (1993) Handbook of Carbon, Graphite, Diamond and Fullerenes, Noyes
Publications, Park Ridge, New Jersey.
Sittig, M. (1980). Carbon and Graphite Fibers. Noyes Data Corporation, Park Ridge. Mechanical
Behavior of Diamond and Other Forms of Carbon, Materials Research Society Symposium
Proceedings, Vol. 383, ed. Dory M.D. et al., Materials Research Society, Pittsburgh,
Pennsylvania, 1995.
Part III
Chapter 1
General Concepts of Biocompatibility
D.F. Williams
1.1 Introduction
The host responses to biomaterials are extremely varied, involve a range of different
mechanisms and are controlled by factors that involve characteristics of host, mate-
rial and surgical procedure. These responses themselves constitute a significant
component of the phenomenon of biocompatibility. In this section, the broad con-
cepts of biocompatibility are critically reviewed with particular reference to the role
that the human host response plays in determining the performance of the biomate-
rial and of the device in which it is used. Particular emphasis is given to the influ-
ence of biocompatibility in the clinical applications of devices. It should be
remembered, however, that biocompatibility phenomena are extremely difficult to
interrogate remotely or to study in an active way, so that accurate information of the
details of biomaterial–human tissue interactions is not readily available. As Black
(1) has pointed out with reference to observations on the host response in general,
we are usually limited to detecting events long after they have occurred by examin-
ing end-points, usually histopathologically, after the host is dead. This is largely the
case with experiments on biocompatibility in animals, but is an even more relevant
observation with the human clinical experience. All comments in this section must
therefore be interpreted with this in mind.
that inertness is a very relative term and there is indeed no such thing as an inert
biomaterial.
The third reason why biocompatibility cannot be equated with inertness is that
there are several, and indeed an increasing number, of applications which involve
intentionally degradable materials. The two most widely quoted situations here are
absorbable sutures and implantable drug delivery systems but many more circum-
stances where degradable scaffolds and matrices could form an essential component
of a device are envisaged. If biocompatibility is predicated on inertness, then
degradable materials cannot, by definition, be biocompatible. This clearly does not
make sense and suggests that the concept of biocompatibility needs to be altered.
The fourth reason is even more compelling, especially when considering bioma-
terials used in devices for tissue reconstruction. If a device is made from materials
which are inert and which do not interact with the body in any way, then it is unlikely
that it can be truly incorporated into the body. For effective long term performance
in the dynamic tissue environment, it is far more preferable for there to be functional
incorporation, which implies that the device should be stimulating the tissues to be
reactive to it positively rather than negatively. Thus biocompatibility should not be
concerned with avoiding reactions but selecting those which are the most beneficial
to device performance.
On the basis of these ideas, biocompatibility was redefined a few years ago (2),
as ‘the ability to perform with an appropriate host response in a specific situation.’
Clearly this definition encompasses the situation where inertness is still required for
the most appropriate response in some situations is indeed no response. A tradi-
tional bone fracture plate is most effective when it is attached mechanically to the
bone and does not corrode; no response of the tissue to the material is normally
required. Even here, however, we have to concede that a material that could actively
encourage more rapid bone healing might be beneficial so that a specific osteoin-
ductive response would be considered appropriate.
More importantly, the definition allows a material to stimulate or otherwise
favour a specific response, including cell activation, where that response optimizes
the performance of the device. It will be obvious that the required response will vary
with the particular application, which clearly implies that the response, both desired
and actual, will vary with the different types of tissue encountered by biomaterials.
The above definition also stipulates that the biocompatibility of a material has to
be qualified by reference to the specific application. The response to some very
common and popular biomaterials may vary quite considerably and some of the
major problems of implantable medical devices have been caused by a misunder-
standing about transferability of biocompatibility data. To recognize the very effec-
tive performance of a material under one set of conditions but then to assume that
the same material can perform equally well under entirely different circumstances
is inherently dangerous since it takes into account neither the variations one might
expect to see in the host response from site to site nor the fact that what is appropri-
ate for one situation may not be appropriate for another.
566 D.F. Williams
The above definition of biocompatibility helps to explain the subject area but cannot
describe exactly what it is. For this purpose we have to consider the various compo-
nents that are involved in biocompatibility processes. Biocompatibility refers to the
totality of the interfacial reactions between biomaterials and tissues and to their
consequences. These reactions and consequences can be divided into four catego-
ries. These involve different mechanisms and indeed quite separate sectors of sci-
ence but are, nevertheless, inter-related.
The first component is that of the protein adsorption mentioned above. This pro-
cess is initiated as soon as a material comes into contact with tissue fluids such that
relatively quickly the surface of the biomaterial is covered with a layer of protein.
The kinetics and extent of this process will vary from material to material which
will in any case be a dynamic phenomenon with adsorption and desorption pro-
cesses continuously taking place. Under some circumstances, this layer is extremely
important in controlling the development of the host response since cell behavior
near the material may depend on interactions with these proteins. For example,
thrombogenicity is controlled by a number of events including the interaction
between plasma proteins and surfaces, these proteins being able to influence the
attachment of platelets to the surface. In other circumstances, the effects of this
protein layer are far from clear.
The second component of biocompatibility is that of material degradation. It is
emphasized here that degradation is a component of biocompatibility rather than a
separate phenomena. There is still confusion over this since it is often perceived that
degradation, which occurs on the material side of the interface, is the counterpart to
biocompatibility which is equated with the other (tissue) side. This is not correct
since degradation is the counterpart to the local host response, both being contribu-
tory to the biocompatibility of the system.
Degradation phenomena are covered elsewhere in this Handbook and will not be
discussed in detail here. It is necessary to point out, however, that descriptions of
material degradation mechanisms have to take the special, and indeed unique, fea-
tures of the tissue environment into account. Whatever its location, a biomaterial
will continuously encounter an aqueous environment during its use. This is not sim-
ply a saline solution, however, but a complex solution containing a variety of anions
and cations, a variety of large molecules some of which are very reactive chemi-
cally, and a variety of cells which again may be in passive or active states. There are
occasions when a degradation process can be explained, mechanistically and quali-
tatively, by the presence of an electrolyte. This is the situation with most metals
when they suffer from corrosion in physiological environments(3). Even here, how-
ever, it is known that the kinetics of corrosion may be influenced by the organic
species present, especially the proteins, and it is even possible for the corrosion
mechanism to be somewhat different to that found in non-biological situations.
With other groups of materials, however, and especially with polymers, kinetics,
mechanisms and consequences of the degradation are fundamentally related to the
1 General Concepts of Biocompatibility 567
extensive and may intervene between the material and tissue it is meant to be in
contact with (for example bone in the case of joint prostheses) but perhaps more
importantly it can change the immediate tissue environment from one of quiescent
fibrosis to that of active chronic inflammation. This is rarely the appropriate response
and, as noted above, is likely to generate an even more aggressive environment.
In the context of the definition of biocompatibility, therefore, it is important that
the interaction between the material and the tissues is one which leads to an accept-
able balance between inflammation and repair. A few points may serve to explain
this further and qualify appropriateness. First, the nature of the host response and
those features which constitute acceptability will vary very considerably from one
host to another and from one location (or set of circumstances) to another within a
particular host. It is often forgotten that host variables are as important as material
variables in the determination of biocompatibility. This is particularly important
when the wide variety of tissue characteristics is considered. Obviously bone is very
different from nerve tissue or a vascular endothelium and there will be very consid-
erable difference in the details of their responses. Not all tissues of the same variety
will be able to respond in the same way and it should always be remembered that
host variables such as age and overall health status will have a major effect.
Secondly, the importance of time and the sequence of events should never be
underestimated. While the above model describes the sequence from surgical inter-
vention to inflammation to repair, such that the process may undergo rapid resolution,
with a resulting long lasting equilibrium, the inflammation may be restimulated at any
time and rarely can we guarantee long term survival. In this context the third point
becomes important, for any feature of the interaction between material and tissue and
material can be responsible. In many situations it is the chemical reactivity, repre-
sented by degradation processes, which drives the inflammatory response, but it can
equally well be a process by which fragments of the material are extracted by physical
or mechanical means. The release of wear debris from orthopaedic prostheses is a
good example here, since the presence of such particles in sufficient numbers can have
a profound effect on the tissue response, which is mediated by the mechanisms of
inflammation but where the clinical results, manifest by loosening of the prostheses,
may not be seen for quite some time(8). As far as the response to debris is concerned,
in general the effects of released fragments will be quite different to those of bulk
materials, both by virtue of physical factors as well as changed chemical factors.
Thirdly, the identification of these events and their importance leads to various
possibilities for the control of biocompatibility. In the balance of inflammation and
repair we have the possibilities of controlling that balance by aiming to eliminate or
at least minimize those events which are undesirable for one set of conditions or
alternatively enhancing or optimizing those events which are most desirable. This
has led to the emergence of the concept and indeed introduction of bioactive materi-
als which have been defined as biomaterials that are designed to elicit or modulate
biological activity.
As a final point about the local host response, it has to be recognized that there
are significant regional and tissue-specific variations to the phenomena. These
1 General Concepts of Biocompatibility 569
c annot be described here, but it is important to mention the particular case with
blood. When a biomaterial comes into contact with blood, there are many different
mechanisms by which the blood can interact with the material, most of which are
preferably avoided. The most important of these are those processes, alluded to in
an earlier paragraph, that are responsible for the clotting of blood. This is a vital
defense mechanism which prevents death by uncontrolled bleeding under everyday
circumstances, but unfortunately in the context of biomaterials, the two processes
which can, either separately or together, initiate the formation of a blood clot, that
is contact phase activation of clotting proteins and platelet activation, are them-
selves initiated by contact with foreign surfaces. Thrombogenicity, defined as the
property of a material which promotes and/or induces the formation of a thrombus,
is clearly an important feature of biocompatibility.
Turning now to the last component of biocompatibility, we have to recognize that
if there is an interfacial reaction, there is no reason why the products of that reaction
and their effects have to be confined to the locality of that interface and the presence
of a benign local response is not necessarily indicative of the absence of any sys-
temic or remote site effects. The possibility of systemic effects arising from the
presence of biomaterials has long been recognized, although extreme difficulties
exist with their identification and interpretation. Indeed, at the present time, there
are few systemic effects that can be readily identified with biomaterials. The trans-
formation of a thrombus into an embolus derived from an intravascular device has
obvious implications and we can imagine and often demonstrate the systemic con-
sequences of using overtly cytotoxic materials. However, the more intriguing specu-
lations refer to the putative implant-related carcinogenicity and even more
speculative implant-related immune responses. At this stage we have to be con-
cerned about such possibilities but have to put the subject into context. While we
cannot deny that there are possible mechanisms for biomaterials to induce tumors,
the evidence that they do so in human clinical experience is very sparse. While it is
possible for some hypersensitivity responses to be seen to implants, the evidence for
any clinically ignificant response from the immune system to biomaterials is even
less available.
1.4 Conclusions
This review attempts to outline the main concept that currently prevail in the
subject area of biocompatibility. Clearly it is a complex subject, about which we
are still relatively ignorant, not least because it involves a juxtaposition between
two quite different sectors of science, the materials sciences and the molecular/
cellular biological sciences. Based on these concepts, however, a better under-
standing is now emerging so that our biomaterials can be chosen, and where
necessary treated, in order to determine that the tissues of the host do indeed
respond appropriately to them.
570 D.F. Williams
Additional Reading
References
J.M. Anderson
2.1 Introduction
Soft tissue responses to biomaterials for medical devices are generally viewed from
the inflammation and wound healing perspectives and are usually considered as
parts of the tissue or cellular host responses to injury. Placement of a biomaterial or
medical device in the soft tissue environment involves injection, insertion, or sur-
gical implantation, all of which injure the tissues or organs involved. Early host
responses are dynamic and change with time (Table 2.1). It is important to consider
this time variable in determining the host response or biocompatibility of a
material.
Four general types of response may occur following the implantation of a biomate-
rial. These are a minimal response, a chemically induced response, a physically
induced response, and cellular/tissue necrosis [1].
A minimal response is generally called fibrous encapsulation and the implant is
encapsulated within fibrous tissue containing mainly collagen with a few fibro-
blasts. At the tissue/implant interface, a one to two cell layer of macrophages and
foreign body giant cells is present which constitutes the foreign body reaction.
Fig. 2.1 The temporal variation in the acute inflammatory response, chronic inflammatory
response, granulation tissue development, and foreign body reaction to implanted biomaterials.
The intensity and time variables are dependent upon the extent of injury created in the implantation
and the size, shape, topography, and chemical and physical properties of the biomaterial.
esis is generally linked to the extent or degree of fibrous capsule formation and the
potential for solid state tumorigenesis is reduced with increasing foreign body
reaction.
2.3 Inflammation
Table 2.2 Cells and Intravascular (blood) cells Blood plasma proteins
Components of Vascularized
Erythrocytes (RBC) Coagulative Proteins
Connective Tissue
Neutrophils Complement Proteins
Monocytes Albumin
Eosinophils Fibrinogen
Lymphocytes Gamma-Globulins
Basophils Others
Platelets Extracellular matrix
components
Connective tissue cells Collagens
Mast Cells Elastin
Fibroblasts Proteoglycans
Macrophages Fibronectin
Lymphocytes Laminin
and the inflammatory response are intimately linked and, in fact, early responses to
injury involve mainly blood and blood vessels. Therefore, both cellular and humoral
elements, i.e., plasma proteins, etc., are considered as cells and components of vas-
cularized connective tissue.
The wound healing response is initiated by the action of monocytes and macro-
phages, followed by proliferation of fibroblasts and vascular endothelial cells, i.e.,
capillaries, at the implant site. The proliferation of fibroblasts and the formation of
capillaries constitute granulation tissue. Modified fibroblasts, i.e., myofibroblasts,
which have contractile properties which assist in wound site closure are transiently
present in granulation tissue. As fibroblasts predominate over macrophages in the
healing response, collagens and other extracellular matrix molecules are deposited
in the implant site. The extent of the wound healing response is generally dependent
on the extent or degree of injury or defect created by the implantation procedure.
Wound healing progresses by primary union (or first intention) if the healing is
clean such as a surgical incision in which the wound edges have been approximated
by surgical sutures, clips, or staples. Healing under these conditions occurs with a
minimal loss of tissue. Wound healing by secondary union (or secondary intention)
occurs when there is a large tissue defect that must be filled or there has been an
extensive loss of cells and tissue. In wound healing by second intention, regenera-
tion of specific organ or tissue cells cannot completely reconstitute the original
architecture and more granulation tissue is formed resulting in larger areas of fibro-
sis or scar formation. Thus, the surgical procedure to create the implant site may
influence the extent or degree of the wound healing response.
576 J.M. Anderson
Repair of implant sites involves two distinct processes: regeneration, which is the
replacement of injured tissue by parenchymal cells of the same type, or replacement
by fibrous connective tissue that forms a capsule [7]. These processes are generally
controlled by either (i) the proliferative capacity of the cells in the tissue or organ
receiving the implant and the extent of injury as it relates to tissue destruction or (ii)
persistence of the tissue framework of the implant site. The regenerative capacity of
cells permits their classification into three groups: labile, stable (or expanding), and
permanent (or static) cells. Labile cells continue to proliferate throughout life, sta-
ble cells retain this capacity but do not normally replicate, and permanent cells can-
not reproduce themselves after birth of the host.
Perfect repair, with restitution of normal structure, theoretically occurs only in
tissues consisting of stable and labile cells, whereas all injuries to soft tissues com-
posed of permanent cells may give rise to fibrosis and fibrous capsule formation
with very little restitution of the normal tissue or organ structure. Tissues composed
of permanent cells (e.g., nerve cells, skeletal muscle cells, and cardiac muscle cells)
most commonly undergo an organization of the inflammatory exudate, leading to
fibrosis. Tissues composed of stable cells (e.g., parenchymal cells of the liver, kid-
ney, and pancreas), mesenchymal cells (e.g., fibroblasts, smooth muscle cells,
osteoblasts, and chondroblasts), and vascular endothelial and labile cells (e.g., epi-
thelial cells and lymphoid and hematopoietic cells) may also follow this pathway to
fibrosis or may undergo resolution of the inflammatory exudate, leading to restitu-
tion of the normal tissue structure. The condition of the underlying framework or
supporting stroma of the parenchymal cells following an injury plays an important
role in the restoration of normal tissue structure. Retention of the framework may
lead to restitution of the normal tissue structure, whereas destruction of the frame-
work most commonly leads to fibrosis. It is important to consider the species-
dependent nature of the regenerative capacity of cells. For example, cells from the
same organ or tissue but from different species may exhibit different regenerative
capacities and/or connective tissue repair. An example of species differences in cell
proliferation and regeneration is the endothelialization process, proliferation of
endothelial cells, on the luminal surface of vascular grafts which does not occur in
humans but does occur in other mammals including other primates.
2 Soft Tissue Response 577
2.6 Summary
Inflammation, wound healing, foreign body response, and repair of implant sites are
usually considered components of the general soft tissue response to biomaterials or
medical devices. The extent or degree and temporal variations in these responses are
dictated by the inherent biocompatibility characteristics of the biomaterial or medi-
cal device. Factors which may play a role in the soft tissue response include the size,
shape, topography, and chemical and physical properties of the biomaterial. As the
implantation procedure involves injury to vascularized connective tissue, blood
responses and interactions may play a role in the general soft tissue response. The
extent, degree or type of soft tissue response is generally considered to be tissue-
specific, organ-specific, and species-specific. Thus, a given biomaterial may be con-
sidered to be biocompatible in one shape or form but not in another and in one tissue
but not in another depending on the given application.
Additional Reading
biological effects of implants, and methods of test for biological performance. The
fourth part, Methods of test for biological performance, is unique to biomaterials
texts and provides the reader with in vitro and in vivo test models and methods as
well as perspectives on the design, selection, standardization, and regulation of
implant materials.
Cohen, I.K., Diegelmann, R.F. and Lindblad, W.J. (eds) (1992) Wound Healing:
Biochemical and Clinical Aspects, W.B. Saunders Co., Philadelphia.
This is an edited volume containing 35 chapters. The volume addresses the fol-
lowing areas: Biological processes involved in wound healing (6 chapters), struc-
tural and regulatory components of wound healing (7 chapters), factors affecting
tissue repair (7 chapters), repair of specific tissues (7 chapters), and clinical man-
agement of healing tissues (7 chapters). This is an excellent volume which provides
an up-to-date and in-depth perspective of various aspects of wound healing. The
references lists provided at the end of each chapter are extensive. The strength of the
volume is its biological perspective and little is provided on biomaterials. The chap-
ter by Frederick Grinnell on cell adhesion does offer a biomaterial perspective.
Gallin, J.A., Goldstein, I.M. and Snyderman, R. (eds) (1992) Inflammation:
Basic Principles and Clinical Correlates, 2nd ed, Raven Press, New York.
This is an edited volume containing 58 chapters by individual authors. The
volume is divided in the following areas: Soluble components of inflammation (10
chapters), cytokines (5 chapters), cellular components of inflammation (21 chap-
ters), responses to inflammation (3 chapters), clinical correlates (13 chapters), and
pharmacologic modulation of inflammation (4 chapters). Each chapter is a critical,
in-depth review of the indicated subject and the references are extensive. This is
an excellent volume for those wanting an in-depth overview of the inflammatory
process and its components. No information is provided on biomaterial/inflamma-
tion interactions.
Greco, R.S. (ed.) (1994) Implantation Biology: The Host Response and
Biomedical Devices, CRC Press, Boca Raton.
This is an edited volume containing 23 chapters. Three chapters deal with bioma-
terials in general, 6 chapters address specific blood and tissue interactions with bio-
materials, 10 chapters address the use of biomaterials in specific surgical disciplines,
and 3 chapters address tissue engineering and genetic manipulation of cells. The
reference list for each chapter is extensive. This is an excellent overview of how
biomaterials interact with the host and the specific use of biomaterials in indicated
applications.
Harker, L.A., Ratner, B.D. and Didisheim, P. (eds) (1993) Cardiovascular
Biomaterials and Biocompatibility: A Guide to the Study of Blood-Tissue-Material
Interactions, Cardiovascular Pathology, 2 (3 Suppl), 1S–224S.
This is the third edition of a standard National Institutes of Health reference
previously entitled Guidelines for Blood-Material Interactions - Report of the
National Heart, Lung, and blood Institute Working Group. The volume contains 20
chapters and 3 appendices. The chapters address the following areas:
Pathophysiologic mechanisms, materials and their physicochemical characteriza-
tion, safety testing of materials and devices, and blood-vessel-material interactions.
2 Soft Tissue Response 579
References
1. Williams, D.F. and Roaf, R (1973) Implants in Surgery, W.B. Saunders Company Ltd., London,
pp. 233–35.
2. Anderson, J.M. (1993) Mechanisms of inflammation and infection with implanted devices.
Cardiovascular Pathology, 2 (3 Suppl.), M319-M321.
3. Anderson, J.M. (1994) In vivo biocompatibility of implantable delivery systems and biomate-
rials. European Journal of Biopharmaceutics, 40, 1–8.
4. Anderson, J.M. (1994) Inflammation and the foreign body response. Problems in General
Surgery, 11(2), 147–160.
