Chapter: Biomaterials Science: An Introduction to Materials in Medicine

Full text

Provided for non-commercial research and educational use only.
Not for reproduction, distribution or commercial use.
This chapter was originally published in the book Biomaterials Science. The copy
attached is provided by Elsevier for the author's benefit and for the benefit of the
author's institution, for non-commercial research, and educational use. This includes
without limitation use in instruction at your institution, distribution to specific
colleagues, and providing a copy to your institution's administrator.
All other uses, reproduction and distribution, including without limitation commercial reprints, selling or
licensing copies or access, or posting on open internet sites, your personal or institution’s website or
repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier’s
permissions site at:
http://www.elsevier.com/locate/permissionusematerial
From More R. B., Haubold A. D., & Bokros J. C. (2013). Pyrolytic carbon for long-term
medical implants. In B. D. Ratner, A. S. Hoffman & F. J. Schoen (Eds.),
Biomaterials Science (pp. 209–222). Elsevier Inc., Academic Press.
ISBN: 9780123746269
Copyright © 2013 Elsevier Inc. All rights reserved.
Academic Press
Author's personal copy

CHAPTER I.2.8 Pyrolytic Carbon for Long-Term Medical Implants 209
Biomaterials Science, Third Edition, 2013, 209–222
Shu, X. Z., Liu, Y., Luo, Y., Roberts, M. C., & Prestwich, G. D.
(2002). Disulfide cross-linked hyaluronan hydrogels. Biomac-
romolecules, 3(6), 1304–1311.
Shin, H., Jo, S., & Mikos, A. G. (2003). Biomimetic materials for
tissue engineering. Biomaterials, 24(24), 4353–4364.
Shu, X. Z., Ghosh, K., Liu, Y., Palumbo, F. S., Luo, Y., et al.
(2004). Attachment and spreading of fibroblasts on an RGD
peptide-modified injectable hyaluronan hydrogel. J. Biomed.
Mater. Res. A, 68(2), 365–375.
Shu, X. Z., & Prestwich, G. D. (2004). Therapeutic Biomaterials
from Chemically Modified Hyaluronan. In H. G. Garg, & C.
A. Hales (Eds.), Chemistry and Biology of Hyaluronan (pp.
475–504). Amsterdam: Elsevier.
Shu, X. Z., Ahmad, S., Liu, Y., & Prestwich, G. D. (2006). Synthe-
sis and evaluation of injectable, in situ cross-linkable synthetic
extracellular matrices for tissue engineering. J. Biomed. Mater.
Res. A, 79(4), 902–912.
Sondrup, C., Liu, Y., Shu, X. Z., Prestwich, G. D., & Smith, M. E.
(2006). Cross-linked hyaluronan-coated stents in the prevention
of airway stenosis. Otolaryngol Head Neck Surg., 135(1), 28–35.
Stenzel, K. H., Miyata, T., & Rubin, A. L. (1974). Collagen as a
biomaterial. Annu. Rev. Biophys. Bioeng., 3(0), 231–253.
Stone, K. R., Steadman, J. R., Rodkey, W. G., & Li, S. T. (1997).
Regeneration of meniscal cartilage with use of a collagen scaf-
fold. Analysis of preliminary data. J. Bone. Joint Surg. Am.,
79(12), 1770–1777.
Suh, J. K., & Matthew, H. W. (2000). Application of chitosan-
based polysaccharide biomaterials in cartilage tissue engineer-
ing: A review. Biomaterials, 21(24), 2589–2598.
Urry, D. W., Hugel, T., Seitz, M., Gaub, H. E., Sheiba, L., et al.
(2002). Elastin: A representative ideal protein elastomer. Phi-
los. Trans. R. Soc. Lond. B Biol. Sci., 357(1418), 169–184.
van der Giessen, W. J., van Beusekom, H. M., Eijgelshoven,
M. H., Morel, M. A., & Serruys, P. W. (1998). Heparin-
coating of coronary stents. Semin. Interv. Cardiol., 3(3–4),
173–176.
Vanderhooft, J. L., Mann, B. K., & Prestwich, G. D. (2007). Syn-
thesis and characterization of novel thiol-reactive poly(ethylene
glycol) cross-linkers for biomaterials. Biomacromolecules, 8,
2883–2889.
VandeVord, P. J., Matthew, H. W., DeSilva, S. P., Mayton, L.,
Wu, B., et al. (2002). Evaluation of the biocompatibility of
a chitosan scaffold in mice. J. Biomed. Mater. Res., 59(3),
585–590.
White, M. J., Kohno, I., Rubin, A. L., Stenzel, K. H., & Miyata, T.
(1973). Collagen films: Effect of cross-linking on physical and
biological properties. Biomater. Med. Devices Artif. Organs,
1(4), 703–715.
Yuan, N., Tsai, R., Ho, M., Wang, D., Lai, J., et al. (2008). Fab-
rication and characterization of chondroitin sulfate-modified
chitosan membranes for biomedical applications. Desalina-
tion, 23(1–3), 166–174.
CHAPTER I.2.8 PYROLYTIC CARBON
FOR LONG-TERM MEDICAL IMPLANTS
Robert B. More,1 Axel D. Haubold,2 and Jack C. Bokros3
1Integra Life Sciences, Austin, TX, USA
2Bed Rock Ranch, Decatur, TX, USA
3On-X Life Technologies, Inc, Austin, TX, USA
INTRODUCTION
Carbon materials are ubiquitous and of great interest
because the majority of substances that make up living
organisms are carbon compounds. Although many engi-
neering materials and biomaterials are based on carbon
or contain carbon in some form, elemental carbon itself
is also an important and very successful biomaterial. Fur-
thermore, there exists enough diversity in their structure
and properties for elemental carbons to be considered
as a unique class of materials beyond the traditional
molecular carbon focus of organic chemistry, polymer
chemistry, and biochemistry. Through a serendipitous
interaction between researchers during the late 1960s
the outstanding blood compatibility of a special form of
elemental pyrolytic carbon deposited at high tempera-
ture in a fluidized bed was discovered. The material was
found to have not only remarkable blood compatibility,
but also the structural properties needed for long-term
use in artificial heart valves (LaGrange et al., 1969).
The blood compatibility of pyrolytic carbon was rec-
ognized empirically using the Gott vena cava ring test.
This test involved implanting a small tube made of a
candidate material in a canine vena cava and observing
the development of thrombosis within the tube in time.
Prior to pyrolytic carbon, only surfaces coated with
graphite, benzylalkonium chloride, and heparin would
resist thrombus formation when exposed to blood for
long periods. The incorporation of pyrolytic carbon in
mechanical heart valve implants was declared “an excep-
tional event” (Sadeghi, 1987) because it added the dura-
bility and stability needed for heart valve prostheses to
endure for a patient’s lifetime. The objective of this chap-
ter is to present the elemental pyrolytic carbon materials
currently used in the fabrication of medical devices, and
to describe their manufacture, characterization, and
properties.
ELEMENTAL CARBON
Elemental carbon is found in nature as two crystal-
line allotrophic forms: graphite and diamond. Ele-
mental carbon also occurs as a spectrum of imperfect,
turbostratic crystalline forms that range in degree of
crystallinity from amorphous to the perfectly crystalline
allotropes. Recently a third crystalline form of elemental
carbon, the fullerene structure, has been discovered. The
crystalline polymorphs of elemental carbon are shown in
Figure I.2.8.1.
The properties of the elemental carbon crystalline
forms vary widely according to their structure. Diamond
with its tetrahedral sp3 covalent bonding is one of the
hardest materials known. In the diamond crystal struc-
ture, covalent bonds of length 1.54 Å connect each
carbon atom with its four nearest neighbors. This tetra-
hedral symmetry repeats in three dimensions throughout
Disclaimer: Due to the run on format of the book you may notice the end of the previous chapter and beginning of the next chapter in your e-offprint.
Author's personal copy

