Fibers, 5.

Synthetic Inorganic 1

Fibers, 5. Synthetic Inorganic

Erich Fitzer, Institut für Chemische Technik der Universität Karlsruhe, Karlsruhe, Federal Republic of
Germany (Chaps.1 and 5)
Rudolf Kleinholz, Gevetex Textilglas-GmbH, Herzogenrath, Federal Republic of Germany (Section 2.1)
Hartmut Tiesler, Grünzweig + Hartmann und Glasfaser AG, Ladenburg, Federal Republic ofGermany
(Section 2.2)
Martyn Hugh Stacey, ICI Chemicals & Polymers Ltd., Runcorn, United Kingdom (Chaps. 3 and 6)
Roger De Bruyne, N.V. Bekaert S.A., Zwevegem, Belgium (Chap. 4)
Ignace Lefever, N.V. Bekaert S.A., Zwevegem, Belgium (Chap. 4)
Michael Heine, Institut für Chemische Technik der Universität Karlsruhe, Karlsruhe, Federal Republic of
Germany (Chap. 5)

1. Survey . . . . . . . . . . . . . . . . . . 3 2.1.5. Toxicology and Occupational Health 21

1.1. General Description of Inorganic 2.1.6. Economic Aspects . . . . . . . . . . . 22
Fibers . . . . . . . . . . . . . . . . . . 3 2.1.7. Producers . . . . . . . . . . . . . . . . 22
1.1.1. Commercial Forms and Terminol- 2.2. Glass Fibers for Insulation . . . . 22
ogy . . . . . . . . . . . . . . . . . . . . 3 2.2.1. Physical Properties . . . . . . . . . . 22
1.2. Crystalline Structure 2.2.2. Chemical Properties . . . . . . . . . . 23
and Mechanical Properties . . . . 4 2.2.3. Raw Materials and Chemical Com-
1.3. Testing Methods for Inorganic position . . . . . . . . . . . . . . . . . 23
Fibers . . . . . . . . . . . . . . . . . . 6 2.2.4. Production . . . . . . . . . . . . . . . . 24
1.4. Inorganic Fibers for Thermal In- 2.2.5. Environmental Protection . . . . . . 27
sulation . . . . . . . . . . . . . . . . . 7 2.2.6. Quality and Analysis . . . . . . . . . 28
1.5. Inorganic Fibers for Reinforce- 2.2.7. Applications . . . . . . . . . . . . . . 28
ment Purposes . . . . . . . . . . . . . 8 2.2.8. Toxicology and Occupational Health 29
1.5.1. Principles of the Fiber Reinforce- 3. Refractory Fibers . . . . . . . . . . 29
ment . . . . . . . . . . . . . . . . . . . 8 3.1. Properties and Raw Materials . . 30
1.5.2. Commercial Continuous Inorganic 3.1.1. Physical Properties . . . . . . . . . . 30
Fibers . . . . . . . . . . . . . . . . . . . 9 3.1.2. Chemical Properties . . . . . . . . . . 33
1.5.3. Surface Chemistry, Surface Treat- 3.1.3. Resources, Raw Materials . . . . . . 33
ment, and Surface Properties . . . . 11 3.2. Production . . . . . . . . . . . . . . . 34
1.5.4. High-Temperature Behavior of Inor- 3.2.1. Boron Monofilament . . . . . . . . . 34
ganic Fibers . . . . . . . . . . . . . . . 12 3.2.2. Silicon Carbide Fibers . . . . . . . . 35
2. Glass Fibers . . . . . . . . . . . . . . 13 3.2.3. High Alumina Fibers . . . . . . . . . 36
2.1. Textile Glass Fibers . . . . . . . . . 13 3.2.4. Zirconia Fibers . . . . . . . . . . . . . 37
2.1.1. Chemical Composition and Proper- 3.3. Environmental Protection . . . . . 38
ties . . . . . . . . . . . . . . . . . . . . 14 3.4. Uses . . . . . . . . . . . . . . . . . . . 38
2.1.2. Manufacture . . . . . . . . . . . . . . 16 3.4.1. High-Temperature Insulation and
2.1.2.1. Glass Filaments . . . . . . . . . . . . 16 Filtration . . . . . . . . . . . . . . . . . 38
2.1.2.2. Glass Staple Fibers . . . . . . . . . . 18 3.5. Miscellaneous Uses . . . . . . . . . . 39
2.1.3. Textile Glass Products: Processing 3.6. Economic Aspects . . . . . . . . . . 39
and Application . . . . . . . . . . . . 18 3.7. Toxicology and Occupational
2.1.3.1. Product Designation . . . . . . . . . . 18 Health . . . . . . . . . . . . . . . . . . 40
2.1.3.2. Correlation Between Processing and 4. Metal Fibers . . . . . . . . . . . . . . 40
Application . . . . . . . . . . . . . . . 18 4.1. Production . . . . . . . . . . . . . . . 40
2.1.3.3. Description of Textile Glass Prod- 4.1.1. Bundle Drawing . . . . . . . . . . . . 40
ucts . . . . . . . . . . . . . . . . . . . . 19 4.1.2. Melt Spinning . . . . . . . . . . . . . 41
2.1.3.4. Applications . . . . . . . . . . . . . . 20 4.1.3. Other Processes . . . . . . . . . . . . 42
2.1.4. Product Standards . . . . . . . . . . . 21 4.2. Bundle-Drawn Fibers . . . . . . . . 43

c 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

10.1002/14356007.a11 001
2 Fibers, 5. Synthetic Inorganic

4.2.1. Form and Dimensions . . . . . . . . 43 5.3. Production . . . . . . . . . . . . . . . 56

4.2.2. Composition . . . . . . . . . . . . . . 43 5.3.1. PAN Precursor Fiber . . . . . . . . . 57
4.3. Uses . . . . . . . . . . . . . . . . . . . 43 5.3.2. Stabilization . . . . . . . . . . . . . . . 58
4.4. Manufacturers . . . . . . . . . . . . 45 5.3.3. Process Parameters . . . . . . . . . . 59
4.5. Melt Spun Fibers . . . . . . . . . . . 45 5.4. The Pitch Fiber Process . . . . . . 59
4.5.1. Properties . . . . . . . . . . . . . . . . 45 5.4.1. Isotropic and Anisotropic Pitch
4.6. Uses . . . . . . . . . . . . . . . . . . . 45 Fiber Precursors . . . . . . . . . . . . 59
5. Carbon Fibers . . . . . . . . . . . . . 45 5.4.2. The Mesophase Formation . . . . . . 60
5.1. Properties . . . . . . . . . . . . . . . . 47 5.4.3. Process Steps . . . . . . . . . . . . . . 60
5.1.1. Mechanical Properties . . . . . . . . 47
5.5. Comparison of the Carbon Fibers-
5.1.2. Chemical Composition . . . . . . . . 47
Based on PAN and on Pitch . . . . 63
5.1.3. Structure . . . . . . . . . . . . . . . . . 48
5.6. Commercial Availability and Eco-
5.1.4. Physical Properties . . . . . . . . . . 51
nomic Aspects . . . . . . . . . . . . . 64
5.1.5. Chemical Properties . . . . . . . . . . 51
5.2. History . . . . . . . . . . . . . . . . . 54 5.6.1. Companies and Trade Names . . . . 64
5.2.1. Rayon-Based Carbon Fibers . . . . . 54 5.6.2. World Capacity . . . . . . . . . . . . . 64
5.2.2. PAN-Based Carbon Fibers . . . . . . 55 5.7. Future Importance . . . . . . . . . . 65
5.2.3. Pitch-Based Carbon Fibers . . . . . 55 6. Refractory Ceramic Fibers . . . . 67
5.2.4. Vapor-Grown Carbon Fibers . . . . 56 7. References . . . . . . . . . . . . . . . 67

1. Survey cept for metal fibers, they are also character-

1.1. General Description of Inorganic
Fibers
The inorganic fibers are treated in five chapters:
Glass Fibers, Refractory Fibers, Metal Fibers,
Carbon Fibers and Refractory Ceramic Fibers.
In this introduction, the similarities with or-
ganic synthetic fibers, as well as special charac-
teristics, are treated generally.
Unlike organic fibers, there are no natural
sources for inorganic fibers, with one exception,
asbestos fibers, which are a separate keyword
(→Asbestos). As synthetic products, inorganic
fibers have obvious similarities with the syn-
thetic organic fibers, first of all in the morphol-
ogy and terminology of the commercially avail-
able forms and in their field of technical appli-
cation. Synthetic organic (Chardonnet silk, a re-
generated cellulose fiber (see Fibers, 1. Survey)
and inorganic fibers were first fabricated a hun-
dred years ago. The critical difference between
inorganic and organic fibers is seen in the differ-
ing properties, which correspond in general to
those of organic polymers on the one hand and
to those of inorganic and metallic materials on
the other hand.
Nonmetallic inorganic fibers are stronger,
stiffer, higher melting, and thus more heat re-
sistant, and in all cases nonflammable, but, ex-
ized by brittleness. Based on these general prop-
erties, they are less important as textiles per
se; however, they are superb for reinforcement
of other materials, in the so-called composites
(→Composite Materials).
They are excellent as thermal insulation at
high temperatures and in corrosive environ-
ments.
1.1.1. Commercial Forms and Terminology
The term fiber can be used for a formed mate-
rial that has a length-to-diameter ratio of at least
10 : 1, and with a cross-sectional area below
0.005 mm2 and a thickness below 0.25 mm [1].
This standardized nomenclature also includes
whiskers, which are mainly single-crystal ma-
terials, and are treated under a separate keyword
(→Whiskers). Nonmetallic fibers include both
continuous (endless) and short fibers. Diameters
range between 5 and 15 µm, but some are much
thicker, for instance, the fibers made by chem-
ical vapor deposition on a substrate fiber such
as B or SiC fibers which have diameters of ca.
100 – 150 µm. The diameter is of critical impor-
tance in regard to the tensile strength, because
all materials showing brittle fracture increase in
strength as the diameter decreases (Fig. 1).
 Fibers, 5. Synthetic Inorganic 3

to produce monofilament fibers. Such monofil-

ament fibers are used to reinforce polymers,
although of decreasing importance. However,
such fibers may be useful in the reinforcement
of light metals. The CVD technique can be used
with success to fabricate yarns of high-melting
substrate fibers with a coating on the individual
monofilament, e.g., SiC or B4 C on boron or TiN,
SiC, etc. on carbon.

1.2. Crystalline Structure and

Figure 1. Tensile strength of monofilaments (carbon fibers)
Mechanical Properties
as a function of fiber diameter. a) Hercules AS4 (Hercules);
b) Torayca T300 (Toray); c) Torayca M40 (Toray) Glass and mineral fibers are characterized by
their amorphous structural state, while metal and
Short fibers can be fabricated by a melt spin- some oxide fibers such as the DuPont A1203
ning process directly to staple fibers (see pp. 16 fiber are characterized by their polycrystallinity.
and 23 – 24) or by chopping (cutting) of con- In the case of most refractory ceramics pro-
tinuous fibers. Commercial short-fiber products duced from organic precursors, and also in
include wool, mats, and felts. The spinning of the case of carbon fibers, the structure is not
short fibers to continuous yarn, as with natural fully amorphous, but the crystals in the fibers
organic fibers, is generally not practiced for syn- are so imperfect that the term polycrystallinity
thetic inorganic fibers. can be used only with reservation. As a con-
Continuous fibers are fabricated by melt spin- sequence, postheat-treatment of such fibers
ning, if the fiber materials are meltable, such as is above 1000 ◦ C easily initiates crystallization
the case for glass or mineral fibers. The advanced processes as demonstrated on quite pure Si3 N4
inorganic fibers, such as carbon fibers and also fiber (experimental-type Tonen in 1988, [2])
those of refractory ceramics, are made from or- in Figure 2. The room temperature strength of
ganic precursor fibers via thermal degradation. the fabricated fiber with final pyrolysis tem-
For the spinning of the precursors the meth- peratures of 1000 ◦ C in nitrogen is 2450 MPa,
ods for fabrication of synthetic organic fibers an impressive value, but after heat treatment at
are used. This precursor technique can produce
both monofilaments and multifilaments, as de-
scribed in “Organic Fibers” (Fibers, 4. Synthetic
Organic). Only filament yarns which consist of
up to 300 000 filaments are commercially avail-
able (see “multifilament tow” in Chap. 5.6.2).
For reinforcement in advanced composites, tows
of 3000 to 12 000 monofilaments (3 – 12 K tows)
are used.
Meltspun glass fibers are also fabricated as
filament tow and applied in rowings of some
thousands of monofilaments. The terms rovings
and tows are synonymous for filament yarn. Fab-
rics from filament yarns are commercially avail-
able. They are fabricated by weaving, braising,
or similar techniques without binder to textiles,
tubes, filter sieves, cables, etc. The chemical va-
por deposition (CVD) method of fabrication, a Figure 2. Composition and X-ray diffraction of an
method developed to produce boron fibers and experimental-type silicon nitride fiber with the composition
still used for B and SiC fibers, can only be used in mass fraction: 59.2 % Si, 39.1 % N, 0.5 % C, 1.1 % O,
0.1 % H; θ = Bragg angle [2]
4 Fibers, 5. Synthetic Inorganic
1600 ◦ C, a perfect β-Si3 N4 crystal structure de-
velops, and room-temperature strength drops be-
low 1000 MPa, a considerable lowering of the
mechanical properties. Figure 3 shows further
examples where the decrease in strength due to
thermally activated crystallization is even more
extreme.
Figure 3. Decrease of tensile strength of various inorganic
fibers made by the precursor method (room temperature ten-
sile strength) HTT = heat treatment temperature
The well-known strength decrease of glass
fibers at elevated temperatures is caused by fa-
cilitated flow processes because of lowered vis-
cosity (Fig. 4).
Figure 4. Decrease in tensile strength and composition of
glass fibers at elevated test temperatures
The drop in strength by postheat-treatment
is also observed for carbon fibers and ex-
plained by thermal healing of structural de-
fects within the polyaromatic layers, thus break-
ing cross-linkages between them. This heal-
ing process enormously increases the tensile
modulus of the fibers because of the extreme
anisotropy of the graphite-like carbon structure.
In no case can three-dimensional crystallinity,
as in the graphite structure, be achieved by
postheat-treatment of carbon fibers, not even
above 2700 ◦ C. Therefore, carbon fibers never
consist of polycrystalline graphite and the term
“graphite fibers,” as used commercially in the
United States, is scientifically unjustified and
misleading.
A characteristic of inorganic fibers, whether
amorphous or polycrystalline is that they all ex-
hibit brittle fracture without plastic deformation.
Figure 5 shows some examples. Only metallic
fibers behave differently.
Figure 5. Stress – strain diagram of some inorganic fibers
The strain-to-failure (elongation at fracture)
for the brittle fibers follows precisely the ratio
strength to tensile modulus. This explains why
high-modulus inorganic fibers show the lowest
values of strain to failure, values below 0.5 %.
For reinforcement of polymers, as in advanced
composites, an elongation up to 2 % is required
by aircraft designers. This requirement is met by
the low-modulus glass fibers but not by the ad-
vanced carbon and ceramic fibers. Glass fibers,
therefore, are privileged in engineering applica-
tion if fracture toughness and impact resistance
are required without the need for extreme stiff-
ness. If high tensile moduli have to be main-
tained, the elongation can be improved only by
increasing the fiber strength (Fig. 5).
It has been well known since Griffith that
the fracture in brittle materials is initiated by ten-
sile stresses far below the intrinsic strength of
the material due to notches of a minimum – the
so-called critical length [3]:
 Fibers, 5. Synthetic Inorganic 5

length, are tested as monofilaments in a paper

E·γ
σ=
C mercial testing machine can be used for this
The probability for the presence of such a
notch decreases with decreasing volume of the
tested sample. For fibers having the same diam-
eter, the measured values for the tensile strength
increase with shorter gauge length and if the
same testing length is used, with smaller diam-
eter of the same fiber type. Figure 1 shows the
tensile strengths of monofilaments of commer-
cial fibers as a function of the fiber diameter.
As a consequence, the efforts to develop new
high strength inorganic fibers for reinforcement
are directed to smaller diameter filaments, for
instance 5 µm. However, this way to improve
strength and elongation is limited by the dan-
ger of carcinogenicity of all fiber materials thin-
ner than 3 µm that have length to diameter ratio
larger than 3 ( l/d> 3)
if the fiber material is persistent in vivo. A
hypothesis concerning cancerogenic potency on
geometric parameters was published in 1978
(Fig. 6) [4].
frame as demonstrated in Figure 7. Each com-
purpose. A gauge length of 25 mm is the stan-
dard [5]. The critical problem is the determina-
tion of the monofilament cross section. In the
case of circular morphology, the measurement
of the monofilament diameter by a scanning mi-
croscope gives reproducible results. The diam-
eters of the monofilaments in a bundle can vary
± 10 % [6]. If the fibers have irregular morphol-
ogy, only integration of the cross sectional areas
of the monofilaments in samples cut perpendic-
ularly to the fiber axis using optical or electronic
microscopy comes into question. Figure 8
The only commercial method is the strand
method [7].

Figure 6. Carcinogenicity of thin inorganic fibers [4]

As a further consequence, the testing method,

especially the gauge length of the fiber to be
tested for tensile strength, has to be strictly stan-
dardized.

1.3. Testing Methods for Inorganic
Fibers
Inorganic fibers are characterized mainly by
tensile strength. Longfibers, at least 30 mm in
Figure 7. Paper frame method for testing room temperature
tensile strength of thin monofilaments
A) Paper frame with the monofilament test sample fixed by
an adhesive on the frame; B) The paper frame in the holding
equipment of the testing machine; C) The paper frame after
cutting; D) The test specimen after measurement
6 Fibers, 5. Synthetic Inorganic
Figure 8. Distribution of tensile strength of monofilaments
of a carbon fiber tow [6]
The cross section of the bundle may also be
determined by the mass per length of the tow
and by the pycnometer density or alternatively
by the He density of the fiber material. The pre-
condition for this cross section determination is
a circular morphology.
The term “dtex” (1 tex = 1 g per 1000 m) is
used to characterize the mass of a filament fiber;
this is the same as for synthetic organic fibers
(→ Fibers, 4. Synthetic Organic).
A special test is the loop test, which is ap-
plicable for thicker monofilaments (see Fig. 9).
This test can be modified to the knot test for fast
qualitative characterization of the flexibility of
filament fibers.
Figure 9. Loop test for low modulus A) and high modulus
B) monofilaments
  indicating
 flexibility of inorganic fibers
Dloop ∼ dε = d·E σ
fiber fiber
D = loop diameter, d = fiber diameter, ε = fiber elon-
gation, E = fiber modulus, σ = fiber strength
The tensile modulus is measured during the
tensile strength test or nondestructively by the
ultrasonic method, which gives good compara-
tive results.
The compressive strength of the various inor-
ganic fibers is a further characteristic property,
but can only be measured if the fibers are inbed-
ded in a unidirectional composite with a resin
as matrix and a fiber volume fraction of 60 %.
Isotropic fibers have high compressive strength,
but the fibers with anisotropic structure and high
preferred orientation parallel with the fiber axis
show lower compressive strengths. The worst
compressive strengths appear for the ultra-high-
modulus types; here, the mesophase pitch-based
carbon fibers are surpassed in a negative sense
by the high-modulus organic fibers.
The determination of the linear thermal ex-
pansion of fibers is straight-forward; however,
the determination in the perpendicular direc-
tion is extremely difficult. For carbon fibers
this is important because of the high degree of
anisotropy and the variety of degrees in pre-
ferred orientation. Data from direct optical mea-
surements are compiled in Table 1. An indi-
rect method is the comparative dilatation mea-
surement using the pure resin and the fiber-
reinforced composite.
Table 1. Thermal expansion of inorganic fibers
Fiber type Thermal expansion, K−1 ∗
α · 10−6 α⊥ · 10−6
Glass fiber
E glass 5.4 5.4
S glass 1.6 1.6
Carbon fiber (−0.5) – 0.2 10 – 15
CVD Monofilaments
SiC fibers 3.1 – 4.5 3.1 – 5.2
Oxide fibers 7.0 – 8.6 7.0 – 8.6
∗ parallel () or perpendicular (⊥) to the fiber axis
1.4. Inorganic Fibers for Thermal
Insulation
The use of inorganic fibers for insulation is based
not only on the low thermal conductivity of the
material itself (one to two orders of magnitude
less than that of metals) but also on the large vol-
ume between the fibers. These two aspects are
discussed in detail in the chapter 2. Application
of glass fibers is limited as shown in Table 13, but
refractory oxide fibers allow use at much higher
temperatures. Saffil fibers Al2 O3 and ZrO2 wool
from ICI are examples (see Section 3.2.4). Up-
per application temperature is ca. 1650 ◦ C. Even
carbon felt can be used as thermal insulation; in
 Fibers, 5. Synthetic Inorganic 7
Table 2. Thermal conductivity of insulation layers from inorganic fibers

Measured at ◦ C in W/mK

Al2 O3 -Blanket (Saffil) 200 0.1

1000 0.26
1400 0.5
Zirkonia-Blanket (Saffil) 200 0.05
1000 0.22
1400 0.42
Silica-Wool 200 0.04
(Quarz and Silica) 800 0.14
Mullite-Texture (Nextel) RT 0.09
200 0.12
1000 0.87
Zirkonia-Insulation-Plate
Type ZYZ3 (Zircar) 400 0.09
800 0.11
1100 0.14
1400 0.19
1650 0.23
Carbon-Felt (Sigri)
HTT = 1000 ◦ C 20 0.06
HTT = 2000 ◦ C 20 0.08

For comparison thermal conductivity of various materials, measured at room-temperature in W/mK

Cu Al Steel C-Fiber Glass Polyethylene
370 175 46 12 – 23 0.9 0.23

Effect of insulation layers

Type Thickness Hotface Coldface

Mullite 1 layer (ca. 3 mm) 1000 ◦ C 500 ◦ C

(in air) 3 layers (ca. 9 mm) 1000 ◦ C 300 ◦ C
Carbon Felt 40 mm 1600 ◦ C 360 ◦ C
(in argon) 40 mm 1200 ◦ C 260 ◦ C

spite of the high thermal conductivity character-
istic of all carbon materials, the upper applica-
tion temperature in a non-oxidizing environment
is 2000 ◦ C (see Table 2).
Mostly short fibers, such as wools, felts, tex-
tiles, and similar fabrics, are used. Mechanical
properties and variation in individual fiber thick-
ness is of little importance.
fibers, where the rule of mixtures can be ap-
plied (Fig. 10A). Comparison of the precalcu-
lated and actual composite properties reveals the
efficiency of the translation of the fiber prop-
erties into the composite. With good adhesion
between fibers and matrix, efficiencies of 80 –
100 % can be achieved, especially as far as the
moduli are concerned. Exact precalculations are
described in [8] and elsewhere.