5. Black, J. (1992) The inflammatory process, in Biological Performance of Materials
-Fundamentals of Biocompatibility, 2nd edn, Marcel Dekker, Inc.,New York, pp. 125–147.
6. Marchant, R.E., Anderson, J.M. and Dillingham, E.O. (1986) In vivo biocompatibility studies.
VII. Inflammatory response to polyethylene and to a cytotoxic polyvinylchloride. Journal of
Biomedical Materials Research, 20, 37–50.
7. Cotran, R.Z. et al. (1994) Inflammation and repair, in Pathologic Basis of Disease, 5th edn,
Cotran, R.Z., Kumar, V. and Robbins, S.L. (eds), W.B. Saunders Co., Philadelphia, pp. 51–92.
8. Gallin, J.I., Goldstein, I.M. and Snyderman, R. (eds) (1992) Inflammation. Basic Principles
and Clinical Correlates, 2nd edn, Raven Press, New York.
9. Spector, M., Cease, C. and Tong-Li, X. (1989) The local tissue response to biomaterials. CRC
Critical Reviews in Biocompatibility, 5 (4), 269–295.
10. Weissman, G., Smolen, J.E. and Korchak, H.M. (1980) Release of inflammatory mediators
from stimulated neutrophils. New England Journal of Medicine, 303, 27–34.
11. Henson, P.M. (1980) Mechanisms of exocytosis in phagocytic inflammatory cells. American
Journal of Pathology, 101, 494–511.
12. Henson, P.M. (1971) The immunologic release of constituents from neutrophil leukocytes:
II. Mechanisms of release during phagocytosis, and adherence to nonphagocytosable surfaces.
Journal of Immunology, 107, 1547–57.
Chapter 3
Hard Tissue Response
T.O. Albrektsson
3.1 Introduction
The initial tissue response when a biomaterial is implanted in the body is dependent
on release of specific growth factors. It has been indicated by Frost [1] that the
inevitable bone injury resulting from surgery and the presence of an implant will
release various types of growth factors that will sensitize cells and promote cellular
mitosis. This is a general healing response that will result in growth of all sorts of
local connective tissues, bone as well as various types of soft tissue.
The balance between these tissue varieties is controlled by the action of chemical
mediators which issue ‘instructions’ for the amount of bone and soft tissue to be
formed as an appropriate healing response. This delicate balance can easily be dis-
turbed inadvertently and may cause the undesirable end-result of an interfacial soft
tissue embedment of the implant or, in the case of fracture healing, formation of a
pseudoarthrosis. The discussion in this section will focus on various modes of
implant fixation, such as cementation, ingrowth and osseointegration (Figure 3.1).
Bone cement, a two component acrylic, is frequently used for implant fixation in the
cases of hip and knee arthroplasties. Bone cement is toxic with localized as well as
general adverse tissue reactions [2]. Therefore, the good long-term results reported
with cemented arthroplasties seem to be quite puzzling. However, it must be
Fig. 3.1 Different classification criteria are used for implants in bone. Here we distinguish between cemented implants, which are dependent upon cure in place
luting or filling agents; cement-free (ingrowth implants) which are dependent on bone tissue growth into surface features for fixation; and osseointegrated
implants, which are dependent upon collagen and cellular processes and their physiochemical interactions with the implant surface for fixation. In each case, a
T.O. Albrektsson
‘ridge’ feature on the implant surface is shown in cross-section; the width of each band of the diagram is several millimeters (cells (osteocytes) are shown larger
than actual size).
3 Hard Tissue Response 583
understood that the strength of the cemented bone interface is not related to the live
state of bone tissue: in reality, the curing bone cement invading the trabecular net-
work results in a substantial cancellous bone interdigitation (Figure 3.1). The inter-
facial bone usually dies from the combined trauma of heat and monomer leakage.
The width of this necrotic zone varies depending on the extent of the surgical trauma
and the type of bone cement used, but is at least a few millimeters [2]. Ironically, a
revitalization of the interfacial bone may prove disastrous, as the bone will then run
an increased risk of resorption. In examining interfacial tissues reactions to retrieved
cemented orthopaedic implants, most studies have examined the tissue after removal
of the implant. This leads to an uncertainty in examining the proportion of true bone
to implant contact.
We have been able to investigate bone implants in situ, at the resolution level of
light microscopy, utilizing the techniques developed by Donath et al. [3]. Computer
assisted calculations, using custom developed software, permit determination of the
proportion of bone to implant contact, the amount of bone in the region adjacent to
the implant and comparisons of the bone density adjacent to and in the immediate
region of the implant (‘outfolded mirror image’) (Figure 3.2).
These data correlate positively to biomechanical tests: the more bone in the inter-
facial region of the implant, the greater the torque necessary to remove the implant
when we apply controlled rotational forces to the implant body. Specimens of 10μm
thickness have been investigated at the light microscopic level of resolution [4].
Other studies using cutting and grinding techniques have investigated the implant in
Fig. 3.2 Computerized
histomorphometric
approach to evaluation of
the bone-implant interface.
The bone-to-implant
contact percentage is the
linear contact area between
bone and implant in the
inside zone; percent bone
ingrowth is the ratio of
bone ‘inside’ to that
‘outside’ in the ‘mirror’
zone.
584 T.O. Albrektsson
In the absence of cement, fixation may be obtained by active bone tissue ingrowth
into the implant surface irregularities of medullary stems (Figure 3.2). As pointed
out by Black [5], one disadvantage of cement-free implants is the 3–4 week waiting
time before the bone-implant interface can support significant shear loading; this is
in contrast to a cemented interface which, if successful, has adequate shear strength
within one hour of implantation. Osseointegrated implants have even lower interfa-
cial shear strength at the the same 3–4 week post operative time point [6].
In the clinical experience, however, the results of cement-free joint arthroplasties
utilizing ingrowth fixation have not matched those of cemented devices. The appro-
priate bone ingrowth is often disturbed and incomplete, leading to early failure of
fixation. Cement-free knee arthroplasty components have been observed to migrate
by up to 2 mm in the first post-operative year, while cemented devices of the same
design migrate only half as much or less [7]. The failure rates of cement-free hip
arthroplasties have been so substantial in comparison to cemented devices that the
former mode of implant insertion has been restricted to young patients for whom
cemented devices give poorer results at most clinics [8].
The outcome of cement-free orthopaedic implants depends, as does the outcome
of craniofacial osseointegrated implants, on the more or less simultaneous control
of a number of different factors including biocompatibility, design and surface con-
ditions of the implant, the state of the host bed, the surgical technique and the load-
ing conditions after implant insertion.
Herein is the explanation for the dubious results of many current designs of
cement-free hips and knees: the implant material used, the design and surface finish
have been well adapted to engineering demands, but not so well matched to the
biological needs.
Ideal implant characteristics with respect to bone anchorage are quite different
from those of gliding surfaces. The surgical technique utilized when inserting a
cement-free hip or knee of current designs is, of necessity, traumatizing: resulting in
an impaired bone healing response. Current stem-type cement-free hips migrate
when loaded, as do tibial components of artificial knees, leading to a change in the
chemical mediators and subsequent increase of soft tissue formation. It is, therefore,
not surprising to learn from studies of retrieved cement-free hips that there is, as a
rule, only sparse bony ingrowth into retrieved acetabular cups [9], femoral cups [2,
10] or femoral stems [9, 11, 12]. More modern types of cement free hip prostheses
with titanium meshwork or HA-coated surfaces have had slightly better clinical
3 Hard Tissue Response 585
results than those implanted and retrieved during the 1980s and it may well be pos-
sible that there is more abundant bone ingrowth in some of the more recent designs.
Nevertheless, from multicenter studies it seems quite clear that summed five-year
failure rates for cement-free arthroplasties are greater than those for cemented
devices [8]. This observation illustrates that the anchorage problems associated with
cement free arthroplasties are far from solved.
3.4 Osseointegration
bone-cement interface consisting of soft tissue or mostly dead bone have demon-
strated significant clinical longevity, in many cases exceeding ten to fifteen years.
Furthermore, the so called osseointegrated interface is still in need of a proper
definition. First described by Brånemark [17] as a bone response that occurred
everywhere around the implant circumference of c.p. titanium screws in placed in
bone, osseointegration is regarded today as a more nonspecific tissue response
resulting in a mix of interfacial soft and hard tissues. In reality, bone anchorage of
foreign bodies is a more general type of tissue response that occurs to c.p. titanium
alone [3]. The only definition of osseointegration that has stood up to a critical
analysis is based on a clinical finding of implant stability: ‘A process whereby clini-
cally asymptomatic rigid fixation of alloplastic materials is achieved and main-
tained, in bone, during functional loading’ [18]. The continued usage of a term such
as osseointegration is motivated by the proven clinical results in the case of cranio-
facial implants and the hope to replicate these findings in the case of orthopaedic
implants in the future. However, from a strict histological point of view osseointe-
gration remains poorly defined.
Osseointegrated implants have resulted in a clinical breakthrough in two differ-
ent clinical applications in the craniofacial skeleton: one of these being oral implants,
irrespective of whether treating total or partial edentulousness [19, 20]; the other
being skin penetrating extra-oral implants. The clinical results of screw-type, c.p.
titanium oral implants in mandibles or maxillas for 5 years or more of follow up
have been in the 90–99% range[21]. The results of skin penetrating implants in the
temporal bone region have been similar, but not in the orbit region, where the host
bed has been irradiated. Now, 20 years since their clinical introduction in 1977,
permanent skin-penetrating, osseointegrated, screw-shaped titanium implants are
regarded as routine clinical treatments for facial disorders or certain types of hear-
ing impairments [22, 23].
Press-fit fixation represents one approach to the osseointegration of orthopedic
implants. The design of press-fit joint replacements is based on three dimensional
geometric data with the intention of fitting the implant as closely as possible to the
host bone. The objective of this design approach is to transfer load across the
implant-bone interface to as wide an area of the bone as possible [24]. In theory,
press-fit fixation may lead to osseointegration of the implanted device. However, as
it is difficult to mimic precisely the resulting intravital loading patterns, osseointe-
gration of initially stable press-fit components is threatened by subsequent bone
remodeling processes. Too stiff a device may cause ‘stress shielding,’ leading to
bone resorption. Conversely, too high local stresses may lead to pressure necrosis
and resorption prior to remodeling. Finally, bone resorption may lead to local inter-
facial movements: these predispose to soft-tissue formation and may cause a subse-
quent failure of fixation.
Polymers are not one hundred percent stable under biological conditions, leading
to highly variable clinical durability. Ultrahigh molecular weight polyethylene
(UHMWPE), although relatively stable, has shown poor outcomes when press-
fitted in knee replacement arthroplasties [25]. The poor fixation of such devices may
3 Hard Tissue Response 587
The implant-bone interface can fail for various reasons but the most common is so-
called ‘aseptic loosening’: i.e. loosening not associated with infection. The caus-
ative factors for aseptic loosening may be classified as mechanical or biological. For
example, early post-operative failures of cemented, cement-free (ingrown) as well
as osseointegrated implants may be attributed to lack of initial fixation due to
mechanical failure of the bone-cement (when present) or bone-implant interface.
Such failures may relate to incidents of overload. However, early failures may also
occur subsequent to overheating of the tissues during surgery, leading to bony
necrosis, collapse and subsequent mechanical loosening. Overheating due to a com-
bination of surgical trauma and polymerization heat release may occur at bone-
cement interfaces; surgical trauma, with associated heating, alone may prevent bone
and other tissue elements from invading cement-free implant-bone interfaces,
whether designed for ingrowth or osseointegration. Biological failure is also possi-
ble in the longer term, associated with generation or release of cytotoxic products,
such as cement monomer or metallic corrosion products or through induction of
specific immune responses [33] (Figure 3.3).
In clinical use, cylindrical shaped oral implants which lack any additional reten-
tion features, such as threads, grooves, etc., are prone to gradual failure due to ongo-
ing bone ‘saucerization’ (gradual loss of bone at the implant-bone-mucosal boundary)
[21]. In the case of orthopaedic implants with medullary structural elements (such as
a femoral stem), Willert and Semlitsch were first to propose that bone loss occurred
secondary to biological response to small particles, such as wear debris. [34]
Macrophages (MP) and foreign body giant cells (GC) ingest these undigestible par-
ticles of metal, polymer or ceramic and release factors which stimulate osteolytic
activity by cells in membranes associated with the implant-bone interface (Figure 3.3)
[35]. With modern histological staining techniques, especially the use of oil red 0
[33], it has now become possible to appreciate the large amounts of small and sub-
micron UHMWPE particles found in the vicinity of loose orthopaedic joint replace-
ment implants. Many investigators believe that biological response to these particles
is the leading biological cause of osteolysis leading to gradual, late implant failure.
3.6 Conclusions
Fig. 3.3 Theoretical failure modes for an osseointegrated calcium hydroxyapatite-coated femoral stem of a total hip replacement prosthesis [29; by permis-
sion]. Key: GC: multinucleated giant cell; MP: macrophage; Ca: calcium; Ph: phosphorous; OAF: osteoclastic activating factor; IL1: interleukin 1, TNF: tumor
necrosis factor; PGE2: prostaglandin E2.
589
590 T.O. Albrektsson
placement but then demonstrate an increasing amount of interfacial bone tissue. The
main proportion of osseointegrated implant failures occur during the first one or two
years and result from failure to achieve a proper osseointegration. This trend is quite
different from most orthopaedic studies reported in the literature, which show clini-
cal failure rates increasing with time. Cemented arthroplasties of today have resulted
in a significantly better clinical outcome than cement-free ones. It must be pointed
out that the results of orthopaedic implants published in the literature are generally
based on the number of revisions alone. This means that the actual number of suc-
cessful arthroplasties is lower than the figures quoted in the literature. Furthermore,
there is a patient drop-out of more than 20% in most clinical reports. From a bio-
logical viewpoint it is important to strive for improved cement-free implants so that
their clinical results will at least match those of the cemented arthroplasties. New
types of osseointegrated hip and knee constructions have been designed and are
presently in clinical trials. Altering prosthetic design (in comparison with current
practice), and improving surgical instruments and procedures may well overcome
some of the hurdles in the development of new osseointegrated arthroplasty devices.
Additional Reading
Lee, A.J.C. and Ling, R.S.M. (1984): loosening, in Complications of Total Hip
replacement, (ed. R.S.M. Ling), Churchill-Livingstone, Edinburgh, pp. 110–145.
Still useful account of the biological and mechanical features of loosening lead-
ing to clinical failure of the various interfaces formed between rigid biomaterials
and bone. Extensive bibliography.
Manley, M.T. (1993): Calcium phosphate biomaterials: A review of the litera-
ture, in Hydroxylapatite Coatings in Orthopaedic Surgery, (eds G.T.R. Geesink and
M.T. Manley), Raven Press, New York, pp. 1–24.
The title speaks for itself. Additional chapters in the same book provide experi-
mental and clinical results of osseointegration, specifically adhesion fixation.
Spector, M. (1987): Historical review of porous-coated implants. J. Arthroplasty,
2(2), 163–177, 1987.
Historical review of experimental and clinical results of ingrowth fixation, with
extensive bibliography.
References
1. Frost, H.M. (1989) The biology of fracture healing. Clinical Orthopaedics and Related
Research, 248, 283–293.
2. Morberg, P. (1991) On bone tissue reactions to acrylic cement. PhD Dissertation, Biomaterials
Group, University of Göthenberg, Sweden, pp. 1–142.
3 Hard Tissue Response 591
3. Donath, K., Laass, M. and Günzl, H.-J. (1992) The histopathology of different foreign body
reactions in oral soft tissue and bone tissue. Virchows Archiv A Pathologia Anatomica, 420,
131–137.
4. Johansson, C. (1991) On tissue reactions to metal implants. PhD Dissertation, Biomaterials/
Handicap Research, University of Göteborg, Göteborg, Sweden, pp. 1–232.
5. Black, J. (1988) Orthopaedic Biomaterials in Research and Practice. New York, Churchill
Livingstone, pp. 267–284.
6. Steinemann, S.G., Eulenberger, J., Maeusli, P.-A. et al. (1986) Adhesion of bone to titanium.
Adv. in Biomaterials 6, 40–44.
7. Ryd, L. (1986) Micromotion in knee arthroplasty. A Roentgen stereophotogrammetric analysis
of tibial component fixation. Acta Orthop. Scand., Suppl. 220, 1–80.
8. Malchau, H., Herberts, P. and Anhfelt, L. (1993) Prognosis of total hip replacement in Sweden.
Follow-up of 92,675 operations performed in 1978–1990, Acta Orthop Scand., 64, 497–506.
9. Collier, J.P., Mayor, M.B., Chae, J.C. et al. (1988) Macroscopic and microscopic evidence of
prosthetic fixation with porous coated materials. Clin. Orthop. Rel. Res., 235, 173–180.
10. Willems, W.J., Eulderbrink, F., Rozing, P.M. et al. (1988) Histopathologic evaluation in failed
Gerard double cup arthroplasty. Clin. Orthop. Rel. Res., 228, 123–133.
11. Engh, C.A., Bobyn, J.D. and Glassman, A.H. (1987) Porous coated hip replacement. The fac-
tors governing bone ingrowth, stress shielding and clinical results. J. Bone Joint Surg., 69B,
45–55.
12. Cook, S.D., Thomas, A.K. and Haddad, R.J. (1988) Histologic analysis of retrieved human
porous coated joint components. Clin. Orthop. Rel. Res., 234, 90–101.
13. Brånemark, P.-I., Hansson, B.-O., Adell, R. et al. (1977) Osseointegrated implants in the treat-
ment of the edentulous jaw. Scand. J. Plastic Reconst. Surg., Suppl 16, 1–116.
14. Albrektsson, T. (1979) Healing of bone grafts. In vivo studies of tissue reactions at autografting
of bone in the rabbit tibia. PhD Dissertation, Laboratory for Experimental Biology, Göteborg
University, Göteborg, Sweden, pp. 1–90.
15. Pilliar, R.M. (1986) Implant stabilization by tissue ingrowth. In Tissue Integration in Oral and
Maxillofacial Reconstruction, D. van Steenberghe (ed.), Amsterdam, Excerpta Medica,
pp. 60–76.
16. Wennerberg, A. (1995) On surface topography of implants. PhD Dissertation, Biomaterials/
Handicap Research, University of Göteborg, Göteborg, Sweden, pp. 1–202.
17. Brånemark, P.-I. (1985) Introduction to osseointegration, in Tissue Integrated Prostheses (eds
P.-1. Brånemark, G. Zarb and T. Albrektsson), Quintessence Co, Chicago, pp. 11–76.
18. Zarb, G. and Albrektsson, T. (1991) Osseointegration – A requiem for the periodontal liga-
ment? – An editorial. Int. J. Periodontal and Restorative Dent., 11, 88–91.
19. Albrektsson, T., Dahl, E., Enbom, L. et al. (1988) Osseointegrated oral implants. A Swedish
multicenter study of 8139 consecutively inserted Nobelpharma implants. J. Periodont., 59,
287–296.
20. Lekholm, U., vanSteenberghe, D., Herrmann, I. et al. (1994) Osseointegrated implants in par-
tially edentulous jaws: A prospective 5-year multicenter study. Inter. J. Oral & Maxillofacial
Implants, 9, 627–635.
21. Albrektsson, T. (1993) On the long-term maintenance of the osseointegrated response.
Australian Prosthodontic J., 7, 15–24.
22. Tjellström, A. and Granström, G. (1994) Long-term follow-up with the bone anchored hearing aid:
A review of the first 100 patients between 1977 and 1985. Ear Nose and Throat J., 2: 138–140.
23. Jacobsson, M., Tjellström, A., Fine, L., et al. (1992) A retrospective study of osseointegrated
skin-penetrating titanium fixtures used for retaining facial prostheses. Int. J. Oral &
Maxillofacial Implants, 7, 523–528.
24. Poss, R., Robertson, D.D., Walker, P.S. et al. (1988) Anatomic stem design for press-fit and
cemented application. In Non-cemented Total Hip Arthroplasty, R. Fitzgerald, Jr (ed.),
New York, Raven Press, pp. 343–363.
25. Freeman, M.A.R., McLoed, H.C. and Levai, J.P. (1983) Cementless fixation of prosthetic
components in total arthroplasty of the knee and hip. Clin. Orthop. Rel. Res., 176, 88–94.
592 T.O. Albrektsson
26. Søballe, K. (1993) Hydroxyapatite ceramic coating for bone implant fixation. Acta Orthop.
Scand. (Suppl 225), 64, 1–58.
27. Han, C.H., Johansson, C., Wennerberg, A. et al. (1995) A quantitative comparison of commer-
cially pure titanium and titanium-6 aluminum-4 vanadium implants in rabbit bone. Proc. Fifth
Biomaterials Club Meeting, Ischgl, Austria, Albrektsson T. and Tjellstrom, A. (eds), p. 25.
28. Søballe, K., Hansen, E.S., Rasmussen, H.B. et al. (1993) The effect of osteoporosis, bone
deficiency, bone grafting, and micromotion on fixation of porous coated vs. hydroxyapatite-
coated implants. In Hydroxyapatite Coatings in Orthopedic Surgery, Geesink, R.G.T. and
Manley, M.T. (eds), New York, Raven Press, pp. 107–136.