210 SECTION I.2 Classes of Materials Used in Medicine
Biomaterials Science, Third Edition, 2013, 209–222
the crystal (Pauling, 1964). In effect, the crystal is a giant
isotropic covalently-bonded molecule; therefore, dia-
mond is very hard.
Graphite, with its anisotropic layered in-plane hex-
agonal covalent bonding and interplane van der Waals
bonding structure, is a soft material. Within each planar
layer, each carbon atom forms two single bonds and one
double bond with its three nearest neighbors. This bond-
ing repeats in-plane to form a giant molecular (graphene)
sheet. The in-plane atomic bond length is 1.42 Å, which
is a resonant intermediate (Pauling, 1964) between the
single bond length of 1.54 Å and the double bond length
of 1.33 Å. The planer layers are held together by rela-
tively weak van der Waals bonding at a distance of 3.4 Å,
which is more than twice the 1.42 Å bond length (Pauling,
1964). Graphite has low hardness and a lubricating prop-
erty because the giant molecular sheets can readily slip
past one another against the van der Waals bonding. Nev-
ertheless, although large-crystallite-size natural graphite is
used as a lubricant, some artificially produced graphites
can be very abrasive if the crystallite sizes are small and
randomly oriented.
Fullerenes have yet to be produced in bulk, but their
properties on a microscale are entirely different from
those of their crystalline counterparts. Fullerenes and
nanotubes consist of a graphene layer that is rolled up
or folded (Sattler, 1995) to form a tube or ball. These
large molecules, C60 and C70 fullerenes and (C60 + 18j)
nanotubes, are often mentioned in the literature (Sattler,
1995) along with more complex multilayer “onion skin”
structures.
There exist many possible forms of elemental car-
bon that are intermediate in structure and properties
between those of the allotropes diamond and graphite.
Such “turbostratic” carbons occur as a spectrum of
amorphous, through mixed amorphous, graphite-like
and diamond-like, to the perfectly crystalline allotropes
(Bokros, 1969). Because of the dependence of properties
upon structure, there can be considerable variability in
properties for the turbostratic carbons. Glassy carbons
and pyrolytic carbons, for example, are two turbostratic
carbons with considerable differences in structure and
properties. Consequently, it is not surprising that carbon
materials are often misunderstood through oversimpli-
fication. Properties found in one type of carbon struc-
ture can be totally different in another type of structure.
Therefore, it is very important to specify the exact nature
and structure when discussing carbon.
PYROLYTIC CARBON (PyC)
The biomaterial known as pyrolytic carbon is not found
in nature; it is manmade. The successful pyrolytic carbon
biomaterial was developed at General Atomic during the
late 1960s using a fluidized-bed reactor (Bokros, 1969).
In the original terminology, this material was consid-
ered a low-temperature isotropic carbon (LTI carbon).
Since the initial clinical implant of a pyrolytic carbon
component in the DeBakey–Surgitool mechanical valve
in 1968, 95% of the mechanical heart valves implanted
worldwide have at least one structural component made
of pyrolytic carbon. On an annual basis this translates
into approximately 500,000 components (Haubold,
1994). Pyrolytic carbon components have been used in
more than 25 different prosthetic heart valve designs
since the late 1960s, and have accumulated a clini-
cal experience in the order of 16 million patient-years.
Clearly, pyrolytic carbon is one of the most successful,
critical biomaterials both in function and application.
Among the materials available for mechanical heart
valve prostheses, pyrolytic carbon has the best combi-
nation of blood compatibility, physical and mechanical
properties, and durability. However, the blood compat-
ibility of pyrolytic carbon in heart valve applications is
not perfect; chronic anticoagulant therapy is needed for
patients with mechanical heart valves. Whether the need
for anticoagulant therapy arises from the biocompat-
ible properties of the material itself or from the particu-
lar hydrodynamic interaction of a given device and the
blood remains to be resolved.
The term “pyrolytic” is derived from “pyrolysis,”
which is thermal decomposition. Pyrolytic carbon is
formed from the thermal decomposition of hydrocar-
bons such as propane, propylene, acetylene, and meth-
ane, in the absence of oxygen. Without oxygen the
typical decomposition of the hydrocarbon to carbon
dioxide and water cannot take place; instead a more
complex cascade of decomposition products occurs that
ultimately results in a “polymerization” of the individual
carbon atoms into large macroatomic arrays.
Diamond Graphite
Fullerene
Bucky Ball
FIGURE I.2.8.1 Allotropic crystalline forms of carbon: diamond,
graphite, and fullerene.
Author's personal copy

CHAPTER I.2.8 Pyrolytic Carbon for Long-Term Medical Implants 211
Biomaterials Science, Third Edition, 2013, 209–222
Pyrolysis of the hydrocarbon is normally carried
out in a fluidized-bed reactor such as the one shown
in Figure I.2.8.2. A fluidized-bed reactor typically con-
sists of a vertical tube furnace that may be induction
or resistance heated to temperatures of 1000 to 2000°C
( Bokros, 1969). Reactor diameters ranging from 2 cm
to 25 cm have been used; however, the most common
size used for medical devices has a diameter of about
10 cm. These high-temperature reactors are expensive
to operate, and the reactor size limits the size of device
components able to be produced.
Small refractory ceramic particles are placed into the
vertical tube furnace. When a gas is introduced into the
bottom of the tube furnace, the gas causes the particle
bed to expand. Interparticle spacing increases to allow
for the flow of the gas. Particle mixing occurs and the
bed of particles begins to “flow” like a fluid. Hence the
term “fluidized bed.” Depending upon the gas flow rate
and volume, this expansion and mixing can be varied
from a gentle bubbling bed to a violent spouting bed. An
oxygen-free, inert gas, such as nitrogen or helium, is used
to fluidize the bed, and the source hydrocarbon is added
to the gas stream when needed.
At a sufficiently high temperature, pyrolysis or ther-
mal decomposition of the hydrocarbon can take place.
Pyrolysis products range from free carbon and gaseous
hydrogen to a mixture of Cx Hy decomposition species.
The pyrolysis reaction is complex and is affected by the
gas flow rate, composition, temperature, and bed surface
area. Decomposition products, under the appropriate
conditions, can form gas-phase nucleated droplets of
carbon/hydrogen, which condense and deposit on the
surfaces of the wall and bed particles within the reactor
(Bokros, 1969). Indeed, the fluidized-bed process was
originally developed to coat small (200–500 micrometer)
diameter spherical particles of uranium/thorium carbide
or oxide with pyrolytic carbon. These coated particles
were used as the fuel in the high temperature gas-cooled
nuclear reactor (Bokros, 1969).
Pyrolytic carbon coatings produced in vertical tube
reactors can have a variety of structures, such as laminar
or isotropic, granular or columnar (Bokros, 1969). The
structure of the coating is controlled by the gas flow rate
(residence time in the bed), hydrocarbon species, tem-
perature, and bed surface area. For example, an inad-
equately fluidized or static bed will produce a highly
anisotropic, laminar pyrolytic carbon (Bokros, 1969).
Control of the first three parameters (gas flow rate,
hydrocarbon species, and temperature) is relatively easy.
However, until recently, it was not possible to measure
the bed surface area while the reactions were taking
place. As carbon deposits on the particles in the fluidized
bed, the diameter of the particles increases. Hence the
surface area of the bed changes, which in turn influences
the subsequent rate of carbon deposition. As surface area
increases, the coating rate decreases, since a larger sur-
face area now has to be coated with the same amount of
carbon available. Thus, the process is not in equilibrium.
The static-bed process was adequate to coat nuclear fuel
particles without attempting to control the bed surface
area because such thin coatings (25–50 μm thick) did not
appreciably affect the bed surface area.
It was later found that larger objects could be sus-
pended within the fluidized bed of small ceramic parti-
cles and also become uniformly coated with carbon. This
finding led to the demand for thicker, structural coatings,
an order of magnitude thicker (250–500 μm). Bed sur-
face area control and stabilization became an important
factor (Akins and Bokros, 1974) in achieving the goal of
thicker, structural coatings. In particular, with the dis-
covery of the blood-compatible properties of pyrolytic
carbon (LaGrange et al., 1969), thicker structural coat-
ings with consistent and uniform mechanical properties
were needed to realize the application to mechanical
heart valve components. Quasi-steady-state conditions
as needed to prolong the coating reaction were achieved
empirically by removing coated particles and adding
uncoated particles to the bed while the pyrolysis reaction
was taking place (Akins and Bokros, 1974). However,
the rates of particle addition and removal were based
upon little more than good guesses.
Three of the four parameters that control carbon
deposition could be accurately measured and controlled,
but a method to measure and control bed surface area
was lacking. Thus, the quasi-steady-state process was
more of an art than a science. If too many coated par-
ticles were removed, the bed became too small to support
the larger components within it and the bed collapsed.
If too few particles were removed, the rate of deposi-
tion decreased, and the desired amount of coating was
not achieved in the anticipated time. Furthermore, there
Add
particles
Bed
reaction
Feed
rate
Withdraw
rate
Pressure
sensor
Hydrocarbon
gas
Controller
Remove
particles
FIGURE I.2.8.2 Fluidized-bed reactor schematic.
Author's personal copy