1.5. Inorganic Fibers for Reinforcement

Purposes
1.5.1. Principles of the Fiber Reinforcement
Controlled inclusion of fibers into a homoge-
neous matrix produces a fiber-reinforced com-
posite. The properties of the composite can be
precalculated from the properties of the tow
components, their volume fractions, and the ge-
ometrical arrangement of the fibers within the
isotropic matrix. The simplest case is the unidi-
rectional (UD) reinforcement with continuous
Figure 10 (see next page). Principles of reinforcement in
unidirectional composites
A) Rule of mixture for strength data in fiber direction
E c = Young’s modulus of composite, E F,M = Young’s mod-
ulus of fiber and matrix, x F,M = Volume fraction of fiber
and matrix, η E = Yield of fiber modulus in the composite, a)
Fiber; b) Composite; c) Matrix
B) Angle dependence of strength a) in unidirectional long
fiber reinforced composites and b) in isotropic short fiber
composites [9]
C) Stiffening effect by short carbon fibers in alu-
minum and thermoplastics a) C-Fiber/Thermoplastics; b) C-
Fiber/Aluminum
8 Fibers, 5. Synthetic Inorganic
Estimation of the strengths by precalcula-
tion, even in unidirectional reinforced samples,
is more difficult, because the fracture processes
overlap with the reinforcement effect. If we con-
sider a polymer matrix with strain-to-failure ex-
ceeding that of the fibers, which is generally the
case, the crack starts in the fibers. The degree of
adhesion between fibers and matrix and the na-
ture of the interphase contribute to the extent of
the reinforcement effect. In addition, chemical
and physical compatibility must be considered.
Nevertheless, the simple rule of mixtures is still
useful and reasonably valid.
The precalculated reinforcement for UD
composites is achieved only in the fiber di-
rection. For other directions, reinforcement is
less effective. The strength of a UD composite
across the fiber axis is that of the pure polymer
(Fig. 10B) [9].
Short fibers arranged statistically in all di-
rections, produce reinforcement of 1/6 of that
precalculated for UD composites by the rule
of mixtures, but this strengthening is isotropic.
However, it should be demonstrated in each case
whether or not reinforcement is worthwhile for
a specific material combination. For example,
as shown in Figure 10C, addition of short car-
bon fibers to an aluminum matrix is not effec-
tive [10]. Precise stiffness and strength informa-
tion is a prerequisite for the successful design of
fiber-reinforced matrices.
1.5.2. Commercial Continuous Inorganic
Fibers
Mechanical Properties For a better under-
standing of the character, properties, and the
importance of today’s inorganic fibers and fu-
ture trends, a brief history is helpful. The fiber
reinforcement of polymers started during the
1930s but was not commercially successful until
the 1940s, only after the interface problem had
been solved by means of coupling agents. In the
1950s the U.S. Air Force and later the NASA
supported development of reinforcing fibers to
achieve light weight and high stiffness for future
airplanes. Boron, a light but high-melting ele-
ment, was selected as the most promising candi-
date. A new production technology was devel-
oped: chemical vapor deposition on thin tung-
sten wires. The boron fibers fabricated in this
way are monofilaments with diameters between
120 and 150 µm and have more the character of
thin wires than that of fibers. Therefore chemists
started to develop filament fibers from the neigh-
boring element carbon, which is also light and
has a high melting point, and from which fibers
have been demonstrated 100 years earlier as fila-
 Fibers, 5. Synthetic Inorganic 9
Table 3. Mechanical and physical properties of the representative inorganic fibers

Fiber type Room temperature properties

Density, Tensile Tensile

g/cm3 strength, modulus,
MPa GPa

Glass fiber 2.50 – 2.62 3400 – 4500 70

Carbon fiber 1.76 – 2.10 2000 – 7000 240 – 700
CVD ∗ Monofilaments
SiC fibers 2.55 – 3.50 2000 – 3700 200 – 420
Oxide fibers (whisker) 3.90 1200 – 1400 340 – 400

∗ CVD = chemical vapor deposition

ments for electric light bulbs. The high stiffness fibers, and Al2 O3 fibers. The physical and
was achieved by hot working, similar to the pro- mechanical properties are compiled in Table 3.
cessing of metal wires. Real technical success The specific strengths and tensile moduli are
was achieved, however, only after this high-tem- presented in a more schematic way in Fig. 11
perature forming process became superfluous [12]. All mechanical properties are determined
upon introduction of the preferred orientation in at room temperature and under standard condi-
the process step before thermal decomposition tions. Composites with these stiffnesses are brit-
of the polymer fiber to the carbon fiber. These tle, however. The impact resistance of the vari-
carbon fibers, which can be handled like glass ous fibers is therefore an important criterion for
fibers, have completely replaced boron fibers in their application. Carbon fibers behave worst in
their function of stiffening polymer composites. this respect in spite of the other overwhelmingly
They are widely used today in combination with excellent properties, as shown in Table 4 (see
glass fibers in so-called advanced hybrid com- next page). One way to overcome this brittle-
posites. Boron fibers on wire substrates have ness is to mix the carbon fibers with glass fibers
also been used to reinforce light metals. But in (Fig. 12). Organic polyaramide fibers (Kevlar
this field, too, a new competitor has appeared, from DuPont or Twaron from ENKA) are in-
namely, CVD-SiC fibers, thick monofilaments, cluded in Table 4 and Figure 12 to allow com-
much like thin wires. Silicon carbide is oxida- parison.
tion resistant in air up to 1400 ◦ C, but has a den-
sity of 3.5. It is heavier than the light metals or
boron. It is a superior fiber component in SiC
and Si3 N4 ceramics, especially for toughening.
The troublesome CVD technique and the stick-
ing of the monofilaments, both dangerous for the
workers, have led to the development of a new
competitor, SiC and Si3 N4 filament yarns, which
are fabricated much like carbon fibers by ther-
mal degradation of polymeric precursor fibers,
carbosilanes, and silazanes. For oxide compos-
ites, refractory oxide fibers are best. Fabrication
techniques starting from organometallic precur-
sor fibers as shown above, have so far produced
only low-strength fibers, but improvement can
be expected. Fabrication from ultrafine crys-
talline powder produces higher strength fibers
[11].
Commercially available today are glass
Figure 11. Mechanical room temperature properties of in-
fibers, carbon fibers (HT and HM types), CVD organic fibers in a stress – modulus diagram [12]
boron, SiC monofilaments, SiC and Si3 N4
10 Fibers, 5. Synthetic Inorganic
Table 4. Qualitative comparison of mechanical properties of glass, polyaramid, and carbon fibers

Static load Dynamic load

Strength Modulus Impact Fatigue Ilss Temperature Density

Tensile Flexural Compressive Tensile Flexural Compressive resistivity

Glass
+ ◦ + − − − + − + ◦ −
Aramide
◦ − − ◦ ◦ ◦ ◦ ◦ ◦ − +
Carbon
+ + ◦ + + + − + + + ◦

+ = better than average; ◦ = average; − = less than average

Ilss = interlaminar shear strength

layer to protect the fiber surface and to im-

Figure 12. Improvement of impact strength of carbon fiber-
reinforced composites by polyaramide (Kevlar) and glass
hybridization, respectively
1.5.3. Surface Chemistry, Surface
Treatment, and Surface Properties
All commercial inorganic fibers are sized, i.e.,
coated with a thin organic, usually polymeric,
prove the handling of the brittle fiber mate-
rial. The most common sizing materials are
poly(vinyl alcohol) and uncured epoxy or poly-
imide resins. Glass fibers are coated by coupling
agents, usually silanes, to improve adhesion in
the matrix resins, which are usually unsaturated
polyesters and, in advanced composites, epoxy
resins. Glass fibers have silanol groups on the
surface, which can react easily with functional
groups in the resin. As such, interfaces within the
composites are very sensitive to attack by water,
and the glass fibers must be protected by lin-
ear organic molecules that have a hydrophobic
group on one end of the molecule. The mecha-
nism of protection is indicated in Figure 13 [13].
Carbon fibers are thoroughly wetted by the
various polymers if they possess such functional
surface groups as acid and basic surface oxides
[14]. Also, the adhesion in the composite is ex-
plained by chemical interchange with these sur-

Figure 13. The effect of coupling agents for glass fibers in polymers (schematic) [13]
 Fibers, 5. Synthetic Inorganic 11

face groups. The surface oxides are formed on Even in the case of reinforcement of aluminum,
the carbon fibers as the last step during fabrica- boron fibers must be coated by diffusion barri-
tion. All commercial carbon fibers are surface- ers, thin layers of SiC or B4 C deposited from
treated and sized in this way. The process pa- the vapor as the last step of the fabrication pro-
rameters of the surface treatment strongly influ- cess (Fig. 15) [16, 17]. SiC fibers are compatible
ence the degree of surface oxidation, and thus the with polymers, and generally with most inor-
character of the interfaces within the composite. ganic materials, if 1000 ◦ C is not exceeded. Re-
Surface treatment is, however, usually a propri- fractory oxide fibers would be ideally suitable
etary process kept secret by the supplier. The for combination with light metals if properties
suitability of the commercially surface treated and economic considerations justify it.
and sized fibers must be tested and confirmed
by the user.
Adhesion between fiber and matrix plays an
important role in respect to the impact resistance
of the composite. Dependant on the strain-to-
failure of the fiber, impact resistance can be im-
proved or worsened by a high degree of adhe-
sion, which in turn is determined by the degree
of surface oxidation [15]; Figure 14 shows some
correlations.

Figure 14. Impact strength of carbon fiber-reinforced uni-

Figure 15. Boron fiber fabrication on tungsten substrate by
directional composites as a function of fiber strain for the
the CVD process (chemical vapor deposition) [16]
untreated and surface-treated state
U: Untreated; S: Surface treated

The chemical compatibility with the matrix

is another criteria for inorganic reinforcement
fibers. Glass fibers are compatible with poly-
mers, with most oxides, but not with concrete.
Carbon fibers are also compatible with poly-
mers, and in addition they are compatible with
concrete. They are not compatible with most
metals, with the exception of lead and tin. They
are incompatible with oxides above 400 ◦ C. Sur-
face coatings as diffusion barriers improve the
chemical compatibility. Boron fibers are com-
patible with polymers, but react with metals.
1.5.4. High-Temperature Behavior of
Inorganic Fibers
Glass fibers are limited to temperatures below
ca. 300 ◦ C. Carbon fibers are not resistant to
oxidation above 400 ◦ C, but are thermally sta-
ble up to 2000 ◦ C in non-oxidative environ-
ments. They can be applied in all high-tem-
perature polymers without restriction. For the
highest temperatures, the performance of PAN-
based carbon HM-fiber must be considered as
exemplary. They can be used in composites with
12 Fibers, 5. Synthetic Inorganic
carbon as matrix without limitation (carbon –
carbon composites [18]), but diffusion barriers
are needed in the cases of matrices of oxides,
metals, and non-oxidic refractory compounds
[19]. Some examples are given in Figures 16,
17, and 18 [20]. All other refractory fibers lose
their room-temperature strength values if heat-
treated above 1200 ◦ C because of crystallization
processes (see Figs. 3 and 4). Inorganic fibers re-
sistant to high temperatures are a fertile field for
further development.
Figure 16. Reinforcement effect by carbon fibers with sili-
con carbide coatings in quartz glass [20]
σ F , σ C , σ M = stress of fibre, composite, and matrix respec-
tively; εF , εC , εM = strain of fibre, composite, and matrix
respectively
Figure 17. Strengthening and toughening of mullite by sil-
icon carbide-coated carbon fiber
A) Al2 O3 Matrix; a) 10 %, b) 54 % and c) 32 % coated
B) Mullite – Matrix; a) 28 %; b) 41 % and c) 34 % coated
2. Glass Fibers
2.1. Textile Glass Fibers
Textile glass fibers have gained in importance in
spite of the availability of other, newer reinforc-
ing fibers, such as carbon or aramid fibers [21–
25]. A large part of this growth is a result of glass
fiber-reinforced plastics (GRP). The market for
GRP in the Federal Republic of Germany has
grown from ca. 10 000 t in 1960 to ca. 200 000 t
in 1985.
Glass fibers that can be processed by con-
ventional textile techniques such as weaving are
called textile glass fibers, to contrast them with
insulation glass fibers. Textile glass fiber is a
collective term for fine fibers possessing an ap-
proximately uniform, mostly circular cross sec-
tion made from molten glass. The terminology
of textile glass and processing aids is standard-
ized in national and international standards.
Depending on the manufacturing process,
textile glass fibers are subdivided into (1) glass
filaments and (2) glass staple fibers, which are
defined as follows in ISO 6355:
1) A glass filament is a textile glass fiber of prac-
tically unlimited length of a given diameter
drawn from molten glass.
2) A glass staple fiber is a textile fiber of limited
length (spun fiber) and of a given diameter.
Textile glass is used both for textile purposes,
e.g., decoration, insulation, filtration, as well as
for reinforcement of a matrix to form a com-
posite. Matrices can be plastics (both thermo-
setting and thermoplastic), bitumen, rubber, ce-
ment, gypsum, or other materials. Of all glass
filaments ca. 90 % are used to reinforce a ma-
trix.
Textile glass fibers can also be categorized
by the type of products and the diameter of the
filaments:
1) Textile products, such as textile glass yarns
and plied yarns with filament diameters of
5 – 13 µm
2) Plastic products such as mats, rovings,
chopped strands, and milled fibers with fil-
ament diameters of 9 – 24 µm
The textile products are mostly woven and are
then available as fabrics for reinforcing plastics
 Fibers, 5. Synthetic Inorganic 13

Figure 18. Toughening of alumina by silicon carbide coated carbon fibers a) Uncoated C-Fibers; b) SiC-Coated C-Fibers

and other applications such as thermal insulation
and decoration.
Plastic products can be manufactured to parts
by hand lay-up, winding, molding, etc. (Chap-
ter 4).
History.The technique of ornamenting glass
vessels with glass yarns was used in ancient
Egypt around 1600 b.c. and developed fur-
ther by the Romans. In the Middle Ages, this
technique–based on drawing heated rods into
fine filaments–was further refined in Venice and
Murano.
The first indications of the use of glass yarns
in fabrics are found in the work of the physi-
cist Ferchault de Reaumur in 1713. The first
known factory for the manufacture of glass yarns
was founded in 1866 in Vienna by Jules de
Brunfaut. The inventor of the bushing, how-
ever, was Louis Schwabe, who first produced
glass yarns by drawing the melt through fine
holes. In 1908, a rather coarse, endless glass fiber
was demonstrated in Hamburg by W. v. Paczin-
sky. True industrial production of textile glass
fibers was realized in the United States in 1930
and in Germany in 1939.
2.1.1. Chemical Composition and Properties
Chemical Composition. The chemical
composition of glass fibers essentially deter-
mines the properties and the applications. Ta-
ble 5 shows the known types of textile glass
fibers and their compositions.

Table 5. Chemical composition of textile glass (nominal values, wt %)

Component Types of glass

Multi-purpose Acid resistant Alkali resistant High strength Good dielectric properties

E A C E-CR [31] R S D Quartz

SiO2 53 – 55 70 – 72 60 – 65 58 60 60 60 – 65 73 – 74 100
Al2 O3 14 – 15.5 0 – 2.5 2–6 12 – 13 0.7 25 20 – 25
CaO 20 – 25 5–9 14 21 5 6 0.5 – 0.6
MgO 20 – 25 1–4 1–3 4.5 9 10
B2 O 3 6.5 – 9 0 – 0.5 2–7 <0.1 0 – 1.2 22 – 23
F 0 – 0.7 <0.15
Na2 O ≤1 12 – 15 8 – 10 0.6 14 0 – 1.1 1.3
ZrO2 ≤1 18
K2 O ≤1 1 0.4 1.5
Fe2 O3 ≤1 0.3 0.1
TiO2 ≤1 2.1
14 Fibers, 5. Synthetic Inorganic

The following types are recognized:
E glass. An alumino-borosilicate glass with a
mass fraction of alkali ≤1 %, which is gen-
erally used in glass-reinforced plastics and in
electrical applications
A glass. An alkali-lime glass with little or no ad-
dition of boron trioxide and a mass fraction
of alkali ≥1 %, for special applications
E-CR glass. An alumino-lime silicate glass with
a mass fraction of alkali ≤1 %, especially
suited for the reinforcement of plastics be-
cause of acid resistance
C glass. An alkali-lime glass with high boron tri-
oxide content, with good chemical resistance
for special applications
D glass. A borosilicate glass, a specialty glass
with high dielectric properties
R glass. An aluminosilicate glass without added
calcium and magnesium oxides, glass that
meets high mechanical requirements
S glass. An aluminosilicate glass without added
calcium oxide but with a mass fraction of
magnesium oxide of ca. 10 %
The most widely used are E-glass fibers,
which are encountered in all areas of indus-
try and everyday life, especially in combination
with plastics (Chapter 4). In glass staple fibers,
C glass is used primarily [29].
More recently, acid-resistant E-CR-glass
fibers have been used in special projects such as
pipes for flue-gas desulfurization plants. Alkali-
resistant glass fibers are primarily intended for
cement reinforcement.
High-strength textile glass fibers such as R-
and S-glass fibers are preferred in applications
where high tensile strength is needed, e.g., in
containers under high internal pressure.
The raw materials for the components are
largely natural minerals. They are subject to
stringent quality control when received, since
the raw materials must be within narrow, well-
defined limits to produce textile glass fibers hav-
ing uniform properties.
Properties. Textile glass fibers are distin-
guished by particularly high tensile strengths.
Measurements on filaments drawn directly from
the bushing yield values of 3400 MPa for E glass
and 4400 MPa for R glass (Table 6).
The modulus of elasticity, which is 1/3 that
of steel and quite adequate for many applica-
tions, and the density of 2.6 g/cm3 , which is high
compared to plastics, are not sufficient for large-
surface, light-weight parts (airplane construc-
tion). For such parts, fibers of lower density and
higher modulus of elasticity (aramid and carbon
fibers) were developed.
Textile glass fibers have a linear stress – strain
curve up to the break, so that the break elonga-
tions εB can be calculated from the modulus of
elasticity E and the tensile strengths σ B (Fig. 19,
see next page). Good resistance to weathering
and heat, nonflammability, good dielectric prop-
erties, low thermal expansion, and, depending
on the type of textile glass, good resistance to
corrosion, are additional characteristics of tex-
tile glass fibers. The thermal conductivity of tex-
tile glass products is determined by the density
of fiber packing. The properties of textile glass
fibers to a great extent determine the properties
of the composites (→Composite Materials).

Table 6. Properties of textile glass (nominal values)

Property Types of glass

Multi-purpose Acid resistant Alkali resistant High strength Good dielectric properties

E A C E-CR [31] R S D Quartz

3
Density, g/cm
Bulk glass 2.61 2.46 2.46 2.55 2.16 2.2
Fiber glass 2.59 2.45 2.70 2.68 2.53 2.48 2.14
Modulus of elasticity,
MPa 73 000 74 000 71 000 72 000 76 000 86 000 85 500 55 000 6200 – 7200
Tensile strength, MPa
Freshly drawn fiber 3 400 3 100 3 100 3 330 3 700 4 400 4 590 2 500 400 – 500
Fiber in composite 2 500 2 400 3 750 1 550
Dielectric constant
at 1 MHz 5.8 – 6.4 6.8 7 5.6 – 6.2 4.9 – 5.3 3.85 3.78
 Fibers, 5. Synthetic Inorganic 15

2.1.2. Manufacture
2.1.2.1. Glass Filaments
Glass filaments are produced continuously by
the bushing fiberizing process. There are two
commercial processes for the manufacture of
glass filaments: the marble melting process and
the direct melt process.
In the older indirect marble melting process,
the production of glass and the glass spinning Figure 19. Stress – strain curves of five fibers
process are separate. E glass is initially produced a) Carbon fiber, high modulus; b) Carbon fiber, high strength;
in the form of glass marbles (15 – 20 mm diame- c) Aramid fiber, high modulus; d) R-glass fiber; e) E-glass
ter), which are then re-melted in a second phase fiber σ B = tensile strength; εB = break elongation
in the glass forming process (two-stage process).
of heated feeders (channels) in which the bush-
This process is still in use today for making fine
ings are placed. In this one-step process, there
filament yarns.
is no separation between glass manufacture and
In the direct melt process, molten glass flows
glass forming (Fig. 20). Temperatures in the pro-
directly from the melting furnace into a system

Figure 20. Direct melting

a) Mixing silos; b) Furnace feed; c) Furnace; d) Heat exchanger; e) Melt; f) Bushing; g) Sizing applicator; h) Cake winder;
i) Spun cake
16 Fibers, 5. Synthetic Inorganic

duction of E glass are 1310 – 1390 ◦ C in the fur-

nace and 1225 – 1260 ◦ C in the feeders [24].
In principle, the fiber forming process in-
volves attenuating glass filaments continuously
from a Pt – Rh bushing by applying a mechanic
drawing force at offtake speeds of up to 80 m/s
(Fig. 21). Bushings may have 400, 800, 1200,
1600, or more holes of 1 – 2.5 mm diameter. A
single filament is drawn from each hole. The
filaments are collected into a bundle (strand)
and wound as a spun cake or directly pro-
cessed into rovings, continuous strand mats, or
chopped strands. The deformation during draw-
ing is high: for example, with bushing-tip diam-
eters of 2 mm and a filament diameter of 10 µm.
A typical ratio of cross sections is 40 000 : 1 Figure 22. Bushing
(Fig. 22). a) Bushing nipple; b) Drawing bulge D = hole
diameter; d f = diameter of glass filament

Figure 21. Spin-draw process (schematic)
a) Bushing; b) Water spray; c) Sizing application; d) Strand
guide; e) Spun cake
Formation of the filaments involves two basic
processes:
1) efflux of molten glass from the bushing tips
under the condition that
Fp +Fg >Fη +Fσ
2) drawing of the meniscus at the bushing tip
to a continuous filament by a force under the
condition that
Fd +Fg >Fa +Fσ
where
F p = force due to hydrostatic pressure
F g = force of gravity
F σ = force due to surface tension
F η = force due to viscosity
F d = drawing force
F a = sum of frictional forces
The viscosity of the melt must be increased
sharply by cooling with air and water to enable
a filament to be formed in the shortest possible
time after the melt leaves the bushing tip.
An aqueous solution, called size or sizing,
is applied to the cooled filaments; this solu-
tion consists basically of film formers, lubri-
cants, and, if necessary, coupling agents. After
the forming packages are dry, the sizing (0.5 –
1.5 wt %) has two functions:
1) Reducing the glass – glass coefficient of fric-
tion from ca. 1.0 to 0.2 to make further pro-
cessing of the yarn possible. Due to the high
coefficient of friction, textile glass strands
without size cannot be processed further
since they cannot be pulled off the cakes.
2) Imparting properties that are prerequisite for
further application of the textile glass prod-
ucts as plastic reinforcement (plastic prod-
ucts) or as textile products with optimum
weaving and desizing properties.
 Fibers, 5. Synthetic Inorganic 17

2.1.2.2. Glass Staple Fibers C = C glass

D = discontinuous
The principles discussed with regard to the draw- 11 = nominal fiber diameter in µm
ing of glass filaments also apply to glass staple 200 = tex ∗
fibers, but the manufacturing processes them- Z140 = Z-twist, 140 turns per meter
selves are different.
∗ One tex is equivalent to 1 g/1000 m. For
In the bushing blowing process, the molten
glass filament products the relationship between
effluent is fibrillated by streams of gas. Adhe-
the number of holes in the bushing N, the fila-
sive sizing is added, and the fibers are collected
ment diameter d f , and the tex number is:
on a rotating perforated drum or on a suction belt
(nonwovens) under reduced pressure. tex = 0.002d2f ×N
In the drum drawing process, fibers are drawn
either from bushings or from rods by a rotat- where d f is in µm. The density of glass G is
ing drum and scraped off. During the drawing taken to be 2.5 g/cm3 .
process, sizing is applied to the fibers for later The other textile glass products are desig-
cementing. The fibers broken by scraping are nated in a similar manner.
swirled in a funnel. Typical properties of products used in textiles
In both processes the fibers are finally con- and plastics are summarized in Tables 7 and 8.
solidated in a finishing tube and drawn off as a
sliver. The sliver can be converted directly into
fabrics, yarns, or plied yarns. 2.1.3.2. Correlation Between Processing and
Application

2.1.3. Textile Glass Products: Processing Textile glass fibers can be processed as follows:
and Application
1) Textile processes such as weaving, knitting,
2.1.3.1. Product Designation and braiding, suitable for textile glass yarns,
plied yarns, and rovings
Glass filament and glass staple yarns have a stan- 2) Plastic processes for reinforcing thermo-
dard form of designation. setting and thermoplastic resins, suitable for
Glass Filament Yarns (ISO 2078 and textile glass fabrics, layers, knits, mats, rov-
ISO 3598) EC9-34 Z28 – sizing x ings, milled fibers, tissues, and chopped
strands
E = E glass 3) Processes for reinforcing other matrix mate-
C = continuous rials, e.g., bitumen, paper, rubber, gypsum,
9 = filament diameter in µm and plaster, suitable for textile glass fabrics,
34 = tex ∗ yarns, surface mats, rovings, and chopped
Z28 = Z-twist, 28 turns per meter strands
Glass Staple Yarns(ISO 2078 and ISO 3598) The processes used in plastic reinforcement
CD11-200 Z140 are summarized in Table 9 (see next page).