29. Gottlander, M. (1994) On hard tissue reactions to hydroxyapatite-coated titanium implants.
PhD Dissertation, Biomaterials/Handicap Research, University of Göteborg, Göteborg,
Sweden, pp. 1–202.
30. Lemons, J.E. (1991) Bone-biomaterial interfaces in retrieved implants, in The Bone Biomaterial
Interface (ed. J.E. Davies), University of Toronto Press, Toronto, pp. 419–424.
31. Albrektsson, T., Eriksson, A.R., Friberg, B., et al. (1993) Histologic investigations on 33
retrieved Nobelpharma implants. Clinical Materials 12 (1), 1–9.
32. Steflik, D.E., Parr, G.R., Singh, B.B. et al. (1994) Light microscopic and scanning electron
microscopic analyses of dental implants retrieved from humans. J. Oral Implantol., 20(1),
8–24.
33. Campbell, P. (1995) On aseptic loosening of total hip replacements: The role of UHMWPE
wear particles. PhD Dissertation, Biomaterials/Handicap Research, Göteborg University,
Göteborg, Sweden, pp. 1–225.
34. Willert, H., Semlitsch, M. (1976) Reactions of the articular capsule to joint prostheses. In
Biocompatibility of Implant Materials, D.F. Williams (ed.), London, Sector Publishing,
pp. 157–169.
35. Goldring, S.R., Schiller, A.L., Roelke, M., et al. (1983) The synovial-like membrane at the
bone-cement interface in loose total hip replacements and its proposed role in bone lysis.
J. Bone Joint Surg., 65A, 575–584.
Chapter 4
Immune Response
K. Merritt
4.1 Introduction
There is increasing concern about the role of specific immune response to implanted
materials. This section discusses the general principles governing immune responses
and outlines techniques for their measurement and evaluation. This is a necessarily
brief presentation of the issues, and the reader is encouraged to pursue the topic
through relevant references provided for further study.
The specific immune response is the normal response of vertebrates when a foreign
substance is introduced into the body. This is a desirable protective response which
detoxifies, neutralizes, and helps to eliminate such substances.
However, in some cases, responses to seemingly innocuous substances may
cause harm to the host. Such effects are usually termed allergic or hypersensitivity
responses. The responses have been classified into four types: Type I, Type II, Type
III, Type IV.
These four responses share elements of a common mechanism, triggered by the
presence of a foreign material termed an antigen. The antigen is initially processed
by a cell, usually either a monocyte or macrophage, but occasionally a skin den-
dritic cell also referred to as an antigen processing cell (APC). The APC engulfs the
antigen, processes it (usually by enzymatic digestion or attempted digestion), and
transfers or presents it to another cell, usually a lymphocyte termed a T helper cell.
The T helper cell then presents the processed antigen to another T lymphocyte,
K. Merritt (*)
17704 Stoneridge Dr., Gaithersburg, MD 20878, USA
Both the T- and B-cells are small lymphocytes, which circulate in the blood, and
are found in lymphoid tissue. They arise from a common stem cell and then undergo
processing in the thymus to become T-cells or an unknown site, probably the bone
marrow, to become B-cells. They are difficult to distinguish and a great deal of work
has been done to facilitate identification of these cells in order to elucidate specific
immune responses. Identification of T-cells has been greatly aided by the recogni-
tion that there are unique substances on the cell surface that can be recognized by
the use of monoclonal antibodies generated using murine cells. These antigens are
referred to as cluster differentiation markers (CDs) and are given numbers. There
are a large number of them, more are being identified, and the importance of each is
being evaluated. However, all T-cells express CD3 which is consequently referred
to as a Pan(= all) T-cell marker. CD2 may also be a Pan T-cell marker. Additionally,
the T helper cell expresses CD4 while the T cytotoxic cell expresses CD8. New CDs
continue to be identified and their importance evaluated.
The B-cell retains small amounts of antibody on its surface which can be used for
identification. The B-cell response results in further differentiation to plasma cells
which produce antibody in large quantities. Antibody is a four chain immunoglobu-
lin which has combining sites specific for a single antigen. Antibody is soluble,
circulates in the plasma, and the plasma or serum drawn from a vertebrate, or the
antibody produced in cell culture for monoclonal antibodies, can be stored frozen
and will last virtually forever.
There are five general classes of immunoglobulins produced as antibodies. In order
of concentration in normal human blood from highest to lowest are IgG, IgM, IgA, IgE,
and IgD. IgD is a surface marker on B-cells and will not be discussed further. IgE is the
4 Immune Response 595
whether or not antibody is increased in the patient population and whether or not its
presence can be correlated with failure of the device.
Most current tests are based on immobilization of the antigen to a solid sur-
face such as polystyrene. The general procedure is outlined in Figure 4.1.
Detection of bound antibody involves the use of enzyme (EIA or ELISA assays)
or a radioisotope labeled antibody (RIA). (The tests are simple to perform, but
great care is needed to wash away excess materials and use the appropriate con-
centrations. Individuals wishing to initiate such testing procedures are encour-
aged to attend workshops (often offered by American Society for Microbiology,
American Type Culture Collection, etc.) or obtain training in a clinical immunol-
ogy lab, and also explore detailed manuals on procedures.) Antiserum to human
antibodies, (all classes or individual types) with enzyme labels, can be purchased
from several biological supply houses. Isotope labeled antibody is available to
licensed laboratories.
Problems with interpretation of results:
Positive results
If negative controls are negative and patient samples are positive, then the interpre-
tation is that the patient made antibody to that antigen. However, the question
remains as to whether or not the antigen is the correct one and if contaminating
antigens are contributing to the reaction. This is difficult to determine but inhibition
studies with well characterized antigens are helpful.
In vivo testing
Type I sensitivity is associated with histamine and vasoactive amine release with
vascular responses. Often such sensitivity is determined by ‘patch’ (skin) testing.
The positive reaction will occur in a few minutes as a wheal and erythematous (red-
dening) flare response in the skin. This test is hampered by availability of antigen
for testing but the biomaterial applied directly to skin or a mucosal surface may
stimulate a response. Caution needs to be taken in interpreting these tests, however,
since this is a short term application.
4 Immune Response 597
Negative results
Negative results are the desired response in evaluating biomaterials for clinical use,
but they are also difficult to evaluate. If there are no responses recorded except in the
positive control sample, this is indicative that patient does not produce antibody to
that antigen. Again the question of appropriateness of the antigen and its concentra-
tion on the solid support remains.
In vivo testing
A negative skin test presents the same problem: Was the antigen correct and/or was
it applied to a correct site?
The procedures for detecting cell mediated responses are much more complicated
and difficult than for antibody determination. Most of the assays require the use of
living cells and thus tests must be done shortly after obtaining the cells. Controls
may have to be done at a different time which complicates the comparisons.
The two most common in vitro test procedures used are lymphocyte proliferation
and cell migration inhibition tests. The basic theory behind both of these is that
T-cells have receptors on their surface which will each respond to a specific antigen.
In the course of the response, soluble substances (cytokines or lymphokines), prin-
cipally blastogenic factor and migration inhibition factor, which act on other cells,
including other T-cells, are produced and released.
(leukocyte inhibition test) requires living lymphocytes and living migrating cells
which may be obtained from fresh, whole blood. The results of the LIF test are avail-
able in 18–24 hours. blood does not contain enough monocytes to evaluate inhibition
of their migration and this indicator cell is usually obtained from the peritoneal cav-
ity of other animals, typically the mouse or the guinea pig. It is possible to stimulate
human lymphocytes in culture for 24–48 hours and then harvest the culture fluid and
add it to the macrophages obtained from the animals. Migration (or inhibition of
migration) of cells is observed by placing them into tissue culture medium solidified
with purified agarose and observing them with a microscope in 18–24 hours or by
packing them into capillary tubes and observing migration from the tubes in a few
hours giving the appearance of ‘ice cream cones’. The factor for LIF and the one for
MIF may be slightly different and thus the two separate terms remain.
Choice of test
There is no evidence that one test is better than the other. The choice is usually
based on laboratory preference. This author uses blood cells and migration in aga-
rose since it requires little equipment, is rapid, and small breaks in aseptic procedure
are tolerated. The evidence is that the stimulated T-cells produce a group of cyto-
kines. Thus detecting migration inhibition factor or lymphocyte proliferation
implies the presence of the others and one is not more specific or sensitive.
Fig. 4.2 Quantitation of antigen by competitive (inhibition) assay. An antigen is fixed (attached)
and binds a specific antibody from solution. However, additional antigen is provided in solution
and antibody binding to the bound antigen is reduced in direct proportion to the concentration of
free antigen. The bound antibody is then detected by binding a second, labeled (enzyme, isotope,
etc.) antibody.
In vivo studies
The classical test for cell-mediated immunity (CMI) (Type IV) in the early days of
immunology was the skin test. Antigens were applied to the skin or injected under
the skin and swelling was observed in 24–72 hours if there was CMI. This was dif-
ferentiated from Type I, IgE mediated responses, which occur rapidly (in minutes)
and usually disappears in 24 hours. CMI begins in 24 hours, has a swelling with
some resemblance to the wheal but does not show a flare. This author does not advo-
cate skin testing for responses to biomaterials since the testing is difficult to do
correctly, has the potential for producing sensitivity to the test agents and the results
are easily misinterpreted. In general, the use of the actual biomaterial (rather than
extracts, corrosion products or constituents) is contraindicated since mechanical
irritation may be read as a false positive or or the biomaterial may fail to release the
antigen and thus produce a false negative.
Skin testing is an excellent diagnostic procedure in patients with clinical suspicion
of hypersensitivity. However, skin testing with haptens, such as metal ions, involves a
risk of sensitization. For the immune response to be detected, the hapten must bind to
dermal cells or proteins. However, such binding produces a complete antigen which
may stimulate an immune response. Since this immune response takes time to develop,
the skin test will be negative, but future tests may then be positive. Thus repeated test-
ing increases the likelihood of inducing sensitivity and should be avoided. Bulk bio-
materials will probably not give adequate release of soluble materials in the 24–48
period of testing so may produce false negative results.
Histochemical techniques
There are a number of studies now examining tissues removed from sites adjacent
to implants. It is possible, using immunological techniques resembling those out-
lined in Figure 4.1, to identify cell types and perhaps cell products produced at the
600 K. Merritt
site. The major interest is in the detection and typing of lymphocytes by use of
antisera prepared against the CD markers described earlier. The same type of assay
is being initiated for detection of the cytokines in the tissue. The techniques are
simple but not all antisera work and thus a variety of antisera are used. The required
antisera are available commercially.
Interpretation of the results is again a problem. Tests using ‘home made’ monoclo-
nal antibodies are suspect until the antisera is made available to other investigators for
conformation. The use of well characterized antisera from companies which supply to
others is better at this stage. Since the tissue section is examined and scored by an
observer, the ‘data’ from these studies are not really available for analysis by the
reader. Computerized image analysis techniques are still not widely used. Thus, in
evaluating the results, possible bias of the investigator must be taken into account.
Detection of immune responses to haptens is the same as that described above, but
there are some special techniques now being used. A hapten-carrier complex can be
prepared in vitro by combination in solution with a large protein such as albumin or
a smaller molecule such as glutathione. These can then be used to coat a solid sub-
strate. Alternatively, the protein carrier can be coated onto a substrate, the hapten
added, and then the assay performed. This probably increases the amount of hapten
that is available for antibody binding and minimizes that which is lost in the tertiary
folding of the protein.
4.6.1 Latex
The term ‘latex’ actually is confusing since the name is given to some materials
because of the way they are processed and not because of their source. The biomate-
rial latex used to fabricate gloves, condoms, etc. is a rubber (elastomer) extracted
from a plant (Hevea brasiliensis). As such, there is a great deal of antigenic protein
contamination. Allergies to latex are usually of the Type I, IgE mediated response,
with an immediate reaction (in minutes) that can be life threatening. Since latex is
encountered in many household objects such as household gloves, balloons, etc.,
sensitivity to it is a frequent pre-existing condition. Latex material cleaned of protein
seems to be nonallergenic. Other types of immune response to latex have not been as
frequent or of much concern. Latex is not used as a long term implanted material and
thus the long term responses are not noted. The population at greatest risk are the
health care workers with the increased use of examining and surgical gloves.
4 Immune Response 601
4.6.2 Collagen
This is another material that is an extract of material of natural origin with bovine
and orvine skin or tissue the favored source. This is a foreign protein and thus
capable of stimulating a variety of immune responses. Antibodies of IgE, IgM, and
IgG classes have been observed. Cell mediated immune responses have also been
observed. As with latex, the important precaution is to remove as much foreign
material as possible. Since collagen across mammalian species has a similar struc-
ture, it is possible to remove contaminating proteins and leave a relatively nonal-
lergenic material. Chemical treatment and cross linking of the collagen can further
reduce antigenicity. Collagen products need to be carefully evaluated for their abil-
ity to initiate immune responses, but it is possible to produce safe products.
The use of chemically defined synthetic polymers is associated with minimal human
immune responses. These materials are based on carbon, hydrogen, nitrogen and
oxygen which are basic building blocks of the biological system. Thus the genera-
tion of antigenic material would be unlikely. Nevertheless, there are some poly-
meric materials with additional chemical moieties that are of concern.
B. Polyurethanes
This is a complex group of polymers. Their propensity to stimulate an immune
response is very small since there are few molecular groups which would be per-
ceived as foreign by the host, perhaps explaining why clinical immune responses to
polyurethanes have not been reported.
C. Poly(methyl)methacrylate
Acrylics are in widespread use in activities of daily living. As with metals, there are
documented cases of contact dermatitis from the use of these materials, especially
self curing glues containing methacrylate monomer that is very skin sensitizing,
602 K. Merritt
usually stimulating a Type IV response. The use of these materials for implants
generally exposes the patient to the monomer for only a brief period of time as the
bone cement or dental acrylic cures in situ. Acrylics which are fully polymerized
before use will not be associated with an allergic response. Reports of sensitization
responses of patients to acrylics are rare and the health care workers at most risk are
the personnel, such as the surgeon and dental laboratory technician, handling the
monomer frequently.
D. Metals
A number of metallic elements and alloys are used extensively in implants, external
medical devices and are encountered in activities of daily living. Allergy to metals
as a contact dermatitis (Type IV response) is well known in individuals in contact
with metal salts, corroding metals, and jewelry or snaps and fasteners. Reactions
have been seen to metals used in dental, orthopaedic, and general surgery. The con-
tact dermatitis from topical use resolves when the device is removed. The role of the
immune response in reactions to metals implanted into the deep tissue remains con-
troversial. Cell mediated immune responses have been associated, in some studies
but not in others, with pain and swelling at the implant site and loosening of the
device. Antibodies to metals in patients with metallic implants have recently been
reported, but again the consequences of this response remain unknown. Concern
remains about the chronic use of metals that are known human sensitizers, such as
chromium, nickel and cobalt.
The immune response is apparently intended to neutralize, detoxify, and help elimi-
nate a foreign material. However, sometimes the immune response can inadver-
tently cause harm. This will be discussed in various categories in the next section.
1. Damage to the implant. The inflammation which is part of the initiation of the
immune response is an oxidative response. Materials subject to oxidative attack,
such as polyethylenes and polyurethanes, may be degraded.
2. Damage to adjacent tissues. Products, particularly from Type II and IV responses,
may initiate swelling and other vascular responses at the site. This may resolve
with no further harm, or it may cause tissue necrosis and/or loss of tissue mass
with loosening or movement of the device.
3. Systemic responses. Immune responses of Type I and II generate vasoactive sub-
stances which may circulate and cause vascular collapse. This is seen in response
to latex materials and drugs which bind to platelets, mast cells, or eosinophils,
resulting in an immune response and release of these vasoactive substances.
4. Autoimmune diseases: This is the most controversial area of consequences of
immune response to implants. Autoimmune diseases are technically the result
4 Immune Response 603
4.8 Conclusions
There has been a rapid growth in our knowledge of the immune response and how to
evaluate and quantitate it. As these techniques are applied to the population in contact
with biomaterials, we will learn more about its importance in performance of the
material. We will also learn more about how to process the materials to minimize the
immune response. However, it is important to remember that the immune response is
a protective response and detection of immune responses to products of biomaterials
does not necessarily indicate clinical problems. On the other hand, implants are for-
eign material and will stimulate host responses, some of which may cause harm to the
host or implant.
Thus the important issue is to distinguish between those immune responses
which are normal and help to render antigens less biologically active from those
which are harmful to the host. It is clear that IgE (Type I) responses are harmful.
Detection of a Type I response to products of biomaterials indicates potential
problems in the clinical setting. Responses of the IgG type are generally protec-
tive and may not be predictive of further problems unless there is continual release
of antigenic material leading to a Type III-response. Biomaterial wear and degra-
dation products that bind to platelets or mast cells pose a potential for adverse
Type II responses.
The most commonly observed is the Type IV (cell-mediated) response. This
is a protective response in walling off the stimulating agent and in killing cells
which have the antigen on the surface, thus eliminating the antigen. However,
the tissue reaction accompanying this response may cause harm to the host
through soft and hard tissue necrosis. The difference between protection and
allergy from Type IV responses is still unclear and careful evaluation of patients
is required.
604 K. Merritt
Additional Reading
General Immunology
Golub, E.S. Green, D.R. (1990) Immunology, A Synthesis, 2nd Edition, Sinauer Associates, Inc.,
Sunderland: Good general text.
Roitt, I. (1971) Essential Immunology, Blackwell Scientific Publications, London: Good descrip-
tion of types I–IV reactions.
Annual Review of Immunology, Annual Reviews Inc., Palo Alto, CA: Yearly publication with
timely reviews.
Immunology Today: Elsevier Science Inc. Tarrytown, NY: Monthly: Good review articles.
Antigen Presentation
Celada, A. and Nathan, C. (1994) Macrophage activation revisited. Immunology Today, 15, 100–
102: good review of macrophages.
Chicz, R.M. and Urban, R.G. (1994). Analysis of MHC presented peptides: applications in autoim-
munity and vaccine development. Immunology Today, 15, 155–160: good review on a compli-
cated subject.
CD markers
Kemeny, D.M., Noble, A. Holmes, B.J. et al. (1994) Immune regulation: a new role for the CD8+
T cell. Immunology Today, 15, 107–110: Good description of the function of the CD8+ T-cell
which is a key cell in Type IV responses.
Sclossman, S.F., Boumsell, L., Gilkes, L.W. et al. (1994) CD antigens 1993. Immunology Today,
15, 98–99: good description of recently reported CDs.
Cytokines/interleukins
Miyajima, A., Kitamura, T., Harada, N. et al. (1992) Cytokine receptors and Signal Transduction.
Annual Reviews of Immunology, 10, 295–331: Review of function and methods of
stimulation.
Mizel, S.B. (1989) The interleukins. FASEB J. 3, 2379–2388: good detailed review.
Effects of cytokines/interleukins
Goldring, M.B., and Goldring, S.R. (1990) Skeletal tissue response to cytokines. Clin. Orthop. Rel.
Res., 258, 245–278: review of cytokines and orthopaedics.
Stashenko, P., Obernesser, M.S., and Dewhirst, F.E. (1989) Effect of immune cytokines on bone.
Immuno Invest., 18, 239–249: one of the few reviews focussing on bone.
Burholm, A.; AI-Tawil, N.A.; Marcusson, J.A. et al. (1990): The lymphocyte response to nickel
salt in patients with orthopedic implants. Acta Orthop. Scand., 61(2): 248–250: Example of use
of LTT test.
Elves, M.W., Wilson, J.N. and Kemp, H.B.S. (1975) Incidence of metal sensitivity in patients with
total joint replacements. Brit. Med. J., 4, 376–378: Second one of the original 1975 articles
pointing to a possible problem. Skin test used.
Evans, E.M., Freeman, M.A.R., Miller, A.J. et al. (1974) Metal Sensitivity as a Cause of Bone
Necrosis and loosening of the Prosthesis in Total Joint Replacement, J. Bone and Joint
Surg., 56B (4), 626–642: One of the original articles pointing to a possible problem. Skin
test used.
Goh, C.L. (1986) Prevalence of contact allergy by sex, race, and age. Contact Dermat., 14, 237–
240: discusses normal population
Grimsdottir, M.R., Gjerdet, N.R. and Hensten-Pettersen, A. (1992) Composition and in vitro cor-
rosion of orthodontic appliances. Am. J. Orthod. Dentofac. Orthop., 101, 525–532: Discusses
sensitivity and stainless steels. Release of nickel related to many metallurgical aspects and not
necessarily to nickel content of the metal.
Lalor, P.A., Revell, P.A., Gray, A.B. et al. (1991) Sensitivity to titanium. J. Bone and Joint Surg.,
73B(1), 25–28: Description of possible titanium sensitivity. Patch test vehicle of unknown
composition, larger cobalt-chromium component than titanium component in device. Of inter-
est and important, but not conclusive.
Menne, T.; and Maibach, H.I. (1989) Systemic contact allergy reactions. Immunol Allergy Clin.
N.A., 9, 507–522: Discusses extension from contact dermatitis to systemic reactions.