212 SECTION I.2 Classes of Materials Used in Medicine
Biomaterials Science, Third Edition, 2013, 209–222
were considerable variations in the mechanical proper-
ties of the coating from batch to batch. It was found that
in order to consistently achieve the hardness needed for
wear resistance in prosthetic heart valve applications, it
was necessary to add a small amount of β-silicon car-
bide to the carbon coating. The dispersed silicon carbide
particles within the pyrolytic carbon matrix added suf-
ficient hardness to compensate for potential variations
in the properties of the pyrolytic carbon matrix. The
β-silicon carbide was obtained from the pyrolysis of
methyl- trichlorosilane, CH3SiCl3. For each mole of sili-
con carbide produced, the pyrolysis of methyltrichlorosi-
lane also produces three moles of hydrogen chloride gas.
Handling and neutralization of this corrosive gas added
substantial complexity and cost to an already complex
process. Nevertheless, this process allowed consistency
for the successful production of several million compo-
nents for use in mechanical heart valves.
A process has been developed and patented that
allows precise measurement and control of the bed sur-
face area. A description of this process is given in the pat-
ent literature and elsewhere (Emken et al., 1993, 1994;
Ely et al., 1998). With precise control of the bed sur-
face area, it is no longer necessary to include the silicon
carbide. Elimination of the silicon carbide has produced
a stronger, tougher, and more deformable pure pyro-
lytic carbon. Historically, pure carbon was the original
objective of the development program because of the
potential for superior biocompatibility (LaGrange et al.,
1969). Furthermore, the enhanced mechanical and phys-
ical properties of the pure pyrolytic carbon now possi-
ble with the improved process control allow prosthesis
design improvements in the hemodynamic contribution
to thromboresistance (Ely et al., 1998).
Structure of Pyrolytic Carbons
X-ray diffraction patterns of the biomedical-grade
fluidized-bed pyrolytic carbons are broad and diffuse
because of the small crystallite size and imperfections.
In silicon-alloyed pyrolytic carbon, a diffraction pat-
tern characteristic of the β form of silicon carbide also
appears in the diffraction pattern along with the car-
bon bands. The carbon diffraction pattern indicates a
turbostratic structure (Kaae and Wall, 1996) in which
there is order within carbon layer planes, as in graphite;
but, unlike graphite, there is no order between planes.
This type of turbostratic structure is shown in Figure
I.2.8.3 compared to that of graphite. In the disordered
crystalline structure, there may be lattice vacancies
and the layer planes are curved or kinked. The ability
of the graphite layer planes to slip is inhibited, which
greatly increases the strength and hardness of the pyro-
lytic carbon relative to that of graphite. From the Bragg
(A)
(C)
(B)
FIGURE I.2.8.3 Structures of: (A) diamond; (B) graphite; and (C) turbostratic pyrolytic carbon.
Author's personal copy

CHAPTER I.2.8 Pyrolytic Carbon for Long-Term Medical Implants 213
Biomaterials Science, Third Edition, 2013, 209–222
equation, the pyrolytic carbon layer spacing is reported
to be 3.48 Å, which is larger than the 3.35 Å graphite
layer spacing (Kaae and Wall, 1996). The increase in
layer spacing relative to graphite is due to both the layer
distortion and the small crystallite size, and is a com-
mon feature for turbostratic carbons. From the Scherrer
equation the crystallite size is typically 25–40 Å (Kaae
and Wall, 1996).
During the coating reaction, gas-phase nucleated
droplets of carbon/hydrogen form that condense and
deposit on the surfaces of the reactor wall and bed par-
ticles within the reactor. These droplets aggregate, grow,
and form the coating. When viewed with high-resolu-
tion transmission electron microscopy, a multitude of
near-spherical polycrystalline growth features are evi-
dent, as shown in Figure I.2.8.4 (Kaae and Wall, 1996).
These growth features are considered to be the basic
building blocks of the material, and the shape and size
are related to the deposition mechanism. In the silicon-
alloyed carbon, small silicon carbide particles are present
within the growth features, as shown in Figure I.2.8.5.
Based on a crystallite size of 33 Å, each growth feature
contains about 3 × 109 crystallites. Although the mate-
rial is quasi-crystalline on a fine level, the crystallites are
very small and randomly oriented in the fluidized bed
pyrolytic carbons so that overall the material exhibits
isotropic behavior.
Glassy carbon, also known as vitreous carbon or
polymeric carbon, is another turbostratic carbon form
that has been proposed for use in long-term implants.
However, its strength is low and the wear resistance is
poor. Glassy carbons are quasi-crystalline in structure,
and are named “glassy” because the fracture surfaces
closely resemble those of glass (Haubold et al., 1981).
Vapor-deposited carbons are also used in heart valve
applications. Typically, the coatings are thin (<1 μm) and
may be applied to a variety of materials in order to con-
fer the biochemical characteristics of turbostratic carbon.
Some examples are vapor-deposited carbon coatings on
heart valve sewing cuffs and metallic orifice components
(Haubold et al., 1981).
Mechanical Properties
Mechanical properties of pure pyrolytic carbon, silicon-
alloyed pyrolytic carbons, and glassy carbon are given
in Table I.2.8.1. Pyrolytic carbon flexural strength is
FIGURE I.2.8.4 Electron micrograph of pure pyrolytic carbon
microstructure showing near-spherical polycrystalline growth fea-
tures formed during deposition (Kaae and Wall, 1996).
FIGURE I.2.8.5 Electron micrograph of silicon-alloyed pyrolytic
carbon microstructure showing near-spherical polycrystalline growth
features formed during deposition (Kaae and Wall, 1996). Small sili-
con carbide particles are shown in concentric rings in the growth
features.
TABLE I.2.8.1 Mechanical Properties of
Biomedical Carbons
Property Pure PyC
Typical Si-alloyed
PyC
Typical
Glassy
Carbon
Flexural strength
(MPa)
493.7 ± 12 407.7 ± 14.1 175
Young’s modu-
lus (GPa)
29.4 ± 0.4 30.5 ± 0.65 21
Strain-to-failure
(%)
1.58 ± 0.03 1.28 ± 0.03 -
Fracture
toughness
(MPa √m)
1.68 ± 0.05 1.17 ± 0.17 0.5–0.7
Hardness (DPH,
500 g load)
235.9 ± 3.3 287 ± 10 150
Density (g/cm3) 1.93 ± 0.01 2.12 ± 0.01 <1.54
CTE (10−6 cm/
cm°C)
6.5 6.1 -
Silicon content
(%)
0 6.58 ± 0.32 0
Wear resistance Excellent Excellent Poor
Author's personal copy

214 SECTION I.2 Classes of Materials Used in Medicine
Biomaterials Science, Third Edition, 2013, 209–222
high enough to provide the necessary structural stabil-
ity for a variety of implant applications, and the density
is low enough to allow for components to move eas-
ily under the applied forces of circulating blood. With
respect to orthopedic applications, Young’s modulus
is in the range reported for bone (Reilly and Burstein,
1974; Reilly et al., 1974), which allows for compliance
matching. Relative to metals and polymers, the pyro-
lytic carbon strain-to-failure rate is low; it is a nearly
ideal linear elastic material and requires consideration
of brittle material principles in component design.
Strength levels vary with the effective stressed volume
or stressed area, as predicted by classical Weibull sta-
tistics (De Salvo, 1970). The flexural strengths cited
in Table I.2.8.1 are for specimens tested in four-point
bending, third-point loading (More et al., 1993) with
effective stressed volumes of 1.93 mm3. The pyrolytic
carbon material Weibull modulus is approximately 10
(More et al., 1993).
Fracture toughness levels reflect the brittle nature
of the material, but the fluidized-bed isotropic pyro-
lytic carbons are remarkably fatigue resistant. Recent
fatigue studies indicate the existence of a fatigue thresh-
old that is very nearly the single-cycle fracture strength
(Gilpin et al., 1993; Ma and Sines, 1996, 1999, 2000).
Fatigue-crack propagation studies indicate very high
Paris-law fatigue exponents, on the order of 80, and
display clear evidence of a fatigue-crack propagation
threshold (Ritchie et al., 1990; Beavan et al., 1993;
Cao, 1996).
Crystallographic mechanisms for fatigue-crack initi-
ation, as occur in metals, do not exist in the pyrolytic
carbons (Haubold et al., 1981). In properly designed
and manufactured components, and in the absence of
externally induced damage, fatigue does not occur in
pyrolytic-carbon mechanical heart valve components. In
the 30 years of clinical experience, there have been no
clear instances of fatigue failure. Few pyrolytic carbon
component fractures have occurred, less than 60 out of
more than four million implanted components (Haubold,
1994), and most are attributable to induced damage from
handling or cavitation (Kelpetko et al., 1989; Kafesjian
et al., 1994; Richard and Cao, 1996).
Wear resistance of the fluidized-bed pyrolytic car-
bons is excellent. The strength, stability, and durability
of pyrolytic carbon are responsible for the extension of
mechanical valve lifetimes from less than 20 years to
more than the recipient’s expected lifetime (Schoen et al.,
1982; Schoen, 1983; More and Silver, 1990; Wieting,
1996).
Pyrolytic carbon in heart valve prostheses is often
used in contact with metals, either as a carbon disk
in a metallic valve orifice or as a carbon orifice stiff-
ened with a metallic ring. Carbon falls with the noble
metals in the galvanic series (Haubold et al., 1981),
the sequence being silver, titanium, graphite, gold,
and platinum. Carbon can accelerate corrosion when
coupled to less noble metals in vivo. However, testing
using mixed potential corrosion theory and potentio-
static polarization has determined that no detrimental
effects occur when carbon is coupled with titanium or
cobalt–chrome alloys (Thompson et al., 1979; Griffin
et al., 1983). Carbon couples with stainless steel alloys
are not recommended.
STEPS IN THE FABRICATION OF
PYROLYTIC CARBON COMPONENTS
To convert a gaseous hydrocarbon into a shiny, polished
black component for use in the biological environment
is not a trivial undertaking. Furthermore, because of the
critical importance of long-term implants to a recipient’s
health, all manufacturing operations are performed to
stringent levels of quality assurance under the auspices
of US Food and Drug Administration Good Manufactur-
ing Practices and International Standards Organization
ISO-9000 regulations. As in the case of fabrication of
metallic implants, numerous steps are involved. Pyrolytic
carbon is not machined from a block of material, as is
the case with most metallic implants, nor is it injection
or reactive molded, as are many polymeric devices. An
overview of the processing steps leading to a finished
pyrolytic carbon coated component for use in a medical
device is shown in Figure I.2.8.6, and is further described
in the following sections.
Substrate Material
Since pyrolytic carbon is a coating, it must be depos-
ited on an appropriately shaped, preformed substrate
(preform). Because the pyrolysis process takes place at
high temperatures, the choice of substrates is severely
Steps in the fabrication of
pyrolytic carbon components
Validate substrate
material
Machine preform
Coat preform
Machine to size
Polish
Assemble
FIGURE I.2.8.6 Schematic of manufacturing processing steps.
Author's personal copy