Table 7. The tex number of textile glass yarns as a function of filament diameter and number of filaments in a strand

Filament diameter Number of filaments per strand

µm Nominal U.S. 50 100 200 400 800 1600

µm System

5.2 5 D 2.8 5.5 11

6.4 6 DE (17) 34 68 136
7.4 7 E 11 22
9.2 9 G 34 68 136 272
11.3 11 J 51 102 204 408
13.0 13 K 136 272 544
18 Fibers, 5. Synthetic Inorganic
Table 8. The tex number of glass yarns for plastic reinforcement as a function of filament diameter and number of filaments in a strand

µm Nominal µm U.S. System Number of filaments per strand

50 ∗ 100 ∗ 200 ∗ 400 800 1600 2000 4000

10.0 10 H 10 20 40 80 160 320

11.3 11 J 12.5 25 50
14.2 14 L 160 320 640
17.3 17 N 480 960 1200 2400
19.4 19 OP 600 1200
24.5 24 ST 2400 4800

∗ Strands separated during spinning

Table 9. Processing of glass-reinforced plastics (GRP)

Process
Manual processes
Hand lay-up
Fiber/resin spray-up
Intermediate processes
Injection process
Vacuum injection process
Wet molding
Industrial processes
Processing of semi-finished
products
Prepregs
SMC ∗
Molding compounds
Rotation process
Reinforcement of polyurethane
Filament winding
Continuous impregnation
Pultrusion
Injection molding of glass
fiber-reinforced granulate
Compression molding of glass
mat-reinforced
thermoplastic plates
∗ SMC = sheet molding compounds
2.1.3.3. Description of Textile Glass Products
There are separate standards for special prod-
ucts:
textile glass rovings
textile glass mats
textile glass fabrics
glass filament yarns
glass staple fiber yarns and rovings
Textile glass products
Glass-reinforced
thermosetting resins:
chopped strand mat, woven
fabric, surface mat
roving
surface mat, continuous mat
surface mat, continuous mat
surface mat, continuous mat,
woven fabric
(compression molding,
injection molding)
woven fabric
roving
chopped strands
roving, woven fabric
milled fiber, continuous mat
roving, woven fabric,
chopped strand mat
continuous mat, chopped
strand mat, woven fabric
roving, woven fabric,
continuous mat, chopped
strand mat
Glass-reinforced
thermoplastic resins:
milled fiber, roving, chopped
strands
continuous mat, roving
ISO 2797
ISO 2559
ISO 2113
ISO 2078 and ISO 3598
ISO 2078 and ISO 3598
The entire line of products is presented in Fig-
ure 23 (see next page).
Textile glass yarns usually have a protective
twist and are sold with textile or plastic sizes.
Textile sizes based on starch/fats are especially
suitable for high-speed looms, e.g., air looms.
For certain applications, e.g., circuit boards for
printed circuits, the textile finish applied to the
fiber must be removed and replaced by an adhe-
sive finish (coupling agent) via a special treat-
ment given to the fabric.
Textile glass fabrics come in different
weaves, depending on the application.
Textile glass rovings with untwisted, approx-
imately parallel filaments or strands are offered
with inside or outside take-off. Textile glass rov-
ings are either assembled into a larger strand
comprising several strands or produced by a di-
rect process below the bushing. In the second
process roving strands of various tex numbers
are obtained by changing the number of holes in
the bushing or by changing the filament diam-
eters. Spun rovings of textile glass are suitable
for the pultrusion process because of their good
mechanical properties in the transverse direc-
tion.
Textile glass mats are made from chopped
strands (chopped strand mats) or of strands laid
up directly below the bushing (continuous strand
mats). Consolidation in both cases is accom-
plished by chemical binders. Chopped strand
mats are used primarily for hand lamination;
continuous strand mats for all types of injection
molding, compression molding, and continuous
impregnation (Table 5).
Chopped strands with lengths of 3, 4.5, 6, or
12 mm have been found useful in the production
of molding compounds and as reinforcement of
thermoplastics. As shown in Figure 5, chopped
strands are either cut directly below the bushing

Figure 23. Textile glass products
Stages 1 – 3 are in the hands of the textile glass producers, while Stage 4 is in the hands of the processors, e.g., weavers
or produced from intermediate products, e.g.,
textile glass roving or glass strands.
Milled textile glass fibers with fiber lengths
of 0.2 mm are suitable as reinforcement of ther-
moplastics and polyurethane foams. In addition,
they are used in the production of primers, fire-
protection coatings, and the like.
Tissues of glass filaments or glass staple
fibers are produced both by dry and wet pro-
cessing. In both cases the fibers are chemically
bonded.
Because of their bulky nature, textured rov-
ings or yarns have turned out to be good replace-
ment products for asbestos, e.g., in the manufac-
ture of clutch facings. Textured yarns are also
used in decorative fabrics.
Layersare planar structures of bonded, non-
woven textile glass yarns with crisscross or par-
allel yarns. Layers are primarily used for rein-
forcing paper or bitumen or in the preparation
of prepregs (preimpregnated sheets).
Textile glass knit goodsare more elastic than
woven fabrics of textile glass. One area of appli-
cation is medical bandages of pre-impregnated
Fibers, 5. Synthetic Inorganic 19
textile glass knits, which take the place of plaster
casts.
Warp beams are used in the preparation of
warps for weaving and consist of a multiplic-
ity of yarns wound in parallel. Beams of strands
with plastic sizing are used for unidirectional re-
inforcement of plastics for the bumper systems
of motor vehicles.
2.1.3.4. Applications
As seen in Table 10 (on next page) there is a
distinct trend toward the use of GRP in motor
vehicles, both passenger cars and trucks. Glass
fiber reinforced thermosetting resins are used in
passenger cars primarily for body parts (hoods,
tail gates) and for structural elements (bumpers,
bumper supports). Glass fiber reinforced ther-
moplastics are used for operating parts within
the engine compartment, noise suppressors be-
low the engine, hub caps, etc.
Textile glass fabrics for epoxy resin prepregs,
the basic material of circuit boards, is not consid-
ered to be GRP but rather is considered among
 20 Fibers, 5. Synthetic Inorganic
Table 10. Percentage of textile glass fibers used to reinforce resins by area of application in the Federal Republic of Germany, 1973 and
1985 [32]
Area Thermosetting resins (except circuit boards) Thermoplastic resins

1973 1985 1973 1985

Construction 24 16 0 1
Industry and agriculture 32 24 23 12
Transport 9 25 25 55
Consumer goods 6 4 14 6
Sport and leisure articles 6 3 0 0
Electronic industry 17 25 36 23
Other 6 3 2 3

other applications. However, it is the most im-
portant such other application, representing one-
third of the amount used for other applications.
The following other applications are of im-
portance:
1) Thermal insulation and heat protection (an
asbestos replacement) with woven glass fab-
rics of glass filament and glass staple fiber
yarns for primary insulation jackets, heat
shielding curtains, fire blankets, pipe insu-
lation, etc.
2) Wall coverings of staple fiber yarns
3) Plaster reinforcement of textile glass mesh
fabrics and textile glass nonwovens
4) Reinforcement of grinding wheels with wo-
ven glass fabrics
5) Filter materials for high-temperature filtra-
tion with woven glass fabrics and nonwovens
6) Bitumen reinforcement with blend fabrics of
glass staple fiber yarns, glass filament yarns,
and textile glass nonwovens
7) Gaskets with glass cord and glass twine (an
asbestos replacement)
8) Decorative fabrics of textured glass filament
yarns and glass staple fiber yarns
9) Expansion joints of impregnated woven glass
fabrics
10) Cable insulation and reinforcement with tex-
tile glass yarns
11) Reinforcement of adhesive tapes with textile
glass yarns
2.1.4. Product Standards
The textile glass standards described in Section 4
contain supplementary basis sheets with speci-
fications. Beyond that, there are customer ac-
ceptance standards, which list minimum require-
ments. In order for the values determined to be
comparable and reproducible, standardized test
procedures are used in which all testing condi-
tions are specified precisely.
Processors of textile glass products should
store the products in dry rooms at temperatures
below ca. 25 ◦ C at a relative humidity level bet-
ween 55 and 65 %. Exposure to direct sunlight is
to be avoided. Prior to processing, the material
should be equilibrated under the temperature-
humidity conditions present in processing.
2.1.5. Toxicology and Occupational Health
Textile glass products do not represent a health
or environmental hazard, mainly because of the
following properties:
1) Textile glass fibers do not contain poisonous
and/or hazardous components.
2) Fiber structure, because of the chemical com-
position, is affected by body fluids in such a
way as to reduce the carcinogenic potential.
3) Textile glass fibers do not split longitudinally,
which largely prevents their penetration into
the alveoli of the lungs.
Extensive systematic examinations of many
thousands of people working for decades in the
production and processing of textile glass prod-
ucts have shown that continuing exposure to
textile glass fibers does not produce increased
morbidity or mortality rates. Likewise, innumer-
able animal experiments (inhalation and inges-
tion of textile glass fibers as well as skin contact)
have revealed no fibrinogenic or carcinogenic
changes in test animals, a contrast to asbestos.
 Fibers, 5. Synthetic Inorganic 21

2.1.6. Economic Aspects mechanical or pneumatic processes. For insula-

tion, nontextile fibers are used exclusively. Such
The growth of textile glass fiber production fibers are also called glass wool, slag wool, rock
is illustrated by comparing the total figure of wool, or mineral wool.
1 135 000 for 1985 (Table 11) with that for 1973, While the manufacture of glass filaments
865 000 t. A 56 % increase was achieved. goes back to antiquity, it was not until 1840
Table 11. Production of textile glass fibers by final use and
that the production of nontextile man-made min-
geographic region, 1985 [33] eral fibers (MMMF) was demonstrated in Wales
Area Production, t [34]. This was an effort to utilize blast furnace
slag and at the same time to find a replacement
Uses in glass Other uses, including for natural asbestos. In the United States, rock
fiber reinforced fabrics for circuit
resins boards wool production is documented to have taken
place in 1893 [35]. An early patent for glass
United States,
Mexico, Canada 520 000 80 000 wadding by the centrifugal process was granted
Western Europe 240 000 65 000 in 1912, but industrial exploitation of the pro-
Japan and Korea 165 000 35 000
Eastern Europe,
cess did not occur until 1930 [36]; the blowing
Former States of 155 000 40 000 process for glass wool, still in use today, was
USSR, China patented in 1933 [37].
Others 50 000
Total 1 130 00 220 000 The hope to replace asbestos could not be re-
alized. Instead, these products found their own
areas of application. Synthetic resin binders
were developed around 1950, which made possi-
2.1.7. Producers ble the production of dimensionally stable prod-
ucts and assured a uniform quality for insulation.
The principal producers of textile glass fibers are
listed in Table 12.
2.2.1. Physical Properties
Table 12. Companies producing textile glass fibers and the
location of their plants
Fibrous insulating materials of man-made min-
Company Location of plants
eral fibers are a highly disperse system of finely
Asahi Fiberglass Japan divided fibrous glass with a specific surface of
Certain Teed Corp.
Fiberglass Ltd.
United States
Great Britain
0.1 – 0.2 m2 /g. The fibers are oriented predom-
Glaswerk Schuller ∗ Federal Republic of Germany inantly in the horizontal direction. For this rea-
Manville Corp. United States son, the material is isotropic in two dimensions
Nippon Electric Japan
Nippon Glassfiber Japan but not in the third. It is an open-pore system
Nitto Boseki Japan with little resistance to flow. This is the basis for
Owens Corning Fiberglass United States, Belgium, France
Pittsburgh Plate Glass (PPG) United States
its thermal and acoustic properties.
Scandinavian Glassfiber Sweden The fiber diameters are distributed over a
Silenka The Netherlands broad spectrum. The logarithm of the fiber diam-
Skloplast Former Czechoslovakia
VEB Glasseidenwerk Oschatz Former German Democratic eter approximately follows a normal distribution
Republic (Fig. 24, see next page). The minimum fiber di-
Vetrotex Federal Republic of Germany ∗∗
(Gevetex), France, Italy, Spain
ameter is process dependent and is between 0.18
Vitrofil Italy and 0.30 µm [38]. However, the properties of the
individual fibers are less important than those of
∗ Glaswerk Schuller specializes in staple fibers.
∗∗ The Vetrotex subsidiary Gevetex operates the plant in Germany. the final insulating material, i.e., the fibers and
binder as a whole.
Table 13 (see page after next) lists the limit-
ing temperature in use which is affected by the
2.2. Glass Fibers for Insulation chemical composition, the fiber diameter, the
amount and type of binder as well as the bulk
Glass fibers for insulation are fibers of vitre- density and the actual conditions of installation.
ous structure produced from silicate melts by
22 Fibers, 5. Synthetic Inorganic
Figure 24. Distribution of fiber diameters in glass and rock wools
The most important property of insulation is
the thermal conductivity, which is shown in Fig-
ure 25 for 14 mineral wools. The thermal con-
ductivity of a particular fibrous material is de-
termined by its bulk density B and its average
fiber diameter D. The mechanical properties are
primarily determined by bulk density and binder
content; in any case, however, fibrous insulating
materials cannot be used as load-bearing struc-
tural units.
2.2.2. Chemical Properties
Mineral wool insulation is inert to organic sol-
vents, even at elevated temperatures. Extracta-
bles are mineral oils, emulsifiers, silicones, and
low molecular mass components of binders.
Water and aqueous solutions attack the glass
surface, the attack magnified by the large fiber
surface. Attack by water also loosens the fiber –
resin bond. Calciferous products (slag wool,
rock wool) are soluble in acids and are attacked
by acid flue gases.
2.2.3. Raw Materials and Chemical
Composition
Glass wool (wadding, fibers) are produced from
mixtures of minerals and industrial chemicals
that produce the required final composition. In
 Fibers, 5. Synthetic Inorganic 23
Table 13. Chemical composition, fiberization process, and limiting temperature in use of glass fibers for insulation

Glass wool Rock wool Slag wool

1 2 3 4 5 6 7 8 9

Composition, wt %
SiO2 57.4 64.6 53.9 52.7 46.3 47.6 46 46.8 40.2
Al2 O3 3.6 2.9 3.6 8.0 13.3 10.2 15 10.2 15.6
Fe2 O3 0.35 0.35 0.35 1.8 5.6
6.5 ∗ 0.9 ∗
FeO 0.10 0.10 0.09 1.4 6.3 7.3 1.0
MnO 0.05 0.2 0.3 0.3
TiO2 0.1 0.6 2.9 1.9 1.5 0.7
ZrO2 4.0
MgO 4.0 3.1 4.6 4.0 9.7 12.0 12 2.3 4.35
CaO 8.6 6.6 23.4 20.9 10.9 14.8 16 37.7 34.9
BaO 1.8 2.4
Na2 O 11.2 14.0 9.4 10.3 2.6 1.6 2.5 0.9
K2 O 1.7 1.1 1.0 0.6 1.3 1.6 0.6 1.5
B2 O 3 5.0 5.9 3.0 0
F 2.3 0 0 0

Fiberization
process ∗∗ C C A A A B B D D

Limiting temperature
in use, ◦ C 570 550 700 700 750 750 750 730 700

∗ Fe2 O3 and FeO.

∗∗ A) Owens process; B) Rock wool process; C) Rotary process; D) Centrifugal blowing process. These four processes are
illustrated in parts A, B, C, and D, respectively, of Figure 26.

addition to quartz sand, the minerals used in-
clude limestone, dolomite, phonolite, feldspar,
nepheline-syenite, kernite, colemanite, ulexite,
and heavy spar (barite). The industrial chemi-
cals include sodium carbonate, sodium sulfate,
and barium carbonate. The use of recycling glass
has gained considerable importance.
In contrast, rock wool and slag wool are, as a
rule, prepared from a single mineral raw material
of suitable composition, such as basalt, rhyolite,
or diabase. To improve meltability, limestone is
added as required. Table 1 shows examples of
chemical compositions.
The mineral fiber is the predominant compo-
nent of the insulating materials:
90 – 99.8 % vitreous mineral fibers, average fiber diameter 3 – 6 µm
10 % max. phenol – formaldehyde resins, frequently modified
with urea or melamine
1 % max. aliphatic mineral oil
0.2 % max. emulsifiers, preferably nonionic polyethoxylenes
0.5 % max. polymethylsiloxanol
2.2.4. Production
Melting. Two types of melting units are in
use for the production of silicate melt: glass
melting tank furnaces and cupola furnaces.
The glass melting tank furnaces are similar in
construction to those used in other areas of glass
manufacture [39]. Preferred are cross-fired recu-
perative furnaces and U-shaped flame furnaces,
but all-electric melters are also used if energy
costs are right. Maximum productivity is 300 t/d,
and up to 3.5 tm−2 d−1 can be achieved. Cupola
furnacesare used to produce molten rock. They
require coarse feed and yield a less uniform melt.
The traditional fuel is coke but oil- or gas-fired
units are also used. Furnace productivity rarely
exceeds 150 t/d.
Fiber Production. Numerous processes and
process variants have been developed for fiber-
izing silicate melt. Of these, only four have with-
stood the test of time (Fig. 26).
Owens Process. In the Owens blowing pro-
cess (Fig. 26A), the melt, in the form of a thick
primary strand, emerges from holes at the bot-
tom of a platinum crucible (the bushing) and is
subsequently entrained by a downward-directed
stream of gas from the blow-jet. The blowing
medium may be steam or compressed air. The
primary strand is drawn to form fibers with an av-
erage diameter of 3 – 6 µm (Fig. 27). The length
24 Fibers, 5. Synthetic Inorganic
Figure 25. Thermal conductivity of glass and rock wools as a function of temperature
of the fibers is about 3 – 10 cm. The product con-
tains slugs. Glass and rock melts are processed
in this way.
Rock Wool Process. This centrifugal cascad-
ing process (Fig. 26B) is also a one-step fiber-
izing process. A single stream of melt hits the
face of a horizontal rotor which distributes it to
two or three additional rotors; from which it is
thrown in the form of fine fibers by centrifugal
force. The fiber length is in the range 3 – 10 cm.
The product contains slugs. Rock melts are used
in this process.
Rotary Process. The rotary process
(Fig. 26C) is a two-step process. The melt, in
the form of a thick stream, passes through the
hollow shaft of a rotating centrifuge into a ro-
tating throw-basket of platinum or special steel,
which distributes the melt evenly on the inside
of a rotating spinner. Primary fibers issuing from
holes in the face of this spinner, are entrained by
the hot combustion gases of a circular burner and
drawn to long fibers with diameters of 3 – 6 µm.
The product is free of slugs (Fig. 28), but the pro-
cess can be used only for “soft, long glasses,”
not for rock melts. The process is highly variable
with respect to process parameters but is easy to
control.
Other Processes. The centrifugal blowing
process, shown in Figure 26D, is of little signif-
icance in the production of slag wool. The rod
drawing process and a variant of the blowing
process serve to produce textile glass fibers for
glass mats. Special glass microfibers are made
by the flameattenuation process. However, these
 Fibers, 5. Synthetic Inorganic 25

Figure 26. Fiberization processes

A) Owens process [40]; B) Rock wool process [41]; C) Rotary process [41]; D) Centrifugal blowing process [41]
a) Melt; b) Bushing; c) Spout; d) Melt stream; e) Primary fiber; f) Jet; g) Centrifugal spinner; h) Burner; i) Fiber stream

Figure 28. SEM photograph of glass wool produced by the

rotary (TEL) process
processes have limited productivity and con-
sume large amounts of energy for fiberization,
Figure 27. Scanning electron microscopy (SEM) photo-
graph of rock wool produced by the Owens process
26 Fibers, 5. Synthetic Inorganic
and therefore they are not used to produce fi-
brous insulating materials.
Product Forming. The fibers are formed
into a continuous sheet by pneumatic pro-
cesses, while an aqueous solution of binder is
sprayed onto the freely falling individual fibers.
The binders are slightly pre-condensed phenol –
formaldehyde resins. They are often modified by
aminoplastic resins (urea, melamine, melamine
ether, dicyanodiamide) to suppress the tendency
of phenolics to exothermally autoxidize. In the
industry these aminoplastic resins are called an-
tipunking agents.
Figure 29 shows schematically the overall
manufacturing process. The binder resins are
cured at temperatures around 180 – 250 ◦ C in a
continuous flow dryer heated by circulated air.
For shaped articles or pipe insulation, for ex-
ample, special ovens that operate on a similar
principle are used. On the other hand, presses
with purely convective heat transfer are rarely
used since they require long cycle times.
Mineral-fiber insulating materials can be cut
to any desired dimension with knives, saws,
shears, and milling cutters.
2.2.5. Environmental Protection
Characteristic emissions arise in the various
stages of producing insulating material from
Figure 29. Mineral-fiber insulation, from raw material to product [42]
The rotary process is used to form fibers.
mineral fibers. Emissions from the melting units
are similar to those of other glass melting instal-
lations. Electric precipitators have proved useful
in dust removal but greatly strain the fluegas con-
ditioning system. They can be combined with the
usual wet and semi-wet installations for sorption
of fluorine and desulfurization.
The binder resins should contain a mini-
mum of monomers. Lubricants, mineral oils, hy-
drophobic agents, and other organic additives
must be selected with an eye to their volatility.
Scrubbing systems with wet electric precipita-
tors or deep-bed aerosol separators are suitable
for minimizing emissions.
Wastewater from the manufacturing plant is
biodegradable. Treatment with municipal waste-
water is always preferable. The high water re-
quirements of the off-gas conditioning system
favors installation of closed circulating systems
with the result that manufacturing plants can be
operated free of wastewater.
Local ordinances must be observed in remov-
ing process waste. The insulating material itself,
like other building materials, is inert and does
not have the potential for endangering water.
Both the dust from the filter units and the process
waste is suitable for internal recycling. The fiber-
ization units are serious noise emitters; there-
fore, the manufacturing plant must be housed in
massive buildings.
The overall effect of mineral-fiber fibrous in-
sulating material on the environment is positive.