Merritt, K. (1984) Role of medical materials, both in implant and surface applications, in immune
response and in resistance to infection. Biomaterials, 5 (1), 47–53.: Review article. Out of date
now but covers literature through 1983.
Merritt, K. (1986) Biochemistry/hypersensitivity/clinical reactions. in: Lang B, Morris, J. and
Rassoog, J. (eds) Proc. International Workshop on Biocompatibility, Toxicity, and
Hypersensitivity to Alloy Systems used in Dentistry. Ann Arbor, U. MI; pp 195–223.: Review
article. Covers the literature through 1984. Good discussion of the problem in the discussion
section of the symposium
Merritt, K. (1986) Chapter 6. Immunological testing of biomaterials, Techniques of Biocompatibility
Testing, D.F. Williams (ed.), Vol. II, CRC Press, Boca Raton: Description of possible test
methods.
Merritt, K.; and Brown, S.A. (1980): Tissue reaction and metal sensitivity: An animal study. Acta
Orthop. Scand. 51 (3), 403–411: Example of use of LIF test.
Rostoker, G., Robin, J., Binet, O. et al. (1987) Dermatitis due to orthopaedic implants. J. Bone
Joint Surg., 69A, 1408–1412: Example of a reaction to implant.
Rudner, E.J., Clendenning, W.E., Epstein, E. et al. (1975) The frequency of contact dermatitis in
North America 1972–1974. Contact Derm. 1, 277–280: Incidence of contact dermatitis.
Shirakawa, T., Kusaka, Y. and Morimoto, K. (1992) Specific IgE antibodies to nickel in workers
with known reactivity to cobalt. Clin. Exp. Allergy, 22 (2), 213–218: Measuring IgE and nickel
cobalt interactions.
Trobelli, L., Virgili, A., Corassa, M. et al. (1992) Systemic contact dermatitis from an orthodontic
appliance. Contact Dermatitis, 27, 259–260: Example of reaction to dental application of
metals.
Yang, J., and Merritt, K. (1994) Detection of antibodies against corrosion products in patients after
Co-Cr total joint replacements. J. Biomed. Mater. Res., 28, 1249–1258: Method for measuring
antibodies to metals
Hanke, C.W., Higley, H.R., Jolivette, D.M. et al. (1991) Abscess formation and local necrosis after
treatment with Zyderm or Zyplast collagen implant. J. Amer. Acad. Dermatol. 25, 319–326:
Deals with some adverse responses to collagen materials which may be related to the immune
response. Points to possible problems.
Meade, K.R., Silver, F.H. (1990) Immunogenicity of collagenous implants. Biomaterials, 11, 176–
180: Discusses immunogenicity problem and cross linking. Good place to begin reading.
Naim, J.O., Lanzafame, R.J. and van Oss, C.J. (1993). The adjuvant effect of silicone gel on anti-
body formation in rats. Immunol Inv., 22, 151–161: Shows that the gel is better than Freund’s
adjuvant in stimulating the response to BSA in rats. Caution on use of gel.
Slater, J.E. (1989) Rubber anaphylaxis. New Eng. J. Med. 320, 1126–1130: Good methods. Good
literature review, real cases reacting to anaesthesia mask.
Sussman, G.L., Tarlo, S. and Dolovich, J. (1991). The spectrum of IgE responses to latex. J. Am.
Med. Assoc. 265, 2844–2847: Latex gloves on health workers causing allergic responses in
patients. Can do skin test with latex to check patients or use non-latex gloves.
Warpinski, J.R., Folgert, J., Cohen, M. et al. (1991) R.K. Bush. Allergic reaction to latex: a risk
factor for unsuspected anaphylaxis. Allergy Proc. 12, 95–102: Clinical symptoms of Type I
allergy. Identifies IgE antibodies against latex (gloves, balloons, condoms). IgE against pro-
teins from latex.
Wolk, L.E., Lappe, M., Peterson, R.D. et al. (1993) Immune response to polydimethylsiloxane
(silicone): screening studies in a breast implant population. FASEB J., 7, 1265–1268: Very
important study with a good technique. Hopefully more studies will be done with this tech-
nique. Valid test of antibody to silicone.
M. Rock
5.1 Introduction
The widespread use of temporary and permanent implants in the post World War
II era has had a dramatic impact on the practice of medicine and on the life of dis-
abled and ill individuals. Nowhere has this been more obvious than in the frequent
use of implants to stabilize fractures and replace diseased joints which has revolu-
tionized orthopedic practice and afforded millions of patients levels of function
that previously could not be achieved. Although the metal alloys used in these
implants exhibit excellent resistance to corrosion, oxidation of these large compo-
nents ultimately produce free ions, chlorides, oxides, and hydroxides which, in
combination with particulate metal matter released by wear and fretting, are
released into the surrounding environment. Efforts to improve these alloys have
included compositional as well as processing changes. Additionally, modifications
have been made to the plastic articulating components in efforts to produce a much
more consistent ultrahigh molecular weight polyethylene. The perceived need to
improve implant wear and corrosion resistance and alter design has been largely
motivated by the excessive soft tissue staining noted by orthopedic surgeons at the
time of removal or revision of clinically failed joint arthroplasty. The presence of
particulate metal matter, polyethylene, and even fragments of polymethyl methac-
rylate in local tissue has been confirmed histologically and by direct analysis [1–
4]. In spite of all of the modifications made in implant composition, implant
fixation, and articulation, biomaterial degradation and release of these products
persist [4–7].
M. Rock (*)
Department of Orthopaedics, Mayo Clinic,
200 First Street Southwest, Rochester, MN 55905, USA
The body’s response to the local presence of debris is dependent on the size, amount,
and composition as well as rate of accumulation. The body attempts to neutralize
these foreign particles by precipitating granulomatous foreign body reactions and/
or removal through local lymphatic channels. If the local accumulation of debris
exceeds the body’s ability to neutralize and/or transport, the debris migrates from
the site to remote areas including the bone-implant interface, possibly contributing
to if not initiating the phenomena of loosening and osteolysis (Figure 5.1).
Of equal or possibly even greater concern is the detection of metal ions, metallic
debris, polyethylene, and even methylmethacrylate in areas remote from the implant
including circulating serum, excreted urine, and regional draining lymph nodes.
Elevated serum levels of metal ions consistent with the composition of the implanted
alloy have been confirmed in animal models [8] and in human patients after total hip
arthroplasty [9]; identifying serum levels of cobalt, chromium, nickel, and titanium
that are two and three fold higher than preoperative determinations. These figures
represent significant elevations both over means for contemporary control groups
and for the individual patients before operation [9]. However, since they are within
the widely accepted normal range for these metallic ions in the unimplanted human
controls, it is assumed that toxic levels of these foreign materials do not materialize.
However, when analyzing the serum to urine concentration in patients subjected to
conventional total hip arthroplasty, it has become apparent that the urinary concen-
tration of chromium in particular does not rise with the same magnitude and time
course as the serum level [9]. This observation parallels that made in the accounts
of industrial overexposure to Cr6+ and suggests that metal ions accumulate in organs
and tissues remote from implantation. Such accumulation is unlike that resulting
from normal systemic circulation. This was previously suggested by Steineman [10]
who calculated the potential release of metallic ions of 0.15 to 0.3 micrograms per
cm2 per day which would translate to between 11 and 22 milligrams per year in
patients with total hip replacement. This incidentally coincides with or exceeds the
total body burden of such metallic ions in a 70 kilogram man [11].
Evidence for metallic debris accumulating in distant organs has also been con-
firmed by Langkamer et al. [12] who identified wide spread dissemination of par-
ticulate wear debris from hip prosthesis to lymph nodes, liver, and spleen. He
reported increases above normal levels in these organs of 30 fold for aluminum,
chromium, and iron in the lymph node, and 10 fold in the spleen and liver.
These findings suggest that concentrations of metal ions and debris at remote
sites may reach such proportions as to precipitate altered cellular dynamics in
organs principally of the lymphoreticular system. It would only be logical to assume
that local concentrations of such debris at the site of implantation would be even
higher, although attempts at quantifying the effects of local concentrations have
been fraught with inaccuracies mostly due to sampling error and the need to distin-
guish between bioavailable and non bioavailable metal species.
What is potentially more disturbing is that these figures for serum concentrations
and the identification of this debris in remote organs have come primarily from
patients who have received conventional polymethyl methacrylate cemented com-
ponents. With the advent of using uncemented porous coated implants, particularly
in younger patients, these figures would be expected to increase, creating the dis-
tinct possibility of toxic levels in the serum, tissues and organs that will respond
with altered cellular dynamics and function.
5.3 Neoplasia
Perhaps one of the greatest concerns with debris dissemination locally and within
the systemic circulation is the possibility of inducing malignant neoplasia. This is
thought to be possible by one of two mechanisms:
(i) A ‘solid-state’ mechanism has been proposed, whereby a large foreign object
implanted in vivo possibly stimulates mutagenesis of local cells, thereby creat-
ing tumor by its mere presence. Most large foreign objects upon implantation
will initiate a very marked fibrous reaction. The cells within this fibrous reac-
tion ultimately mutate and become cancer growths.
(ii) The other possibility is that particulate metal matter or other debris have an
innate capacity, upon corrosion or dissolution, to induce cancer through a more
traditional chemical route.
Cancer, the end product of carcinogenesis, is the result of transformation of a
normal cells to ones which grow in an uncontrolled or malignant manner. Cancer is
a genetic disease, which may result from expression of genetic pre-dispositions
present from birth or from later insults to cells of many different types. In particular,
the phagocytosis or pinocytosis of foreign matter (in an attempt to neutralize or
eliminate it) may cause or precipitate malignant conversion. Such conversion, if not
lethal to the cell, may then persist through cell duplication, creating first a cluster of
cells with altered DNA and eventually a clinical malignant tumor. Malignancy is
characterized by rapid, uncontrolled growth, invasion in surrounding tissues and
seeding to form tumors (metastases) in other anatomical locations such as the lung.
610 M. Rock
Some of the more common malignant tumors of musculoskeletal origin are osteo-
sarcoma (OS) of bone and malignant fibrous histiocytoma (MFA) of soft tissue.
Osteosarcoma is the most common tumor of bone: it occurs in children, adoles-
cents and, less frequently, in adults. OS may also occur as a consequence of radia-
tion therapy or in Paget’s disease, an ostensibly benign bone embrittling disease of
the elderly. It frequently appears about the knee (distal femur; proximal tibia), and
in the proximal femur and proximal humerus.
MFA is the most common primary malignant tumor of soft tissues and can occur
in bone in adults over the ages of 50–55. The more common soft tissue type usually
involves the large muscular areas of the body, including the thigh, buttock and upper
arm and shoulder.
In 1961, Sir John Charnley introduced total hip arthroplasty as an alternative in the
management of arthritic hips. No other orthopedic procedure has been adopted with
such enthusiasm. Thirty-five years later we are still witnessing an incremental
increase in the yearly utilization of this operation, attesting to the obvious success
associated with it. According to some investigators, we may be coming into an era
of increased tumor activity in the vicinity of or possibly remote from implantation
sites of these orthopedic appliances.
In 1976, Harris et al. [17] were the first to describe an aggressive granulomatous
lesion around a cemented femoral stem in a total hip replacement. This was a condi-
tion of localized tumor-like bone resorption that appeared radiographically as large
lytic defects within the femur, approximating the cement mantle of the implant.
Initially thought to be neoplastic, these lesions were surgically biopsied and found
to be consistent with well-organized connective tissue containing numerous histio-
cytes, monocytes, and fibroblastic reactive zones. Immunohistologic evaluation
revealed multinucleated giant cells and nonspecific esterasepositive monocyte mac-
rophages. These findings suggest a foreign-body type reaction, and with the subse-
quent isolation of polyethylene, polymethyl methacrylate, and metal debris, it was
theorized that these constituents of the construct likely migrated down around the
implant cement mantle in cemented prostheses and implant-bone interface in non-
circumferentially coated ingrowth implants. Such a reaction suggests an excessive
accumulation of debris at the site of articulation that surpasses the body’s ability to
neutralize and/or transport the material resulting in migration of debris to sites
remote from the source. This rapid appearance of bone loss radiographically which
is often associated with a deteriorating clinical course has been termed type-II asep-
tic loosening [17].
In 1978, two years after the recognition of pseudo tumors of bone induced by the
degradation products of total hip arthroplasty, Arden and Bywaters [20] (Table 5.1)
reported a case of a 56-year-old patient who developed a high-grade fibrosarcoma
of soft tissue 2.5 years after receiving a metal-on-metal McKee-Farrar hip prosthe-
sis. The tumor apparently did not have a direct association with the underlying bone
or any components of the total hip arthroplasty. No formal analysis of the tumor for
debris products was performed. This case drew attention to the possibility of tumors
being initiated in the presence of large orthopedic appliances. It was not until 1984
when this concept became fashionable in large part due to three articles that appeared
simultaneously in the Journal of Bone and Joint Surgery recounting two malignant
fibrous histiocytomas and one osteosarcoma at the site of a total hip arthroplasty
[21–23].
This sudden and rather unexpected evolution prompted editorials [24, 25] in the
same journal addressing the issue of sarcoma and total hip arthroplasty and encour-
aged the orthopedic community worldwide to report such cases to a central registry
612 M. Rock
to obtain more accurate figures on the incidence of such a problem. These tumors
occurred 2, 4, and 5 years after hip replacement that was performed with various
femoral and acetabular components, some with metal-on-metal articulations and
others with metal on polyethylene. In two of these cases, the proximal femur was
extensively involved with tumor that was in direct contact with the component. The
remaining case was a soft-tissue sarcoma not in direct approximation to the prosthe-
sis. Two of these tumors were malignant fibrous histiocytomas, one of bone and one
of soft tissue. The remaining tumor was osteosarcoma. In this particular case, there
was evidence of gray-brown pigmentation both intra- and extracellularly between
the tumor and femoral component. No formal metal analysis was performed. Three
additional cases emerged prior to 1988 at 15 months, 4.5 years, and 2.0 years after
implantation [26–28].
In 1988, five cases were reported occurring at 10 [29, 30] and 11[29, 31, 32]
years after implantation. The sarcomas included two osteosarcomas, two malignant
fibrous histiocytomas, and one synovial sarcoma. Two of these were soft tissue in a
location with no direct association with the implant, yet in the case reported by Tait
et al. [32] there was evidence of nickel within tumor cells. The remaining three
5 Cancer 613
patients all had direct contact with either the cement or implant with the tumor
originating in bone.
In 1990 there were three additional reports in the literature which included an
osteosarcoma developing at the site of a Charnley total hip arthroplasty 8 years [36]
after implantation, malignant fibrous histiocytoma developing 15 years after a
Charnley-Muller total hip arthroplasty [37], and metastatic adenocarcinoma devel-
oping at the site of a Freeman total knee arthroplasty three months after implantation
[38]. In 1992, Jacobs et al. [39] presented a malignant fibrous histiocytoma develop-
ing one half year after implantation of a cementless AML total hip arthroplasty.
In that same journal volume, unpublished but submitted reports of five tumors
occurring around implants were brought to the attention of the orthopedic commu-
nity [41] (Table 5.2). These included malignant fibrous histiocytomas around a
Thompson and a Muller total hip arthroplasties, an osteosarcoma around a Charnley
total hip arthroplasty, a rhabdomyosarcoma of soft tissue in the vicinity of a
Christiansen total hip arthroplasty, and a chondrosarcoma developing in a patient
with Maffucci syndrome having a Charnley total hip arthroplasty. The intervals from
implantation to tumor detection were 9, 3, 10, 9 and 1 years respectively. To this, we
add two previously unreported additional patients, neither of whom had their joint
replacement done at the Mayo Clinic (Table 5.2). The first is that of a 79-year-old
man who nine months previously came to total hip replacement with an uncemented
Harris-Galante component who was found to have a large malignant fibrous histio-
cytoma engulfing the proximal femur and extending to the implant. There was, how-
ever, no evidence of any particulate debris within the tumor cells removed. The
second case was that of a 56-year-old man who developed a soft tissue osteosarcoma
14 months after a left total knee arthroplasty with conventional cemented compo-
nents. The tumor extended down to both the femoral and patellar components.
experiences of tumors around implants nearly eight years before such concern was
voiced with the application of these same metallic alloys in humans [52].
As impressive as these cases may be, they must be put into perspective given the
global use of internal fixation and prosthetic devices. Approximately 300 000 to 350
000 total hip joint replacements are performed worldwide on a yearly basis [53]. It
can be assumed that approximately four million people will have had total hip
arthroplasties performed by the end of 1995. To date, there have been 28 reports of
malignant tumor arising in close proximity to these implants (25 total hip and three
total knee arthroplasties). No direct contact was noted in four. If we assume a mini-
mal latency of five years to suggest association between presence of implant and
tumor, 15 of the 28 could have association. As such, the incidence of sarcomas in
total joint replacement would be approximately 1 in 250 000. There are approxi-
mately 3000 new primary bone tumors and 5000 soft-tissue sarcomas in the United
States per year. This would give an incidence of approximately 1 in 100 000 for the
general population to develop a primary bone sarcoma and 1 in 40 000 to develop a
soft tissue sarcoma a year. This is obviously not stratified for age given that many
primary bone tumors develop in the second and third generation of life, yet it does
afford the opportunity of putting this rather unusual event in perspective.
The prevalence of osteosarcoma among the osseous malignancies in this series is
not entirely unexpected. Of the total osteosarcoma population 15 percent to 20 per-
cent occur after the age of 50 years. Most of these cases are superimposed on Paget’s
disease or in previously irradiated tissue, yet de novo cases of osteosarcoma do
occur in this age group. Malignant fibrous histiocytoma of bone is somewhat less
common. A review of the Mayo Clinic files reveals 71 cases with more than half of
these occurring after age 55. Malignant fibrous histiocytoma of soft tissue is the
most common soft-tissue sarcoma. It is not surprising, therefore, that two of six
soft-tissue tumors in the combined series are of this histogenesis. As such, the
distribution of sarcomas in the combined series could have been predicted from
general population data given the age of the patients and anatomical distribution.
There have been two separate reports that have critically analyzed the cancer
risk after total hip arthroplast [54, 55]. The combined person years of exposure after
operation between the two series was 20 015. The overall cancer incidence among
total hip replacement procedure in both series did not appear to be any different
than what was expected or anticipated. The cancer-observed/expected ratio was
especially low for the first two years following surgery in both series, implying that
patients undergoing this procedure are otherwise generally healthy. In both series,
the observed/expected ratio of developing lymphoma or leukemia was two to three
times higher in patients who had total hip arthroplasty. Additionally, there was a
two fold decrease in breast carcinoma among patients who had total hip
arthroplasty.
616 M. Rock
5.8 Summary
the incidence in the general public. The frequency of occurrence and the associated
individual and group risks of systemic and remote site malignancy remains
unresolved.
Additional Reading
Nyren, O., McLaughlin, J.K. et al. (1995) Cancer risk after hip replacement with
metal implants: A population based study. J. National Can. Inst. 87, 28–33.
An extensive review of risk of cancer in 39 154 total hip replacement patients
who appeared in the Swedish National Cancer Registry between 1965 and 1983. A
review of 60 cancer-specific sites showed an overall, not clinically significant
increase of 3% in incidence, slight increases noted for kidney cancer, prostate can-
cer (in men) and melanoma accompanied by a continuous decline in gastric cancer
for both sexes. This would appear to be the definitive review of the risk for develop-
ing cancers after total hip replacement arthroplasty.
Brand, K.G. and Brand, I. (1980) Risk assessment of carcinogenesis at implanta-
tion sites. Plastic Reconst. Surg. 66, 591–595.
Review of possible foreign body cancer initiation in humans based upon pub-
lished case reports. The authors conclude that, since the clinical use of prosthetic
implants has been popular for more than twenty years and since, extrapolating from
animal experience, at least 25% if not 50% of foreign body tumors should have
appeared by the time of their publication, there is little risk of such non-chemically
mediated tumors occuring in patients.
Gillespie, W.J., Frampton, C.M.A., Henderson, R.J. et al. (1988) The incidence
of cancer following total hip replacement. J. Bone Joint Surg., 70B, 539–542.
A New Zealand study of 1358 patients with total hip arthroplasty, for a total of
14 286 patient years. A significant increase in tumors of the hemopoetic and lym-
phatic systems, accompanied by a significant decrease of cancers of breast (in
women), colon and bowel was observed. The authors suggest that these data are
evidence for increased immune surveillance, allowing or precipitating hemopoetic
and lymphatic tumors but at the same time providing better resistance to the devel-
opment of soft tissue tumors. The first large scale study of this question.
Visuri, T. and Koskenvuo, M. (1991) Cancer risk after McKee-Farrar total hip
replacement. Orthopedics, 14, 137–142.
A study similar to that of Gillespie et al. but on a Finnish patient group (433
patients; 5729 patient years) leading to the same general conclusions. Includes a
historical discussion of the carcinogenic properties of various trace elements.
Jacobs, J.J., Rosenbaum, D.H., Hay, R.M. et al. (1992): Early sarcomatous
degeneration near a cementless hip replacement. A case report and review. J. Bone
Joint Surg., 74B, 740–744.