CHAPTER I.2.8 Pyrolytic Carbon for Long-Term Medical Implants 215
Biomaterials Science, Third Edition, 2013, 209–222
limited. Only a few of the refractory materials, such
as tantalum or molybdenum/rhenium alloys and graph-
ite, can withstand the conditions at which the pyrolytic
carbon coating is produced. Some refractory metals
have been used in heart valve components; for example,
Mo/Re preforms were coated to make the struts for the
Beall–Surgitool mitral valve. It is important for the ther-
mal expansion characteristics of the substrate to closely
match those of the applied coating. Otherwise, upon
cooling of the coated part to room temperature the
coating will be highly stressed and can spontaneously
crack. For contemporary heart valve applications, fine-
grained isotropic graphite is the most commonly used
substrate. This substrate graphite can be doped with
tungsten in order to provide radioopacity for X-ray
visualizations of the implants. The graphite substrate
does not impart structural strength. Rather, it provides
a dimensionally stable platform for the pyrolytic car-
bon coating, both at the reaction temperature and at
room temperature.
Preform
Once the appropriate substrate material has been selected
and prior to making a preform, it must be inspected to
ensure that the material meets the desired specifications.
Typically, the strength and density of the starting material
are measured. Thermal expansion is ordinarily validated
and monitored through process control. The preform,
which is an undersized replica of the finished compo-
nent, is normally machined using conventional machin-
ing methods. Because the fine-grained isotropic graphite
is very abrasive, standard machine tools have given way
to diamond-plated or single-point diamond tools. In the
case of heart valves, numerical control machining meth-
ods are often required to maintain critical component
dimensional tolerances. After the preform is completed,
it is inspected to ensure that its dimensions fall within the
specified tolerances and that it contains no visible flaws
or voids.
Coating
Generally numerous preforms are coated in one furnace
run. A batch to be coated is made up of substrates from
a single lot of preforms. Such batch processing by lot is
required in order to maintain “forward and backward”
traceability. In other words, ultimately it is necessary to
know all of the components that were prepared using a
specific material lot, given either the starting material lot
number (forward) or given the specific component serial
number (backward). The number of parts that can be
coated in one furnace run is dictated by the size of the
furnace and the size and weight of the parts to be coated.
The batch of substrates is placed within the fluidized bed
in the vertical tube furnace and is coated to the desired
thickness. Coating times are generally on the order of
a few hours, but the entire cycle (heat-up, coating, and
cool-down) may take as long as a full day.
A statistical sample from each coating lot is taken
for analysis. At this point, typical measurements include
coating thickness, microhardness, and microstructure.
The microhardness and microstructure are determined
from a metallographically prepared cross-section of the
coated component taken perpendicular to the plane of
deposition. Thus, this test is destructive. An example of a
metallographically prepared cross-section of a pyrolytic
carbon component is shown in Figure I.2.8.7.
Machine to Size
The components used to manufacture medical devices
have strict dimensional requirements. Because of the
inability, until recently, to precisely measure and control
bed size, and indirectly coating thickness, the preforms
were generally coated more thickly than necessary to
ensure adequate pyrolytic carbon coating thickness on
the finished part. The strict dimensional requirements
were then achieved through precision grinding or other
machining operations. Because pyrolytic carbon is very
hard, conventional machine tools again cannot be used.
Diamond-plated grinding wheels and other diamond
tooling are required. The dimensions of final machined
parts are again verified.
Polish
The surface of as-deposited, machined, and polished
components is shown in Figure I.2.8.8. It was found
early on in experiments (LaGrange et al., 1969; Saw-
yer et al., 1975; Haubold et al., 1981) that clean pol-
ished pyrolytic carbon surfaces of tubes when placed
within the vasculature of experimental animals accumu-
lated minimal if any thrombus; and certainly less than
FIGURE I.2.8.7 Metallographic mount cross-section of heat valve
component. The light-colored pyrolytic carbon layer is coated
over the interior, darker-colored granular-appearing graphite sub-
strate.
Author's personal copy

216 SECTION I.2 Classes of Materials Used in Medicine
Biomaterials Science, Third Edition, 2013, 209–222
pyrolytic carbon tubes with the as-deposited surface.
Consequently, the surfaces of pyrolytic carbon have
historically been polished, either manually or mechani-
cally, using fine diamond or aluminum oxide pastes and
slurries. The surface finish achieved has roughness mea-
sured on the scale of nanometers. As can be seen from
Table I.2.8.2 (More and Haubold, 1996), the surfaces
of polished pyrolytic carbon (30–50 nm) are an order
of magnitude smoother than the as-deposited surfaces
(300–500 nm).
Once the desired surface quality is achieved, compo-
nents are again inspected. The final component inspec-
tion may include measurement of dimensions, X-ray
inspection in two orientations to verify coating thick-
ness, and visual inspection for surface quality and flaws.
In many cases, automated inspection methods with
computer-controlled coordinate measurement machines
are used. X-ray inspection can be used to ensure that
minimum coating thickness requirements are met. Two
orthogonal views ensure that machining and grinding of
the coating was achieved uniformly, and that the coating
is symmetrical. The machining and grinding operation
after coating is not without the risk of inducing cracks
or flaws in the coating, which may subsequently affect
the service life of the component. Such surface flaws are
detected visually or with the aid of dye-penetrant tech-
niques. Components may also be proof-tested to detect
and eliminate components with subsurface flaws. With
the advent of bed size control, which allows coating to
exact final dimensions, the concerns about flaws intro-
duced during the machining and grinding operation have
been eliminated.
The polished and inspected components, thus pre-
pared, are now ready for assembly into devices, or
are packaged and sterilized in the case of standalone
devices. Shown in Figure I.2.8.9 are the three pyrolytic
carbon components for a bileaflet mechanical heart
valve. The components were selected and matched
for assembly using the data generated from the final
dimensional inspection to achieve the dimensional
requirements specified in the device design. In Fig-
ure I.2.8.10, the pyrolytic carbon components for a
TABLE I.2.8.2 Surface Finish (Ra, Average, and
Rq, Root Mean Square) of Pyrolytic
Carbon Heat Valve Componentsa
Specimen Ra (nm) Rq (nm) Comments
Glass microscope
slide
17.14 26.80
On-X leaflet 33.95 42.12 Clinical
Sorin Bicarbon
leaflet
40.12 50.63 Nonclinical
SJM leaflet 49.71 62.74 Clinical
CMI (SJM) leaflet 67.98 85.56 Nonclinical
Sorin Monoleaflet 99.59 128.10 Clinical
DeBakey–Surgitool
ball
129.78 157.93 Nonclinical
As-coated slab 389.07 503.72
aComponents/prepared by: On-X/Medical Carbon Research Institute, Austin,
TX, USA; Sorin/Sorin Biomedica, Saluggia, Italy; SJM/Saint Jude Medical,
Saint Paul, MN, USA; CMI (SJM)/CarboMedics, Austin, TX, USA; DeBakey-
S/CarboMedics, San Diego, USA (circa 1968). “Clinical” was from as-pack-
aged valve; “nonclinical” lacks component traceability.
FIGURE I.2.8.9 Components for On-X bileaflet heart valve.
(A)
(B)
FIGURE I.2.8.8 Scanning electron microscope micrographs of:
(A) as-coated; and (B) as-polished surfaces.
Author's personal copy