There is no other building material that offers
a similar potential for minimizing the consump-
tion of primary energy for the same raw material
and energy requirements. Average energy amor-
tization times of 0.3 years and ecological amor-
tization times of less than 0.1 years have been
calculated.
2.2.6. Quality and Analysis
Each country has very detailed standards for
fibrous insulating materials. International stan-
dardization is in progress, e.g., in the areas of
thermal insulation in high-rise construction, fire
protection, and industrial acoustics [43].
Figure 30 shows a scheme of analysis for
mineral-fiber insulating material. Glass analy-
sis alone is, as a rule, not sufficient to identify
Figure 30. Scheme of analysis for mineral-fiber insulating materials
Fibers, 5. Synthetic Inorganic 27
the producer. However, analysis of characteris-
tic trace elements generally allows identification
of the producer. Certain finishing specialities of
the final product may permit identification.
From the ecological point of view, an elu-
ate analysis is essential. The eluate is obtained
by shaking the insulating material at room tem-
perature with a tenfold amount of water for 24
hours.
For the determination of formaldehyde, the
perforator method DIN-EN 120 and the gas anal-
ysis method of DIN 52693 are suitable. Analo-
gous methods are specified in other countries.
2.2.7. Applications
Mineral-fiber insulating materials are finished
products for use in protection from heat, cold,
28 Fibers, 5. Synthetic Inorganic
fire, and noise. The major end use area is high-
rise construction. Considerable quantities are
used in industrial and machine construction for
thermal insulation. Fibrous insulating materials
are also important as noise absorbers in build-
ings and elsewhere.
Economic Significance. World production
of mineral-fiber insulating material reached a
maximum of ca. 6×106 t in 1980 and has stag-
nated since (Table 14). About 75 % of the total
is produced in the industrialized countries with
temperate and subarctic climate. Consumption
is closely tied to investment and building activ-
ity. About 60 % of the demand for insulation in
industrialized countries is met by mineral-fiber
insulating material.
Table 14. Production of mineral-fiber insulating materials in
various geographic areas, 1973 and 1983.
Area Production
1973 1983
t % t %
North America
Western Europe
1 600 000 43
1 200 000 32
2 350 000 42 46.9
1 750 000 31 45.8
Eastern Europe and 600 000 16 800 000 14 33.3
the Former States of
USSR
Japan 200 000 5 400 000 7 100
Australia 50 000 1 mineral and synthetic aluminosilicate fibers by
} 120 000 } 3 } 150
Other 250 000 5
World total 3 750 000 100 5 600 000 100 49.3
2.2.8. Toxicology and Occupational Health
Synthetic mineral fibers are hard, abrasive mate-
rials. Constant handling can damage the epider-
mis, especially the finger tips and other exposed
areas of the body. Therefore, work gloves should
be worn, and the skin should be kept smooth by
applying oily, protective skin lotions.
The fibrous dust formed during mechanical
loading, cutting, sawing, or rubbing may irritate
the skin or the mucous membranes. Especially
the coarse fibers, with diameters greater than
about 12 µm, and the nonfibrous components are
able to penetrate the epidermis, thus causing in-
juries and facilitating the entry of germs. The
irritations are slight and rapidly heal with nor-
mal body hygiene.
Ordinances pertaining to protection of the
workplace prescribe a general permissible limit
of dust; in most countries, this limit is 5 –
10 mg/m3 . In the Federal Republic of Germany
the limit is 6 mg/m3 . Epidemiological investiga-
tions suggest that there is no health risk associ-
ated with the handling of mineral-fiber insulat-
ing material if these limits are observed.
Suspicions of a carcinogenic potential were
derived from animal experiments with ex-
tremely fine glass fibers carried out in 1972 –
1978. Up to now these suspicions have not been
confirmed with fibers from insulating materials.
Animal experiments involving dust inhalation
gave no indication of carcinogenicity. Exten-
sive epidemiological investigations could nei-
ther confirm nor deny these suspicions. Further
research is needed.
Fibrous insulation materials are free of
preservatives, fungicides, or biocides. They do
not contain asbestos or silicogenic material.
Increase, Formaldehyde is usually absent or present only
%
in trace amounts (<10 mg/kg).
3. Refractory Fibers
Refractory fibers are used at temperatures in ex-
cess of 1000 ◦ C. They are distinguishable from
their greater durability at high temperature and
by the fact that they have polycrystalline mi-
crostructures rather than being amorphous. They
are relatively new products, having only been de-
veloped during the last 20 years, and additional
varieties continue to be announced at a signif-
icant rate. The majority of fibers of this type
have been developed for thermal insulation at
high temperatures or for advanced composites
having the special physical properties required
in aerospace applications. These materials are
expensive since they are still manufactured on a
small scale.
Historical Background.Until the 1960s, re-
fractory fibers were either naturally occurring
fibrous minerals of the asbestos type or melt-
spun glass fibers such as E glass, rockwool, or
kaolin-derived fibers. Their maximum use tem-
peratures were limited by their melting points,
kaolin fibers having the highest use temperature
of 1250 ◦ C. Since then more refractory grades

of melt-spun kaolin fibers, capable of use up to
1400 ◦ C, have appeared.
The development of several types of novel
oxide fibers since 1970 has had several moti-
vations. At the time, an awareness was devel-
oping that a safe, noncarcinogenic, cheap as-
bestos replacement might be possible. Asbestos
was widely used in medium-temperature insu-
lation, in textiles, and in friction materials so
that replacement in these uses appeared to be
an attractive technical target. About the same
time the worldwide rise in oil prices increased
the demand for energy-saving insulation, espe-
cially in high-temperature furnaces operating
above 1250 ◦ C. Consequently the need for im-
proved fiber refractoriness and freedom from
nonfibrous impurities (combined with an aware-
ness that very fine fibers like asbestos might be
carcinogenic) lead to the development of new
high-alumina and zirconia staple fibers. Simulta-
neously, continuous textile-grade multifilament
fibers have also been developed to satisfy the
need for high-temperature resistant textile appli-
cations such as furnace curtains, fire barriers, hot
gas filters, electrical insulation, and protective
fire-fighting suits. In the 1980s renewed interest
in the potential of metal-matrix composites is
also sustaining the further development of novel
high-alumina fibers.
During the 1960s the needs of the U.S.
aerospace program for high-modulus, low-
density reinforcing fibers led to the develop-
ment of entirely new non-oxide fibers; initially
to CVD boron fiber and later to carbon fibers. To
a large extent the development of cheap, high-
quality, carbon fibers in the 1970s has overtaken
that of boron fiber since the price of the latter
has not been reduced significantly. A major dis-
advantage of carbon and boron fibers is their
ready oxidation at moderate temperatures, and
so during the late 1960s and 1970s various kinds
of oxidation-resistant silicon carbide fibers have
been evolved. For example, the makers of CVD
boron have turned to CVD silicon carbide as
a potentially cheaper and better product. An-
other driving force in the development of con-
tinuous textile-grade silicon carbide fibers (such
as Nicalon) has been the desire to have good
high-temperature properties combined with re-
sistance to oxidation and to attack by molten
metals. At present, therefore, advanced oxide
Fibers, 5. Synthetic Inorganic 29
fibers and silicon carbide variants represent the
two main categories of refractory fibers.
The chief materials of interest are boron
[7440-42-8], silicon carbide [409-21-2], high-
alumina [1344-28-1] (0 – 30 % SiO2 /Al2 O3 ),
and zirconia-based [1314-23-4] products. For
each of these materials there is a choice of sev-
eral grades from a range of suppliers. Fibers are
available in a variety of primary forms such as
continuous multifiber yarn, short staple random
fibers, or as whiskers (→Whiskers). Typically,
products have diameters within the range 1 –
150 µm. Some of these primary products may be
converted easily to secondary forms such as wo-
ven cloths, ropes, papers, boards, prepregs, etc.
All products on the market have complex com-
positions and microstructure so that each can
only be referred to uniquely by its trade name.
These are listed in Table 15.
Other materials that are or have been avail-
able in fiber form are boron carbide (B4 C), sili-
con nitride [12033-89-5] (Si3 N4 , whiskers), and
boron nitride (BN). For these materials the sup-
ply situation varies since their applications tend
to be highly speculative.
3.1. Properties and Raw Materials
3.1.1. Physical Properties
The behavior of fibrous materials is strongly de-
pendent on the shape and diameter of the fibers.
Typically the finer fibers are in diameter, the
more flexible they become and the more diffi-
cult it is to pack them into high volume frac-
tions. Fine random staple and whisker fibers
are usually low bulk density materials, while
thick monofilament or multifilament continuous
fibers, which are both more rigid and less flexi-
ble, are easily packed into high volume fractions
(60 % for example).
These characteristics determine utility and
the methods of handling. Fine staple fibers are
easily used as low density thermal insulation
and for the manufacture of low volume fraction
composites, but may present dust problems in
handling. Large diameter continuous fibers are
more readily woven into textile forms for high-
temperature use and are also more readily made
into unidirectional high-volume fraction com-
posites. On the other hand, the main handling
Table 15. Fiber properties ∗

30
Manu- Fiber name Chemical Type (fiber Dimensions Mechanical properties Compress- Fiber Thermal properties Crystal
facturer type count) ive strength, density Phases
MPa kg/m3 ,

Parti- Diameter, Length Modulus, Tensile Coeffi- Melting Maximum

culate µm GPa strength, cient of point or use temper-
◦
MPa thermal liquidus, C ature,
◦
expansion, C
106 K−1

Avco Boron boron filament 140 continuous 390 3000 – 5000 6600 2400 4.7 – 7.0 2200 500 Amorphous
(tungsten)
Avco SCS2 silicon filament 140 continuous 420 3400 – 4500 4000 3000 2830 1600 β-SiC
carbide
Nippon Nicalon silicon yarn (500) 15 continuous 176 – 196 2450 – 2950 >3000 2600 3.1 ca. 1400 β-
Carbon SiC/graphite

carbide
Ube Tyranno silicon− yarn 10 – 13 continuous 220 2800 – 3000 2300 – 2500 3.1 >1300 Amorphous
Industries
Fibers, 5. Synthetic Inorganic

Ti−C−O
Arco Silar SCS9 silicon whisker 10 – 20 % 0.6 10 – 80 µm 700 7000 3200 2700 α-SiC
Metals-
Silag carbide
Tokai Tokomax silicon whisker <1 % 0.1 – 0.5 50 – 200 µm 400 – 700 3000 – 3190 2700 1600 β-SiC
Carbon 14 000
carbide
Tateho SCW#1 silicon whisker 0.1 – 0.5 10 – 40 µm 480 20 700 3210 β-SiC
Chemical carbide
Los silicon whisker 3 – 10 10 – 25 mm 520 – 650 2000 – β-SiC
Alamos 20 000
carbide
Tateho SNW#1 silicon whisker 0.5 – 2 50 – 300 µm 380 14 000 3180 1900
nitride
Chemical
3M Nextel 312 24 % Silica/ yarn (600, 11 continuous 152 1720 2700 3.5 1800 1200 “inverse
mullite”
14 % 900, 1200)
Boria/
Alumina
3M Nextel 440 28 % Silica/ yarn (600, 11 continuous 207 – 240 1720 3100 5.0 1890 1430 η- or
γ-Al2 O3
Table 15. (Continued)

Manu- Fiber name Chemical Type (fiber Dimensions Mechanical properties Compress- Fiber Thermal properties Crystal
facturer type count) ive strength, density Phases
MPa kg/m3 ,

Parti- Diameter, Length Modulus, Tensile Coeffi- Melting Maximum

culate µm GPa strength, cient of point or use temper-
◦
MPa thermal liquidus, C ature,
◦
expansion, C
106 K−1

2 % Boria/ 900, 1200)

Alumina
Carbo- Fibermax 27 % Silica/ staple 2 – 3.5 150 850 1890 1650
rundum
Alumina
Denka Alcen 20 % Silica/ staple; yarn 2 – 3; 10 170 1700 3300 1890 1600 mullite/α-
Al2 O3

Alumina
Sumitomo 15 % Silica/ yarn (1000) 17 continuous 200 1500 2800 – 4400 3200 8.8 1890 1250 γ-Al2 O3
Alumina
ICI Catalytic 5 % Silica/ staple <1 % 1–5 5 cm 100 1000 1800 1980 η-Al2 O3
Saffil Alumina
ICI Saffil RF 5 % Silica/ staple <1 % 1–5 5 cm 300 2000 3300 1980 δ/θ-Al2 O3
Alumina
ICI Saffil HA 5 % Silica/ staple <1 % 1–5 5 cm >300 1500 3400 1980 1600 α-
Al2 O3 /mullite

Alumina
Du Pont Fiber FP >99 % yarn (200) 20 continuous 380 >1400 6900 3900 2040 1320 α-Al2 O3
Alumina
3M Nextel Z11 32 % Silica/ yarn 14 continuous 76 1310 3700 2000 1000 tetragonal
ZrO2
Zirconia
Zircar ZYBF-2 8 % Yttria/ yarn/staple 3 – 10 continuous/ 6000 2580 2200 tetragonal
Products ZrO2
Zirconia 1.6 mm

∗ In air, fiber strength severely affected.

Fibers, 5. Synthetic Inorganic
31
32 Fibers, 5. Synthetic Inorganic
risk is skin irritation caused by penetration of
the rigid fibers.
The physical properties of most importance
are fiber diameter, length, shape, density, mod-
ulus, tensile and compressive strength, expan-
sion coefficient, thermal conductivity, and max-
imum working temperature. These are summa-
rized from available manufacturers brochures in
Table 15.
Only data on fibers with diameters greater
than 1 µm are reliable. Data for finer whiskers
are implicit from data on composites. Neverthe-
less, single crystal whiskers of silicon carbide
have the best mechanical properties (modulus
around 600 GPa, strengths up to 20 000 MPa).
Polycrystalline Avco silicon carbide and boron
and Du Pont α-alumina are of comparable stiff-
ness, but Du Pont alumina is weaker than the
others (modulus about 400 GPa and strengths
1500 – 5000 MPa). Finally polycrystalline tran-
sition aluminas, mullites, and Nicalon have
moduli around 200 GPa and strengths around
2000 MPa. In composite applications specific
modulus and strength are more important, and
here boron possesses the greatest advantage
since its density is the lowest. Silicon carbide and
transition aluminas follow at ca. 3000 kg/m3 ,
while α-alumina and zirconia-based fibers fol-
low with densities of ca. 4000 kg/m3 . Figure 31
shows SEM microphotographs of three different
refractory fibers.
Figure 31. SEM microphotographs of several refractory fibers; scale: —−→—10 µm
A) Silicon carbide whiskers; B) Aluminum oxide staple fiber (Saffil, ICI); C) Titanium-modified silicon carbide fiber (Tyranno,
Ube Industries)
3.1.2. Chemical Properties
Refractory fibers as a class of materials are of
low reactivity. The chief area of concern is their
resistance to oxidation at high temperatures and,
if used in metal – matrix composites, their inert-
ness to corrosion by metals during fabrication
and use. Reported thermal data are also given in
Table 1, including claimed maximum use tem-
peratures. Boron fibers are the most reactive to-
wards oxidative and metallic attack. Several at-
tempts have been made to coat boron fibers with
inert layers, which has led to the BorSic product
(SiC-coated boron). The chemical vapor deposi-
tion process for making boron fibers lends itself
to laying down additional coatings. In France
boron carbide-coated boron has also been tested
[44].
The properties of silicon carbide fibers and
whiskers in composites have been thoroughly
reviewed by Andersson and Warren [45, 46].
Their resistance to metal attack exceeds that of
boron by 200 – 300 ◦ C. High alumina fibers are
inert in air and light alloys, but are slowly at-
tacked by hot strong acids or alkalis. Zirconia
fibers have exceptional thermal stability and are
also resistant to hot alkali.
3.1.3. Resources, Raw Materials
The raw materials for refractory fibers are usu-
ally specially selected intermediates. Since the
elements needed are among the most abundant,
 Fibers, 5. Synthetic Inorganic 33

there is no shortage of primary sources. The in-

termediates, however, are only readily available
in industrialized countries. In the case of silicon
carbide whiskers, a process has been commer-
cialized which uses rice hulls as raw material
[47].

3.2. Production

Because of their relative novelty, the processes

for making refractory fibers are almost all pro-
prietory, and full details have never been pub-
lished. A summary of the patent literature [48]
since 1970, relating to both the production of
refractory fibers and their uses in composites,
has appeared. There is a special section on re-
fractory fibers in Bela von Falkai’s book [49].
For further useful information refer to [50, 51].

3.2.1. Boron Monofilament

Carlsson [52] has reviewed general methods

of making boron fibers. In the 1960s, Avco
(USA) developed a chemical vapor deposition
(CVD) process to make 140-µm boron monofil-
aments having an 8-µm tungsten core [52, 53] on
a scale of 15 t/a. According to the company’s lat-
est brochure, the present capacity is only slightly
greater. More recently, a cheaper variant was
made available, in which the core is a 30-µm
carbon fiber, but this has now been withdrawn.
This development was initially financed by the
U.S. Government to develop advanced compos-
ites for use in the space shuttle and in aerospace
applications [53].
The process deposits amorphous boron with
a nodular morphology from a BCl3 – H2 mixture
onto the electrically heated tungsten core at ca.
1300 ◦ C.
BCl3 + 3/2 H2 → B + 3 HCl
The operation is carried out in single-stage
or two-stage tube reactors (Fig. 32), the filament
passing at 3 – 4 m/min through mercury seals,
which also act as electrical contacts. Control of
deposition temperature is critical because too
high values lead to crystallization of the boron,
which reduces the strength of the filaments. Aux-
iliary heating techniquescan increase quality
Figure 32. Apparatus with mercury seals for the prepara-
tion of boron fibers by two-stage chemical vapor deposition
[76].
a) Apertures (25 – 254 µm diameter); b) hydrogen feed tube;
c) BCl3 feed tube; d) Mercury pool; e) Tungsten wire sub-
strate; f ) Gas effluent; g) Power supply (d.c.)
and output. A range of products with various
diameters has been produced. Initially, the tung-
sten core made up a large proportion of raw
materials costs, but later the price of tungsten
dropped, and larger diameter fibers were devel-
oped so that this cost factor decreased in impor-
tance. The introduction of carbon-cored B fila-
ment has lead to further cost reduction [53]. Var-
ious surface treatments, such as acid polishing,
have been investigated to increase the strength
from 2 GPa to 5 – 7 GPa. Also surface coatings
such as silicon or boron carbide are applied by
34 Fibers, 5. Synthetic Inorganic

the CVD technique [54] to reduce the reactivity

with aluminum alloys and oxidation at temper-
atures above 300 ◦ C.
In the commercial plant, hundreds of reactors
are fed in parallel with the gaseous reactants.
Since the conversion of gas per pass is low, the
unconverted gases are recovered and recycled,
while the byproduct HCl is sold.
More recently, more rapid boron deposi-
tion methods have been reported: these meth-
ods operate at high temperature and high pres-
sure and do not require resistive heating of the
substrate. Pulse discharges have been used for
boron coating of filaments [55]. The excited
gas species are formed by inductive or capac-
itive coupling of pulsed high electric power to
a boron-containing gas mixture at a pressure of
0.1 – 0.3 MPa (Fig. 33). The gases admitted to
the deposition zone are hydrogen, mixed with
boron halides, boranes, or alkyl borates. The
deposition process is performed at 20 – 350 ◦ C
on filaments suspended in the deposition zone:
tungsten filaments, graphite yarn, and polyacry-
lonitrile yarns were coated with boron to a thick-
ness of about 50 µm. The residence time of the
substrate within the deposition zone was about Figure 33. Apparatus for the production of boron fibers
5 min. by capacitive coupling of pulsed high electric power [55].
a) Power supply (pulsed high electric power); b) Oil-cooled
reaction tube; c) Let-off spool; d) Seal; e) Take-up spool; f )
Oil reservoir; g) Pump
3.2.2. Silicon Carbide Fibers
Because of the highly attractive physical prop-
since been commercialized by Nippon Carbon
erties of fully crystalline silicon carbide and the
Co. as Nicalon [58]. A second generation car-
low cost of raw materials, more effort has been
bide fiber, namely Tyranno fiber from Ube In-
devoted to generating silicon carbide in fibrous
dustries, contains titanium in addition to silicon
forms than for any other material [45, 46, 56].
and is reported to have superior strength and
Silicon Carbide Monofilament. This mate- chemical resistance.
rial is available in composites from Avco and The Nicalon fiber process begins with the
is probably made by a CVD process [57] simi- conversion of dimethyldichlorosilane (an inter-
lar to that used for boron filament. The reagents mediate in silicone manufacture) to poly(di-
are an alkylchlorosilane or a mixture of alkane methylsilane) by dechlorination with sodium.
and chlorosilane condensed onto a tungsten or The poly(dimethylsilane) is heated under Ar in
carbon core. The product has a columnar grain an autoclave at 470 ◦ C for 8 – 14 h to convert
structure with the growth direction normal to the it to a poly(carbosilane) of complex structure
fiber axis. In order to increase the strength of the (see [58]). After vacuum distillation to 280 ◦ C
fiber it is usually coated with a thin carbon layer to remove low molecular mass fractions, the
(CVD process), which fills in any grain bound- poly (carbosilane) (M r 1500) is then melt-spun
ary faults at the filament surface. at 350 ◦ C as a multifilament tow. The raw fibers
are then cross-linked by oxidation in air for
Silicon Carbide Continuous Yarn. Sili- 30 min up to 190 ◦ C and heat-treated to 1200 ◦ C
con carbide continuous yarn was invented by in inert atmosphere to convert them to the fin-
S.Yajima in the 1970 – 1975 period and has

ished fiber having maximum strength. The fibers
have a complex constitution consisting of a
fine-grained mixture of graphite, β-SiC, and
glassy silica. The fiber stability is limited above
1200 ◦ C, and the modulus (which is a function
of fiber diameter) is only 50 % of the value of
CVD polycrystalline silicon carbide.
Silicon Carbide Whiskers (→Whiskers,
Chap. 4.1.). These have traditionally been made
by high-temperature synthesis between SiO2
and a carbon source under reducing conditions.
A more recent development uses as raw material
rice hulls, which contain Si and are processed at
1500 – 1660 ◦ C [47, 59]. They are usually single
crystals, although there is some confusion in the
literature as to the direction of the growth axis.
Very high strength β-silicon carbide whiskers
(6 µm diameter) are grown by a vapor – liquid –
solid (VLS) method at 1400 ◦ C from methane
and silicon monoxide using powdered iron cat-
alyst [60, 61]. Properties are given in Table 1.
3.2.3. High Alumina Fibers
High alumina fibers containing less than 30 wt %
SiO2 cannot be readily spun from melts be-
cause the temperatures are too high. Fibers in
this class are either made from water-soluble
or dispersible precursors (sol – gel products) or
from an organometallic precursor (only Sumit-
omo’s alumina). Fibers that are available range
from fine diameter (3 µm) random staple (ICI
Saffil alumina fiber) through to relatively coarse
(20 µm) continuous fibers (Du Pont FP alumina).
This class includes mullite [1302-93-8] fibers
(27 % SiO2 ) and 3 M Nextel alumina – boria –
silica fibers. Single-crystal whisker products are
also available (→Whiskers, Chap. 4.3.).
Continuous Yarn Fibers. Du Pont, 3 M,
and Sumitomo have supplied products of this
type for several years. ICI, Denka Kagaku, and
Mitsui mining announced new alumina-based
continuous or semicontinuous fibers.
In each case a different precursor is used,
but the processes are otherwise broadly similar
(Fig. 34, see next page). First a viscous precursor
solution is prepared and then converted to green
(or gel) fibers by extrusion through multi-hole
Fibers, 5. Synthetic Inorganic 35
spinnerets. The gel fibers are converted to refrac-
tory fibers by heat treatment, usually in multi-
stage furnace treatments. Major features of these
processes are the ability to make any composi-
tion and the large shrinkage experienced during
calcination.
According to patents [62, 63], Du Pont’s
FP fibers may be made from colloidal α-
alumina (e.g., Alcoa XA16 fractionated to 90 %
<0.5 µm diameter) dispersed by HCl treat-
ment and blended with an aluminum salt,
such as the chlorohydrate (commercially avail-
able basic aluminum chloride [1327-41-9],
Al2 (OH)5 Cl · 2 – 3 H2 O), acetate, or nitrate. The
viscous mixture is extruded into air to give gel
fibers, which are then subjected to two- or three-
stage calcination to give the final fiber (either
flame-polished in an oxypropane flame or at
1400 – 1500 ◦ C in a furnace). The fibers have
a diameter of 20 µm and are polycrystalline
α-alumina with a grain size <0.5 µm. Their
strength can be improved by a surface coating
ca. 5 nm thick of silica derived from a silicone
[64].
According to a patent [65] 3 M’s Nex-
tel fibers are made from a silica hydrosol
blended with aluminum formoacetate and boric
acid together with organic improvers such as
polyvinylpyrollidone (PVP), poly(vinyl alco-
hol) (PVA), lactic acid, or glucose. After concen-
tration to a viscosity between 45 and 500 Pa · s,
the solution is extruded through a multi-hole
spinneret into air to give gel fibers. These gel
fibers are calcined in air to at least 1000 ◦ C (poly-
crystalline η- or γ-alumina <60 nm grain size)
or to 1400 ◦ C (polycrystalline mullite >50 nm
grain size) if they are to be used up to 1400 ◦ C.
The finished fibers have an oval cross section
with a mean diameter of 11 µm and are suitable
for weaving cloths.
Sumitomo has described their process [66,
67]: first triethylaluminum or triethoxyalu-
minum is hydrolyzed with a controlled amount
of water to a polyalkyl(oxy)aluminoxane (de-
gree of polymerization of 10 – 200), which is
then dissolved in an organic solvent. An alkyl sil-
icate is added (to give 15 wt % SiO2 in the final
product) to prepare the viscous solution (0.1 –
500 Pa · s) for spinning. The fibers are dry-spun
into air. These gel fibers are then calcined in air to
700 ◦ C to decompose them to oxide and further
fired at 1000 ◦ C to crystallize them to a dense
36 Fibers, 5. Synthetic Inorganic