A review of a patient who developed a malignant fibrous histiocytoma at the site
of a cementless total hip replacement five months after implantation and suc-
cumbed of diffuse metastases, as is typical for such patients, within one year of
618 M. Rock
References
1. Coleman, R.R., Herrington, S. and Scales, J.T. (1973) Concentration of wear products in hair,
blood, and urine after total hip arthroplasty. Brit. Med. J. 1 1527–1529.
2. Lux, F. and Zeisler, R. (1974) Investigations of the corrosive deposition of components of
metal implants and of the behavior of biologic trace elements in metallosis tissue by means of
instrumental, multi-element activation analysis. J. Radiol. Anal. Chem., 19, 289–297.
3. Rock, M.G. and Hardie, R. (1988): Analysis of local tissue response in 50 revision total hip
arthroplasty patients. Trans. Soc. Biomater., XI, 43.
4. Agins, H.J., Alcock, N.W., Bansal, M. et al. (1988) Metallic wear in failed titanium alloy total
hip replacements. A histological and quantitative analysis. J. Bone Joint Surg. 70A, 347–356.
5. Buchert, B.K., Vaughn, B.K., Mallory, T.H. et al (1986): Excessive metal release due to loos-
ening and spreading of sintered particles on porous coated hip prosthesis. Report of two cases.
J. Bone Joint Surg., 68A, 606–609.
6. Jacobs, J.J., Skipor, A.K., Black, J. et al. (1991), Release in excretion of metal in patients who
have a total hip replacement component made of titanium base alloy. J. Bone Joint Surg. 73A,
1475–1486.
7. Witt, J.D. and Swann, M. (1991) Metal wear and tissue response in failed titanium alloy total
hip replacements. J. Bone Joint Surg. 73B, 559–563.
8. Woodman, J.L., Jacobs, J.J., Galante, J.O., et al. (1984), Metal ion release from titanium-based
prosthetic segmental replacements of long bones in baboons. A long term study. J. Orthop.
Res., 1, 421–430.
9. Bartolozzi, A. and Black, J. (1985) Chromium concentrations in serum, blood clot and urine
from patients following total hip arthroplasty. Biomaterials, 6, 2–8.
10. Steineman, S.G. (1985) Corrosion of Titanium and Titanium Alloys For Surgical Implant. in:
Lutergering, G., Swicker, U., Bunk, W. (eds), Titanium, Science, and Technology, Volume 2,
DG für Metal. e.V. Oberuresel, Berlin, 1373–1379.
11. Harvey, A.M., Johns, R.S., Owens, A.H. et al. (1972) The Principles and Practice of Medicine.
New York: McGraw-Hill, pp. 68–102.
12. Langkamer, V.G., Case, C.P., Heap, P. et al. (1992) Systemic distribution of wear debris after
hip replacement. A cause for concern? J. Bone Joint Surg., 74B, 831–839.
13. Doll, R. (1958) Cancer of lung and the nose. Nickel workers. Brit. J. lndust. Med., 15,
217–223.
14. Swanson, S.A.V., Freeman, M.A.R. and Heath, J.C. (1973) Laboratory tests on total joint
replacement prosthesis. J. Bone Joint Surg., 55B, 759–773.
15. Laskin, D.M. (1954): Experimental production of sarcomas by methylacrylate implant. Proc.
Soc. Exper. Biol. Med., 87, 329–333.
16. Carter, R.L. and Rowe, F.J.C. (1969): Induction of sarcomas in rats by solid and fragmented
polyethylene; Experimental observations and clinical implications. Brit. J. Cancer, 23,
401–407.
17. Harris, W.H., Schiller, A.L., Scholler, J.M. et al. (1976) Extensive localized bone resorption in
the femur following total hip replacement. J. Bone Joint Surg. 58A, 612–618.
18. Castleman, B., McNeely, B.U. (eds) (1965) Case 38–1965, Case records of the Massachusetts
General Hospital. New Engl. J. Med., 273, 494–504.
19. Rushforth, G.F. (1947) Osteosarcoma of the pelvis following radiotherapy for carcinoma of
the cervix. Brit. J. Radiol. 47, 149–152.
5 Cancer 619
20. Arden, G.P. and Bywaters, E.G.L. (1978), in Surgical Management of Juvenile Chronic Poly
Arthritis. Arden, G.P., Ansel, B.M. (eds), Academic Press, London, pp. 269–270.
21. Bágo-Granell, J., Aguirre-Canyadell, M., Nardi, J. et al. (1984) Malignant fibrous histiocy-
toma of bone at the site of a total hip arthroplasty. A case report. J. Bone Joint Surg. 66B,
38–40.
22. Penman, H.G. and Ring, P.A. (1984) Osteosarcoma in association with total hip replacement.
J. Bone Joint Surg. 66B, 632–634.
23. Swann, M. (1984) Malignant soft tissue tumor at the site of a total hip replacement. J. Bone
Joint Surg. 66B, 269–231.
24. Hamblen, D.L. and Carter, R.L. (1984) Sarcoma and Joint Replacement (Editorial). J. Bone
Joint Surg. 66B, 625–627.
25. Apley, A.G. (1989) Malignancy and joint replacement. The tip of an iceberg? (Editorial).
J. Bone Joint Surg. 71B, 1.
26. Weber, P.C. (1986) Epithelioid sarcoma in association with total knee replacement. J. Bone
Joint Surg. 68B, 824–826.
27. Ryu, R.K.N., Bovill, E.G. Jr, Skinner, H.B. et al. (1987) Soft tissue sarcomas associated with
aluminum oxide ceramic total hip arthroplasty. A case report. Clin. Orthop. 216, 207–212.
28. Vives, P., Sevestre, H., Grodet, H. et al. (1987) Histiocytome fibreux malin du fémur après
prosthèses totale de hanche (Malignant fibrous histiocytoma of the femur following total hip
replacement). Rev. Chir. Orthop. 73, 407–409.
29. Van der list, J.J.J. (1988) Malignant epithelioid hemangioendothelioma at the site of a hip
prosthesis. Acta Orthop. Scand. 59, 328–330.
30;31. Lamovec, J., Zidar, A., Cucek-Plenicar, M. et al. (1988): Synovial sarcoma associated
with total hip replacement. A case report. Addendum: Osteosarcoma associated with a
Charnley-Muller hip arthroplasty. J. Bone Joint Surg. 70A, 1558–1560.
32. Tait, N.P., Hacking, P.M. and Malcolm, A.J. (1988) Case report, malignant fibrous histiocy-
toma occurring at the site of a previous total hip replacement. Brit. J. Radiol., 61, 73–76.
33. Martin, A., Bauer, T.W., Manley, M.T. et al. (1988) Osteosarcoma at the site of a total hip
replacement. J. Bone Joint Surg. 70A, 1561–1567.
34. Haag, M. and Adler, C.P. (1989) Malignant fibrous histiocytoma in association with hip
replacement. J. Bone Joint Surg. (Brit), 71B, 701.
35. Mazabraud, A., Florent, J. and Laurent, M. (1989) (A case of epidermoid carcinoma developed
in contact with a hip prosthesis)(Fr.)(au. transl.). Bull. Cancer, 76, 573–581.
36. Brien, W.W., Salvati, E.A., Healey, J.H. et al. (1990) Osteogenic sarcoma arising in the area of
a total hip replacement. A case report. J. Bone Joint Surg., 72A, 1097–1099.
37. Troop, J.K., Mallory, T.H., Fisher, D.A., et al. (1990) Malignant fibrous histiocytoma after
total hip arthroplasty: A case report. Clin. Orthop. Rel. Res.,253, 297–300.
38. Kolstad, K. and Högstorp, H. (1990): Gastric carcinoma metastasis to a knee with a newly
inserted prosthesis: A case report. Acta Orthop. Scand. 61, 369–370.
39. Jacobs, J.J., Rosenbaum, D.H., Hay, R.M. et al. (1992): Early sarcomatous degeneration near
a cementless hip replacement. A case report and review. J. Bone Joint Surg., 74B, 740–744.
40. Solomon, M.I. and Sekel, R. (1992) Total hip arthroplasty complicated by a malignant fibrous
histiocytoma. A case report. J. Arthroplasty, 7, 549–550.
41. Goodfellow, J. (1992) Malignancy and joint replacement (Editorial). J. Bone Joint Surg., 74A,
645.
42. McDougall, A. (1956) Malignant tumor at site of bone plating. J. Bone Joint Surg. 38B,
709–713.
43. Delgado, E.R. (1958) Sarcoma following a surgically treated fractured tibia: A case report.
Clin. Orthop. 12, 315–318.
44. Dube, V.E. and Fisher, D.E. (1972) Hemangioendothelioma of the leg following metallic fixa-
tion of the tibia. Cancer 30, 1260–1266.
45. Tayton, K.J.J. (1980) Ewing’s sarcoma at the site of a metal plate. Cancer 45, 413–415.
46. McDonald, I. (1981) Malignant lymphoma associated with internal fixation of a fractured
tibia. Cancer 48, 1009–1011.
620 M. Rock
47. Dodion, P., Putz, P., Amiri-Lamraski, M.H. et al. (1983) Immunoblastic lymphoma at the site
of an infected vitallium bone plate. Histopathol., 6, 807–813.
48. Lee, Y.S., Pho, R.W.H. and Nather, A. (1984) Malignant fibrous histiocytoma at site of metal
implant. Cancer 54, 2286–2289
49. Hughes, A.W., Sherlock, D.A., Hamblen, D.L. et al. (1987) Sarcoma at the site of a single hip
screw. A case report. J. Bone Joint Surg. 69B, 470–472.
50. Ward, J.J., Thornbury, D.D., Lemons, J.E. et al. (1990) Metal-induced sarcoma: A case report
and literature review. Clin. Orthop. 252, 299–306.
51. Khurana, J.S., Rosenberg, A.E., Kattapuram, S.V. et al. (1991) Malignancy supervening on an
intramedullary nail. Clin. Orthop. 267, 251–254.
52. Sinibaldi, K. (1976): Tumors associated with metallic implants in animals. Clin. Orthop. Rel.
Res., 118, 257–266.
53. Moss, A.J. (1991) Use of selected medical devices in the United States. Advance Data from
Vital and Health Statistics of the National Center for Health Statistics, 191, 1–24.
54. Gillespie, W.J., Frampton, C.M., Henderson, R.J. et al. (1988) The incidence of cancer follow-
ing total hip replacement. J. Bone Joint Surg. 70B, 539–542.
55. Visuri, T. (1992) Cancer risk after McKee-Farrar total hip replacement. Orthopedics, 14, 137–
142, 1992.
56. Nyren, O., McLaughlin, J.K. et al. (1995) Cancer risk after hip replacement with metal
implants: A population based study. J. National Can. lnst. 87, 28–33.
57. Li, X.Q., Stevenson, S., Hom, D.L. et al. (1993): Relationship between metallic implants and
cancer: A case-control study in a canine population. Vet. Comp. Orthop. Traumatol. 6, 70–74.
Chapter 6
Blood–material Interactions
S.R. Hanson
6.1 Introduction
There are several factors which have precluded the rational engineering design of
devices based on first principles. While thousands of materials have been put forward as
‘biocompatible’ or non-thrombogenic, based on in vitro studies and animal testing, the
relevance of these tests for outcomes in humans remains uncertain. Device responses in
vivo depend upon actual device configuration and resulting flow geometry as well as
upon intrinsic materials’ properties. In many applications, mechanical and physical
property requirements may dominate materials’ selection. For example, membranes
used in dialyzers and oxygenators must be both solute and gas permeable; chronic vas-
cular grafts and heart valves must be mechanically durable and chemically stable for
years; heart assist devices require flexible pumping chambers. Thus, the use of in vitro
assays or simplified in vivo flow geometries, as in many animal models, has not proven
adequate to predict actual device performance in patients. Furthermore, most animals
and humans, as individuals, differ markedly from one another in both blood chemistry
and in blood response to foreign materials [6]. It is deemed unethical to perform screen-
ing tests in humans, hence relatively few materials have undergone clinical evaluation
and only limited human comparative data are available. In the case of chronic implants,
devices removed at autopsy provide only a single set of observations which cannot be
related to dynamic blood–material interactions prior to explantation.
Another limitation is our recognition that all blood–material interactions of
clinical consequence are preceded by complex interactions between the biomate-
rial surface and circulating blood proteins. Plasma contains more that 100 identi-
fied proteins with specific functions and varying biologic properties [7]. These
proteins interact with surfaces in a complex, interdependent and time-dependent
fashion that remains poorly understood, except in low dilution, simplified model
systems [8]. These reactions may vary from individual to individual depending
upon coagulation status, the use antithrombotic or other drug therapies, or the
administration of contrast media for fluoroscopic imaging. A partial listing of
variables which may affect device outcomes following blood exposure is given
in Table 6.1.
Despite these limitations, the design engineer may be guided by previous success-
ful applications of materials in a variety of device configurations, and by certain
generalizations which have resulted from these studies. Devices which are com-
monly used include catheters, cannulae, guide wires, stents, shunts, vascular grafts,
heart valves, heart and ventricular assist devices, oxygenators, and dialyzers. With
respect to these devices it is important to consider those events which can lead to
serious clinical complications. These complications include: (1) thrombosis, (2)
thromboembolism, (3) consumption (ongoing destruction) or activation of circulat-
ing hemostatic blood elements, and (4) activation of inflammatory and immunologic
pathways. An appreciation for the biologic mechanisms of these events is essential
for understanding the blood-compatibility of devices, and may be briefly described
as follows. Thrombus forms as the localized accumulation of blood elements on,
within, or associated with a device, and thrombus which is actively deposited can
accumulate to the extent of producing device dysfunction or blood vessel occlusion.
Interruption of normal blood flow may produce ischemia (relative lack of oxygen)
and infarction (tissue death due to total oxygen deprivation) in distal circulatory beds
leading to heart attacks and strokes. Thrombus structure may be complex, and is
6 Blood–material Interactions 623
distinguished from that of whole blood clots which are often formed under static flow
conditions. Thus, clots are relatively homogeneous structures containing red blood
cells and platelets trapped in a mesh of polymerized protein (fibrin), while thrombus
formed under arterial flow conditions and high fluid shear rates (‘white thrombus’)
may be composed primarily of layers of fibrin and platelets (small procoagulant cells
occupying only about 0.3% of the total blood volume). Under conditions of low fluid
shear, as found in veins, thrombus may more closely resemble the structure of whole
blood clots (‘red thrombus’). Thromboembolism is the dislodgement by blood flow
of a thrombus which is then transported downstream, ultimately blocking vessels
which are too small for the thrombus to traverse. Thromboembolism is a common
cause of stroke (cerebrovascular infarction) and peripheral limb ischemia. Often the
balance between dynamic processes of thrombus deposition and its removal by
embolic and lytic mechanisms will produce platelet consumption (ongoing destruc-
tion) and a net reduction in circulating platelet levels. Other clotting factors may be
consumed as well [9]. Finally, certain devices, particularly those having large surface
areas, may activate enzyme systems (e.g., complement) leading to inflammatory or
immunologic responses [10]. With these issues in mind we will now review the per-
formance of various classes of biomaterials in actual device configurations.
These materials, which preferentially adsorb or absorb water (hydrogels), were ini-
tially postulated to be blood compatible based on the view that many naturally occur-
ing phospholipids, comprising the cell membranes of blood contacting tissues, are
also hydrophilic. Thus, water, as a biomaterial, was expected to show minimal inter-
action with blood proteins and cells [14, 15]. Interestingly, in animal studies highly
hydrophilic polymers based on acrylic and methacrylic polymers and copolymers, as
well as poly(vinyl alcohol) are all found to consume platelets excessively although
they accumulate little deposited thrombus [12, 16]. The materials have variable
thrombogenicity, but little capacity to retain adherent thrombus, i.e., they shed depos-
ited platelets as microemboli. Thus, while surface-grafted hydrogels (which are
mechanically weak) are currently used to improve catheter lubricity and as reservoirs
for drug delivery, they have not received widespread application for improving
blood-compatibility.
Another hydrophilic polymer that has received considerable attention is
poly(ethylene oxide) [17, 18]. While poly(ethylene oxide) surfaces have been
shown (like hydrogels) to have relatively low interactions with blood proteins and
cells in in vitro studies and in some animal models, the suitability of such poly-
6 Blood–material Interactions 625
mers for actual device applications and long-term implants has not been
established.
6.5 Metals
Metals, as a class, tend to be thrombogenic, and are most commonly applied in situ-
ations requiring considerable mechanical strength, such as in the struts of mechanical
heart valves and as endovascular stents [3, 19] or electrical conductivity, as in pacing
electrodes. Stents are metallic mesh devices placed within blood vessels to preserve
vessel patency after procedures to expand the vessel lumen diameter (e.g., after bal-
loon angioplasty). Metals most commonly employed are stainless steel (316L type)
and tantalum; however, both are thrombogenic [19, 20]. Catheter guide wire throm-
bogenicity, although readily documented, has been less of a clinical problem because
of the usually short period of blood exposure involved in most procedures.
In early canine implant studies, the thrombogenicity of a wide series of metallic
implants was seen to be related to the resting electrical potential of the metal which was
generated upon blood contact [21]. Metals with negative potentials tend to be antithrom-
bogenic, while stainless steel tends to be neutral. Copper, silver, and platinum are posi-
tive and extremely thrombogenic. Indeed, the use of copper coils inserted into canine
arteries continues to be a widely used model for inducing a thrombotic response [22].
Theoretically, reduced thrombogenicity of metallic stents and heart valve com-
ponents might be achieved by thin film polymer coatings, although the clinical
effectiveness of this strategy has not been demonstrated. Thus, chronic systemic
anticoagulation is generally employed in patients with prosthetic heart valves (with
metallic components) and stents.
6.6 Carbons
Cardiac valves with components fabricated from low temperature isotropic carbons
(pyrolytic carbon) are successfully used clinically [23]. These materials are appropriate
for such applications as mechanical valves which require long-term chemical inertness,
smoothness, and wear-resistance. The reasons for the marked improvement in the per-
formance (reduced thrombosis and thromboembolic stroke rates) of these newer vs.
older style heart valves are not entirely understood, but are undoubtedly multifactorial
and related to improved patient management and valve design, as well as to the nature
of the carbon surface. The specific benefits conferred by pyrolytic carbons with respect
to blood cell and protein interactions, resulting in a very low frequency of clinical com-
plications, remain to be defined. The use of carbon coatings has been proposed for other
devices, i.e., vascular grafts, although such devices have not yet been used clinically.
626 S.R. Hanson
Polymeric thin films of widely varying chemical composition may be deposited onto
polymers, metals, and other surfaces using the method of plasma polymerization (also
called ‘glow-discharge’ polymerization) [24]. This method is advantageous since very
thin films (e.g., 100 nm) may selectively modify the surface chemistry of devices, but
not their overall mechanical properties or surface texture. Plasma polymers form
highly cross-linked, covalent, inert barrier layers which may resist the adsorption of
proteins, lipids, and other blood elements. Plasma reacted coatings, based on hydro-
carbon monomers such as methane, may produce durable diamond like coatings.
Plasma polymer coatings have been proposed for vascular grafts and stents, based on
promising animal studies [25], but are not used clinically at the present time.
6.8 Membranes
Blood contacting membranes are used for gas exchange (e.g., blood oxygenators) or
solute exchange (e.g., dialyzers). The large membrane surface area, which may
exceed 2 m2, and the complexity of cardiopulmonary bypass circuits can produce
consumption and dysfunction of circulating blood elements such as platelets, lead-
ing to an increased risk of bleeding as well as thromboembolism [26]. The activa-
tion of inflammatory and immune response pathways (complement system) by
dialysis and oxygenator membranes may also produce significant complications
[27]. Complement activation by dialysis membranes may be related in part to the
availability of surface hydroxyl groups, particularly on cellulosic membranes.
Complement activation may be greatly attenuated by the use of other membrane
materials such as polysulfone and polycarbonate. Complement activation by bioma-
terial membranes has been reviewed [27].
The use of biological and bioactive molecules as device surface coatings may confer
thromboresistance. Such coating materials include phospholipids and heparin.
Phospholipids such as phosphorylcholine, a normal constituent of cell membranes,
may orient polar head groups towards the aqueous phase and locally organize water
molecules, much like hydrogel surfaces. These surfaces may minimize protein and
complement interactions [28]. In preliminary animal studies, phosphorylcholine
coated stents, guide wires, and vascular grafts have shown improved thromboresis-
tance. This approach is being actively developed for clinical applications.
6 Blood–material Interactions 627
6.11 Conclusion
Acknowledgement This work was supported by Research Grant HL 31469 from the Heart, Lung
and blood Institute, the National Institutes of Health.
Additional Reading
Colman, R.W., Hirsch, J., Marder, V.J. and Salzman, E.W. (eds)(1994) Hemostasis
and Thrombosis: Basic Principles and Clinical Practice, 3rd edn, J.B. Lippincott,
Philadelphia.
This book is highly recommended. This state-of-the-art text covers in detail
essentially all important hematological aspects of cardiovascular device blood com-
patibility. In particular, Chapter 76, Interaction of blood with artificial surfaces,
which considers many theoretical, experimental, and animal studies, and Chapter
77, Artificial devices in clinical practice, which describes clinical device thrombo-
embolic complications, are of great practical value.