CHAPTER I.2.8 Pyrolytic Carbon for Long-Term Medical Implants 217
Biomaterials Science, Third Edition, 2013, 209–222
replacement metacarpophalangeal total joint prosthe-
sis are shown.
Assembly
The multiple components of a mechanical heart must
be assembled. The brittleness of pyrolytic carbon poses
a significant assembly problem. Because the strain-to-
failure is on the order of 1.28% to 1.58%, there is a
limited range of deformation that can be applied in
order to achieve a proper fit. Relative fit between the
components defines the capture and the range of motion
for components that move to actuate valve open-
ing and closing. Furthermore, component obstructive
bulk and tolerance gaps are critical to hemodynamic
performance.
In designs that use a metallic orifice, the metallic com-
ponents are typically deformed in order to insert the
pyrolytic carbon occluder disk. For the all-carbon bileaf-
let designs, the carbon orifice must be deflected in order
to insert the leaflets. As the valve diameter decreases, and
as the section modulus of the orifice design increases, the
orifice stiffness increases. The possibility of damage or
fracture during assembly was a limiting factor in early
orifice design. For this reason, the orifices in valve designs
using silicon-alloyed pyrolytic carbon were simple cylin-
drical geometries, and the smallest sizes were limited to
the equivalent of a 19 mm diameter tissue annulus. The
simple cylindrical orifice designs are often reinforced
with a metallic stiffening ring that is shrunk on after
assembly. The stiffening ring ensures that physiologi-
cal loading will not produce deflections that can inhibit
valve action or result in leaflet escape.
The increased strain-to-failure of pure pyrolytic car-
bon, relative to the silicon-alloyed carbon, allows designs
with more complex orifice section moduli. This allows
designers to utilize hydrodynamically efficient shapes
such as flared inlets and to incorporate external stiffen-
ing bands that eliminate the need for a metallic stiffening
ring. The increased strain-to-failure of On-X carbon has
been used to advantage in the On-X mechanical heart
valve design (Ely et al., 1998).
Cleaning and Surface Chemistry
Pyrolytic carbon surface chemistry is important because
the manufacturing and cleaning operations to which a
component is subjected can change and redefine the sur-
face that is presented to the blood. Oxidation of carbon
surfaces can produce surface contamination that detracts
from blood compatibility (LaGrange et al., 1969; Bokros
et al., 1969). Historically, the initial examinations of
pyrolytic carbon biocompatibility assumed de facto that
the surface needed to be treated with a thromboresistant
agent such as heparin (Bokros et al., 1969). It was found,
however, that the non-heparin-coated surface was actu-
ally more blood compatible than the treated surface.
Hence, the efforts toward surface coating with heparin
were abandoned.
In general it is desired to minimize the surface oxy-
gen and any other non-carbon surface contaminants.
From X-ray photoelectron spectroscopy (XPS) analyses,
a typical heart valve component surface has 76–86% C,
12–21% O, 0–2% Si, and 1–2% Al (King et al., 1981;
Smith and Black, 1984; More and Haubold, 1996). Pol-
ishing compounds tend to contain alumina, and some
alumina particles may become imbedded in the carbon
surface. Other contaminants that may be introduced at
low levels (<2% each) are Na, B, Cl, S, Mg, Ca, Zn, and N.
The XPS carbon 1s peak when scanned at high resolu-
tion can be deconvoluted to determine carbon oxidation
states. The carbon 1s peak will typically consist primar-
ily of hydrocarbon-like carbon (60–81%), ether alcohol/
ester-like carbon (10–24%), ketone-like carbon (0–6%),
and ester/acid-like carbon (1–12%) (More and Haubold,
1996). Each manufacturing, cleaning, and sterilizing
operation potentially redefines the surface. The effect of
modified surface chemistry on blood compatibility is not
well characterized, so this adds a level of uncontrolled
variability when considering the blood compatibility
of pyrolytic-carbon heart valve materials from differ-
ent manufacturing sources and different investigators.
In general, the presence of oxygen and surface contami-
nants should be eliminated.
BIOCOMPATIBILITY OF PYROLYTIC
CARBON
The suitability of a material for use in an implant is a
complex issue. Biocompatibility testing is the focus of
other chapters. In the case of pyrolytic carbon, its suc-
cessful history interfacing with blood in mechanical heart
valves attests to its suitability for this application. A note
of caution, however, is in order. Until about a decade
ago, the pyrolytic carbon used so successfully in mechan-
ical heart valves was produced by a single manufacturer;
the material, many applications in the biological envi-
ronment, and the processes for producing the material,
were all patented. Since the expiration of the last of these
patents in 1989, other sources for pyrolytic carbon have
FIGURE I.2.8.10 Replacement metacarpophalangeal total joint
prosthesis components, Ascension Orthopedics, Austin TX, USA.
Author's personal copy

218 SECTION I.2 Classes of Materials Used in Medicine
Biomaterials Science, Third Edition, 2013, 209–222
appeared that are copies of the original General Atomic
material. When considering alternative carbon mate-
rials, it is important to recognize that the proper com-
bination of physical, mechanical, and blood-compatible
properties is required for the success of the implant
application. Furthermore, because there are a number of
different possible pyrolysis processes, it should be rec-
ognized that each can result in different microstructures
with different properties. Just because a material is car-
bon, a turbostratic carbon or a pyrolytic carbon does not
qualify its use in a long-term human implant (Haubold
and Ely, 1995). For example, pyrolytic carbons prepared
by chemical vapor deposition processes, other than
the fluidized-bed process, are known to exhibit anisot-
ropy, nonhomogeneity, and considerable variability in
mechanical properties (Agafonov et al., 1999). Although
these materials may exhibit biocompatibility, the poten-
tial for variability in structural stability and durability
may lead to valve dysfunction.
The original General Atomic-type fluidized-bed
pyrolytic carbons all demonstrate negligible reactions
in the standard Tripartite and ISO 10993-1 type bio-
compatibility tests. Results from such tests are given in
Table I.2.8.3 (Ely et al., 1998). Pure pyrolytic carbon
is so non-reactive that it can serve as a negative control
for these tests.
It is believed that pyrolytic carbon owes its demon-
strated blood compatibility to its inertness and to its
ability to quickly absorb proteins from blood without
triggering a protein denaturing reaction. Ultimately,
the blood compatibility is thought to be a result of the
protein layer formed upon the carbon surface. Baier
observed that pyrolytic carbon surfaces have a relatively
high critical surface tension of 50 dyn/cm, which imme-
diately drops to 28–30 dyn/cm following exposure to
blood (Baier et al., 1970). The quantity of sorbed protein
was thought to be an important factor for blood com-
patibility. Lee and Kim (1974) quantified the amount of
radiolabeled proteins sorbed from solutions of mixture
proteins (albumin, fibrinogen, and gamma-globulin).
While pyrolytic carbon does adsorb albumin, it also
adsorbs a considerable quantity of fibrinogen, as shown
in Figure I.2.8.11. As can be seen in Figure I.2.8.11, the
amount of fibrinogen adsorbed on pyrolytic carbon sur-
faces is far greater than the amount of albumin on these
surfaces, and is comparable to the amount of fibrinogen
that is sorbed on silicone rubber. The mode of albumin
adsorption, however, appears to be drastically different
for these two materials. Albumin sorbs immediately on
the pyrolytic carbon surfaces, whereas the build-up of
fibrinogen is much slower. In the case of silicone rubber,
both proteins sorb at a much slower rate. It appears that
TABLE I.2.8.3 Biological Testing of Pure PyC
Test Description Protocol Results
Klingman maximization ISO/CD 10993–10 Grade 1; not significant
Rabbit pyrogen ISO/DIS 10993–11 Nonpyrogenic
Intracutaneous injection ISO 10993–10 Negligible irritant
Systemic injection ANSI/AAMI/ISO 10993–11 Negative – same as controls
Salmonella typhimurium reverse
mutation assay
ISO 10993–3 Nonmutagenic
Physico-chemical USP XXIII, 1995 Exceeds standards
Hemolysis – rabbit blood ISO 10993–4/NIH 77–1294 Nonhemolytic
Elution test (L929 mammalian cell culture) ISO 10993–5, USP XXIII, 1995 Noncytotoxic
Time (min)
0 10 20 30 40 50 60 70 80 90 100 110
Amount of adsorbent (µg/cm2)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Fibrinogen on silastic rubber
Albumin on silastic rubber
Fibrinogen on PyC
Albumin on PyC
FIGURE I.2.8.11 Fibrinogen and albumin adsorption on pyrolytic carbon (PyC) and Silastic silicone rubber.
Author's personal copy