Figure 34. Flow chart of continuous refractory oxide fiber production [77].
a) Precursor solution; b) Metering pump; c) Multi-hole spinneret; d) Fiber bundle; e) Hot air distributor; f ) Draw rolls;
g) Transfer conveyor; h) Pre-fire resistance furnace; i) High-temperature tube furnace; j) Transfer rolls; k) Multi-filament fiber
tow

polycrystalline fiber having a γ-alumina and/or
a mullite structure with a grain size of several
nanometers). Production fibers are made with a
diameter of 17 µm.
So far no details are available for methods of
producing the newest continuous fibers.
Staple Fibers. Staple fiber products include
ICI Saffil alumina, Carborundum’s Fibermax
(and the equivalent Toshiba Monofrax), and
several products available from Japan such as
Denka’s Alcen.
In sol – gel processes [68], an aqueous
spinning solution is formulated from basic
aluminum chloride (aluminum chlorohydrate,
Al2 (OH)5 Cl · x H2 O), a water-dispersible silox-
ane, or a silica sol, and a viscosity modifier such
as PVP, PVA, or poly(ethylene oxide). Fibers are
either blow-spun through multi-hole spinnerets
or centrifugally spun into warm dry air and col-
lected as a random mat on a conveyor belt. They
are decomposed at 200 – 500 ◦ C and calcined
in one or two stages to recrystallize and sinter
the initially amorphous material. Various prod-
ucts are available, with different finishing tem-
peratures and containing different crystalline-
transition alumina phases (Table 1). Centrifu-
gally spun products contain more nonfibrous
material than blow-spun products.
3.2.4. Zirconia Fibers
Zircar fibers (yttria stabilized) are made by the
relic process, where a continuous cellulose fiber
is impregnated with suitable salts and then cal-
cined in air to decompose the salts to oxides and
burn away the cellulose. Yttria is added to main-
tain the zirconia in the tetragonal phase, pre-
venting the destructive conversion to monoclinic
phase on cooling. Alternative phase stabilizers

that have been investigated are calcium oxide
and magnesia. An alternative process, which has
been used by ICI to make staple fibers, is the
sol – gel route. This is described by others in
the literature [69] and generally uses zirconyl
acetate as the raw material. Other details are
similar to those used for alumina fibers. Nextel
Z-11 continuous zirconia – silica [14940-68-2]
fibers from 3 M are made much like the alu-
mina – silica – boria products [70].
3.3. Environmental Protection
Since these products have only been made in
small quantities so far and are physically and
chemically inert under ambient conditions, they
pose no significant threat to the environment.
Ultimate disposal is likely to be by burial.
The byproducts of fiber manufacture are either
acids, which are scrubbed and sold or neutral-
ized before disposal, or carbon oxides and small
amounts of volatile organic compounds.
3.4. Uses
The principal uses of refractory fibers are in
high-temperature insulation and composites.
3.4.1. High-Temperature Insulation and
Filtration
By reason of their refractoriness, those fibers
which are not attacked by oxygen or which form
passive oxide layers (oxides, SiC) can be used
in furnaces as high-temperature insulation or
structural elements. High-alumina fibers dom-
inate these applications. In general, the staple
and whisker materials that have diameters be-
low 6 µm are best for low-density insulation hav-
ing a low thermal conductivity. Conductivities at
1400 ◦ C as low as 0.3 – 0.4 Wm−1 K−1 can be
reached. Use of such low thermal mass insula-
tion allows faster cycle times in batch furnaces,
giving energy savings of 20 – 40 %.
In furnace applications, the short fiber prod-
ucts can be used in a variety of forms. When re-
silience is required, blanket or loose wool can be
employed in several ways. In one method, strips
Fibers, 5. Synthetic Inorganic 37
of blanket are glued to a rigid backing (stack-
bonding) to give a surface with high gas ero-
sion resistance. This technique can be applied to
existing brick-lined gas-fired furnaces (veneer-
ing) to give substantial energy savings. Staple
fibers can also be converted by wet-lay vacuum-
forming processes to rigid forms. For such prod-
ucts the refractory fibers are often blended with
amorphous alumino-silicate fibers so as to uprate
the maximum working temperature of the amor-
phous fibers, e.g., from 1250 to 1400 or 1500 ◦ C.
In these forms, sacrificial organic binders, such
as starch, are often combined with a refractory
binder such as a silica sol. It is also possible to
make papers by blending small proportions of
fine glass fibers into the slurry.
Thermal insulation for nuclear reactors
presents a requirement of high resilience under
high acoustic vibration loadings at modest tem-
peratures. Very low shot level staple fibers are
superior for this application.
The yarn and monofilament materials are best
used as woven cloths [71], which can be em-
ployed as flame barriers and radiation shields or
as bag filters for hot-gas filtration. Where den-
sity is not a prime consideration, Nextel Z11 is
claimed to have superior firewall properties. The
long fiber products can be converted to string,
woven cloths, ropes, tapes, thread, braiding for
electrical wiring, etc. Such yarns, ropes, etc. can
be employed as expansion joint packings, door
seals, and the like. Both staple and continuous
fibers can be used for space vehicle insulation
systems, e.g., tiles as on the space shuttle.
Both short and long fiber products can be used
in advanced composites. Large volumes of data
are available for composite properties using ei-
ther resin matrices, metal matrices, or ceramic
matrices. These present a range of opportunities.
The refractory properties of the fibers are
most effectively exploited in metal – matrix
or ceramic – matrix composites because good
properties are then available above 200 ◦ C in
many such systems. Both types of material are
presently undergoing intensive development to
evaluate their potential for replacement of con-
ventional metal alloys.
Resin-Matrix Composites. In resins, re-
fractory fibers are in competition with carbon
fibers and Kevlar fibers. In situations where the
composite is required to have high compressive
38 Fibers, 5. Synthetic Inorganic
strength, the use of boron, silicon carbide, or alu-
mina fibers is preferable. This has found utility
in parts of sporting goods such as golf clubs, ten-
nis rackets, and fishing rods. Aerospace applica-
tions have also been explored [53]. The U.S. F14
fighter had boron – epoxy horizontal-stabilizer
skins. Where electrical insulating properties as
well as mechanical properties are important,
then alumina or silicon carbide fibers can be
more effective than glass or Kevlar fibers.
Metal – Matrix Composites. Metal –
matrix composites (MMCs) where the matrix
is an aluminum, magnesium, titanium, or nickel
alloy, should be able to replace parts currently
made of steel, titanium, or nickel alloys. Fabrica-
tion and shaping of MMCs presents formidable
problems. Corrosion of fibers by the metal dur-
ing fabrication and use is the chief problem.
Aluminas are superior in this respect to silicon
carbide and boron, which, for example, is only
stable to 200 ◦ C in 2024 aluminum alloy. How-
ever, boron fibers are usable when SiC-coated.
Because of high costs, nearnet-shape forming
methods such as squeeze-casting, extrusion, and
forging are preferred. Aluminum-based MMCs
have reached commercialization in a few appli-
cations. Boron filament – aluminum composites
have been utilized for cargo bay spars in the
space shuttle. In the future, other fibers, such as
silicon carbide (continuous and whiskers) [72,
73] and alumina [74], will find uses in aerospace
and automotive engineering applications where
low density must be combined with rigidity,
strength, heat, and/or wear resistance.
The use of small amounts (<30 vol %) of
short alumina fibers in diesel engine pistons has
been commercialized by Toyota. Here the effect
of fibers is to increase the wear resistance of the
piston alloy in the region of the ring grooves so
that high nickel alloy inserts are not required. In
more advanced applications such fibers in piston
crowns can increase resistance to fatigue crack-
ing so that more efficient engines become pos-
sible. High volume fractions of unidirectional
long alumina fibers are currently being evalu-
ated in connecting rods for internal combustion
engines to decrease the mass of moving parts,
which should lead to increased engine efficiency.
BorSic and silicon carbide monofilaments
have been successfully incorporated into Ti-6Al-
4V alloy [44] with the aim of increasing the
maximum operating temperature of this alloy,
but this matrix is generally too aggressive to be
compatible with other types of fibers.
Ceramic – Matrix Composites. When ma-
terials are required to exceed the maximum
working temperature of nickel-based superal-
loys, ceramic – matrix composites are forecast
to fill the gap. If the more severe processing
and property requirements can be met, they will
be utilized in advanced gas-turbine blades and
in advanced internal combustion engine compo-
nents such as pistons and exhaust turbo-charger
rotors.
Currently much effort is devoted to improv-
ing the fracture toughness of monolithic engi-
neering ceramics such as glasses, glass ceram-
ics, silicon carbide, silicon nitride, alumina, and
partially-stabilized zirconia by the incorporation
of fiber reinforcement. This is the most severe
application of refractory fibers since they are re-
quired to resist corrosion by the matrix during
fabrication of the composite at temperatures of-
ten exceeding 1200 ◦ C. Some encouraging re-
sults have been reported in the literature [57],
but a real application has not yet been success-
ful.
3.5. Miscellaneous Uses
Miscellaneous applications of refractory fibers
have developed, such as the use in nuclear tech-
nology. Here, boron fibers find uses in fabricat-
ing flexible neutron shields.
Zirconia fibers find utility because of their
great resistance to alkali attack and are recom-
mended for use as battery separators. Nextel Z-
11 has outstanding flame resistance and is pre-
ferred for fire-wall uses.
Porous fibers, such as ICI catalytic grade Saf-
fil alumina, have been found to be effective as
catalyst supports in domestic flameless heaters.
3.6. Economic Aspects
Refractory fibers are specialities made on a rel-
atively small scale where raw materials costs
make little contribution to the selling price. The
major costs are related to processing and capital
costs.

Of all the fibers considered, short staple alu-
mina fibers are the product available in the great-
est quantity and at lowest price. World capac-
ity probably exceeds several hundred tonnes per
year. Prices range from $ 30 to $ 60 per kg (ca.
30 – 60 ¤/kg).
Prices for continuous fibers or whiskers vary
from $ 80 – 900 per kg. The cheapest is Nextel
312 at $ 80 per kg with Nextel 440 at $ 260 per
kg. SiC whiskers, Nicalon, Sumitomo alumina,
and DuPont FP fibers all cost ca. $ 450 – 550 per
kg, while Avco boron fiber (tungsten core) is
$ 825 – 900 per kg. These products are all prob-
ably made on a 10 – 20-t/a scale. Du Pont has
projected that at 500 t/a FP alumina fiber would
cost $ 88 per kg.
Secondary forms of the above materials–such
as vacuum-formed shapes for staple fiber or wo-
ven cloths for continuous fibers–can cost double
the basic fiber cost. Some products are available
as epoxy prepregs, and there is an appropriate
markup in price for such intermediates.
3.7. Toxicology and Occupational
Health
Fibers with diameters exceeding 3 µm are not
respirable and, like glass fibers, are difficult to
handle only because of skin irritation. For such
materials, operators handling the fibers are rec-
ommended to wear loose-fitting clothing.
Fine fibers with diameters below 3 µm are
respirable. Because of the known ability of
asbestos dust to cause both asbestosis and
mesothelioma, considerable effort has been de-
voted to assessing other fibrous mineral and ce-
ramic dusts for similar effects. In general, sus-
picion has now focussed on those materials that
are durable in the body and have diameters be-
low 0.5 µm and lengths greater than 8 µm [75].
Products that have diameters significantly above
this limit have been found to be inert biologi-
cally. According to a proposal, such materials
should be classified as nuisance dusts, for which
an MAK of 6 mg/m3 is appropriate. In the UK
and the United States the MAK is 10 mg/m3 . The
finer materials, especially whiskers, remain un-
der suspicion. Similar products have been with-
drawn from the market because animal tests have
proved them to have positive carcinogenic po-
tential. For such materials an MAK equivalent
Fibers, 5. Synthetic Inorganic 39
to that of asbestos has been recommended until
more definitive information is available.
4. Metal Fibers
The definition for a metal fiber is not straight-
forward, especially when metal fibers are classi-
fied according to nature, structure, properties, or
manufacturing processes. What cannot be con-
sidered a metal fiber is easier to state:
A single drawn wire used separately; any-
thing much thicker than 100 µm; and steel wool.
Metal fibers can usually be handled in the same
way as textiles, paper, or polymers, and have
small, equivalent diameters.
In this chapter, emphasis is placed upon
bundle-drawn and melt-spun fibers. For metal
whiskers see →Whiskers. Other methods of
manufacturing fibers are mentioned briefly.
4.1. Production
4.1.1. Bundle Drawing
Very fine wires can be produced by wire draw-
ing. Production costs are extremely high, how-
ever, and increase exponentially as the diame-
ter decreases. To overcome this problem bun-
dle drawing was developed as early as 1936
[78]. Instead of drawing wires singly, a num-
ber (in some cases several thousand) are bundled
and drawn simultaneously [79] (Fig. 35, see next
page). The sole difficulty remaining is to sepa-
rate the individual fibers with a suitable material
prior to bundling so that the final fibers can be
individualized again. Quite often used is another
metal that has sufficiently similar properties to
the wire metal to be made into fibers, to with-
stand the drawing and annealing operations, but
a lower chemical resistance to allow the fibers to
be freed from this matrix in a leaching operation.
Whereas this method greatly reduces drawing
and annealing costs, it only produces bundles of
fibers and not reels of single wires such as is
needed to produce fine tungsten wire to serve as
core material for boron fiber manufacturing.
40 Fibers, 5. Synthetic Inorganic

4.1.2. Melt Spinning [80–82] Melt Extrusion. Melt extrusion (melt spin-
ning) involves the solidification of a liquid jet of
The basic idea behind all the from-the-melt metal extruded through an orifice (Fig. 36). The
production methods is the possible economic stream of molten metal is inherently unstable,
advantage of single-step processes over con- mainly due either to its high surface tension and
ventional time- and energy-consuming, labor- low viscosity or to a violent drop in its viscosity
intensive fiber-drawing processes. These from- at its spinning temperature.
themelt processes can be roughly divided into
two classes: melt extrusion and melt extraction.
Dating back to the 1930s, from-the-melt fibers
developed rapidly in the 1960s after Duwez de-
veloped the splat cooling technique: a revolution
in metallurgy, resulting in what are now gen-
erally known as the RST (Rapid Solidification
Techniques), sparked intense interest in glassy
or amorphous metals. The Battelle Development
Corporation has pioneered all these processes in
one way or another, and is still doing so.

Figure 36. Melt spinning of metal fibers.

a) Crucible; b) Nozzle

The stream must be stabilized. There are a

number of variants, which can be grouped in
four divisions:
1) Jet stabilization in a controlled atmosphere,
triggering off a surface reaction and oxide
film formation (e.g., Monsanto, Michelin,
Bekaert, Brunswick).
2) Jet stabilization by sheathing in glass and
spinning the resulting composite. This so-
called Taylor process (Fig. 37) has been suc-
cessful on a semi-industrial scale (Glass De-
velopment, U.K.), but has virtually been
abandoned in spite of the excellent fiber qual-
Figure 35. Manufacturing alternatives of bundle-drawn ity, mainly because of low production capac-
metal fibers

ity, hence high cost. Fibers down to 1 µm are
possible.
3) Stabilization by heat transfer to a liquid (in-
rotating-liquid spinning) (e.g., Allied Signal,
Michelin, Unitika).
4) Solidification by direct contact with a
moving-solid, heat-extracting surface. Al-
though round fibers cannot be made by this
technique, it was and still is the only one that
has developed into a full-blown industrial
process (Allied Signal, 1976). In this respect,
it is the process responsible for the commer-
cialization of amorphous metals (ribbons and
sheets).
Figure 37. Taylor process for the production of metal fibers
Most of the processes failed to become fully
industrial because they could not be upgraded
from workable single-fiber processes to multi-
fiber operations. The most promising method
is the in-rotating liquid spinning, established in
1981 and the only method that enables fibers
with circular cross sections to be manufactured
in continuous lengths.
Improvements remain to be realized with re-
spect to production costs (development of the
Fibers, 5. Synthetic Inorganic 41
multi-hole process), raw-material prices (up to
10 % boron), and some vital mechanical prop-
erties, especially torsion characteristics and fa-
tigue behavior.
Melt Extraction. The melt extraction pro-
cesses avoid the need for orifices because the
molten metal is extracted from a melt by a
wedge-shaped edge of a rotating disk: CME
(crucible melt extraction) or PDME (pendant
drop melt extraction), see Figure 38. The pro-
cess was invented and developed by Battelle and
exploited on an industrial scale by Fiber Tech-
nologies (U.K.).
Figure 38. Crucible melt extraction (CME) for the produc-
tion of metal fibers.
a) V-edged wheel; b) Wiping wheel
4.1.3. Other Processes
Metal fibers have also been made by chemical
vapor deposition, mechanical methods, and elec-
trolytic procedures. Other less important meth-
ods include slurry drying, decomposition, and
reduction; in-situ decomposition of inorganic
and organic compounds; and physical vapor de-
position.
Chemical Vapor Deposition (Whiskers).
Metallic whiskers are generally made from the
gas phase by decomposing a compound of the
metal (e.g., pentacarbonyl iron). Whiskers of
iron, copper, and tin have been manufactured.
Mechanical Methods. Mechanical shaving
methods, similar to those for steelwool produc-
tion can be used, starting from either wire or
stacked thin foils. The fibers so produced have ir-
regular cross sections, either triangular or rectan-
gular with a surface roughness and poor mechan-
ical properties. The shaving methods were im-
42 Fibers, 5. Synthetic Inorganic

proved by the introduction of superimposed vi-

brations, which result in more homogeneous
cross sections and properties.

Electrolytical Procedures. After car-

bonization, organic fibers can be coated by elec-
trolysis to produce metal sheaths. The organic
and carbon residues can also be decomposed to
leave hollow fibers. Metal fibers of Ni, Cr, or Cu
as well as their alloys, can be formed. Another
process consists of depositing continuous fibers
on a suitably grooved mandrel.

4.2. Bundle-Drawn Fibers

4.2.1. Form and Dimensions

Bundle-drawn fibers can be produced in the

form of continuous bundles, broken bundles
(slivers), cut fibers, spun yarns, threads, cables,
web, sintered web, needle felt, filter media, etc.
All these morphologies are obtained by tex- Figure 39. Microphotographs of bundle-drawn metal fibers
tile industry-like operations combined with met- A) Scanning electron microscope (×1250) B) Optical mi-
allurgical techniques, processes, and methods. croscope (×1100)
a) 22 µm; b) 12 µm; c) 8 µm; d) 4 µm
The fiber diameters depend to a large extent
on the metallurgical and structural purity of the
starting material (wire rod), itself a function of
the alloying elements and the manufacturing and
purification procedures during wire-rod produc-
tion.
Figure 40. Very thin bundle-drawn metal fibers, diameter
2 µm (×625)
4.2.2. Composition

The following are typical alloys lending them-

selves to the bundle-drawn fiber manufacturing: 4.3. Uses
Stainless AISI 316L
Stainless AISI 302 Anti-Static Applications. Since metal
Carpenter 20Cb3 fibers are good electrical conductors, they were
Inconel 601 first used in anti-static carpeting to dissipate
FeCrAlloy the static charges. Very soon the same principle
Hastelloy X was applied to clothing for medical use as well
Nichrome as for work with explosives or sensitive elec-
titanium tronic components of instruments, followed by
anti-static filterbags for use with explosive dusts
Standard diameters are 4, 8, 12, and 22 µm [83]. Cast floors, e.g., as in operation rooms,
(Fig. 39). Of course, larger diameters are possi- are made sufficiently conductive to eliminate
ble, such as 40, 50, and even 100 µm. For some the possibility of sparks and electric discharges.
alloys 12 µm is the lower limit. For very pure The same also applies to bed linen for use with
alloys, fibers 2 µm and even finer are possible oxygen tents, etc.
(Fig. 40).

Filter Media [84]. Filters are one of the most
promising areas for the use of metal fibers, since
the existing products, such as woven wire mesh
and sintered metal powders, no longer fulfill all
criteria:
1) Fine pore size from 20 µm to a few microm-
eters
2) Large permeability (4 – 10 times higher than
at present)
3) High dirt holding capacity, therefore longer
operational life (fourfold or more than at
present)
4) Acceptable and competitive price in the ap-
plication niches
Only sintered metal fiber filter media can sat-
isfy these requirements successfully. The princi-
pal areas of metal fiber filter media applications
include the following:
polymer fiber and foil manufacturing
hot gas filtration
aircraft fuel de-icing and filtration
filtration in the fields of hydraulics and robotics
fine chemicals and pharmaceuticals
crossflow filtration
generation of nuclear energy: gas cooled reac-
tors, cooling
water condensate, elimination of radioactive
aerosols from waste gas flows during emer-
gencies
Sound Absorption. Sintered metal fiber
sheets can be deployed as highly effective sound
barriers under circumstances where other ab-
sorbers are inefficient, e.g., in high-temperature
environments and at high-energy sound levels.
This is particularly the case for aircraft and
turbine noise reduction.
Gas Burner Panels and Tubes [85].
Highly porous sintered FeCrAlloy sheets and
tubes produce an efficient radiant surface for
burning natural gas. The advantages over the
usual ceramic burners are improved life cycle,
lower NOx , better malleability, and sturdiness.
Thermal Insulation and Flame Arresting.
The heat conductivity of porous metal fiber
structures is comparable to that of an aver-
age quality glass fiber mat, and such products
have solved high-temperature insulation prob-
lems such as those involving the exhaust gases
Fibers, 5. Synthetic Inorganic 43
of the space shuttle thrust rockets. These sin-
tered metal fiber structures also make highly
permeable, extremely efficient, versatile flame
arrestors.
Medical Applications. Titanium fibers are
bio-compatible and can consequently be applied
as porous ingrowth materials for implants to
evoke bone development in the pores, guarantee-
ing perfect fixing and load transmission. Stain-
less steel fibers can be used to reinforce dental
bioglass.
Electromagnetic Interference (EMI) [86].
With the proliferation of electronic appliances
and instruments in our daily life, electromag-
netic interference must be controlled. More and
more appliances are shielded against outside
interference, as well as against creating inter-
ference themselves. Present-day housings for
electronic devices are made from composite
materials of polymer matrix and metal fibers.
The presence of a low concentration of conduc-
tive metal fibers of the correct alloy, structure,
diameter, and length give these casings a total
attenuation of more than 80 dB, this without
altering the other properties of the polymer ma-
trix.
Radar Camouflage. As a special case of
EMI, an amount of the correct metal fiber in
radar camouflage netting that matches as closely
as possible the absorption and reflection char-
acteristics of radar waves for the environment
has the effect of concealing an object. Structures
containing metal fiber can be adjusted to vary
from zero to almost 100 % reflection or absorp-
tion.
Microwave Detection. Every metallic fiber
structure has a unique microwave signature, or
fingerprint, that can be detected, registered, per-
sonalized, stored, and recognized. This feature
can serve for the identification and verification
of valuables, security papers, passports, credit
cards, identification cards, etc.
High-Voltage Protection. The effect of the
strong electromagnetic field on persons work-
ing on live high-tension lines can be harmful to
their health. A protective suit made from cloth
containing up to 25 % metal fibers acts as an ef-
ficient shield.
44 Fibers, 5. Synthetic Inorganic

Capillary Uses. Since bundle-drawn metal

fibers are finely grooved on their surfaces and
form very fine tubes in between themselves, a
fiber wick exhibits large capillary action. This
effect is used in heat pipes, space technology,
etc.

Other Uses. Many other uses are under con-

sideration or are being developed. A few impor-
tant ones should be mentioned: catalysts and cat-
alyst carriers, heat-exchanger media, diffusion
barriers, coalescing and demisting media, gas –
liquid separator media, high-temperature seals,
and the heating of textiles (e.g., curtains).