Harker, L.A., Ratner, B.D. and Didisheim, P. (eds)(1993) Cardiovascular
Biomaterials and Biocompatibility, Cardiovascular Pathology, 2(3) (suppl.),
1S–224S.
In this volume, 20 chapters by expert authors treat all aspects of biomaterials and
blood compatibility including pathologic mechanisms, material characterization,
blood-material interactions and device performance. This volume updates an
excellent earlier book Guidelines for blood-Material Interactions, National
Institutes of Health, Washington, DC, Publication No. 85–2185 (1985).
Szycher, M. (ed.) (1983) Biocompatible Polymers, Metals, and Composites,
Technomic Publishing Co., Lancaster, Pennsylvania.
Many of the same issues of blood–material interactions are broadly covered
while selected polymer and device applications are described in additional detail. of
particular interest are Section I (Fundamental Concepts in blood/Material
Interactions) and Section II (Strategies for Hemocompatibility).
6 Blood–material Interactions 629
References
1. Salzman, E.W., Merrill, E.W. and Kent K.C. (1994) Interaction of blood with artificial surfaces,
in Hemostasis and Thrombosis: Basic Principles and Clinical Practice, 3rd edn, R.W. Colman,
J. Hirsch, V.J. Marder, and Salzman, E.W. (eds), J.B. Lippincott, Philadelphia, pp. 1469–85.
2. Harker, L.A., Ratner, B.D. and Didisheim, P. (eds) (1993) Cardiovascular Biomaterials and
Biocompatibility, Cardiovascular Pathology 2(3)(supplement), 1S–224S.
3. Szycher, M. (ed.) (1983) Biocompatible Polymers, Metals, and Composites, Technomic
Publishing Co., Lancaster, Pennsylvania.
4. Williams, D.F. (ed) (1981) Biocompatibility of Clinical Implant Materials, CRC Press, Boca
Raton, Florida.
5. Clagett, G.P. and Eberhart, R.C. (1994) Artificial Devices in Clinical Practice, in Hemostasis
and Thrombosis: Basic Principles and Clinical Practice, 3rd edn R.W. Colman, J. Hirsch,
V.J. Marder, and Salzman, E.W. (eds), J.B. Lippincott, Philadelphia, pp. 1486–1505.
6. Grabowski, E.F., Didisheim, P., Lewis, J.C. et al. (1977) Platelet adhesion to foreign surfaces
under controlled conditions of whole blood flow: human vs. rabbit, dog, calf, sheep, pig,
macaque, and baboon. Transactions - American Society for Artificial Internal Organs, 23,
141–51.
7. Lentner, C. (ed.) (1984) Geigy Scientific Tables (vol. 3): Physical Chemistry, Composition of
blood, Hematology, Somatometric Data, Ciby-Geigy, Basle.
8. Brash, J.L. and Horbett, T.A. (eds) (1987) Proteins at Interfaces. Physicochemical and
Biochemical Studies, American Chemical Society, Washington, DC.
9. Harker, L.A. and Slichter, S.J. (1972) Platelet and fibrinogen consumption in man. New
England J Med. 287(20), 999–1005.
10. Bennett, B., Booth, N.A., Ogston D. (1987) Potential interactions between complement, coag-
ulation, fibrinolysis, kinin-forming, and other enzyme systems, in: Haemostasis and
Thrombosis (2nd edn), A.L. bloom and D.P. Thomas (eds), Churchill Livingstone, New York,
pp. 267–82.
11. Andrew, M. (1995) Developmental hemostasis: relevance to thromboembolic complications in
pediatric patients. Thrombosis and Hemostasis, 74(1), 415–25.
12. Hanson, S.R., Harker, L.S., Ratner, B.D. et al. (1980) In vivo evaluation of artificial surfaces
using a nonhuman primate model of arterial thrombosis. J Laboratory Clinical Med. 95,
289–304.
13. Silver, J.H., Myers, C.W., Lim, F. et al. (1994) Effect of polyol molecular weight on the physi-
cal properties and haemocompatibility of polyurethanes containing polyethylene oxide macro-
glycols. Biomaterials 15(9), 695–704.
14. Hoffman, A.S. (1974) Principles governing biomolecular interactions at foreign surfaces.
J. Biomedical Materials Res. (Symp.) 5(1), 77–83.
15. Andrade, J.D., Lee, H.B., Jhon, M.S. et al. (1973) Water as a biomaterial. Transactions-
American Society for Artificial Internal Organs 19, 1–7.
16. Strzinar, I. and Sefton, M.V. (1992) Preparation and thrombogenicity of alkylated polyvinyl
alcohol coated tubing. J. Biomedical Materials Research 26, 577–92.
17. Merrill, E.W. (1993) Poly(ethylene oxide) star molecules: synthesis, characterization, and
applications in medicine and biology. J. Biomaterials Science, Polymer Edition 5(1–2), 1–11.
18. Llanos, G.R. and Sefton, M.V. (1993) Does polyethylene oxide possess a low thrombogenic-
ity? J. Biomaterials Science, Polymer Edition, 4(4), 381–400.
19. Sigwart, U., Puel, J., Mirkovitch, V. et al. (1987) Intravascular stents to prevent occlusion and
restenosis after transluminal angioplasty. New England J. Med, 316(12), 701–6.
20. Scott, N.A., Nunes, G.L., King, S.B. et al. (1995) A comparison of the thrombogenicity of
stainless steel and tantalum coronary stents. American Heart J., 129, 866–72.
21. Saywer, P.N., Stanczewski, B., Lucas, T.R., et al. (1978) Physical chemistry of the vascular
interface, in Vascular Grafts, P.N. Sawyer and M.J. Kaplitt (eds), Appleton -Century-Crofts,
New York, pp. 53–75.
630 S.R. Hanson
22. Rapold, H.J., Stassen, T., Van de Werf, F., et al. (1992) Comparative copper coil-induced
thrombogenicity of the internal mammary, left anterior descending coronary, and popliteal
arteries in dogs. Arteriosclerosis and Thrombosis, 12(5), 634–44.
23. Schoen, F.J. (1983) Carbons in heart valve prostheses: Foundations and clinical performance,
in M. Szycher (ed.), Biocompatible Polymers, Metals, and Composites, Technomic Publishing
Co., Lancaster, Pennsylvania, pp. 239–61.
24. Yasuda, H.K. (1985) Plasma Polymerization, Academic Press, Orlando.
25. Yeh, Y.S., Iriyama, T., Matsuzawa, Y., et al. (1988) blood compatibility of surfaces modified
by plasma polymerization. J. Biomedical Materials Research, 22, 795–818.
26. Harker, L.A., Malpass, T.W., Branson H.E., et al. (1980) Mechanism of abnormal bleeding in
patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associ-
ated with selective alpha-granule release. blood, 56(5), 824–34.
27. Hakim, R. (1993) Complement activation by biomaterials. Cardiovascular Pathology, 2(3)
(suppl), 187S-198S.
28. Yu, J., Lamba, N.M., Courtney J.M., et al. (1994) Polymeric biomaterials: Influence of phos-
phorylcholine polar groups on protein adsorption and complement activation. International
Journal of Artificial Organs, 17(9), 499–504.
29. Hardhammer, P.A., van Beusekom H.M., Emanuelsson, H.U., et al. (1996) Reduction in
thrombotic events with heparin-coated Palmaz-Schatz stents in normal porcine coronary arter-
ies. Circulation, 93(3), 423–30.
30. Schneider, P.A., Kotze, H.F., Heyns, A. duP., et al. (1989) Thromboembolic potential of syn-
thetic vascular grafts in baboons. J. Vascular Surgery, 10, 75–82.
31. Dasse, K.A., Poirier, V.L., Menconi, M.J., et al. (1990) Characterization of TCPS textured
blood-contacting materials following long-term clinical LVAD support. In: Cardiovascular
Science and Technology: Basic and Applied: II, JC Norman (ed.), Oxymoron Press, Boston,
MA, pp. 218–220.
32. Kormos, R.L., Armitage, J.M., Borovetz, H.S., et al. (1990) Univentricular support with the
Novocor left ventricular assist system as a bridge to cardiac transplantation: An update in
Cardiovascular Science and Technology: Basic and Applied: II, JC Norman (Ed), Oxymoron
Press, Boston, MA, pp. 322–324.
Chapter 7
Soft Tissue Response to Silicones
S.E. Gabriel
Although the term ’silicone’ refers to a group of organic silicone compounds, the
one most commonly used in medicine is composed of a polymer known as dimethy-
polysiloxane (DMPS). In silicone gel the polymer is cross-linked; the more cross-
linking, the more solid is the gel. Liquid silicone consists of glucose-linked DMPS
polymer chains. Silicones first became commercially available in 1943, with the
first subdermal implantation of silicone occurring in the late 1940s [1–3]. Silicones
have since been developed for a wide variety of medical applications, most notably
in joint and breast prostheses.
There is a large body of literature attesting to the chemical and physical inertness
of silicone [4–12]. Recently, there has been increasing interest in the possible
adverse effects of silicones used in implantation. Much of the literature describing
the adverse effects of silicone has been in reference to direct silicone injection.
The following discussion will review the immunologic effects of prostheses used in
breast reconstruction and augmentation.
Systemic reactions have been reported following the introduction of silicone into
the body. In one instance, a severe systemic reaction consisting of a febrile illness,
acute renal insufficiency, respiratory compromise, pulmonary infiltration, delirium,
anemia, and thrombocytopenia has been reported following implantation of a sili-
cone gel envelope prosthesis. Improvement followed implant removal. Silicone was
identified by mass spectrophotometry in this case [36]. Another case involved the
injection of a large quantity of free silicone under the breasts by an unauthorized
individual. The patient expired within 10 hours of injection. Silicone was identified
7 Soft Tissue Response to Silicones 633
It has been estimated that two million American women have undergone breast
augmentation or reconstruction since the introduction of the silicone gel-filled elas-
tomer envelope-type prosthesis in the early 1960s [60, 61]. The reported cases of
systemic sclerosis among this population raise the important question of whether
the association between systemic sclerosis and silicone breast implants is a real one.
Unfortunately, almost all of the evidence to date is derived from case reports, which
are the very weakest form of data bearing on the question of causality (Table 7.2).
Indeed, the most important evidence for establishing a cause-effect relationship is
the strength of the research design used to study that relationship [62]. Randomized
control trials provide the strongest evidence but are seldom ethical in studies of
causation because they involve randomly assigning individuals to receive or not to
receive a potentially harmful intervention. In addition, the long latent periods and
large numbers of subjects needed to answer most cause and effect questions in clini-
cal medicine make it impractical to utilize this research design.
Well conducted prospective cohort studies are the next strongest design because
they minimize the effects of selection bias, measurement bias, and known confound-
ers. Such a study would involve following a large population of women, preferably
for one or more decades, looking for the outcomes of interest (e.g., connective tissue
disorders). A relative risk for connective tissue disorders among those women who
elect to have breast implantation compared to those who do not can then be calcu-
lated. Although this is a powerful research design, it is usually impractical because
of the necessary long follow-up period. This problem can be circumvented by a ret-
Table 7.1 Cases of Systemic Sclerosis after Augmentation Mammoplasty Reported in the English-Language Literature 7
Extent of
Age at Age at Interval to Systemic Raynaud Antinuclear
Patient Diagnosis, y Mammoplasty, y Type of Implant Onset, y Sclerosis Phenomenon Systemic Involvement Antibodies* Reference
1 52 50 Silicone bag gel 2 Diffuse No No — (55)
2 41 20 Silicone bag gel 21 Diffuse Yes Pulmonary, + (55)
gastrointestinal,
3 63 53 Silicone bag gel 10 Limited Yes Pulmonary, + (55)
4 37 32 Silicone bag gel 5 Diffuse No Pulmonary, + (55)
5 45 25 Paraffin injection 19 Diffuse Yes No — (42)
6 49 20 Paraffin injection 16 Diffuse Yes Pulmonary, — (42)
gastrointestinal
7 51 25 Paraffin injection 17 Limited Yes No — (43)
Soft Tissue Response to Silicones
- negative.
+ positive.
* Determined using immunofluorescence.
636 S.E. Gabriel
rospective cohort study, which is similar with the exception that the population and
the exposure (breast implantation) is identified in the past, allowing the patients to be
followed to the present for the outcomes of interest. Although this is a very attractive
research design, it requires that the complete exposed and unexposed populations be
identifiable and that follow-up information be available on all individuals.
Case-control studies retrospectively compare the frequency of breast implanta-
tion in women with and without the outcomes of interest. If, for example, connec-
tive tissue disorders were more likely to occur among women with breast implants,
this would constitute some evidence for causation. Case-control studies typically
require less time and resources than cohort studies. However, they are susceptible to
many more biases than cohort studies [62]. The primary reason not to perform a
case-control study here, however, is that a separate case-control study would be
required for each of the outcomes of interest, i.e., a case-control study of systemic
sclerosis, a case-control study of rheumatoid arthritis, etc. The retrospective cohort
design is much more efficient since it can evaluate multiple outcomes in a single
study, as is the need here.
A set of eight criteria has been proposed as a guide to formulate decisions
regarding cause and effect relationships (Table 7.3). The relationship between
breast implants and connective tissue disorders does fulfill the criterion of tempo-
rality since, at least in the published case reports, the connective tissue disorders
all followed breast implantation. There is no evidence describing the magnitude of
the relative risk in this relationship. There is also no evidence for a dose response
relationship, i.e., that women with bilateral implants perhaps have an increased
likelihood of connective tissue disorders compared to women with unilateral
7 Soft Tissue Response to Silicones 637
implants. The evidence regarding the reversibility of these disorders with removal
of implants is variable. Although there have been some reports of improvement of
connective tissue disorders following removal of the implants, this is not consis-
tent and the number of patients involved is small. The relationship does appear to
be consistent, i.e., it has been observed repeatedly by different persons in different
places, circumstances, and times, however it has not yet been assessed using ade-
quate study designs. Perhaps the most compelling evidence is the biologic plausi-
bility of this relationship due to the hypothesis of silicone acting as an immune
adjuvant. The relationship does not appear specific, as silicone implants have lead
to not just one effect, but several, albeit somewhat related, effects. Finally, a cause
and effect relationship is strengthened if there are examples of well established
causes that are analogous to the one in question. Adjuvant induced arthritis can be
considered analogous [63].
In summary, in spite of the anecdotal evidence, until very recently there was a
lack of evidence to either support or refute a cause-and-effect relationship between
silicone breast implants and connective tissue/ autoimmune disorders.
At least seven controlled studies have now been published (Table 7.4), each of
which provided a quantitative assessment of the risk of connective tissue diseases
among women with breast implants [64–70]. The first of these was a case-control
study of augmentation mammoplasty and scleroderma [68]. The aims of this study
were to compare the frequency and temporal relationship of augmentation mam-
moplasty among scleroderma cases and matched controls. Scleroderma patients and
age stratified general practice controls were interviewed using a pretested telephone
questionnaire. Self-reported dates of augmentation mammoplasty were ascertained
as were dates of scleroderma symptoms and diagnoses as relevant. Frequencies of
noriaugmentation mammoplasty silicone exposure between interviewed cases and
controls were expressed in terms of rate ratios and 95% confidence intervals. Rate
ratios were also adjusted for socioeconomic status.
A total of 315 cases and 371 controls were interviewed, of whom 251 and 289,
respectively, were female. The unadjusted rates for augmentation mammoplasty
among interviewed cases and controls were 4/251 (1.59%) and 5/289 (1.73% ),
respectively. The socioeconomic status adjusted rate of augmentation mammo-
plasty in scleroderma patients was 1.54% (95% CI: 0.03–3.04) which is very
similar to the 1.73% rate in interviewed controls. These results indicate that aug-
mentation mammoplasty rates were comparable in cases and controls. In addi-
tion, the rates of exposure to nonmammoplasty silicone mastectomy and breast
lumpectomy were comparable in interviewed cases and controls. This study
failed to demonstrate an association between silicone breast implantation and the
Table 7.4 Summary of controlled studies examining the relationship between breast implants (BI) and connective tissue diseases (CTD)
638
Study Population
Cases Controls Outcome(s)
Reference Study Design (exposed) (unexposed) Examined Main Result Conclusions
68 Case control 315 371 Systemic sclerosis Rates of BI among cases and controls Rates of BI were similar in cases
(SS, scleroderma) were 1.59% and 1.73% and controls.
65 Retrospective 749 1498 Connective Relative risk (cases:controls) of There was no association
cohort tissue and other developing any of these diseases was between BI and the connective
autoimmune diseases 1.06 (95% CI: 0.34–2.97). tissue and other disorders studied.
67 Case control 195 143 Systemic lupus One (0.8%) of the SLE cases and 0 (0%) No association was shown
erythematosus (SLE) of the controls reported having a BI between BI and SLE.
(p=0.57).
64 Case control 349 1456 Rheumatoid Relative risk for a history of BI No increased risk for RA among
arthritis (RA) (cases:controls) was 0.41 (95% CI: women with BI was
0.05–3.13). demonstrated.
60 Multi-center 869 2061 SS (scleroderma) Odds ratio for BI surgery (cases:controls) No significant causal relationship
case control was 1.25 (95% CI: 0.62–2.53). was demonstrated between BI
and the development of SS.
70 Nested case 121700 Connective tissue Five cases with BI were identified among No association was found
control disease 300 patients with RA;0 cases with BI between BI and CTD.
among 123 with SLE, 20 patients with
SS, 3 with Sjögren's syndrome, 13 with
dermato/polymyositis, and 2 with mixed
connective tissue disease.
69 Case control 592 1184 SS (scleroderma) Odds ratio for BI (cases:controls) was No significant association
(preliminary 0.61 (95% CI: 0.14–2.68) between BI and SS was found.
results)
CI = confidence interval
S.E. Gabriel
7 Soft Tissue Response to Silicones 639
power of the study was only able to provide sufficient evidence against a very large
association.
Three additional controlled studies have been presented and are published in
abstract form [64, 66, 70]. As part of a prospective case-control study of the risk of
rheumatoid arthritis, Dugowson et al. recruited 349 women with new-onset rheu-
matoid arthritis and 1456 similarly aged control women. Information about breast
implants was obtained on both cases and controls and age-adjusted risk for a history
of breast implants among cases was compared to that of controls. The relative risk,
i.e., comparing the rate of a history of breast implants among rheumatoid arthritis
cases compared to a similar history among controls, was 0.41 (95% CI: 0.05–3.13)
[64]. These data did not support an increased risk for rheumatoid arthritis among
women with silicone breast implants.
A multi-center, case-controlled study was performed to examine the association
between scleroderma and augmentation mammoplasty [66]. A total of 869 women
with systemic sclerosis recruited from three university-affiliated rheumatology
clinics and 2061 local community controls matched on age in three strata (ages
25–44, 45–64, and ≥65); race and sex were identified by random-digit dialing. Data
on exposure and potential confounding variables were collected from cases and
controls by self-administered questionnaires and telephone interviews, respec-
tively. The frequency of breast implant surgery was compared in both groups and
the odds ratio and 95% confidence intervals for the association of augmentation
mammoplasty with systemic sclerosis, adjusted for age, race, marital status, and
site, was 1.25 (95% CI: 0.62–2.53). These data failed to demonstrate a significant
causal relationship between augmentation mammoplasty and the development of
systemic sclerosis.
Using a nested case-control study, Sanchez-Guerrero et al. examined the asso-
ciation between silicone breast implants and connective tissue diseases among a
cohort of 121 700 registered American nurses followed since 1976 [70]. In 1992, a
questionnaire was sent to nurses who had reported any rheumatic disease from
1976 to 1990 asking about rheumatic symptoms and silicone exposure. The com-
plete medical records were obtained on all participants who confirmed any rheu-
matic or musculoskeletal symptoms. Connective tissue disease cases were classified
according to the American College of Rheumatology or other published criteria.
Ten age-matched controls per case were randomly selected among nurses with no
rheumatic or musculoskeletal complaints. Odds ratios and 95% confidence inter-
vals were used as a measure of association. This study identified 448 cases with
definite connective tissue diseases and 1209 nurses with silicone breast implants.
Five patients had silicone breast implants and any connective tissue disease. The
mean time since implantation was 139 ± 95.81 months among these five patients
and 119.17 ± 76.64 months among all nurses with silicone breast implants. These
five patients were identified among 300 patients with rheumatoid arthritis. No case
with a silicone breast implant was identified among 123 patients with systemic
lupus erythematosus, 20 patients with scleroderma, three with Sjögren’s syndrome,
13 with dermato/polymyositis, and two with mixed connective tissue diseases. In
7 Soft Tissue Response to Silicones 641
conclusion, this study found no association between silicone breast implants and
connective tissue diseases.
Finally, Burns and Schottenfeld are conducting a population-based casecontrol
study examining the relationship between this condition and prior history of breast
implant surgery [69, 72]. The cases and normal population controls are being
assembled from the states of Michigan and Ohio. The cases are being identified
using several sources: a computerized data base of hospital discharge diagnostic
listings during the period 1980–1991; a collaborative network of major medical
centers; a postal survey of certified rheumatologists; and from patient members of
the United Scleroderma Foundation. Although this study is still underway, prelimi-
nary results demonstrate a crude odds ratio of 0.61 (95% CI: 0.14–2.68) for breast
implants among cases compared to controls. These results do not support a causal
relationship between breast implants and systemic sclerosis.