CHAPTER I.2.8 Pyrolytic Carbon for Long-Term Medical Implants 219
Biomaterials Science, Third Edition, 2013, 209–222
the mode of protein adsorption is important, and not the
total amount sorbed.
Nyilas and Chiu (1978) studied the interaction of
plasma proteins with foreign surfaces by measuring
directly the heats of adsorption of selected proteins onto
such surfaces using microcalorimetric techniques. They
found that the heats of adsorption of fibrinogen, up to
the completion of first monolayer coverage, are a factor
of eight smaller on pyrolytic carbon surfaces than on the
known thrombogenic control (glass) surface, as shown
in Figure I.2.8.12. Furthermore, the measured net heats
of adsorption of gamma globulin on pyrolytic carbon
were about 15 times smaller than those on glass. They
concluded that low heats of adsorption onto a foreign
surface imply small interaction forces with no confor-
mational changes of the proteins that might trigger the
clotting cascade. It appears that a layer of continu-
ously exchanging blood proteins in their unaltered state
“masks” the pyrolytic carbon surfaces from appearing as
a foreign body.
There is further evidence that the minimally altered
sorbed protein layers on pyrolytic carbon condition
blood compatibility. Salzman et al. (1977), for exam-
ple, observed a significant difference in platelet reac-
tion with pyrolytic carbon beads in packed columns
prior to and after pretreatment with albumin. With no
albumin preconditioning treatment, platelet retention
by the columns was high, but the release of platelet
constituents was low. However, with albumin pretreat-
ment, platelet retention and the release of constituents
was minimal.
The foregoing observations led to the view that pyro-
lytic carbon owes its demonstrated blood compatibility
to its inertness, and to its ability to quickly adsorb pro-
teins from blood without triggering a protein-denaturing
reaction (Nyilas and Chiu, 1978; Haubold et al., 1981).
However, the assertion that pyrolytic carbon is an inert
material and induces minimal conformational changes in
adsorbed protein was re-examined by Feng and Andrade
(1994). Using differential scanning calorimetry and a
variety of proteins and buffers, they found that pyrolytic
carbon surfaces denatured all of the proteins studied.
They concluded that whether or not a surface denatures
protein cannot be the sole criteria for blood compat-
ibility. Their suggestion was that the specific proteins
and the sequence in which they are denatured may be
important. For example, it was suggested that pyrolytic
carbon may first adsorb and denature albumin, which
forms a layer that subsequently passivates the surface
and inhibits thrombosis.
Chinn et al. (1994) re-examined the adsorption of
albumin and fibrinogen on pyrolytic carbon surfaces and
noted that relatively large amounts of fibrinogen were
adsorbed, and speculated that the adsorbed fibrinogen
was rapidly converted to a non-elutable form. If the
elutable form is more reactive to platelets than the non-
elutable form, then the non-elutable protein layer may
contribute to the passivating effect.
Work on visualizing the carbon surface and plate-
let adhesion done by Goodman et al. (1995) using low
accelerating-voltage scanning electron microscopy, along
with critical-point drying techniques, has discovered that
the platelet spreading on pyrolytic carbon surfaces is
more extensive than previously observed (Haubold et al.,
1981). However, platelet loading was in a static flow sit-
uation that does not model the physiological flow that a
heart valve is subjected to. Hence, this approach cannot
resolve kinetic effects on platelet adhesion. However,
Okazaki, Tweden, and co-workers observed adherent
platelets on valves following implantation in sheep that
were not treated with anticoagulants (Okazaki et al.,
1997). There were no instances of valve thrombosis,
even though platelets were present on some of the valve
surfaces. But the relevance of this observation to clinical
valve thromboses is not clear, because human patients
with mechanical heart valves undergo chronic anticoag-
ulant therapy (Edmunds, 1987), and have a hemostatic
system different from that of sheep.
A more contemporary version of the mechanism of
pyrolytic carbon blood compatibility might be to reject
the assumption that the surface is inert, as it is now
thought by some that no material is totally inert in the
body (Williams, 1998), and to accept that the blood–
material interaction is preceded by a complex, inter-
dependent, and time-dependent series of interactions
between the plasma proteins and the surface (Hanson,
1998) that is as yet poorly understood. To add to the
confusion, it must also be recognized that much of the
Relative surface coverage
Integral net heat of sorption (CAL/MOLEX ´ 106)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
0
2
4
6
8
10
12
14
16
18
20
22
24
26
Fibrinogen on Glass 37°C
Fibrinogen on Glass 25°C
Fibrinogen on PyC 37°C
Fibrinogen on PyC 25°C
FIGURE I.2.8.12 Integral heat of sorption for fibrinogen on glass
and fibrinogen on PyC at two different temperatures (Nyilas and
Chiu, 1978).
Author's personal copy

220 SECTION I.2 Classes of Materials Used in Medicine
Biomaterials Science, Third Edition, 2013, 209–222
aforementioned conjecture depends on the assumption
that all of the carbon surfaces studied were in fact pure
and comparable to one another.
CLINICAL APPLICATIONS
Widespread clinical use of pyrolytic carbon components
for heart valve replacement began in October of 1968
when Dr. Michael DeBakey implanted an aortic valve
with a hollow Pyrolite® carbon ball occluder. Following
this first implant, several million PyC mechanical valve
prostheses have been implanted worldwide, generating
an experience on the order of 20 million patient-years.
Use of PyC to replace polymers in valve prostheses was
declared an “exceptional event” (Sadeghi, 1987) because
the superior durability, stability, and biocompatibility of
PyC enabled valves to endure for the patient’s lifetime.
However, patients with mechanical valve prostheses
require chronic anticoagulation therapy because of the
risk of valve-related hemostatic complications. In efforts
to reduce the risk of hemostatic complications and the
need for chronic anticoagulation, it was hypothesized
that valve-related hemostatic complications were in part
due to flow-induced mechanical trauma to the blood,
and to the presence of PyC itself because of the potential
thrombogenicity of the silicon-carbide alloy constituent
(LaGrange et al., 1969).
In 1992, advances in pyrolytic carbon manufacturing
technology enabled precise control of processing parame-
ters (Bokros et al., 1994). With this precise control it was
possible to produce pure isotropic pyrolytic carbons having
significantly improved properties relative to silicon-alloyed
PyC, thus eliminating the need for the silicon-carbide alloy
(see Table I.2.8.1). Precise process control also enabled a
coat-to-size capability needed to eliminate surface blem-
ishes caused by post-coating machining processes such as
grinding.
The On-X valve shown in Figure I.2.8.9 was specifi-
cally designed to exploit the improved pure PyC which
enables features to reduce flow-induced trauma (Bokros
et al., 1994, 1997, 1998, 2003), and introduced into
clinical practice (FDA, 2001, 2002). The success of the
On-X valve in general clinical experience, particularly
in noncompliant anticoagulant therapy patient popula-
tions (Williams and van Riet, 2006) and animal studies
(Flameng and Meuris, 2002) strongly indicated improve-
ments in mechanical valve-related hemostatic complica-
tions. Data from this experience was used to justify the
first and only FDA approved non-warfarin and reduced
warfarin prospective randomized trial for a mechanical
heart valve. This trail for the On-X Prosthetic Heart
Value, with the objective to reduce anticoagulation
bleeding risk, was initiated at Emory Crawford Long
Hospital, Atlanta in 2006. The trial is currently ongo-
ing at 40 institutions to include 1200 patients. The high
risk aortic group has been fully enrolled, with more than
500 pt-yr experience, maintained with a reduced dose
of Coumadin (INR of 1.5 to 2.0) and aspirin (81 mg/
dy). The low risk aortic group is maintained with plate-
let inhibitors, (75 mg/dy plavix) and aspirin. The mitral
group enrollment was extended to 2012; this group
is maintained at an INR of 2.0 to 2.5 and aspirin (81
mg/dy). A successful outcome for this trial would offer
further improvements in quality of life for a significant
number of On-X mechanical heart valve recipients.
During the past 28 years PyC has also been used as a
loadbearing material for small orthopedic joint replace-
ment implants. Such prostheses relieve pain, correct
deformity, and improve the appearance of joints dam-
aged by disease such as rheumatoid arthritis, osteoar-
thritis, and post-traumatic conditions. However, all
PyC joint replacements require careful patient selection
for good quality bone and soft tissue. As is true with all
implants, prosthesis sizing, alignment, and interactions
with soft tissue are critical considerations during implan-
tation surgery and rehabilitation.
Successful applications for upper limb total joint
prostheses include the metacarpophalangeal (MCP)
joint (Figure I.2.8.10) and the proximal interphalangeal
(PIP) joint (Cook et al., 1999; Bravo et al., 2007). Pure
isotropic PyC is a nearly ideal material for orthopedic
application, with demonstrated advantages over tradi-
tional materials such as polymers, ceramics, and metals
(Stanley et al., 2008) which include:
• Elimination of wear-related failures
• Absence of osteolytic adverse tissue reactions
• Excellent fatigue resistance
• Non-cemented fixation via bone apposition
• Minimization of stress shielding effects and bone
resorption
• Excellent compatibility with joint cartilage and bone
tissues.
This excellent compatibility with cartilage and bone
tissue enables a number of hemiarthroplasty appli-
cations, in which only one component of the joint is
replaced leaving the PyC device bearing and articulat-
ing against native synovial surfaces. Successful devices
for hemiarthroplasty include the MCP and PIP joints,
carpometacarpal (CMC) joints, radial head, lunate, and
inter-positional articulating surface spacers for use in the
CMC joint. Currently, approximately 18,000 PyC small
joint and hemiarthroplasty devices have been implanted
worldwide.
Given the clinical success of the small joint implants,
enhanced compatibility with joint tissue, superior dura-
bility, and potential significantly extended device life-
times, efforts are currently underway to use PyC as a
platform for larger joint implants such as the shoulder,
knee, and hip. In the larger loadbearing joints, a viable
strategy is to use PyC as the bearing surface in conser-
vative resurfacing devices and in total joint modular
devices. The mechanical valve experience has demon-
strated excellent PyC compatibility with traditional
Author's personal copy