4.4. Manufacturers Figure 41. Surface of melt-extruded metal fiber, diameter

There are presently three major metal fiber pro-
ducers: Brunswick Corporation (U.S.A.), Nip-
pon Seisen (Japan), and N.V. Bekaert (Belgium).
Not all of them produce all the various quali-
ties and shapes available; some are integrated
into the endproducts. N.V. Bekaert S.A. makes
the largest variety of alloys, diameters, half-
products, and shapes, as company policy is not
to be integrated.
4.5. Melt Spun Fibers
4.5.1. Properties
The quality and the properties of fibers produced
by most melt extrusion processes are quite sat-
isfactory, and this is particularly the case for the
amorphous or microcrystalline fibers (Fig. 41).
In addition to the somewhat unusual magnetic
and physical properties of the fibers, tensile
strengths up to 7000 N/mm2 have been obtained.
However, poor dynamic behavior (e.g., fatigue)
is still a weakness. Equivalent diameters (gen-
erally the fibers are not round) range from 25 to
300 µm, depending on the process and the melt
composition.
Producing continuous lengths by melt extrac-
tion has proved to be extremely difficult, but by
notching the edge, short fibers and flakes with an
equivalent diameter of 20 – 200 µm can be pro-
duced. Filaments with circular cross sections are
not available commercially.
ca. 0.36 µm (magnification ×250)
4.6. Uses
Melt Extrusion Fibers. If amorphous or mi-
crocrystalline fibers with circular cross section
become commercially available, there would be
an opening in those areas that already are served
by amorphous ribbons: transformers and other
electromagnetic equipment, shielding, sensors,
etc. However, by far the biggest market would
be for structural uses, especially in tire reinforce-
ment, with a potential of 0.5 – 1.0×106 t/a.
Melt Extraction Fibers. Melt extraction
fibers are mainly used for the reinforcement
of concrete and, in particular, of refractory ce-
ments where they compete with conventionally
drawn carbon and stainless steel wires. The
smaller flakes also have some potential as fillers
in polymers for electromagnetic interference
applications, where they must compete with the
bundle-drawn metal fibers.
Economic Aspects. Metal fiber prices are
tending downwards as more and more high vol-
ume applications are being implemented.
5. Carbon Fibers
Carbon fibers are black fibers used as yarns,
felt, or powderlike short monofilaments with di-
ameters smaller than 10 µm. They are mainly

applied to reinforce polymers, much like glass
fibers have been used for decades in fiber glass,
with the principal difference that composites
with carbon fibers are so-called advanced com-
posites or high-performance composite materi-
als. The progress achieved with carbon fibers,
as compared with glass reinforcement fibers, is
based on the superior stiffness of carbon fibers,
combined with high strength and low density.
As can be seen from Figure 42, the density of
polyaramide fibers, introduced into the market
nearly simultaneously with carbon fibers, is even
lower, but strength and stiffness of these organic
fibers do not reach near the top values for carbon
fibers.
Figure 42. Comparison of the reinforcing fibers for ad-
vanced composites with polymer matrix: glass, polyarami-
des, and carbon
Figure 43 shows the superior specific stiff-
ness (Young’s modulus measured in fiber direc-
tion divided by the density) of various types
of carbon fibers in comparison with those of
polyaramide and glass fibers. The extremely
low values for isotropic conventional metals
and the theoretical maximum for the anisotropic
graphite single crystal are also included. This
figure demonstrates not only the superiority of
carbon fibers but also the variability of stiffness
among the various carbon fiber types. In the best
case, up to 80 % of the theoretical modules of a
graphite layer can be realized in a commercial
material.
Fibers, 5. Synthetic Inorganic 45
Figure 43. Specific tensile modulus (Young’s modulus di-
vided by density) of the various carbon fiber types (PAN and
MPP) in comparison to other structural materials
∗ = single crystal (a-direction)
Nomenclature.Terms for description of new
materials should be based on scientific char-
acterization. Trivial terms, introduced locally,
and still used, should be avoided. In UK and
France the term carbon fiber is used correctly.
In USA, however, the scientifically unjustified
term “graphite fiber” is preferred in industry,
and in German, as well as in Eastern Euro-
pean languages, the tradition of describing car-
bon materials as coal has been carried over to
the new material carbon fiber (German: Kohle-
faser). The IUPAC has therefore established a
sub-committee for description and characteriza-
tion of carbon and graphite materials in general
and of carbon fibers in specific. Some tentative
recommendations for the terminology [87], part-
ly still under discussion, are the following:
Carbon fibers are fibers (filaments, yarns,
rowings) consisting of at least 92 wt % carbon,
regularly in the nongraphitic stage, that is, with
two-dimensional long-range order of the carbon
atoms in planar hexagonal networks, but with-
out any measurable crystallographic order in the
third direction (c direction) apart from more or
less parallel stacking.
HM-type carbon fibers are carbon fibers with
a value of the modulus of elasticity in the fiber
direction larger than 30 % of the elastic number
c11 of a graphite single crystal, but with a ratio
of tensile strength to tensile modulus below 1 %.
46 Fibers, 5. Synthetic Inorganic

HT-type carbon fibers are carbon fibers with strength: stiffness ratio strictly controls the frac-
a value of the tensile strength above 3000 MPa ture elongation (strain to failure) of the fibers.
and a ratio of strength to stiffness of ca. 1.5 – In Figure 44 these relationships are demon-
2.0 %. strated graphically in three coordinates. The val-
IM-type carbon fibers belong to the HT type ues of all commercially available carbon fibers
carbon fibers because of comparable values of are grouped around two areas, which describe
tensile strength but are characterized by im- opposing combinations of properties: the high
proved stiffness (Young’s modulus up to 30 % modulus domains (HM-type fiber) and the high-
of the theoretical c11 value with remaining val- strength, high-tenacity domains (HT-type fiber).
ues of the ratio strength to modulus above 1 %).
In addition, the precursor material of the car-
bon fibers is important because the combination
of various fiber properties and the behavior of
the carbon fiber mechanically, physically, and
chemically depend strongly on the raw material.
PAN-based carbon fibers are carbon fibers
obtained from polyacrylonitrile (PAN) precur-
sor fibers by stabilization treatment, carboniza-
tion, and possibly final heat treatment at even
higher temperature.
Isotropic pitch-based carbon fibers are car-
bon fibers obtained from isotropic pitch fibers
by stabilization treatment and carbonization.
Anisotropic pitch-based carbon fibers (MPP-
based carbon fibers) are carbon fibers obtained
from mesogenic pitch after this has been trans-
formed into mesophase pitch (MPP) during the Figure 44. Characterization of the various commercial car-
process of spinning, by stabilization of the spun bon fiber types by tensile strength and Young’s modulus
mesophase pitch fibers, carbonization, and final HT = high tensile; IM = intermediate modulus; HM = high
modulus; MPP = mesophase pitch-based high-modulus
heat treatment at even higher temperature. fibers
Rayon-based carbon fibers are carbon fibers
obtained from rayon precursor fibers by chem-
ical pretreatment and carbonization. They are Nearly all carbon fibers available commer-
isotropic carbon fibers and can be transformed cially have a near-circular cross section. As
into anisotropic carbon fibers with high strength shown in Figure 45 (see next page) the cross
and stiffness by hot stretching above 2800 K. section of the carbon fibers mirrors the fiber pre-
Gas-phase grown carbon fibers are carbon cursor. Currently (1986) wetspun PAN is usually
fibers grown in a hydrocarbon atmosphere with the raw material. Gasphase-grown carbon fibers
the aid of fine particulated solid catalysts such as exhibit the best graphitic order and preferred ori-
iron or another transition metals. They consist of entation, but have no importance for reinforce-
graphitizable carbon and can be transformed into ment.
graphite fibers by heat treatment above 2800 K.

5.1.2. Chemical Composition

5.1. Properties
5.1.1. Mechanical Properties
The strength values differ dramatically among
the various carbon fiber types. Because car-
bon fibers exhibit brittle fracture behavior, the
Carbon fibers consist of elemental carbon. As all
synthetic carbon materials, they are fabricated
from carbon compounds by thermal degrada-
tion (pyrolysis), and the final (i.e., maximum)
heat-treatment temperature controls the nature
and amount of remaining ligand atoms in the
residual carbon. The HM-type fibers consist

Figure 45. Fracture surface of the precursor fibers (rayon, wetspun PAN, mesophase pitch) and of the resulting carbon fibers
(REM)
Carbon yield given in wt %
of nearly pure carbon; however, the HT-types
contain some nitrogen (<4 wt %) and oxygen
(<1 wt %). Hydrogen contents can be neglected
[88,89].
Surface-treated carbon fibers contain, in ad-
dition, solid oxygen complexes on the surface
and organic coatings (sizing), which protect the
carbon fibers during handling and transport.
5.1.3. Structure
Solid elemental carbon has two equilibrium
modifications: diamond with isotropic homopo-
lar bonds between the carbon atoms and sp3
hybridization of the electron levels and the
hexagonal graphite with sp2 hybridization, thus
anisotropic structure. Rhombohedral graphite
is a result of mechanical deformation of the
hexagonal form. The broad variety of disordered
Fibers, 5. Synthetic Inorganic 47
carbons belongs to the graphitic modification
mainly with sp2 hybridization and thus a ten-
dency to form planar polyaromatic molecules
but with only two-dimensional order and no
hexagonal ABAB-shaped arrangement of these
polyaromatics in the third dimension. In this
sense carbon fibers can be described as disor-
dered carbon with more or less ordered polyaro-
matic layers and a hypermolecular anisotropic
arrangement of these polyaromatics parallel to
the fiber axis. Perfectness, size, and arrange-
ment of these polyaromatics control the physical
and especially the mechanical properties of the
fibers. Partial sp3 hybridization was confirmed
in all artificial carbon, and especially in carbon
fibers, and is explained by the carbon atoms on
the edges and in disclinations of the polyaromat-
ics [90, 91].
48 Fibers, 5. Synthetic Inorganic

Figure 46. TEM bright field images (cut perpendicular to the fiber axis) [90, 92]
A) Typical commercial HT type (T 300) B) HM type (SIGRAFIL HM) carbon fiber

Transmission electron microscopy (TEM)

produces bright field images of thin-layer cut-
tings of HT- and HM-type carbon and reveals
the different degree of defect structure (Fig. 46)
[90,92].
The preferred orientation in the direction of
the fiber axis as indicated by the Ruland param-
eter q (0 corresponds to a statistic arrangement
and −1.0 means completely preferred orienta-
tion) is shown in Figure 47 [93]. Some exper-
imental fiber types based on mesophase pitch
reach nearly 100 % of the theoretical Young’s
modulus, i.e., that of a single crystal. Such a
Figure 47. Young’s modulus of the various types of car-
preferred orientation is combined, however, with bon fibers correlated with the degree of preferred orientation
the disadvantage of low shear resistivity. measured by X-ray diffraction and expressed by the Ruland
The layer distance, which is easily measured, parameter [93]
is a good indication of the degree of defect struc- 0 = statistical distribution, 1.0 = orientation of graphite sin-
gle crystal
ture of a carbon fiber. Figure 48 shows the values
for commercial fibers as received and the reduc-
tion of the c/2 value by further heat treatment,
which heals structural defects [94].
 Fibers, 5. Synthetic Inorganic 49

cape of the remaining ligands, which causes

Figure 48. Main interlayer distances c/2 of various com-
mercial carbon fibers and their decrease due to subsequent
heat treatment. The c/2 values of graphitizable petroleum
coke are included for comparison [92, 94, 95]
a) Ideal graphite; b) Petroleum coke; c) P55, MPP-based;
d) AS4, PAN-based; e) T300, PAN-based; f ) T800, PAN-
based
This healing process results in higher density
and increased Young’s modulus, which is shown
in Figure 49B. The strength passes a maximum
at heat-treatment temperatures around 1500 ◦ C
(Fig. 49A). The reduction of strength, with the
minimum around 1800 ◦ C, is caused by the es-
severe structural disturbance. Only minor im-
provement takes place during further annealing.
The variety of curves for PAN-based fibers is
caused by differing oxidative pretreatment dur-
ing the stabilization before carbonization [96,
137].
Figure 50 (see next page) shows the effects on
the ultrastructure of heat treatment. The ordering
process at the highest temperatures is easy to rec-
ognize for a commercial HT fiber by the enlarge-
ment of domains with parallel layer stacking
(Figs. A and B). Such improvement of strength
by annealing is not found in the case of pitch-
based fibers with superior preferred orientation
because of increased shear sensibility by anneal-
ing (Figs. 50 C and D). The much higher ten-
dency of mesophase pitch-based fiber towards
structural rearrangement can also be seen in Fig-
ure 50.
Based on electron microscopic studies, sev-
eral models for structure in carbon fibers have
been proposed. Four of these are shown in Fig-
ure 51 [97–99]. In all cases, parallel stack-
ings of polyaromatic layers are arranged in
the fiber direction in forms such as ribbons,

Figure 49. The effects of heat posttreatment of carbon fibers [96, 137]
A) Tensile strength
SAF, stabilization treatment a) 230 – 260 ◦ C/40 min; b) 230 – 270 ◦ C/40 min; c) 230 – 290 ◦ C/40 min; d) MPP
B) Young’s modulusSAF, stabilization treatment a) 230 – 260 ◦ C/40 min; b) 230 – 270 ◦ C/40 min; c) 230 – 290 ◦ C/40 min;
d) MPP
The effect of varied heat pretreatment during stabilization is included. For all PAN-based fibers, SAF precursor was used
[137]. For comparison [96], the behavior of MPP carbon fibers is included.
50 Fibers, 5. Synthetic Inorganic
Table 16. Physical properties of typical commercial PAN-based carbon fibers.

Density, g/cm3 Electrical TEP, µV/K Susceptibility χ ∗, Specific Thermal

resistance, 10−6 emu/g heat Cp, J/gK conduct. λ,
10−3 Ω m W/mK

⊥ 

PAN-based HT 1.76 – 1.81 1.5 – 2.4 −1.6 to −2.4 −0.3 −0.05 0.71 8 – 20
PAN-based HM 1.82 – 2.00 0.6 – 1.0 17.2 −6.6 −0.63 0.71 60 – 100
Anisotropic 2.00 – 2.15 0.4 – 1.1 22.6 −9 to −10.4 −0.70 0.93 110 – 375
pitch-based
Isotropic 1.60 5 – 10 −0.74 to −5.32 −0.64 to −4.26 15
pitch-based

∗ Perpendicular (⊥) or parallel () to fiber axis

plates, hoses, tubes, or combinations of the var-

ious structural elements.

5.1.4. Physical Properties

Table 16 shows the densities, electrical resistivi-

ties, thermoelectric powers (TEP), magnetic sus-
ceptibilities, heat capacities, and thermal con-
ductivities of the various carbon fiber types.
The reversible thermal expansion is highly
anisotropic (Table 17). In PAN-based HT-type
carbon fibers high anisotropy is also found, as is
the case for graphite single crystals. In contrast,
the reported anisotropy in a mesophase pitch-
based HM-type fiber is much smaller [100]. Ex-
perimental measurement for the direction per-
pendicular to the fiber axis, however, is very
difficult, and a comprehensive study is still not
available.

Table 17. Reversible thermal expansion of some commercially

available carbon fibers.
CTE α , CTE α⊥ , — α⊥ /α —
10−6 K−1 10−6 K−1

Graphite, single crystal 0.5 – 1.0 29 29 – 58

PAN-based HT 0.6 5 – 30 8.3 – 50
PAN-based HM 1.0 15 15
Anisotropic, pitch-based 0.9 – 1.4 6 4.3 – 6.7
Isotropic, pitch-based 3 3 1

Typical mechanical properties of carbon

fibers are listed as a part of Figure 44.
Figure 50. Effect of high-temperature heat treatment on the
ultra structure and especially on the enlargement of the do-
mains with parallel stacking of layers (TEM bright field of 5.1.5. Chemical Properties
cuts parallel to the fiber axis)
A) PAN-based HT fibers (AS4W) as received; B) The same The chemical resistivity of carbon and graphite
after-treatment at 2400 ◦ C; C) MPP-based fiber (P55) as re- materials against aqueous solutions and most
ceived; D) The same after treatment at 2440 ◦ C
 Fibers, 5. Synthetic Inorganic 51

Figure 51. Structural models of carbon fibers

A) Two-dimensional model proposed by Ruland [97] (ribbon model); stretched graphitized rayon-based fibers B) Three-
dimensional model proposed by Johnson [98] (plate/sheet model); PAN-based fibers C) Ring/tube model proposed by Ober-
lin [98] for all carbons resulting from thermal degradation of carbonizable solids [99] D) Hose-shaped layer stacking model
proposed for highly heat-treated carbon fibers [99]

chemicals at ambient temperatures is well-

known and has been utilized in chemical appa-
ratus. The behavior of carbon fibers is similar.
Only fluorine acids as well as strong oxidants
attack carbon fibers strongly. The tendency of
graphite to form intercalation compounds with
halogens, salts, oxidizing acids, and some other
substances is also well-known. The majority of
carbon fibers, which consist of nongraphitizable
carbon is less susceptible to intercalation attack.
Only gasphase grown fibers can be intercalated
easily.
Most destructive for carbon fibers are their
reactivity with metals, above all aluminum at
temperatures below 1000 K, and their reactivity
with solid oxides at temperatures above 1000 K.
This reactivity limits the use of uncoated carbon
fibers for reinforcement of metals and oxides.
However, if reaction and diffusion barriers are
built in – silicon carbide, titanium nitride, and
similar coatings have proved to be successful
barriers (Fig. 52) –carbon fibers can be used to
reinforce metals and oxides [101, 102]. Figure 52. SiC-coated carbon fibers in a SiO2 glass matrix
(parallel to the fibers) [101]
52 Fibers, 5. Synthetic Inorganic

Figure 53. Oxide complexes on surface-treated carbon fibers and their affinity to polar groups in epoxy resins and curing
agents [103]

Solid surface oxides form as an intermediate
stage before formation of volatile carbon oxides.
Such solid surface oxide complexes improve the
adhesion of carbon fibers in polymers and matrix
in a composite [105,106]. The thermal stabil-
ity of surface oxides is limited to temperatures
below 800 K for carboxylic groups and below
1000 K for nonacidic surface oxides. Figure 53
[103] shows the various types of surface oxides
and their affinity to polar groups in the resins,
which initiates at least chemisorptive adhesion
and perhaps even formation of extensive chem-
ical bonding between fiber and resin.
The degree of adhesion is measured by inter-
laminar shear strength (Ilss). Figure 54 shows
the effect of stepwise desorption of the surface

Figure 54. Gases (CO2 , CO) liberated during thermal desorption of surface-treated carbon fibers [89]
A) HM type, Sigrafil HM12S (Sign), 120 h oxidized in 68 % HNO3
B) HT type, showing loss of nitrogen, content of PAN-based carbon fibers above 1000 ◦ C
a) T300S; b) T300U; S: surface treated; U: untreatedIlss = interlaminar shear strength

oxides in surface-treated carbon fibers by heat
treatment in argon and the resulting effects of
such stepwise desorption on the remaining ad-
hesion in an expoxy matrix. The strong improve-
ment of adhesion in surface-treated HM fibers is
associated with the carboxylic groups. The im-
provement of adhesion in HT-type fibers by sur-
face treatment is much smaller, a maximum of
30 % of the total adhesion. The remaining ad-
hesion in HT-type fibers is strongly correlated
with the nitrogen content in the fiber. The ad-
hesion of the fibers decreases abruptly if they
are heat-treated above 1200 ◦ C, which drives off
the nitrogen remaining from the PAN precursor
[88].
The controlled formation of surface oxides
can be attained by three routes:
1) Thermal oxidation at temperatures above
700 K
2) Wet oxidation in boiling nitric acid or in a
similar oxidative liquid
3) Anodic oxidation in aqueous solution
Anodic oxidation is the only method used in-
dustrially for surface oxidation of carbon fibers.
This process is the easiest to control and forms
the greatest amount of surface oxides (Fig. 55).
Figure 55. Amount of surface oxides formed by thermal,
wet, and anodic surface treatment of HM fibers as measured
by weight loss during thermal desorption in argon [104]
a) Thermal oxidation; b) in HNO3 oxidized; c) anodic oxi-
dation
Thermal oxidation, although simple to do, is
dangerous because of the rapid reaction rate and
the danger of fiber damage due to total oxida-
tion. Figure 56 shows in an Arrhenius plot the
experimental weight loss on oxidation of car-
Fibers, 5. Synthetic Inorganic 53
bon fibers in air [104]. The HT-type fiber, which
is given a maximum heat treatment at 1450 ◦ C
during fabrication, is much more sensitive to to-
tal oxidation than the HM-type fibers, which are
heat treated above 2500 ◦ C during fabrication.
Total oxidation can be catalyzed even by alkali-
metal compounds, as is the case for carbon and
graphite materials. Heavy metal oxides are ef-
fective catalysts. Experimental results on the ef-
fect of copper oxide and vanadium oxide on the
relatively resistant HM-type fibers are included
in Figure 56. As a consequence, metallic impu-
rities in carbon fibers have to be strictly avoided
during industrial production.
Figure 56. Thermal oxidation rates of HM- and HT-type
carbon fibers (Arrhenius plots), also showing the catalytic
effects of heavy metal oxides on the oxidation of HM fibers
[104]; r = rate constant
a) HM (Sigrafil); b) HT (Toho); c) HM + CuO;
d) HM + V2 O5
5.2. History
Public disclosures of practical processes for con-
verting textile fibers into useful carbon fibers
have been confined largely to the patent liter-
ature.
5.2.1. Rayon-Based Carbon Fibers
For rayon as the precursor material, a com-
prehensive survey was published by R. Bacon
[107]. One hundred years ago, Edison [108]
described an incandescent-lamp filament made
by carbonizing a filament of natural and re-
generated cellulose material. In the 1950s Ab-
bott [109] developed a process for converting
54 Fibers, 5. Synthetic Inorganic
rayon into a fibrous carbon material for insu-
lation, filtration, and absorption. This material
was carbonized at temperatures up to 1000 ◦ C
and possessed tensile strengths of only 280 MPa.
In 1959, Union Carbide [110–112] started com-
mercial production of similar fibers, and other
companies in the U.S.A. followed.
Various fiber types heat-treated at final tem-
peratures around 1200, 1800, and even above
2500 ◦ C have become commercially available,
e.g., the carbonized VIB type and the WYB type
heat-treated at graphitization temperatures, both
from UCC (Union Carbide). The tensile strength
of such fibers ranged from 300 to 700 MPa,
while their moduli were between 30 and 50 GPa.
The most important application was the rein-
forcement of phenolic resins to form ablative
materials for rocket and missile components.
This early period of carbon fiber production
and application is reviewed by Schmidt and
Hawkins [113].
In these first years of the carbon fiber indus-
try, several companies introduced a pretreatment
process for the precursor fibers to increase car-
bon yields and speed up the processing rates,
especially in the early heat treatment stage. The
Pluton fiber of 3 M [114] was introduced in
1961. A decisive improvement of the mechani-
cal properties of the carbon fibers was the plastic
deformation of the fibers at temperatures above
2500 ◦ C. This UCC process [115] led to the com-
mercial production of the Thornell type, which
had in 1965 a tensile strength of 1260 MPa and
a modulus of 175 GPa (Thornell 25).
This was the period when the US Air Force
Material Laboratory in Dayton, Ohio, strongly
supported further development in the field of
high-modulus carbon fibers [116, 117]. The goal
was to replace boron fibers as soon as it proved
possible to fabricate high-performance carbon
fibers with a minimum modulus of 350 GPa.
In 1970, UCC reached the goal with the com-
mercially available Thornell 75 HM-type car-
bon fiber, which also had a Young’s modulus
of 500 GPa, a value corresponding to 50 % of
the theoretical c11 value. Hitco [118] entered the
market with yarns produced by similar methods.
Although these carbon fibers were satisfactory in
applications, the final decision of the market was
to restrict use to the PAN-based carbon fibers.
5.2.2. PAN-Based Carbon Fibers
The history of PAN-based carbon fibers starts in
1950 with the observation of R. C. Houtz [119]
that polyacrylonitrile fibers are converted into
black, flameproof materials by heat treatment at
300 ◦ C in air. However, products produced by
this method were not used before 1959 [120].
At that time the discovery was made that flex-
ible, flameproof textiles with the original fiber
structure can be obtained only if the exother-
mic reaction between 150 and 300 ◦ C is strictly
controlled and if the high shrinkage during this
thermal treatment is minimized by mechanical
means. In 1960 the process to produce a flame-
proof fiber, starting with Dupont’s AF Orlon
fiber and commercialized as Black Orlon, was
described in detail by Vosburgh [121]. Heat
treatment up to 15 h at 220 ◦ C was needed to
produce a well stabilized fiber with a strength
of 40 % of the value of the original PAN fiber.
The oxygen content of the final product was 10 –
20 wt %.
Surprisingly, this oxidizing-blackening treat-
ment of the PAN fibers was not complemented
by carbonization. Such a combination was first
claimed by Shindo [120] in 1959. A maxi-
mum tensile strength of 100 MPa was obtained.
In 1965, Johnson, Philip, Watts [122] added
to this combined process the need to restrain
shrinkage or alternatively to stretch the PAN
fiber during stabilization treatment. They ob-
tained carbon fibers with Young’s moduli up
to 420 GPa. This process is now the standard
process for fabrication of PAN-based carbon
fibers, although patent protection turned out to
be limited and is completely gone today. A more
detailed history is given by Watts [123], and
a complete literature survey is given by D. J.
Mueller [124].
5.2.3. Pitch-Based Carbon Fibers
The processes for carbon fibers from pitches are
based on the observation of Otani [125] in 1963
that some pitches are spinnable to pitch fibers,
which can be converted into carbon fibers. Pitch
from poly(vinyl chloride) turned out to be the
best, but a variety of other pitches can be used af-
ter specified pretreatment as summarized by Su-
gio Otani and Asao Oya [126]. The mechani-
cal properties of the resulting carbon fibers were