In summary, although numerous anecdotal case reports have suggested an asso-
ciation between silicone breast implants and connective tissue diseases, all seven
controlled epidemiologic studies conducted to date have failed to confirm such an
association. Whether silicone breast implants cause a new and previously unde-
scribed condition is yet to be determined.
Reference
1. Williams, D.F. (1981) Biocompatibility of clinical implant materials. Vol 2. Boca Raton: CRC
Press, pp. 1–272.
2. Blocksma. R. and Braley, S. (1965) The silicone in plastic surgery. Plast. Reconstr. Surg.,
35(4), 366–370.
3. Braley, S.A., Jr (1964) The medical silicones. Trans. Am. Soc. Artif Int. Organs, 10,
240–243.
4. Weiner, S.R. and Paulus, H.E. (1986) Chronic arthropathy occurring after augmentation mam-
maplasty. Plast. Reconstr. Surg., 77(2), 185–192.
5. Ward, T.C. and Perry, J.T. (1981) Dynamic mechanical properties of medical grade silicone
elastomer stored in simulated body fluids. J. Biomed. Mater. Res., 15 511–525.
6. Homsy, C.A. (1970). Bio-compatibility in selection of materials for implantation. J. Biomed.
Mater. Res., 4, 341–356.
7. Rigdon, R.H. and Dricks, A. (1975) Reaction associated with a silicone rubber gel: An experi-
mental study. J. Biomed. Mater. Res., 9, 645–659.
8. Robertson, G. and Braley, S. (1973) Toxicologic studies, quality control, and efficacy of the
silastic mammary prosthesis. Med. Instrum., 7(2), 100–103.
9. Boone, J.L. and Braley, S.A. (1966) Resistance of silicone rubbers to body fluids. Rubber
Chern. Technol. , 39(4), 1293–1297.
10. Boone, J.L. (1966) Silicone rubber insulation for subdermally implanted devices. Med. Res.
Eng., 5, 34–37.
11. Roggendorf, E. (1976) The biostability of silicone rubbers, a polyamide, and a polyester.
J. Biomed Mater. Res., 10, 123–143.
12. Speirs, A.C. and Blocksma, R. (1963) New implantable silicone rubbers. An experimental
evaluation of tissue response. Plast. Reconstr. Surg., 31(2), 166–175.
13. Symmers, W.StC. (1968) Silicone mastitis in ’topless’ waitresses and some other varieties of
foreign-body mastitis. Br. Med 1., 3, 19–22.
642 S.E. Gabriel
14. Ortiz-Monasterio, F. and Trigos, I. (1972) Management of patients with complications from
injections of foreign materials into the breasts. Plast. Reconstr. Surg. , 50(1) 42–47.
15. Parsons, R.W. and Thering, H.R. (1977) Management of the silcone-injected breast. Plast.
Reconstr. Surg., 60(4), 534–538.
16. Wilkie, T.F. (1977) Late development of granuloma after liquid silicone injections. Plast.
Reconstr. Surg., 60(2), 179–188.
17. Pearl, R.M., Laub, D.R., and Kaplan, E.N. (1978) Complications following silicone injections
for augmentation of the contours of the face. Plast Reconstr Surg, 61(6): 888–891.
18. Kozeny, G.A., Barbato, A.L., Bansal, V.K. et al. (1984) Hypercalcemia associated with
silicone-induced granulomas. N. Engl. J. Med., 311(17), 1103–1105.
19. Andrews, J.M. (1966) Cellular behavior to injected silicone fluid: A preliminary report. Plast.
Reconstr. Surg, 38(6), 581–583.
20. Ballantyne, D.L., Rees, T.D. and Seidman, I. (1965) Silicone fluid: Response to massive sub-
cutaneous injections of dimethylpolysiloxane fluid in animals. Plast. Reconstr. Surg., 36(3),
330–338.
21. Ben-Hur, N. and Neuman, Z. (1965) Siliconoma - another cutaneous response to dimethylpo-
lysiloxane. Experimental study in mice. Plast. Reconstr. Surg., 36(6), 629–631.
22. Rees, T.D., Platt, J., and Ballantyne, D.L. (1965) An investigation of cutaneous response to
dimethylpolysiloxane (silicone liquid) in animals and humans - A preliminary report. Plast.
Reconstr. Surg., 35(2), 131–139.
23. Travis, W.D., Balogh, K. and Abraham, J.L. (1985) Silicone granulomas: Report of three cases
and review of the literature. Hum. Pathol., 16(1), 19–27.
24. Mason, J. and Apisarnthanarax, P. (1981) Migratory silicone granuloma. Arch. Dermatol., 117,
366–367.
25. Apesos, J. and Pope, T.L., Jr (1985) Silicone granuloma following closed capsulotomy of
mammary prosthesis. Ann. Plast. Surg., 14(5), 403–406.
26. Rees, T.D., Ballantyne, D.L., Seidman, I. et at. (1967) Visceral response to subcutaneous and
intraperitoneal injections of silicone in mice. Plast. Reconstr. Surg., 39(4), 402–410.
27. Chastre, J., Basset, F., Viau, F. et al. (1983) Acute pneumonitis after subcutaneous injections
of silicone in transsexual men. N. Engl. J. Med. , 308(13), 764–767.
28. Ellenbogen, R. and Rubin, L. (1975) Injectable fluid silicone therapy: Human morbidity and
mortality. J. Am. Med. Assoc. , 234(3), 308–309.
29. Murakata, L.A. and Rangwala, A.F. (1989) Silicone lymphadenopathy with concomitant
malignant lymphoma. J. Rheumatol., 16(11), 1480–1483.
30. Christie, A.J., Weinberger, K.A. and Dietrich, M. (1977) Silicone lymphadenopathy and syno-
vitis. Complications of silicone elastomer finger joint prostheses. J. Am. Med. Assoc., 237(14),
1463–1464.
31. Kircher, T. (1980) Silicone lymphadenopathy. A complication of silicone elastomer finger
joint prostheses. Hum. Pathol., 11(3), 240–244.
32. Benjamin, E. and Ahmed, A. (1982) Silicone lymphadenopathy: A report of two cases, one
with concomitant malignant lymphoma. Diagn. Histopathol., 5, 133–141.
33. Digby, J.M. (1982) Malignant lymphoma with intranodal silicone rubber particles following
metacarpophalangeal joint replacements. Hand, 14(3), 326–328.
34. Endo, L.P., Edwards, N.L., Longley, S. et al. (1987) Silicone and rheumatic diseases. Semin.
Arthritis Rheum., 17(2), 112–118.
35. Varga, J., Schumacher, H.R. and Jimenez, S.A. (1989) Systemic sclerosis after augmentation
mammoplasty with silicone implants. Ann. Intern. Med., 111(5), 377–383.
36. Uretsky, B.F., O’Brien, J.J., Courtiss E.H. et al. (1979) Augmentation mammaplasty associ-
ated with a severe systemic illness. Ann. Plast. Surg., 3(5), 445–447.
37. Nosanchuk, J.S. (1968) Injected dimethylpolysiloxane fluid: A study of antibody and histo-
logic response. Plast. Reconstr. Surg. , 42(6), 562–566.
38. Heggers, J.P., Kossovsky, N. and Parsons, R.W. et al. (1983) Biocompatibility of silicone
implants. Ann. Plast. Surg., 11, 38–45.
7 Soft Tissue Response to Silicones 643
39. Pernis, B. and Paronetto, F. (1962) Adjuvant effect of silica (tridymite) on antibody produc-
tion. Proc. Soc. Exp. Biol. Med., 110, 390–392.
40. Miyoshi, K., Miyamura, T. and Kobayashi, Y. (1964) Hypergammaglobulinemia by prolonged
adjuvanticity in man. Disorders developed after augmentation mammaplasty. Jpn Med. J.,
2122, 9.
41. Yoshida, K. (1973) Post mammaplasty disorder as an adjuvant disease of man. Shikoku Acta
Med., 29, 318.
42. Kumagai, Y., Abe, C. and Shiokawa, Y. (1979) Scleroderma after cosmetic surgery. Four cases
of human adjuvant disease. Arthritis Rheum., 22(5), 532–537.
43. Kumagai, Y., Shiokawa, Y. and Medsger, T.A., Jr (1984) Clinical spectrum of connective tissue
disease after cosmetic surgery. Arthritis Rheum., 27(1), 1–12.
44. Fock, K.M., Feng, P.H., Tey, B.H. (1984). Autoimmune disease developing after augmentation
mammoplasty: Report of 3 cases. J. Rheumatol, 11(1), 98–100.
45. Okano, Y., Nishikai, M. and Sato, A. (1984) Scleroderma, primary biliary cirrhosis, and
Sjogren’s syndrome after cosmetic breast augmentation with silicone injection: A case report
of possible human adjuvant disease. Ann. Rheum. Dis., 43, 520–522.
46. van Nunen, S.A., Gatenby, P.A. and Basten, A. (1982) Post-mammoplasty connective tissue
disease. Arthritis Rheum., 25(6), 694–697.
47. Baldwin, C.M., Jr, and Kaplan, E.N. (1983) Silicone-induced human adjuvant disease? Ann.
Plast. Surg., 10, 270–273.
48. Byron, M.A., Venning, V.A. and Mowat, A.G. (1984) Post-mammoplasty human adjuvant dis-
ease. Br. J. Rheumatol., 23(3), 227–229.
49. Vargas, A. (1979) Shedding of silicone particles from inflated breast implants. Plast. Reconstr.
Surg., 64(2), 252–253.
50. Walsh, F.W., Solomon, D.A., Espinoza, L.R. et al. (1989) Human adjuvant disease. A new
cause of chylous effusions. Arch. Intern. Med. , 149, 1194–1196.
51. Ziegler, V.V., Haustein, U.F., Mehlhorn, J. et al. (1986) Quarzinduzierte sklerodermie
sklerodermie-ahnliches syndrom oder echte progressive sklerodermie? Dermatol. Mon. Schr.,
172(2), 86–90.
52. Taura, N., Usa, T. and Eguchi, K. (1976) A case of human adjuvant disease. J. Jap. Soc. Intern.
Med., 65, 840–841.
53. Alusik, S., Jandova, R. and Gebauerova, M. (1989) Anticardiolipin syndrome in plastic sur-
gery of the breast. Cor. Vasa, 31(2), 139–144.
54. Olesen, L.L., Ejlertsen, T. and Nielsen, J. (1991) Toxic shock syndrome following insertion of
breast prostheses. Br. J. Surg., 78(5), 585–586.
55. Spiera, H. (1988) Scleroderma after silicone augmentation mammoplasty. J. Am. Med. Assoc.,
260(2), 236–238.
56. Brozena, S.J., Fenske, N.A., Cruse, C.W. et al. (1988) Human adjuvant disease following
augmentation mammoplasty. Arch. Dermatol., 124, 1383–1386.
57. Gutierrez, F.J. and Espinoza, L.R. (1990) Progressive systemic sclerosis complicated by severe
hypertension: Reversal after silicone implant removal. Am. J. Med., 89(3), 390–392.
58. Hitoshi, S., Ito, Y., Takehara, K. et al. (1991) A case of malignant hypertension and sclero-
derma after cosmetic surgery. Jpn J. Med., 30(1), 97–100.
59. Sahn, E.E., Garen, P.D., Silver, R.M. et al. (1990) Scleroderma following augmentation mam-
moplasty. Report of a case and review of the literature. Arch. Dermatol., 126, 1198–1202.
60. Deapen, D.M., Pike, M.C., Casagrande, J.T. et al. (1986) The relationship between breast
cancer and augmentation mammaplasty: An epidemiologic study. Plast. Reconstr. Surg.,
77(3), 361–367.
61. May, D.S. and Stroup, N.E. (1991) The incidence of sarcomas of the breast among women in
the United States, 1973–1986. Plast. Reconstr. Surg., 87(1), 193–194.
62. Fletcher, R.H., Fletcher, S.W. and Wagner, E.H. (1988) Cause. In: Collins, N., Eckhart, C.,
Chalew, G.N., eds. Clinical Epidemiology: The Essentials. Baltimore: Williams & Wilkens,
pp. 208–225.
644 S.E. Gabriel
63. Pearson, C.M. (1963) Experimental joint disease. Observations on adjuvant-induced arthritis.
J. Chron. Dis., 16, 863–874.
64. Dugowson, C.E., Daling, J., Koepsell, T.D. et al. (1992) Silicone breast implants and risk for
rheumatoid-arthritis. Arthritis Rheum., 35, S66. (Abstract)
65. Gabriel, S.E., O’Fallon, W.M., Kurland, L.T. et al. (1994) Risk of connectivetissue diseases
and other disorders after breast implantation. N. Engl. J. Med., 330, 1697–1702.
66. Hochberg, M.C., Perlmutter, D.L., White, B. et al. (1994) The association of augmentation
mammoplasty with systemic sclerosis: Results from a multi-center case-control study. Arthritis
Rheum., 37 (Supplement): S369 (Abstract).
67. Strom, B.L., Reidenberg, M.M., Freundlich, B. et al. (1994) Breast silicone implants and risk
of systemic lupus erythematosus. J. Clin. Epidemiol. (in press).
68. Englert, H.J., Brooks, P. (1994) Scleroderma and augmentation mammoplasty - a causal rela-
tionship? Aust. NZ J. Med., 24, 74–80.
69. Burns, C.J., Laing, T.J., Gillespie, B.W. et al. (1996) The epidemiology of scleroderma among
women: Assessment of risk from exposure to silicone and silica. J. Rheumatol., 23(11),
1904–1911.
70. Sanchez-Guerrero, J., Karlson, E.W., Colditz, G.A. et al. (1994) Silicone breast implants (SBI)
and connective tissue disease (CTD). Arthritis Rheum, 37 (Supplement): S282 (Abstract).
71. Tan, E.M., Cohen, AS., Fries, J.F. et al. (1982) The 1982 revised criteria for the classification
of systemic lupus erythematosus. Arthritis Rheum, 25 (11), 1271–1277.
72. Schottenfeld, D., Burns, C.J., Gillespie, B.W. et al. (1995) The design of a population-based
case-control study of systemic sclerosis (scleroderma): Commentary on the University of
Michigan Study. J. Clin. Epidemiol. 48(4), 583–586.
Chapter 8
Vocal Folds
8.1 Introduction
Vocal folds are two infoldings of complex tissue housed in the larynx whose vibration,
known as phonation, results in voice. To create voice, vocal folds are adducted, effec-
tively closing the glottis or the space between them creating a pressure difference
between the top and bottom of the tissue. This pressure difference, combined with
Bernoulli forces and tissue elasticity, pushes the vocal folds apart and brings them
back together again [1]. This rapid, sustained opening and closing cycle produces
vocal fold vibration at various fundamental frequencies ranging from 0 to 300 Hz for
normal vocal levels [2]. The vocal folds are multilayer structures that consist of
muscle, lamina propria, and epithelium. Fundamental frequency is controlled by
laryngeal muscles which alter vocal fold physical properties, such as length and thick-
ness. The extracellular matrix (ECM) of the lamina propria contributes significantly to
vocal quality. Disruption of vocal fold vibration and/or quality through muscular
dysfunction, airflow disruption, or tissue damage results in disturbance of normal
voice production. Vocal fold scarring, a prevalent vocal fold injury, is characterized by
pathophysiologic changes to the lamina propria ECM, directly causing a marked
decrease in voice quality [3].
J. Gaston
Department of Biomedical Engineering, University of Wisconsin–Madison,
5118 WIMR, 1111 Highland Ave., Madison, WI 53705, USA
S.L. Thibeault (*)
Department of Surgery, University of Wisconsin–Madison,
5107 WIMR, 1111 Highland Ave., Madison, WI 53705, USA
Department of Biomedical Engineering, University of Wisconsin–Madison,
5107 WIMR, 1111 Highland Ave., Madison, WI 53705, USA
Department of Communication Sciences and Disorders, University of Wisconsin–Madison,
5107 WIMR, 1111 Highland Ave., Madison, WI 53705, USA
e-mail: [email protected]
8.2 Composition
The vocal fold can be divided into three layers, divided primarily by tissue function.
The top layer consists of squamous cell epithelium, which separates the underlying
tissue from the exposed airway. Below that lies the primary vibratory region, the
lamina propria. The cellular and protein composition of this layer determines the
ease of phonation. The lamina propria can be further subdivided into three layers,
superficial, intermediate, and deep lamina propria (Fig. 8.1). The most interior sec-
tion of the vocal fold consists of the thyroarytenoid muscle (muscularis), which
relaxes and shortens the vocal folds. The large majority of biomechanical character-
ization has been completed on the lamina propria for biomaterial applications; sub-
sequently the remainder of this chapter concentrates on the properties of this specific
layer. The most important proteins, glycosaminoglycan (GAG), and proteoglycans
in the lamina propria, as well as their concentrations at different depths, can be
found in Table 8.1.
SUPERFICIAL
INTERMEDIATE
DEEP
EPITHELIUM
MUSCULARIS
100 80 60 40 15 0
Fig. 8.1 Schematic of the divisions of the vocal fold, including lamina propria subdivisions.
Reprinted with permission from [4]
8 Vocal Folds 647
Table 8.1 Lamina propria ECM components and their relative concentrations with respect to
tissue depth
Superficial lamina Intermediate Deep lamina
ECM component propria lamina propria propria
Fibrous proteins Collagen I Low High Highest
Collagen III Low High Highest
Elastin Low Highest High
Fibronectin Highest Low Low
GAG Hyaluronic acid Very low Highest Low
Proteoglycans Decorin High Very low High
Biglycan Intermediate Very low Highest
Versican Very low High High
8.2.1 Collagen
The primary protein component of the vocal fold lamina propria is collagen, which
constitutes roughly 43 % of the total tissue protein by weight [5]. Immunohistological
staining shows that collagen I and collagen III are the primary collagens of the lamina
propria. Both collagen types show the highest staining in the deep lamina propria, close
to the thyroarytenoid muscle, with the intermediate layer having slightly lower staining
intensity. The high fibrillar collagen content imparts strength and resiliency to the vocal
folds. In general, the collagen fibers are spatially oriented from anterior to posterior,
allowing them to support the force applied by the intrinsic laryngeal muscles [6].
8.2.2 Elastin
Elastin plays a crucial role in vocal fold vibration and the mechanical properties of
the lamina propria. Accounting for approximately 8.5 % of the total lamina propria
tissue protein by weight, it is mainly responsible for imparting elastic recoil to the
human vocal folds [7]. Like collagen, higher concentrations of elastin are observed
in the deep and intermediate layers of the lamina propria, with the concentration
decreasing closer to the epithelial layer. Fibrillin, another elastic protein, appears to
have deposition opposite that of elastin, with the highest concentration in the super-
ficial lamina propria and the lowest concentration in the intermediate layer [7].
propria [7]. The high concentration has a significant effect on vocal fold mechanical
properties. The long chain length and loosely coiled molecular structure of HA
allow it to sequester large volumes of water, resulting in high tissue viscosity. The
high tissue viscosity acts as a shock absorber, resisting tissue compression and dam-
age caused during phonation.
Several other proteins play prominent roles in the vocal fold lamina propria. These
small proteins contribute to ECM organization and typically are found at different
levels in the lamina propria. Fibronectin plays a key role in both matrix organization
and interaction with resident fibroblasts [8]. It is primarily located in the superficial
layer of the lamina propria and the basement membrane that divides the epithelium
from the superficial layer of the lamina propria. Decorin, a small proteoglycan, is
also predominantly found in the superficial lamina propria [9]. It plays a role in col-
lagen fibril assembly [10] and may contribute to the low density of collagen fibers
in the superficial lamina propria. In addition, the ability of decorin to modulate
stress transmission through the extracellular matrix is crucial for vocal fold phona-
tion. Biglycan, a small proteoglycan with a similar role to decorin, is mainly located
in the deep lamina propria. Within the deep lamina propria, it appears to be most
dense near the intermediate layer [9]. Versican, a large chondroitin sulfate proteo-
glycan, is found primarily in the intermediate lamina propria [9]. It is often found
colocalized with HA, though it is capable of binding with a variety of partners.
Extracellular proteins such as fibrillin, fibronectin, and collagen I are capable of
binding versican, and may affect cell adhesion to these structural proteins.
Scarring is the leading cause of poor voice quality after vocal fold surgery. The tis-
sue fibrosis defined by scarring results in disorganized ECM architecture and altered
mechanical properties. Controlled animal studies show that the major fibrous ECM
proteins, collagen and elastin, are altered following scarring of the lamina propria
(Table 8.2). To date, only one ten-patient study has analyzed the histology of scarred
vocal folds [15]. Collagen deposition in the scarred lamina propria was found to be
related to injury depth, with thicker fiber deposition correlated with deeper injury
depth. Elastin was absent or fragmented and disorganized in all but one patient,
consistent with animal studies. HA deposition followed a similar pattern, with HA
being unobservable in all but one patient, where staining was barely visible. Finally,
a quantitative assessment of the collagen, elastin, and HA content could not be per-
formed, as the study did not have a normal control from the patients.