CHAPTER I.2.8 Pyrolytic Carbon for Long-Term Medical Implants 221
Biomaterials Science, Third Edition, 2013, 209–222
implant material metals, ceramics, and polymers. This
PyC materials compatibility lends great versatility in
design for modular devices. We fully expect that PyC
devices will prove more functional, aesthetic, durable,
and complication free than implants with traditional
materials only.
CONCLUSION
Because the blood compatibility of pyrolytic carbon in
mechanical heart valves is not perfect, anticoagulant
therapy is required for mechanical heart valve patients.
However, pyrolytic carbon has been the most success-
ful material in heart valve applications because it offers
excellent blood and tissue compatibility which, com-
bined with the appropriate set of physical and mechani-
cal properties and durability, allows for practical implant
device design and manufacture. Improvements in bio-
compatibility are desired, of course, because when heart
valves and other implants are used, a deadly or disabling
disease is often treated by replacing it with a less patho-
logical, more manageable chronic condition. Ideally, an
implant should not lead to a chronic condition.
It is important to recognize that the mechanism for
the blood compatibility of pyrolytic carbons is not
fully understood, nor is the interplay between the bio-
material itself, design-related hemodynamic stresses,
and the ultimate biological reaction. The elucidation
of the mechanism for blood and tissue compatibility of
pyrolytic carbon remains a challenge.
It is also worth restating that the suitability of carbon
materials from new sources for long-term implants is not
assured simply because the material is carbon. Elemental
carbon encompasses a broad spectrum of possible struc-
tures and mechanical properties. Each new candidate
carbon material requires a specific assessment of biocom-
patibility based on its own merits, and not by reference
to the historically successful General Atomic-type pyro-
lytic carbons.
BIBLIOGRAPHY
Agafonov, A., Kouznetsova, E., Kouznetsova, V., & Reif, T. (1999).
TRI carbon strength and macroscopic isotropy of boron carbide
alloyed pyrolytic carbon. Artif. Organs, 23(7), 80.
Akins, R. J., & Bokros, J. C. (1974). The deposition of pure and
alloyed isotropic carbons and steady state fluidized beds. Car-
bon, 12, 439–452.
Baier, R. E., Gott, V. L., & Feruse, A. (1970). Surface chemical
evaluation of thrombo-resistant materials before and after
venous implantation. Trans. Am. Soc. Artif. Intern. Organs,
16, 50–57.
Beavan, L. A., James, D. W., & Kepner, J. L. (1993). Evaluation
of Fatigue in Pyrolite Carbon. In P. Ducheyne, & D. Christian-
sen (Eds.), Bioceramics Vol. 6, (pp. 205–210). Oxford, UK:
Butterworth-Heinemann.
Bokros, J. C. (1969). Deposition, Structure and Properties of Pyro-
lytic Carbon. In P. L. Walker (Ed.), Chemistry and Physics of
Carbon Vol. 5, (pp. 1–118). New York, NY: Marcel Dekker.
Bokros, J. C., Gott, V. L., LaGrange, L. D., Fadall, A. M., Vos,
K. D., et al. (1969). Correlations between blood compatibility
and heparin adsorptivity for an impermeable isotropic pyro-
lytic carbon. J. Biomed. Mater. Res., 3, 497–528.
Bokros, J. C., Emken, M. R., Haubold, A. D., et al. (1994). Pros-
thetic heart valve. US Patent No. 5,308,361: Issued May 3,
1994.
Bokros, J. C., Ely, J. L., Emken, M. R., et al. (1997). Prosthetic
heart valve. US Patent No. 5,642,324: Issued June 24, 1997.
Bokros, J. C., Ely, J. L., Emken, M. R., et al. (1998). Pros-
thetic heart valve with improved blood flow. US Patent No.
5,772,694: Issued June 30, 1998.
Bokros, J. C., Stupka, J. C., More, R. B., et al. (2003). Prosthetic
heart valve. US Patent No. 6,096,075: Issued August 1, 2003.
Bravo, C. J., Rizzo, R., Hormel, K. B., & Beckenbaugh, R. D.
(2007). Pyrolytic carbon proximal interphalangeal joint
arthroplasty: Results with minimum two-year follow-up evalu-
ation. J. Hand Surg., 32A, 1–11.
Cao, H. (1996). Mechanical performance of pyrolytic carbon
in prosthetic heart valve applications. J. Heart Valve Dis.,
5(Suppl. I), S32–S49.
Chinn, J. A., Phillips, R. E., Lew, K. R., & Horbett, T. A. (1994).
Fibrinogen and albumin adsorption to pyrolite carbon. Trans.
Soc. Biomater., 17, 250.
Cook, S. D., Beckenbaugh, R. D., Redondo, J., Popich, L. S.,
Klawitter, J. J., et al. (1999). Long term follow-up of pyro-
lytic carbon metacarpophalangeal implants. J. Bone and Joint
Surg., 81A(5), 635–648.
De Salvo, G. (1970). Theory and Structural Design Applications of
Weibull Statistics, Report WANL-TME-2688, Westinghouse
Electric Corporation.
Dillard, J. G. (1995). X-ray Photoelectron Spectroscopy (XPS)
and Electron Spectroscopy for Chemical Analysis (ESCA). In
H. Ishida, & L. E. Fitzpatrick (Eds.), Characterization of Com-
posite Materials Vol. 1, (pp. 22). Boston, MA: Butterworth-
Heinemann.
Edmunds, L. H. (1987). Thrombotic and bleeding complications
of prosthetic heart valves. Ann. Thorac. Surg., 44, 430–445.
Ely, J. L., Emken, M. R., Accuntius, J. A., Wilde, D. S., Haubold,
A. D., et al. (1998). Pure pyrolytic carbon: Preparation and
properties of a new material, on-X carbon for mechanical
heart valve prostheses. J. Heart Valve Dis., 7, 626–632.
Emken, M. R., Bokros, J. C., Accuntius, J. A., & Wilde, D. S.
(1993). Precise control of pyrolytic carbon coating. Buffalo,
New York: Presented at the 21st Biennial Conference on Car-
bon. June 13–18, 1993, Extended Abstracts and Program Pro-
ceedings, pp. 531–532.
Emken, M. R., Bokros, J. C., Accuntius, J. A., & Wilde, D. S.
(1994). U.S. Patent No. 5,284,676, Pyrolytic deposition in a
fluidized bed. Feb. 8, 1994.
Feng, L., & Andrade, J. D. (1994). Protein adsorption on low-tem-
perature isotropic carbon: I. Protein conformational change
probed by differential scanning calorimetry. J. Biomed. Mater.
Res., 28, 735–743.
Flameng, W., & Meuris, B. (2002). Performance of bileaflet heart
valve prostheses in an animal model of heart valve thrombo-
sis. 6th Annual Hilton Head Workshop on Prosthetic Heart
Valves. South Carolina: Hilton Head. March 610.
Gilpin, C. B., Haubold, A. D., & Ely, J. L. (1993). Fatigue Crack
Growth and Fracture of Pyrolytic Carbon Composites. In
P. Ducheyne, & D. Christiansen (Eds.), Bioceramics Vol. 6,
(pp. 217–223). Oxford, UK: Butterworth-Heinemann.
Goodman, S. L., Tweden, K. S., & Albrecht, R. M. (1995). Three-
dimensional morphology and platelet adhesion on pyrolytic
carbon heart valve materials. Cells Mater., 5(1), 15–30.
Griffin, C. D., Buchanan, R. A., & Lemons, J. E. (1983). In vitro
electrochemical corrosion study of coupled surgical implant
materials. J. Biomed. Mater. Res., 17, 489–500.
Author's personal copy