in the range 800 – 1800 MPa for tensile strength
and 20 – 50 GPa for Young’s modulus. These
pitch-based carbon fibers were commercialized
in 1970 by Kureha Chemical Industries and are
still the only ones commercially available in the
form of continuous yarns [127]. In 1985, Free-
man [128] reported that the Ashland Company
started production of a variety of carbon fiber
types using petroleum pitch as the raw material.
Coal tar pitch is recommended as an alternative
precursor material by Rommey [129]. All these
pitch-based carbon fibers belong to the isotropic,
low-modulus type.
Fabrication of anisotropic pitch-based carbon
fibers with greatly improved mechanical prop-
erties was achieved by Lewis and Singer et
al. [130] in 1976. This process is based on the
mesophase, a liquid crystal formed within the
pitch before spinning that enables the forma-
tion of a preferred orientation in the pitch fiber,
thus also in the final carbon fiber. Values of ten-
sile strength up to 3000 MPa and an extremely
high Young’s modulus, up to 600 GPa, can be
achieved.
Laboratory developments have continued in
many countries of the Western World. Commer-
cial products are available from AMOCO (pre-
viously Union Carbide) [131].
5.2.4. Vapor-Grown Carbon Fibers
Whiskers and needles of carbon deposited from
the gas phase were described in 1966 by Fitzer
and Schlesinger [132]. Fitzer and Rhee
[133] continued these studies and reported in
1970 growth rates and mechanical properties of
the whiskers and needles. Tensile strengths up
to 1500 MPa and stiffness values up to 300 GPa
were reported. The decisive progress was re-
Table 18. Restrictions on the precursor fiber for carbon fiber fabrication
Preconditions for carbon fiber manufacture Parameters
Precursor
Preservation of the fiber form of the thermosetting or thermoplast fibers
polymer
High carbon yield selection of the polymers
Preferred orientation of the polycarbon highly cross-linked, unmeltable
layers with highest degree of defects in the polymers
ultra structure
Fibers, 5. Synthetic Inorganic 55
ported by Endo [134] in 1974 when he rec-
ognized that the whisker growth follows a cat-
alytic mechanism starting from fine, dispersed
solid metal particles in the substrate. Whisker
lengths up to 250 mm and tensile strength val-
ues up to 26 000 MPa have been reported so far
by Endo [134]. Commercialization has not been
announced, however, probably because of the
completely different structure of vapor-grown
carbon fibers. These consist of soft carbon,
that is, the graphitizable less-disordered variety.
Therefore, high mechanical properties cannot be
expected. They are most suitable for intercala-
tion (high electrical conductivity).
5.3. Production
All commercial production processes for car-
bon fibers are based on carbonization of poly-
mer fiber precursors. The preconditions for use-
ful fiber formation compiled in Table 18 must
be fulfilled. Thus, the number of suitable poly-
mer fibers is limited. The fiber form is always
preserved during carbonization in the case of
nonmeltable polymers, such as thermosetting
resins, which can be thermally degradated di-
rectly. Thermoplastic precursor fibers have to be
converted into a nonmeltable form before car-
bonization. This is called stabilization treatment
and occurs on conversion of the linear polymers
into ladder polymers and on formation of addi-
tional cross-linkages. In the case of pitch precur-
sor fibers, the melting of the mixture of polyaro-
matic compounds must be hindered by oxida-
tion. Three precursor materials are satisfactory:
(1) rayon fibers, (2) PAN fibers, and (3) pitch
fibers. Rayon fibers do not require stabilization,
only preimpregnation is necessary to improve
the carbon yield. PAN and pitch fibers need sta-
Process
direct thermal degradation or stabilization treatment of
meltable precursor before thermal degradation
optimization of the stabilization and carbonization
parameters
stretching of the precursor before or during stabilization
and/or hot-stretching of the carbon fiber
56 Fibers, 5. Synthetic Inorganic
bilizing pretreatment. The carbon yield is lowest
for rayon fibers and highest for stabilized pitch
fibers:
Rayon fibers 25 %
Oxidized pitch fibers 80 %
Polyacrylonitrile 50 %
Oxidized pitch up to 80 %
Phenolic and furan resins up to 85 %
The most important parameter in producing
high-quality carbon fibers is the preferred orien-
tation of the polyaromatic layers parallel to the
carbon fiber axis. Only then can the high strength
of the bonds in the molecular structure of the
graphite lattice (c11 = 1060 GPa, c33 = 33 GPa) be
transferred into stiffness and strength of the car-
bon fiber. The low shear modulus in the graphite
crystal structure (c44 = 4.6 GPa) must be avoided
by having the highest degree of defects in the
ultrastructure of the carbon fiber. In the case of
rayon precursor fibers the needed high degree of
structural imperfectness can be realized, how-
ever the preferred orientation is realizable only
by hot working of the carbon fiber at 3000 K.
Figure 57. The chemistry of stabilization and carbonization: fabrication from homopolymer PAN and stabilized PAN [137,
138, 147]
This hot stretching process is difficult to carry
out on an industrial scale. Therefore, carbon
fibers are no longer produced from rayon.
In PAN- and pitch-based anisotropic carbon
fibers, the preferred orientation is introduced
into the precursor before carbonization. Poly-
acrylonitrile is the most important precursor ma-
terial. More than 90 % of all commercially pro-
duced carbon fibers that are used for reinforce-
ment of polymer material are based on this poly-
mer. Only a minor part is produced from pitch.
5.3.1. PAN Precursor Fiber
Textile PAN fibers usually consist of ternary
mixtures of copolymers with 89 – 95 wt % acry-
lonitrile (AN), 4 – 10 wt % of a non-ionogenic
comonomer plasticizer, and 1 wt % of additive
to chemically bond dyes [135]. For the pre-
cursors for carbon fibers, a lower content of
comonomers, usually below 3 wt %, is used.
 Fibers, 5. Synthetic Inorganic 57

Polymer fibers can be spun by melt, wet, and jugated double bonds. The resulting aromatic
dry spinning processes. In the case of polyacry- polymer has ladder units of about five, corre-
lonitrile as a carbon fiber precursor, the wet spin- sponding to the average number of isotactic se-
ning process is the most important. In 1986 a quences in the nitrile polymer. Cyclization can
new melt spinning process was presented which occur with or without simultaneous dehydrogen-
would offer economic advantages [136]. ation as shown in Figure 58 [137]. In both cases
The solvents dimethylformamide (DMF) and strong exothermic reactions are observed. In in-
dimethylacetamide (DMA) are usually preferred dustrial processes, cyclization is always carried
for wet spinning today. For carbon fiber precur- out in an oxidizing atmosphere because of the
sors, however, the inorganic solvents nitric acid, much higher carbon yield during the subsequent
stannum chloride and thiocyanate are also used. carbonization [139].
In any case, copolymer composition of the pre-
cursor, the spinning technology, and especially
the solvent are almost always strict commercial
secrets. Only Courtaulds (United Kingdom) de-
scribes the use of dimethyl sulfoxide for fiber
precursor fabrication.
The main difference between textile PAN
fibers and PAN precursor fibers is the different
degree of stretching: the stretching ratio for tex-
tile fibers is 1 : 2 to 1 : 10, that for PAN precur-
sor fiber 1 : 12 or more. There is the danger of
overstretching, which introduces defects into the
polymer fiber that initiate defects in the carbon
fibers, thus causing low strength.
The extremely high-strength and high-
tenacity carbon fibers are mainly produced from
precursors more carefully spun and stretched to
a higher degree. Dry-jet wet spinning would be
a good method for this.

5.3.2. Stabilization

Polyacrylonitrile is a thermoplastic polymer that

decomposes above 370 ◦ C. A suitable stabiliza-
tion process is needed before carbonization to
obtain an infusible intermediate that preserves
the orginal fiber structure. This intermediate is
known as inflammable textile (e.g., PAN-Ox).
The stabilization step is the most important
one from the viewpoint of carbon fiber qual-
ity and process economy. The reaction is dif-
fusion controlled, and a residence time of ca.
1 h at temperatures between 200 and 280 ◦ C
is needed. Fiber shrinkage during stabilization
must be avoided (see Section5.2.2).
The chemistry of the stabilization process is
illustrated schematically in Figure 57 (see pre-
vious page). The thermoplastic chain is cyclized
to a naphthyridine ring system and simultane-
ously dehydrated in the backbone to form con-
Figure 58. The exothermic stabilization in nitrogen (cy-
clization) and in air (cyclization and dehydrogenation) as
revealed by DSC (differential scanning calorimetry) [137]
There is a direct correlation between the
amount of oxygen absorbed and the density of
the stabilized fiber. The best mechanical prop-
erties result for a precursor density of 1.4 –
1.42 g/cm3 , which corresponds to oxygen con-
tents of 10 – 14 wt % (Fig. 59).
The stabilization causes the color to change
from the white of polyacrylonitrile to yellow
58 Fibers, 5. Synthetic Inorganic
and brown and finally to the black of the in-
termediate, the oxidized, inflammable polyacry-
lonitrile. The stabilization chemistry has been
treated comprehensively in numerous papers,
e.g., [137, 138, 140–147].
Figure 59. The effect of oxygen uptake during stabilization
on the strength of the final carbon fiber
a) 230 ◦ C isotherm/air; b) 230 ◦ C isotherm, 1st stage N2 ,
2nd stage air; c) 230 → 270 ◦ C nonisotherm/air
5.3.3. Process Parameters
The carbonization step can be carried out much
faster than the stabilization step; however, the
evolution of the volatile byproducts, mostly
H2 O, HCN, CO2 , and N2 , must be taken into
account [148].
Practical temperature – time profiles have
been published [137, 142, 149–151]. The qual-
Figure 60. Hitco process for production of carbon cloth from PAN cloth showing temperature and tension profiles [152]
ity of the final carbon fiber product is controlled
by the precursor fiber, the stabilization treatment
and the oxygen content, and finally by the car-
bonization schedule.
An example of a continuous stabilization and
carbonization for fabrication of carbon cloth
from PAN from patent literature is shown in Fig-
ure 60 [152].
5.4. The Pitch Fiber Process
5.4.1. Isotropic and Anisotropic Pitch Fiber
Precursors
Carbon fibers made from pitch precursor are
available with isotropic or anisotropic structure,
the former made from isotropic pitch, the later
from mesophase pitch. The isotropic pitch-based
carbon fibers have low mechanical properties.
They have been produced industrially from coal
tar pitch by Kureha (Japan), for more than 15
years. The petroleum pitch producer Ashland
(United States) manufactures isotropic carbon
fibers, which are used mostly for thermal and
mechanical insulation but also as fillers.
Anisotropic carbon fibers, which can be used
for reinforcement because of their excellent
mechanical properties, are produced in a process
that takes advantage of the larger planar polyaro-
matic molecules in pitch and their tendency to
form liquid crystals with parallel polyaromatic
layers. The mesophase pitch is melt spun and
stretched in the plastic deformable stage to pro-
duce pitch fibers with preferred orientation.

Figure 61. The coalescence of discotic liquid crystals in pitch
All pitch fibers must be stabilized before car-
bonization, but pitch fibers with preferred orien-
tation consisting to a large extent of mesophase
not only hinder melting during carbonization but
also cross-link the polyaromatics upon oxidation
thus hindering formation of graphitized carbon
fibers.
5.4.2. The Mesophase Formation
Pitch is the residue from distillation of coal tar
or crude oil or from cracking hydrocarbons. All
pitches consist of thousands of aromatic hydro-
carbons along with more or less cyclic aliphatics
and aliphatic side chains. The molecular mass
of the polyaromatic compounds ranges between
200 and 800. All the compounds in a single pitch
have never been identified. Pitches are char-
acterized by fractionation: their solubility in a
sequence of standard solvents (Table 19). The
model molecules shown in the table are sug-
gested by comprehensive analytical investiga-
tions [153]. The melting points of the various
pure compounds cover a range of hundreds of
degrees. Pitches, however, form multi-eutectic
systems with softening temperatures between 50
and 300 ◦ C, much lower than the melting tem-
peratures of the pure aromatic components.
The mesophase is formed during thermal
degradation in the temperature range 400 –
450 ◦ C and is controlled by the hydrogen con-
tent remaining in the aromatics during this heat
treatment. Aromatics with an average molecular
mass around 800 (coal tar pitches) and aromat-
Fibers, 5. Synthetic Inorganic 59
ics with aliphatic side chains and a molecular
mass around 1000 (petroleum pitches) are able
to form nuclei for discotic liquid crystals [153].
The liquid crystals in pitch are spherical with
the molecular orientation shown in Figure 61.
The coalescence of these liquid crystals with
disklike molecules is similar to that of the liquid
crystals with rodlike molecules.
From the viewpoint of the phenomenology,
the formation of liquid crystals is similar to that
of crystallization, starting from observable nu-
clei with diameters larger than 1 µm. The growth
and coalescence proceed until all isotropic pitch
matrix is consumed. Insoluble impurities in the
molten pitch matrix (primarily quinoline insol-
ubles, i.e., the pitch residue) inhibit coalescence
and therefore also the formation of large vol-
umes of anisotropy.
5.4.3. Process Steps
The viscosity of pitch, a precondition for sat-
isfactory melt-spinning behavior, is determined
by pitch composition. Figure 62 shows the ap-
parent viscosities of a typical petroleum pitch
and a typical coal tar pitch as functions of heat
treatment [154]. The behavior of toluene-soluble
(TS) and toluene-insoluble (TI) fractions of the
coal tar pitch and of two mixtures of TS and TI
fractions are included. Softening of the pitches
upon heating is indicated by viscosity decrease;
solidification due to mesophase formation, by
viscosity increase. Spinning is difficult for
Table 19. Characterization of typical coal tar and petroleum pitches by solvent fraction and model molecular structures for the fractions [153]

60
Fibers, 5. Synthetic Inorganic
 Fibers, 5. Synthetic Inorganic 61

high mesophase contents (which produce the

HM fibers), however carbon yield increases with
viscosity (Fig. 63) [154]. These contradicting
conditions cause technological problems.

Figure 62. Apparent viscosity (rotation viscosimeter) of the

petroleum pitch A240, the purified coal tar pitch SP80, frac-
tions resulting from toluene extraction of SP80 (TS = toluene
soluble, TI = toluene insoluble), as well as two mixtures of
TS and TI [154]

Figure 63. Carbonization residues from the pitch samples

described in Figure 62 [154]
a) Toluene soluble; b) SP80 and 4 : 1 mixture; c) 1 : 1 mix-
ture; d) 1 : 4 mixture; e) 1 : 9 mixture; f ) toluene insoluble

Stabilization is carried out between 200 and

300 ◦ C [155], much like in the PAN process.
The stabilization times are dramatically shorter,
however, between 10 and 30 min. The patent lit-
erature indicates that commercial pitch-based
carbon fibers are given a final heat treatment
Table 19. Continued.

around 1800 ◦ C (P55 type). Fiber types with a

relatively high Young’s modulus, in the area of
400 GPa, can be achieved (cf. Fig. 44). A flow
sheet for the entire fabrication process is given
in Figure 64 [156].
62 Fibers, 5. Synthetic Inorganic
Figure 64. Flow sheet for manufacture of carbon fibers from
pitch (Union Carbide Corp. (UCC) process) [156]
5.5. Comparison of the Carbon
FibersBased on PAN and on Pitch
Figure 65 shows the stress – strain behavior
of PAN-based carbon fibers, as commercially
available, of the intermediate modulus fibers in
development, and of isotropic and anisotropic
pitch-based fibers.
Figure 65. Stress – strain behavior of commercially avail-
able PAN- and pitch-based carbon fibers
The pitch process for carbon fibers poten-
tially offers the capacity to produce carbon fibers
covering the whole mechanical property range
of the carbon fiber family. However, these ex-
pectations for pitch-based fibers have not yet
been realized in spite of worldwide, intensive
research.
The overwhelming part of commercial car-
bon fibers is the PAN-based HT type. High
strain-to-failure of the fibers is seen as one of the
key properties for the fracture toughness needed
for broader application of advanced compos-
ites. The best strain values, between 1.5 and
2 %, are possible on industrial scale only with
PAN-based HT-type fibers. Further, manufac-
turing problems are reduced if fabrication starts
from a fully synthetic precursor material, with
strictly controlled quality, instead of from a
semi-natural raw material, such as petroleum or
coal tar pitches, with strongly varying compo-
sitions depending on crude or coal type and the
thermal history.
The applications of ultrahigh modulus carbon
fibers from anisotropic pitch will most likely re-
main restricted to space structures and to carbon
fiber-reinforced carbon fiber material (carbon –
carbon composites). The challenge to increase
the applications of anisotropic pitchbased car-
bon fibers can be satisfied only if production of
medium-quality fiber types can be coupled with
further drastic price reductions.
 Fibers, 5. Synthetic Inorganic 63

5.6. Commercial Availability and Sumitomo, another Japanese PAN-precursor

Economic Aspects
5.6.1. Companies and Trade Names
In the 1960s, the industrial production of car-
bon fibers started in UK, Germany, Japan, and
USA with a total volume below 1000 t/a. Very
soon, however, Japan had the advantage because
of the availability of PAN-fiber plants using the
wet process, whereas in USA and Germany PAN
was produced mostly by the older dry process.
Two Japanese companies specializing in the pro-
duction of synthetic fibers, namely Toray and
Toho Beslon, have succeeded in greatly expand-
ing carbon fiber production. Today, these two
companies dominate the world’s carbon fiber
production by licenses and joint ventures.
Hercules (USA) is independent from the two
Japanese companies, but works jointly with
producer. Courtaulds fabricates precursors and
is thus independent from Japanese imports. It is
also the only company that sells PAN-precursor
fibers suitable for a carbon fiber manufacture the
world over without the limitations of licenses
or joint ventures. In Germany, Sigri is inde-
pendent, but fabricates no precursors. The pro-
ducers, their production, their trade names, and
some commercially available carbon fiber types
from PAN precursors are compiled in Table 20.
5.6.2. World Capacity
The world capacity to produce carbon fibers is
estimated to exceed 5000 t/a, and is increasing
rapidly. The development of world production
capacity during the last 15 years is shown in Fig-
ure 66. The small quantities of anisotropic and

Table 20. Carbon fiber producing companies, trade names, and commercial availability in the world market.

Producer (country) Production Trade name Types Licensed by

capacity, t/a state
1988

HT IM HM

Toho (Japan) 1380 Besfight HTA IM 400 HM 35

ST 3 IM 500 HM 40
HM 45
Toray (Japan) 1500 Torayca T 300 T 800 M 30 Asahi Nippon
T 700 T 1000 M 40 (Japan) 300
M 50
Amoco (USA) 360 Torayca T 300 T 800 M 30 Toray
T 700 T 1000 M 40
M 50
Hercules (USA) 1360 Magnamite As 4 IM-6 HMS Toray
As 6 IM-7 HMU
IM-8
BASF (USA) 450 Celion G 30 – 400 G 40 – 600 G 50 – 300 Toho
G 30 – 500 G 40 – 700 Gy 70
G 30 – 600 Gy 80
Hysol (UK + USA) 700 Grafil XA-S IM-S HM-S
Sigri (FRG) 100 Sigrafil C − +
(PANOX 400)
AKZO (FRG + USA) 860 Tenax HTH Toho
ST 3
Soficar (F) 300 Filkar T 300 T 800 M 30 Toray
T 400 M 40
M 50
Amoco (USA) 230 ∗ Thornel P 55 P 75
P 100
P 125
Mitsubishi Chem. Ind. 500 ∗ Dialead
Ashland (USA) 140 ∗ Carboflex isotropic types
Kureha (Japan) 900 ∗ isotropic types

∗ pitch-based carbon fiber types

64 Fibers, 5. Synthetic Inorganic
isotropic pitch-based carbon fibers are included.
There is disequilibrium between capacity, pro-
duction, and real consumption, but more exact
information is not available. It is estimated that
capacity will grow up to 10 000 t/a and increase
between 30 and 50 % per year.
Figure 66. World production capacity for carbon fibers
Average prices of carbon fibers are shown
in Figure 67. The price for PAN-based carbon
fibers strongly depends on the number of fila-
ments in one tow. The so-called low-filament
tows, with 3000 monofilaments, have been used
for years. Currently (mid-1980s) most prepregs
use tows with 6000 monofilaments. The ten-
dency is to try to increase the filament number
up to 12 000.
Figure 67. The development of carbon fiber prices
PAN-multifilament tows are the cheapest
form of PAN-based carbon fibers: continuous
320 000 monofilament tows as used in tex-
tile application are difficult to handle in high-
performance composites. They are mainly used
for cutting into chopped fibers. However, mul-
tifilament tows with only 40 000 monofilaments
have been introduced. The technological gap
between the low filament tows (12 K and in-
creasing) and the multifilament tows (40 K and
decreasing) is diminishing. This should result in
lower prices.
The price situation for anisotropic pitch-
based carbon fibers is complicated. Around 1981
Union Carbide started to produce high-quality
carbon fibers and sell them at $ 20 per kg, much
below the price of PAN-based fibers at that time.
As shown in Figure 67, the market situation and
production difficulties forced the producer in
1983 to more than double the price. Anisotropic
pitch-based high-modulus fibers are now much
more expensive than PAN-based HT fibers.
The price situation is stable for isotropic
pitch-based carbon fibers as delivered by Kureha
for decades. Ashland, the Major petroleum pitch
producer, has entered the market.
5.7. Future Importance
As has become evident over the last decade,
polymers reinforced with carbon fibers present
the opportunity not only to replace but also to
displace structural metals [157]. The advanced
composites (carbon fiber-reinforced plastics,
CFRPs) are superior to conventional metals in
respect to strength, stiffness, lightness, ther-
mal expansion, fatigue, and corrosion resistance
(Fig. 68).
Comparable Costs. The comparable costs
of strength and stiffness of a unit volume of
an advanced composite reinforced with PAN-
based HT or HM carbon fibers, polyaramide
fibers, glass fibers, former boron, or silicon car-
bide fibers have been estimated and compiled in
Table 21.
From the viewpoint of strength, the glass
fiber reinforcement is cheapest, followed by pol-
yaramide and PAN-based HT fibers. PAN-based
HM fibers are much more expensive. The CVD
boron and silicon carbide fibers cannot be con-
sidered on an economical basis. From the view-
point of stiffness, the carbon fiber-reinforced
composites, although attractive on a price basis,
are limited to the low modulus range.
 Fibers, 5. Synthetic Inorganic 65