8 Vocal Folds 649
Table 8.2 Review of animal studies detailing vocal fold injury with histological and compositional
changes
Animal Type of Timeline
model injury post-injury Collagen Elastin HA
Rabbit [11] Lamina 60 days Disorganized Short, compact Density similar
propria fibers, less dense fibers. Less to normal
forceps than normal dense than controls
biopsy controls normal controls
Rabbit [13] Lamina 6 months Organized fibers, Fibers Density similar
propria thicker than fragmented and to normal
forceps normal control disorganized control
biopsy
Canine [14] Vocal fold 2 months Disorganized Disorganized Density similar
stripping fibers, thicker fibers, less to normal
than normal dense compared control
control to normal
control
Rat [15] Unilateral 8 weeks Increased density N/A Decreased
vocal fold compared to density
stripping normal controls compared to
normal controls
Rat [15] Unilateral 12 weeks Increased density N/A Decreased
vocal fold compared density
stripping normal to compared to
controls normal controls
8.3.1 General
The mechanical properties of the vocal fold are critical for normal phonation. Vocal
fold tissue mechanical properties contribute to the minimum amount of air pressure
needed to produce voice, also known as phonation threshold pressure. Phonation
threshold pressure is the minimum subglottal pressure needed to overcome airway
pressure, forcing the vocal folds open to begin oscillation. This measure is often
taken as an objective indication of the energy required by an individual for vocal
function. The combination of large quantities of collagen, elastin, and hyaluronic
acid, and their respective mechanical properties, imparts viscoelastic properties to
the vocal fold lamina propria. In normal conditions, and different disease states, the
amount and organization of each protein affect the overall viscoelasticity of the
tissue. In the past, the Young’s modulus for the human intermediate and deep layers
of the vocal fold has been calculated at 33.1 ± 10.4 kPa and 135 kPa for strains of
15 % and 25 %, respectively [16]. The wide frequency range covered during vocal
fold phonation means direct measurement of the vocal fold viscoelastic properties
has greater clinical application. Presently there are no methods to measure viscoelas-
tic properties in vivo. For in vitro analysis, shear rheology is the favored method for
quantifying the mechanical properties of excised lamina propria vocal fold tissue.
650 J. Gaston and S.L. Thibeault
8.3.2 Rheology
Elastic shear modulus quantifies the elasticity of the tissue under shear stress, and is
proportional to the energy stored elastically due to strain. With rheological testing, the
elastic shear modulus is a function of frequency, and only increases gradually across
most of the frequency range tested (Fig. 8.2). At frequencies greater than 1 Hz, the elas-
tic shear modulus increases with higher frequency. Shear properties are typically only
measured at a frequency range of .01 to 15 Hz, as shear thinning and the small sample
size prevent meaningful results at higher testing frequencies for vocal fold tissue.
Rheology data published indicates that in general, the female vocal fold lamina propria
had a lower elastic shear modulus compared to the male vocal fold lamina propria.
10000
72 y.o.
62 y.o.
1000
60 y.o.
30 y.o.
m (Pa)
28 y.o.
100
50 y.o.
34 y.o.
36 y.o.
10
31 y.o.
59 y.o.
1
0.01 0.1 1 10 100
Frequency (Hz)
Fig. 8.2 Elastic shear modulus of male human vocal fold lamina propria. Reprinted with permis-
sion from [18]
8 Vocal Folds 651
Dynamic viscosity characterizes the opposition of the tissue to shear flow, and is
proportional to the energy lost in the viscoelastic tissue. Dynamic viscosity is
related to the viscous shear modulus through the angular frequency. Rheological
testing shows that the dynamic viscosity decreases as a function of frequency, and
as such the vocal folds exhibit shear thinning behavior (Fig. 8.3). This behavior is
fundamental to vocal fold movement—as the fundamental frequency of vibration
increases, the lamina propria will shear thin providing movement of this tissue at
very high frequencies.
Only one study has investigated the Poisson ratio of human vocal folds [19]. An
optical method was used, where a marked vocal fold sample was illuminated, cycli-
cally stretched at 1 Hz and 37 °C, and video recorded using a high-speed camera
system. Processing of the recorded images resulted in a Poisson ratio of .392. This
number should be taken with caution; several models consider the human vocal
folds to be transversely isotropic [20, 21], and testing was performed at only one
frequency [19].
1000 72 y.o.
62 y.o.
100 60 y.o.
50 y.o.
10 30 y.o.
h (Pa-s)
28 y.o.
34 y.o.
1
36 y.o.
59 y.o.
0.1
31 y.o.
0.01
0.01 0.1 1 10 100
Frequency (Hz)
Fig. 8.3 Dynamic viscosity of male human vocal fold lamina propria. Reprinted with permission
from [18]
652 J. Gaston and S.L. Thibeault
Numerous studies have shown that scarring of vocal fold tissue results in tissue
mechanics detrimental to normal voice function. Animal models have shown that
vocal fold scarring results in permanently elevated elastic shear modulus and
dynamic viscosity, making the tissue both stiffer and more viscous (Table 8.3).
Rabbit models are particularly useful, as normal vocal fold viscoelastic shear prop-
erties are similar to those of human vocal fold tissue [17]. This can be visualized
further when the viscoelastic moduli are plotted as a function of frequency for rabbit
animal models. Elastic shear modulus of scarred vocal fold tissue is higher than the
normal control tissue for all samples tested (Fig. 8.4a), indicating an increase in
stiffness [22]. The same trend is seen for the dynamic viscosity, with scarred tissue
having a significantly higher modulus than normal controls (Fig. 8.4b). Though no
direct comparison of mechanical properties between normal and scarred human
vocal folds, nearly all mammalian animal models have shown the same patterns
described above [13, 14].
Injectable biomaterials have been developed and used to treat vocal fold lamina
propria with abnormal viscoelastic properties due to ECM disruption. The most
important property for consideration for a vocal fold injectable is that the biomate-
rial has biomechanical properties most similar to the lamina propria, so as to allow
for normal tissue movement. When injected into the vocal fold lamina propria, the
viscoelastic properties of injectable biomaterials should be matched as closely as
possible to normal lamina propra. To this end, several natural and synthetic bioma-
terials have been developed to mitigate dysphonia. Upon injection, biomaterials
with significantly higher viscous and elastic moduli make vocal fold oscillation, and
therefore phonation, difficult. Finally, any biomaterial implanted into the vocal fold
must undergo shear thinning in order to match the basic viscoelastic behavior of the
lamina propria.
Table 8.3 Animal studies detailing vocal fold injury and changes in mechanical properties
Animal Timeline
model Type of injury post-injury Elastic shear modulus Dynamic viscosity
Rabbit [12] Lamina propria 60 days Stiffer tissue, higher More viscous, higher
forceps biopsy modulus modulus
Rabbit [13] Lamina propria 6 months Stiffer tissue, higher More viscous, higher
forceps biopsy modulus in eight out modulus in nine out
of ten samples of ten samples
Canine [14] Vocal fold 2 months Stiffer tissue, higher More viscous, higher
stripping modulus modulus
8 Vocal Folds 653
a
10000 Scar One
Scar Two
Scar Three
Scar Four
Scar Five
1000
Scar Six
Norm One
m (Pa)
Norm Two
Norm Three
Norm Four
100
Norm Five
10
0.01 0.1 1 10 100
b
10000 Scar One
Scar Two
Scar Three
1000
Scar Four
Scar Six
100
Scar Five
Norm One
h (Pa-s)
10 Norm Two
Norm Three
Norm Four
1
Norm Five
0.1
0.01
0.01 0.1 1 10 100
Fig. 8.4 (a) Elastic shear modulus (μ) of rabbit vocal fold. (b) Dynamic viscosity (η) of rabbit
vocal fold tissue. Reprinted with permission from [22]
Several different collagen formulations and adipose have been tested as injectables
for vocal fold scarring. A study utilized rheological testing to compare collagen and
adipose tissue to normal vocal fold tissue, with a specific focus on viscosity [17].
Adipose tissue prepared as a fat graft was found to have the closest viscosity to
normal vocal fold tissue, followed by uncrosslinked collagen (Fig. 8.5).
654 J. Gaston and S.L. Thibeault
TABLE II.
Dynamic Viscosity of Implantable Biomaterials and Human Vocal Fold Mucosal Tissues
Measured at 10 Hz and Extrapolated to 100 Hz.
Dynamic Viscosity (Pa-s)
Fig. 8.5 Comparison of dynamic viscosity of several biomaterial injections for the vocal fold
lamina propria. Reprinted with permission from [17]
HA hydrogels show the most promise as injectable biomaterials for the vocal fold
to date. However, natural HA is rapidly degraded upon injection, typically taking
roughly 3–5 days to dissolve [12]. As such several different HA hydrogel formula-
tions with differing cross-links and chemical moieties have been designed and
investigated for vocal fold applications (Table 8.4).
Hylan B, a non-inflammatory divinyl sulfone cross-linked simple HA hydrogel,
has been used in the past as a space-filling injection. It is no longer in the market,
but has since been succeeded by the Restylane line of HA injectables. Multiple
animal studies show that Hylan-B injections into rabbit vocal fold lamina propria
show mechanical results similar to normal vocal fold [24, 27]. Specifically, 6
months after injection no significant difference in dynamic viscosity was observed
between Hylan-B-injected tissue and normal, noninjected tissue. In the same study,
vocal fold tissue injected with Hylan-B was found to have a lower dynamic viscos-
ity than tissue injected with other biomaterials, such as collagen or Teflon. In an
83-person clinical study, Hylan-B had a more favorable effect than collagen on
vocal fold function [28]. Most notable was the increase in vocal fold mucosal wave
production compared to baseline, indicating positive viscoelastic mechanics. This
result could not be directly corroborated, as rheology can only be performed on
ex vivo tissue.
Hydrogel cross-linking is of particular interest in vocal fold applications, as
different cross-linking methods and groups can alter the mechanical properties of
the hydrogel. Several different HA hydrogels have been formulated that utilize a
polyethylene glycol diacrylate (PEGDA) cross-linker, allowing the gelation time
and swelling ration to be easily modified [29]. PEGDA cross-linking is prominent
8 Vocal Folds 655
in two HA hydrogels for vocal fold applications; HA-DTPH and Carbylan-S. Both
HA-DTPH and Carbylan-S are similar in composure, consisting of an HA backbone
with PEGDA cross-links. The Carbylan-S backbone has additional carboxylate
groups on the HA backbone, resulting in a higher elastic shear modulus than
HA-DTPH [25]. To date, both hydrogel formulations have been primarily examined
as a prophylactic treatment to reduce or prevent vocal fold scarring by injection at
the time of injury. Excised tissue from injured rabbit vocal fold injected with
Carbylan-S had a lower elastic shear modulus than injured tissue treated with
saline controls or injured tissue injected with HA-DTPH-PEGDA [30]. Tissue
injected with either Carbylan-S or HA-DTPH-PEGDA was significantly less
viscous than the saline-treated samples as well. Trichrome staining performed on
sections of the excised larynges showed that the average fibrosis levels for animals
injected with the HA-DTPH-PEGDA were moderate and not significantly different
from saline-treated controls.
656 J. Gaston and S.L. Thibeault
The Carbylan-S HA backbone has also been cross-linked using thiolated gelatin
to generate a hydrogel commercially designated as Extracel® [26]. Extracel® has
undergone several naming changes since its first formulation, and can also be found
as Gycosil® or HyStem®. Extracel® has been investigated as both a prophylactic
treatment for vocal fold tissue injury and as a treatment for previously scarred tis-
sue. A rabbit study showed that 6 months after injury and injection, excised vocal
fold tissue injected with Extracel® had a lower elastic shear modulus and lower
viscous modulus compared to saline controls [12]. While histology was not per-
formed, quantitative PCR showed a decrease in mRNA transcript levels for fibro-
nectin, fibromodulin, and procollagen [12]. A similar study was performed on
rabbits where an injury was induced and allowed to scar for 3 months [23], at which
time Extracel® or saline was injected into the vocal fold. Two months following
injection, the tissue injected with Extracel® had a lower elastic shear modulus and
viscous modulus compared to saline injections. Finally, histological staining showed
that the Extracel®-treated vocal folds had higher procollagen and collagen than
saline, a marked contrast from the prophylactic study.
Additional Reading
Miri AK. Mechanical characterization of vocal fold tissue: a review study. J Voice
2014; 28:657–67.
Caton T, Thibeault SL, Klemuk S, Smith ME. In reference to viscoelasticity of
hyaluronan and nonhyaluronan based vocal fold injectables: Implications for muco-
sal versus muscle use—Reply. Laryngoscope 2007, 117(8): 1506–1508.
Chan RW, Titze IR. Dependence of phonation threshold pressure on vocal tract
acoustics and vocal fold tissue mechanics. Journal of the Acoustical Society of
America 2006; 119:2351–62.
References
1. Benninger MS, Alessi D, Archer S, Bastian R, Ford C, Koufman J et al (1996) Vocal fold scar-
ring: current concepts and management. Otolaryngol Head Neck Surg 115:474–482
2. Titze IR (2000) Principles of voice production. National Center for Voice and Speech
3. Smith E, Gray S, Verdolini K, Lemke J (1995) Effects of voice disorders on quality of life.
Otolaryngol Head Neck Surg 113:P121
4. Butler JE, Hammond TH, Gray SD (2001) Gender-related differences of hyaluronic acid dis-
tribution in the human vocal fold. Laryngoscope 111:907–911
5. Hahn MS, Kobler JB, Zeitels SM, Langer R (2006) Quantitative and comparative studies of
the vocal fold extracellular matrix. II: Collagen. Ann Otol Rhinol Laryngol 115:225–232
6. Gray SD, Titze IR, Alipour F, Hammond TH (2000) Biomechanical and histologic observa-
tions of vocal fold fibrous proteins. Ann Otol Rhinol Laryngol 109:77–85
7. Hahn MS, Kobler JB, Starcher BC, Zeitels SM, Langer R (2006) Quantitative and comparative
studies of the vocal fold extracellular matrix. I: Elastic fibers and hyaluronic acid. Ann Otol
Rhinol Laryngol 115:156–164
8 Vocal Folds 657
8. Hansen JK, Thibeault SL (2006) Current understanding and review of the literature: vocal fold
scarring. J Voice 20:110–120
9. Hahn MS, Kobler JB, Zeitels SM, Langer R (2005) Midmembranous vocal fold lamina propria
proteoglycans across selected species. Ann Otol Rhinol Laryngol 114:451–462
10. Kalamajski S, Oldberg A (2010) The role of small leucine-rich proteoglycans in collagen
fibrillogenesis. Matrix Biol 29:248–253
11. Hirano S, Minamiguchi S, Yamashita M, Ohno T, Kanemaru S, Kitamura M (2009) Histologic
characterization of human scarred vocal folds. J Voice 23:399–407
12. Thibeault SL, Klemuk SA, Chen X, Johnson BHQ (2011) In vivo engineering of the vocal fold
ECM with injectable HA hydrogels-late effects on tissue repair and biomechanics in a rabbit
model. J Voice 25:249–253
13. Rousseau B, Hirano S, Chan RW, Welham NV, Thibeault SL, Ford CN et al (2004)
Characterization of chronic vocal fold scarring in a rabbit model. J Voice 18:116–124
14. Rousseau B, Hirano S, Scheidt TD, Welham NV, Thibeault SL, Chan RW et al (2003)
Characterization of vocal fold scarring in a canine model. Laryngoscope 113:620–627
15. Tateya T, Tateya I, Sohn JH, Bless DM (2005) Histologic characterization of rat vocal fold
scarring. Ann Otol Rhinol Laryngol 114:183–191
16. Caton T, Thibeault SL, Klemuk S, Smith ME (2007) In reference to viscoelasticity of hyaluro-
nan and nonhyaluronan based vocal fold injectables: implications for mucosal versus muscle
use—Reply. Laryngoscope 117:1506–1508
17. Chan RW, Titze IR (1998) Viscosities of implantable biomaterials in vocal fold augmentation
surgery. Laryngoscope 108:725–731
18. Chan RW, Titze IR (1999) Viscoelastic shear properties of human vocal fold mucosa: measure-
ment methodology and empirical results. J Acoust Soc Am 106:2008–2021
19. Alipour F, Vigmostad S (2012) Measurement of vocal folds elastic properties for continuum
modeling. J Voice 26:9
20. Alipour F, Berry DA, Titze IR (2000) A finite-element model of vocal-fold vibration. J Acoust
Soc Am 108:3003–3012
21. Rosa MD, Pereira JC, Grellet M, Alwan A (2003) A contribution to simulating a three-
dimensional larynx model using the finite element method. J Acoust Soc Am 114:2893–2905
22. Thibeault SL, Gray SD, Bless DM, Chan RW, Ford CN (2002) Histologic and rheologic char-
acterization of vocal fold scarring. J Voice 16:96–104
23. Thibeault SL, Duflo S (2008) Inflammatory cytokine responses to synthetic extracellular
matrix injection to the vocal fold lamina propria. Ann Otol Rhinol Laryngol 117:221–226
24. Borzacchiello A, Mayol L, Garskog O, Dahlqvist A, Ambrosio L (2005) Evaluation of injec-
tion augmentation treatment of hyaluronic acid based materials on rabbit vocal folds viscoelas-
ticity. J Mater Sci Mater Med 16:553–557
25. Shu XZ, Liu Y, Prestwich GD (2011) Modified macromolecules and methods of making and
using thereof
26. Duflo S, Thibeault SL, Li WH, Shu XZ, Prestwich GD (2006) Vocal fold tissue repair in vivo
using a synthetic extracellular matrix. Tissue Eng 12:2171–2180
27. Dahlqvist A, Garskog O, Laurent C, Hertegard S, Ambrosio L, Borzacchiello A (2004)
Viscoelasticity of rabbit vocal folds after injection augmentation. Laryngoscope 114:138–142
28. Hertegard S, Hallen L, Laurent C, Lindstrom E, Olofsson K, Testad P et al (2002) Cross-linked
hyaluronan used as augmentation substance for treatment of glottal insufficiency: safety
aspects and vocal fold function. Laryngoscope 112:2211–2219
29. Shu XZ, Liu YC, Palumbo FS, Lu Y, Prestwich GD (2004) In situ crosslinkable hyaluronan
hydrogels for tissue engineering. Biomaterials 25:1339–1348
30. Hansen JK, Thibeault SL, Walsh JF, Shu XZ, Prestwich GD (2005) In vivo engineering of the
vocal fold extracellular matrix with injectable hyaluronic acid hydrogels: early effects on tis-
sue repair and biomechanics in a rabbit model. Ann Otol Rhinol Laryngol 114:662–670
31. Hallen L, Johansson C, Laurent C (1999) Cross-linked hyaluronan (hylan B gel): a new inject-
able remedy for treatment of vocal fold insufficiency—an animal study. Acta Otolaryngol
119:107–111
Index
Endoprostheses G
hip, 151 Gauge length, 357
knee, 151 Glass ceramics, 447–54
Environmental stress cracking, 264 clinical uses, 453
Epiligament/epitenon, 55 physical properties, 450, 451, 453
Epoxy resins Glass fibers, 206–7, 208–9
curing agents, 212–5 advantages, 208
description, 240 chemical composition, 206
end uses, 240 electrical properties, 206–7
properties, 244 grades, 209
Equivalence, 317–35 mechanical properties, 206
Extracellular matrix (ECM) of lamina optical properties, 207
propria, 645 physical properties, 206
thermal properties, 206
Glassy carbons, 556
F fabrication, 556
Face-centered cubic (FCC), 162, 167 mechanical properties, 252
Fatigue, 178, 178–81, 179, 180, 181 shrinkage, 556
cortical bone, 10–11 Glycosaminoglycan (GAG), 646
porosity, 11 Graphite-fiber-reinforced
wear, 456, 458, 473, 475 thermoplastics, properties, 235
Young’s modulus, 11 Graphite fibers, 208, 209
Female, fluid volume, 120 Graphites, 554–5
Femoral articulating surface, 45–6 colloidal, 555
Fiberglass, properties, 238 manufacturing processes, 555
Fiber-reinforced plastic pipe, natural, 554–5
220–1 synthesized, 555
Fibrocartilage, 38, 45–51 Griffith’s theory, 381–4
anisotropy, 45
composition, 45–7, 46, 47
mechanical properties, 42 H
structure, 45–7, 46, 47 Hand lay-up, 220
Fibrosis, 575 Hanks’ buffered saline solution, 154
Filament winding, 224 HAp, 496
storage tanks, 225 bioresorption, 503
Fillers degradation, 499–505
brazing temperatures, 199 dissolution rate, 497
chemical composition, 199 highly porous, 503
Finite element analysis (FEA), ceramic from marine coral, 502
materials, 433–5 precipitation, 535
Flexural strength, dentin, 29 seeds, 536
Fluid volume Haptens, immune response to, 600
female, 120 Havar
male, 120 chemical composition, 165
Fractures, internal fixation mechanical properties, 166
malignancies, 614 Haversian
Fracture toughness remodelling, 11
dentin, 30 systems, 4
enamel, 30 HDPE, properties, 230
Fracture toughness, cancellous bone, 19 Health of donors, cortical bone, 5–6
Fracture work, ceramic materials, Heart
392–6 pacemakers, 152
Friction coefficients, implants, 464–71 valves, 551, 552
Index 667