222 SECTION I.2 Classes of Materials Used in Medicine
Biomaterials Science, Third Edition, 2013, 209–222
Hanson, S. R. (1998). Blood–Material Interactions. In J. Black, &
G. Hastings (Eds.), Handbook of Biomaterial Properties (pp.
545–555). London, UK: Chapman and Hall.
Haubold, A. D. (1994). On the durability of pyrolytic carbon
in vivo. Medi. Prog. Technol., 20, 201–208.
Haubold, A. D., & Ely, J. L. (1995). Carbons used in mechanical
heart valves. Transactions Society for Biomaterials. San Fran-
cisco: 21st Annual Meeting. 275.
Haubold, A. D., Shim, H. S., & Bokros, J. C. (1981). Carbon in
Medical Devices. In D. P. Williams (Ed.), Biocompatibility of
Clinical Implant Materials (pp. 3–42). Boca Raton, Florida:
CRC Press.
Kaae, J. L., & Wall, D. R. (1996). Microstructural characterization
of pyrolytic carbon for heart valves. Cells Mater., 4, 281–290.
Kafesjian, R., Howanec, M., Ward, G. D., Diep, L., Wagstaff, L.,
et al. (1994). Cavitation damage of pyrolytic carbon in mechan-
ical heart valves. J. Heart Valve Dis., 3(Suppl. I), S2–S7.
Kelpetko, V., Moritz, A., Mlczoch, J., Schurawitzki, H., Domanig,
E., et al. (1989). Leaflet fracture in Edwards-Duromedics bile-
aflet valves. J. Thorac. Cardiovasc. Surg., 97, 90–94.
King, R. N., Andrade, J. D., Haubold, A. D., & Shim, H. S. (1981).
Surface Analysis of Silicon: Alloyed and Unalloyed LTI Pyro-
lytic Carbon. In D. W. Dwight, T. J. Fabish, & H. R. Thomas
(Eds.), Photon, Electron and Ion Probes of Polymer Structure
and Properties, ACS Symposium Series 162 (pp. 383–404).
Washington, DC: American Chemical Society.
LaGrange, L. D., Gott, V. L., Bokros, J. C., & Ramos, M. D.
(1969). Compatibility of Carbon and Blood. In R. J. Hegy-
eli (Ed.), Artificial Heart Program Conference Proceedings
(pp. 47–58). Washington, DC: US Government Printing Office.
Lee, R. G., & Kim, S. W. (1974). Adsorption of proteins onto
hydrophobic polymer surfaces: Adsorption isotherms and
kinetics. J. Biomed. Mater. Res., 8, 251.
Ma, L., & Sines, G. (1996). Fatigue of isotropic pyrolytic carbon
used in mechanical heart valves. J. Heart Valve Dis., 5(Suppl. I),
S59–S64.
Ma, L., & Sines, G. (1999). Unalloyed pyrolytic carbon for
implanted heart valves. J. Heart Valve Dis., 8(5), 578–585.
Ma, L., & Sines, G. (2000). Fatigue behavior of pyrolytic carbon.
J. Biomed. Mater. Res., 51, 61–68.
More, R. B., & Haubold, A. D. (1996). Surface chemistry and sur-
face roughness of clinical pyrocarbons. Cells Mater., 6, 273–279.
More, R. B., & Silver, M. D. (1990). Pyrolytic carbon prosthetic
heart valve occluder wear: in vivo vs. in vitro results for the
Björk–Shiley prosthesis. J. Appl. Biomater., 1, 267–278.
More, R. B., Kepner, J. L., & Strzepa, P. (1993). Hertzian Fracture
in Pyrolite Carbon. In P. Ducheyne, & D. Christiansen (Eds.),
Bioceramics Vol. 6, (pp. 225–228). Oxford, UK: Butterworth-
Heinemann.
Nyilas, E., & Chiu, T. H. (1978). Artificial surface/sorbed pro-
tein structure/hemocompatibility correlations. Artif. Organs,
2(Suppl.), 56–62.
Okazaki, Y., Wika, K. E., Matsuyoshi, T., Fukamachi, K.,
Kunitomo, R., et al. (1997). Platelets were early postoperative
depositions on the leaflet of a mechanical heart valve in sheep
without postoperative anticoagulants or antiplatelet agents.
ASAIO J., 42, M750–M754.
Pauling, L. (1964). College Chemistry (3rd ed). San Francisco, CA:
W. H. Freeman and Company.
Reilly, D. T., & Burstein, A. H. (1974). The mechanical properties
of bone. J. Bone Joint Surg. Am., 56, 1001.
Reilly, D. T., Burstein, A. H., & Frankel, V. H. (1974). The elastic
modulus for bone. J. Biomech., 7, 271.
Richard, G., & Cao, H. (1996). Structural failure of pyrolytic car-
bon heart valves. J. Heart Valve Dis., 5(Suppl. I), S79–S85.
Ritchie, R. O., Dauskardt, R. H., Yu, W., & Brendzel, A. M.
(1990). Cyclic fatigue-crack propagation, stress corrosion and
fracture toughness behavior in pyrolite carbon coated graphite
for prosthetic heart valve applications. J. Biomed. Mat. Res.,
24, 189–206.
Sadeghi, H. (1987). Dysfonctions des prostheses valvulaires
cardaques et leur traitment chirgical. Schwiez. Med. Wochen-
schr., 117, 1665–1670.
Salzman, E. W., Lindon, J., Baier, D., & Merril, E. W. (1977).
Surface-induced platelet adhesion, aggregation and release.
Ann. NY. Acad. Sci., 283, 114.
Sattler, K. (1995). Scanning tunneling microscopy of carbon nano-
tubes and nanocones. Carbon, 7, 915–920.
Sawyer, P. N., Lucas, L., Stanczewski, B., Ramasamy, N., Kam-
mlott, G. W., et al. (1975). Evaluation techniques for potential
cardiovascular prosthetic alloys experience with titanium alu-
minum 6–4 ELI tubes. Proceedings of the San Diego Biomedi-
cal Symposium, Vol. 14, 423–427.
Schoen, F. J. (1983). Carbons in Heart Valve Prostheses: Foun-
dations and Clinical Performance. In M. Zycher (Ed.), Bio-
compatible Polymers, Metals and Composites (pp. 240–261).
Lancaster, PA: Technomic.
Schoen, F. J., Titus, J. L., & Lawrie, G. M. (1982). Durability of
pyrolytic carbon-containing heart valve prostheses. J. Biomed.
Mater. Res., 16, 559–570.
Smith, K. L., & Black, K. M. (1984). Characterization of the
treated surfaces of silicon alloyed pyrolytic carbon and SiC. J.
Vac. Sci. Technol., A2, 744–747.
Stanley, J., Klawitter, J., & More, R. (2008). Replacing Joints with
Pyrolytic Carbon. In P. A. Revell (Ed.), Joint Replacement
Technology: New Developments (pp. 631–656). Cambridge,
UK: Woodhead Pub and CRC Press LLC.
FDA (US Food and Drug Administration) (2001). Summary of
Safety and Effectiveness, On-X® Prosthetic Heart Valve. FDA
PMA P000037 May 30.
FDA (US Food and Drug Administration) (2002). Summary of
Safety and Effectiveness, On-X® Prosthetic Heart Valve. FDA
P000037/S1, March 6, 2002, European Primary Trial, updated
May 31, 2003.
Thompson, N. G., Buchanan, R. A., & Lemons, J. E. (1979).
In vitro corrosion of Ti-6Al-4V and Type 316L stainless steel
when galvanically coupled with carbon. J. Biomed. Mater.
Res., 13, 35–44.
Wieting, D. W. (1996). The Björk–Shiley Delrin tilting disc heart
valve: Historical perspective, design and need for scien-
tific analyses after 25 years. J. Heart Valve Dis., 5(Suppl. I),
S157–S168.
Williams, D. F. (1998). General Concepts of Biocompatibility. In J.
Black, & G. Hastings (Eds.), Handbook of Biomaterial Prop-
erties (pp. 481–489). London, UK: Chapman and Hall.
Williams, M. A., & van Riet, S. (2006). The On-X® prosthetic
heart valve: Mid-term results in a poorly anticoagulated popu-
lation. J. Heart Valve Dis., 15, 80–86.
Author's personal copy