Figure 68. The technical advantages of CFRPs over conventional metals

Table 21. The price related to specific tensile strength and specific
stiffness.
carbon-reinforced resins instead of light metals
like aluminum. Commercial aircrafts are follow-
Fiber type Price, Density, Price divided by
DM/kg 103 kg/m3 specific
ing step by step to replace aluminum by carbon
fiber-reinforced composites.
Tensile Young’s The cost-benefit schedule of typical compos-
strength, modulus,
DM/ DM/ ite applications is:
104 kgm 104 kgm
space market 5 000 ¤/kg
E glass 10 2.50 1.5 0.07 air craft 500 ¤/kg
Kevlar 40 1.45 2 0.05 racing car 50 ¤/kg
CPAN (HT) 70 1.75 3 0.05 automotives up to 5 ¤/kg
CMPP 200 1.99 17.5 0.19
CPHN (HM) 450 1.85 27 0.30 The present calculation gives load savings of
B/c 1000 2.50 67 0.50
SiC/c 4000 3.50 400 3 15 – 30 % in aircraft and component costs of 77 –
150 ¤/kg. In automotives, load savings of 20 –
50 % but component costs up to 2.6 ¤/kg. Al-
Technological Fields. The aerospace field though the cost-benefit ratio is lowest in automo-
can expect the highest technical and economic tive applications, this field is the most promis-
benefit from application of these new materials. ing for the future. To produce large cars with
In the past such risky innovations were the only extremely light weight using the design rules of
chance to realize some ambitious projects. The the aerospace industry presents a real challenge
U.S. space shuttle is the best example. Although to the auto industry. The question becomes more
NASA designed this shuttle in the early 1970s and more a question of economy of scale for the
essentially in conventional materials, it was not fiber production on the one hand, and of rational-
realizable without carbon fiber-reinforced doors ized processing for final products on the other.
for the loading room. In this way, carbon fiber- Plans for future production lines have already
reinforced epoxies were tested in practice and been published [158].
accepted as performing structural materials for Another argument supporting a shift from
primary structures in manned flying vehicles. metals to carbon fiber-reinforced composites
Today, military aircraft contain up to 50 wt % is the energy requirements shown in Table 22
66 Fibers, 5. Synthetic Inorganic
[159]. The numbers in the table include both the
energy needed to process the materials and to
process the raw material. The composites ap-
pear advantageous in terms of volumes and even
more so in terms of mechanical properties per
unit volume.
Table 22. Energy requirement in kWh for the production of Al, Fe,
CFRP’s and PVC [159].
Material Density, Energy requirement for
g
cm3
production
kWh kWh
kg L L
Al 2.8 37.3 104.5
Fe 7.9 6.8 53.6
CFRP (UD) 1.5 3.7 5.6
CFRP (2D) 1.5 3.7 5.6
PVC 1.4 7.9 11.1
6. Refractory Ceramic Fibers
High-purity aluminosilicate wool fibers, essen-
tially free from group 1 and group 2 oxides, are
melt-spun from calcined kaolin or pure alumina
and silica mixtures in a fashion similar to that
for glass wool, rock wool, and mineral wool
fibers (see Chap. 2). Five compositions are in
common use: 50/50 wt % alumina/silica (maxi-
mum use temperature of 1250 ◦ C), 60/40 wt %
alumina/silica (maximum use temperature of
1400 ◦ C), and chromia-, boria-, or zirconia-
modified 50/50 wt % alumina/silica (also for use
at 1400 ◦ C). The physical properties of the fibers
are mainly determined by their diameters. most
products have a weight mean diameter in the
range of 2 – 3 µm.
Further details concerning production tech-
nology and composition can be found in [160].
The principle use of such fibers is for hot-face in-
sulation of pottery kilns and metal-working fur-
naces. Production in Europe is of the order of
16 000 t/a. Correspondingly large amounts are
manufactured in the USA and in Japan. World-
wide, the major manufacturers are the Morganite
Group (USA, Japan, UK, FRG), the Carborun-
dum Group (USA and FRG), and the Société
Européenne des Produits Réfractaires (SEPR)
in France [161].
The fibers are manufactured as a low density
blanket (64 –96 kg/m3 ), which is flexible and
can be used as such. Alternatively, rigid boards
and shapes can be created by vacuum-forming
wet slurries of raw fibers and a refractory binder
(usually a silica sol). Products can be designed
to be dimensionally-stable and to have excellent
low thermal conductivity up to 1500 ◦ C when
combined with high-alumina products such as
ICI “Saffil” alumina (see Section 3.4.1) Rigid
boards lend themselves to modular furnace de-
signs with rapid heating and cooling cycles or
where high gas velocities would lead to erosion
of blankets. In all cases, the use of low-density
kWh N
fiber aggregates leads to significant energy sav-
/ mm 2
ings in batch firing regimes compared to furnace
232.2 designs employing dense refractories.
26.8
3.5
6.9
1110.0 7. References
1. ASTM D3878.
2. Tonen, 1988, Data sheet.
3. A. A. Griffith, Philos. Trans. R. Soc. London
221 A (1920) 163.
4. F. Pott Staub.-Reinhalt. Luft 38 (1978) 181.
5. ASTM D3379.
6. H. Böder, D. Gölden, G. Popp, J. Vogelsang,
SAMPE (1983), paper 12.
7. ASTM D4018-81.
8. S. W. Tsai, NASA Tech. Rep. CR-224 (1965).
9. E. Fitzer, W. Hüttner, Bericht über das 93.
Wehrtechn. Symp., Verbundwerkstoffe in der
Wehrtechn. part I, Bundesakademie für
Wehrverwaltung und Wehrtechnik, p 93.04.01.
22 – 24 May 1978, Mannheim.
10. E. Fitzer, M. Heym, Chemistry and Industry,
21 August 1976, 663 – 676.
11. A. K. Dhingra, Philos. Trans. R. Soc. London
A 294 (1980) 411 – 417.
12. R. W. Lang, H. Stutz, M. Heym, D. Nissen,
Lecture held on the symposium organized by
the Fachgruppe “Makromolekulare Chemie”
der Gesellschaft Deutscher Chemiker on
“Neue Polymere – Spezielle Eigenschaften –
Moderne Technologien,” Bad Nauheim
14. – 15. April 1986.
13. R. Weiss, Ph. D. Dissertation University of
Karlsruhe 1984.
14. J. B. Donnet, P. Ehrburger, Carbon 15 (1977)
143.
15. E. Fitzer: Fibre Reinforcement for Composite
Materials, A. R. Bunsell, Amsterdam.
16. E. Fitzer, J. Schlichting, 18. Int.
Chemiefasertagung Dornbirn (1979), in:
Anorganische Fasern – Herstellung,
Eigenschaften, Verwendung – 2.
Werkstofftechnik, 11 (1980) 330 – 341.

17. G. Jacobsen, “Influence of the Interface on
Transverse Strength, Creep Behaviour and
Impact Behaviour of Boron- and Silicon
Carbide-Fibre Reinforced Aluminum,”
ICCM-IV, Tokyo, 25. – 28. 10. 1982, Prog. Sci.
Eng. Compos., 1323 – 1330.
18. E. Fitzer, Carbon 25 (1987) 163 – 190.
19. E. Fitzer, Pure Appl. Chem. 60 (1988) no. 3,
287 – 302.
20. E. Fitzer, P. Schubert, Rev. Int. Hautes Temp.
Refract. 16 (1979) 147 – 155.
General References
21. Tagungshandbücher SPI, Conference reports,
(Society of the Plastic Industry) (SPI), New
York.
22. Tagungshandbücher AVK (conference reports),
Arbeitsgemeinschaft Verstärkte Kunststoffe,
Frankfurt.
23. K. A. F. Schmidt: Textilglas für die
Kunststoffverstärkung, Zechner und Hüthig,
Speyer 1972.
24. K. L. Loewenstein: The Manufacturing
Technology of Continuous Glass Fibers,
Elsevier, Amsterdam-Oxford-New York 1983.
25. H. S. Katz, J. V. Milewski: Handbook of Filers
and Reinforcements for Plastics, van Nostrand
Reinhold, New York 1978.
26. W. Bobeth, W. Böhme, J. Techel:
Anorganische Textilfaserstoffe, VEB Verlag
Technik, Berlin 1955.
27. J. G. Mohr, W. P. Rowe: Fiber Glass, Van
Nostrand Reinhold Co., New York 1978.
Specific References
28. R. Kleinholz, J. Molinier:
“Verstärkungswerkstoffe nach Maß,”
AVK-Tagung, 1985.
29. R. Kleinholz: “Neue Erkenntnisse bei
Textilglasfasern zum Verstärken von
Kunststoffen,” Internationale
Chemiefasertagung 22. Dornbirn, Austria,
1983.
30. J. Eule: “Glasstapelfaser-Vorgarne für die
Herstellung textiler Erzeugnisse,” Textilbetrieb
Würzburg, 100 (1982).
31. J. Bijen: “Durability of Some Glass Fiber
Reinforced Cement Composites,” J. Am.
Concr. Inst. 80 (1983) no. 7, 8.
32. Vetrotex report 2/86, Gevetex, Aachen 1986.
33. Gevetex Textilglas GmbH, Aachen 1986.
34. C. Z. Carroll-Porczinski: Inorganic Fibers,
National Trade Press, London 1958.
35. P. A. Koch, Glastech. Ber. 23 (1950)
204 – 212.
Fibers, 5. Synthetic Inorganic 67
36. G. von Pazciczky, DE 272 441, 1912. F.
Rosengarth, DE 539 738, 1930.
37. Owens Illinois Glass, US 2 133 236, 1933 (G.
Slayter, J. H. Thomas).
38. E. Poeschel, R. König, R. Haag, U. Roth et al.:
Umweltrelevanz künstlicher Fasern als
Substitut für Asbest, Erich Schmidt Verlag,
Berlin 1978.
39. W. Trier: Glasschmelzöfen, Springer Verlag,
Berlin-Heidelberg-New York-Tokyo 1984.
40. W. Giegerich, W. Trier: Glasmaschinen,
Springer Verlag, Berlin-Göttingen-Heidelberg
1964.
41. J. A. Škol’nikov, B. M. Polik, E. P. Kočarov,
E. R. Nigin: Steklannoe štapel’noe volokno,
Isd. Khimija, Moscova 1969.
42. F. V. Tooley: The Handbook of Glass
Manufacture, 3rd ed., vol. 2, Ashlee
Publishing, New York 1984.
43. ISO/DIS 8142, ISO/DIS 8144, ISO/DIS 8145.
44. J. Thebault, R. Pailler, G. Bontemps-Moley,
M. Bourdeau et al., J. Less Common Met. 47
(1976) 221 – 223.
45. C.-H. Andersson, R. Warren, Composites 15
(1984) 16 – 24.
46. C.-H. Andersson, R. Warren, Composites 15
(1984) 101 – 111.
47. J. G. Lee, I. B. Cutler, Am. Ceram. Soc. Bull.
54 (1975) 195 – 198.
48. P. Bracke, H. Schurmans, J. Verhoest:
Inorganic Fibres and Composite Materials,
1st ed., Pergamon Press, Oxford 1984.
49. B. von Falkai (ed.): Synthesefasern,
Section 6.2.4, Verlag Chemie, Weinheim 1981,
pp. 268 – 274.
50. W. C. Miller, Kirk-Othmer, 20 (1982) 65 – 77.
51. G. Winter, Angew. Chem. (1972); Angew.
Chem. Int. Ed. Engl. 11 (1972) 751 – 760.
52. J.-O. Carlsson, J. Mater. Sci. 14 (1979)
255 – 264.
53. V. Krukonis in H. S. Katz, J. V. Milewski
(eds.): Handbook of Fillers and
Reinforcements for Plastics, Chap. 28, Van
Nostrand, New York 1978, pp. 545 – 561.
54. A. M. Tsirlin in W. Watt, B. V. Perov (eds.):
Handbook of Composites, “Strong Fibres,”
vol 1, Chap. 5, Elsevier, Amsterdam 1985,
156 – 194.
55. K. C. Hou, US 3 677 799, 1972.
56. J. Schlichting, Ber. Dtsch. Keram. Ges. 57
(1980) 289 – 298.
57. T. F. Foltz, Ceram. Eng. Sci. Proc. 6 (1985)
1206 – 1220.
68 Fibers, 5. Synthetic Inorganic
58. S. Yajima in W. Watt, B. V. Perov (eds.):
Handbook of Composites, “Strong Fibres,”
vol 1, Chap. 6, Elsevier, Amsterdam 1985,
pp. 201 – 237.
59. Tokai Carbon Co., GB 2 129 025, 1983
(Akiyama, Yamada, Takashata).
60. J. V. Milewski, F. D. Gac, J. J. Petrovic, S. R.
Skaggs, J. Mater. Sci. 20 (1985) 1160 – 1166.
61. J. J. Petrovic, J. V. Milewski, D. L. Rohr, F. D.
Gac, J. Mater. Sci. 20 (1985) 1167 – 1177.
62. Du Pont, US 3 808 015, 1968; GB 1 264 973,
1969 (L. E. Seufert).
63. A. K. Dinghra, Philos. Trans. R. Soc. London.
Ser. A 294 (1980) 411 – 417.
64. Du Pont, GB 1 417 791, 1973 (J. R. Green,
R. F. Tietz).
65. 3 M, US 4 047 965, 1976 (K. A. Karst, H. G.
Sowman).
66. Y. Abe, S. Horikiwi, K. Fujimura, E. Ichiki in
T. Hayashi, K. Kawata, S. Umekawa (eds.):
Prog. Sci. Eng. Compos., ICCM-IV, Tokyo
1982, pp. 1427 – 1434.
67. Sumitomo, GB 1 457 801, 1973; US 4 495 121,
1982.
68. J. D. Birchall, J. A. A. Bradbury, J. Dinwoodie
in W. Watt, B. V. Perov (eds.): Handbook of
Composites, “Strong Fibres,” vol 1, Chap. 4,
Elsevier, Amsterdam 1985, pp. 115 – 154.
69. E. Leroy, C. Robin-Brosse, J. P. Torre in L. L.
Hench, D. R. Ulrich, (eds.): Ultrastructure
Processing of Ceramics, Glasses and
Composites, Wiley-Interscience, New York
1984, pp. 219 – 231.
70. 3 M, US 3 793 041, 1971, (H. G. Sowman).
71. H. G. Sowman, D. D. Johnson, Ceram. Eng.
Sci. Proc. 6 (1985) 1221 – 1230.
72. H. Teranishi, H. Ichikawa, T. Ishikawa, New
Mater. New Processes 2 (1983) 379 – 385.
73. S. V. Nair, J. K. Tien, R. C. Bates, Int. Met.
Rev. 30 (1985) 275 – 290.
74. A. K. Dinghra, CHEMTECH 1981, 600 – 608.
75. F. Pott, Staub Reinhalt. Luft 38 (1978)
486 – 490.
76. United Aircraft Corporation, BE 1 177 854
(1967); US 3 549 424, 1971, 3 574 649, 1971.
77. US 3 760 049, 1973.
78. S. J. Everett, US 2 050 298, 1936.
79. J. A. Roberts in H. S. Katz, J. V. Milewski
(eds.): Handbook of Fillers and
Reinforcements for Plastics, “Metal
Filaments,” Chapter 31, Van Nostrand
Reinhold, New York, 1978, pp. 597 – 616.
80. S. Kavesh: Spinning Wire from Molten Metal,
vol. 74, AIChE no. 180, 1978.
81. H. Jones: “Experimental Methods in Rapid
Quenching from the Melt,” Treatise Mater. Sci.
Technol. 20 (1981).
82. E. A. Feest: “Rapid Solidification for the Wire
Industry,” Wire Ind. June 1985.
83. P. Vansteenkiste: “Prevention of Static Charges
in Dust Filter Fabrics with Stainless Steel
Fibres,” Filtr. Sep. 18 (1981) March-April.
84. R. De Bruyne: “Multipass Test Performance of
Metal Fibre Filter Media,” Proceedings World
Filtration Congress III, vol. 2, Downingtown,
1982, pp. 400 – 408.
85. L. C. Shirvill, I. Lefever: “Metallic Fibre
Surface Combustion Radiant Gas Burners,”
Proceedings 1986 International Gas Research
Conference, Toronto, Canada, 1986.
86. J. Samyn: “Bekiscan, a Fast Nondestructive
Method to Evaluate Emi-Shielding Properties
of Metal Fibre-Filled Poylmers,” Antec ’86
Proceedings, Boston 1986.Chapter 5
87. J. B. Donnet, E. Fitzer, K.-H. Köchling,
Proceedings of 4th International Carbon
Conference (Carbon ’86), Baden-Baden
30. 6. – 4. 7. 1986, pp. 701 – 705.
88. J. Simitzis, Ph. D. Thesis, University of
Karlsruhe, FRG, 1975.
89. H. Jäger, Ph. D. Thesis, University of
Karlsruhe, FRG, 1986.
90. E. Fitzer, F. Rozploch, High Temp. – High
Pressures, 1988 in print.
91. E. Fitzer, E. Gantner, F. Rozploch, D. Steiner,
High Temp. – High Pressures, 1988, in print.
92. E. Fitzer, W. Frohs, M. Heine, F. Rozploch,
Proceedings of 4th International Carbon
Conference (Carbon ’86), Baden-Baden
30. 6. – 4. 7. 1986, pp. 850 – 852.
93. W. Ruland: J. Polym. Sci. Part C 28 (1969) 5.
94. E. Fitzer, W. Frohs, F. Rozploch, High Temp. –
High Pressures 19 (1987) 311 – 322.
95. E. Fitzer, P. Laudenklos, G. Schwarz, Ber.
Dtsch. Keram. Ges. 53 (1976) no. 6, 169 – 174.
96. A. A. Bright, L. S. Singer, Carbon 17 (1979)
59 – 69.
97. W. Ruland, J. Appl. Phys. 38 (1967) 3585.
98. S. C. Bennet, D. J. Johnson: Proc. London Int.
Carbon Graphite Conf. 5th 1 (1978) 337.
99. M. Guigon, A. Oberlin, G. Desarmot, Fibre
Sci. Technol. 20 (1984) 117 – 198.
100. W. Marciniak, F. Rozploch, High Temp. –
High Pressures 11 (1979) 709 – 710.
101. P. Schubert, Ph. D. Thesis, Univ. of Karlsruhe,
FRG, 1977.
102. K. Brennfleck, E. Fitzer, G. Schoch, 1984 in
Extended Abstracts, Carbone 84, Bordeaux,
France (Talence: Centre de Recherche Paul
Pascal) pp. 162 – 163.

103. H.-P. Boehm, Catalysis 16 (1966) 176.
104. J. Cziollek, Ph. D. Thesis, University of
Karlsruhe, FRG, 1983.
105. E. Fitzer, Carbon Fibres and Their
Composites, Springer Verlag 1985.
106. E. Fitzer, R. Weiss, Carbon 25 (1987) no. 4,
455 – 467.
107. R. Bacon, in P. J. Walker, jr. (ed.): Chemistry
and Physics of Carbon, vol. 9, Marcel Dekker,
New York 1973, p. 7.
108. T. A. Edison, US 223 898/470 925, 1879/1892.
109. W. F. Abbott, US 3 053 775, 1962.
110. J. Electrochem. Soc. 106 (1959) 147 C.
111. G. E. Cranch: in Proceedings of the Fifth
Conference on Carbon, vol. 1, Pergamon
Press, New York 1962, p. 589.
112. C. E. Ford, C. V. Mitchel, US 3 107 152, 1963.
113. D. L. Schmidt, H. T. Hawkins, in Reinforced
Plastics ’65, Society of Plastics Engineers
Regional Conference, Seattle, 1965, p. 111.
114. E. M. Peters, US 3 235 353, 1966.
115. R. Bacon, G. E. Cranch, R. O. Moyer, jr., W. H.
Watts, US 3 305 315, 1967.
116. R. Bacon, W. A. Schalamon, Eighth Biennial
Conference on Carbon, Buffalo, N. Y., June
1967, Paper MI-58.
117. W. J. Spry, GB 1 093 084, 1967; US 3 454 362,
1969.
118. D. W. Gibson, G. B. Langlois, ACS Polymer
Preprints 9 (1968) 1376.
119. R. C. Houtz, Tex. Res. J. 20 (1950) 786.
120. A. Shindo, JP 28 287, 1959.
121. W. G. Vosburgh, Tex. Res. J. 30 (1960)
882 – 896.
122. W. Johnson, H. Phillips, W. Watt,
GB 1 110 791, 1965.
123. W. Watt, in A. Kelly et al. (eds.): Handbook of
Composites, “Strong fibres,” vol. 1, Elsevier,
Amsterdam 1985.
124. D. J. Müller, Ph. D. Thesis, University of
Karlsruhe, “Zur Herstellung von
Kohlenstoffaser aus Polyacrylnitril über die
katalytische Cyclisierung des Polymeren,”
1973.
125. S. Otani, Mol. Cryst. Liq. Cryst. 63 (1981)
249.
126. S. Otani, A. Oya, Kogyo Kagahu Zasshi 73
(1970) 493.
127. S. Otani, A. Oya, in Petroleum Derived
Carbon, Am. Chem. Soc. Symp. Ser. 303,
Washington, D.C., 1986, p. 323.
128. W. P. Hettinger, J. W. Newmann, R. P. Krook,
D. C. Boyer, Earth Moving Industry
Conference, Peoria, Illinois, USA, April
8 – 10, 1986 in SAE (Society of Automotive
Engineers) Technical Paper series 860 767.
Fibers, 5. Synthetic Inorganic 69
129. I. Romey, “Preprint,” 2. Int.
Kohlenstofftagung, Baden-Baden 1976, p. 470.
130. J. B. Barr, S. Chwastiak, R. Didchenko, R. T.
Lewis, J. C. Lewis, L. Singer, Appl. Polym.
Symp. 29 (1976) 261.
131. Union Carbide, US 4 005 183, 1977.
132. E. Fitzer, H. Schlesinger, Ber. Dtsch. Keram.
Ges. 43 (1966) no. 3, 209 – 218.
133. E. Fitzer, W. Fritz, B. Rhee, High Temp. –
High Pressures 4 (1972) 219 – 224.
134. T. Koyama, M. Endo, Y. Hishiyama, Jpn. J.
Appl. Phys. 13 (1974) no. 12, 1933 – 1939.
135. A. Nogaj, in B. von Falkai (ed.):
Synthesefasern, Verlag Chemie, Weinheim
1981, pp. 197 ff.
136. G. P. Daumit, Y. S. Ko, in K. Brunsch et al.
(eds.): in High Tech – the Way into the Nineties,
Elsevier, Amsterdam, 1986, pp. 201 – 213.
137. E. Fitzer, W. Frohs, M. Heine, Carbon 24
(1986) 387 – 395.
138. K. Morita et al., Ext. Abstracts Int. Symp. on
Carbon, Toyohashi 1982, p. 321.
139. E. Fitzer, M. Heine, G. Jacobsen, Int. Symp. on
Carbon, Toyohashi, Japan 1982.
140. E. Fitzer, D. Müller, A. K. Fiedler, Int.
Conference on Carbon Fibres, Their
Composites and Application, no. 2, London
1971.
141. E. Fitzer, A. K. Fiedler, D. J. Müller, Circ. Inf.
Tech. Cent. Doc. Sider. 43 (1971) 923 ff,
1022 ff.
142. E. Fitzer, A. K. Fiedler, Preprints Carbon 72,
Baden-Baden, Juni 1972, paper 21.
143. E. Fitzer, D. J. Müller, Chem. Z. 96 (1972)
20 – 26.
144. E. Fitzer, M. Heym, Chem. Ind. 16 (1976)
663–676.
145. E. Fitzer, T. Müller, Proc. of Carbon ’82 Ext.
Abstracts, London 1982, pp. 249 – 252.
146. N. Grassie, R. McGuchan, Eur. Polym. J. 7
(1971) 1357.
147. E. Fitzer, D. J. Müller, Carbon 13 (1975) 63.
148. E. Fitzer, M. Heine, W. Metzler, Proc. of
Carbon ’86, Baden-Baden, DKG 1986,
pp. 853 – 855.
149. Toho, Beslon Co., Ltd. DE-OS 31 32 784,
1981.
150. Toray Industries, JP 48 228, 1975.
151. E. Fitzer, A. Gkogkidis, G. Jacobsen, Extended
Abstracts Int. Symp. on Carbon, Toyohashi
1982, p. 284, Carbon Society of Japan, Tokyo.
152. D. W. McGuffin, R. J. Larsen, R. W. Mitchell,
B. E. Caruso, US 3 803 672, 1974 assigned to
HITCO.
70 Fibers, 5. Synthetic Inorganic
153. T. Trendelenburg, Ph. D. Thesis, University of
Karlsruhe, FRG, 1984.
154. D. Kompalik, Ph. D. Thesis, University of
Karlsruhe, FRG, 1986.
155. E. Fitzer, D. Kompalik, Sprechsaal 119
(1986) no. 10, 911 – 916.
156. Union Carbide, US 4 005 183, 1977.
157. Engineering, Sept. 1982, p. 611.
158. H. Böder, SIGRI GmbH Meitingen/Augsburg,
Vortrag anläßlich Carbon ’86 in Baden-Baden.
159. H. Fleckenstein, VDI-Z. 125 (1983) nr. 4,
123 – 131.
160. W. C. Miller, Kirk-Othmer 20 (1982) 65 – 77.
161. J. Griffiths, “Synthetic Mineral Fibres,” Ind.
Miner. (London), Sept. (1986) 20 – 43.