Alumina Etching

Materials Science and Engineering, 26 (2000) 97±199
Silazane derived ceramics and related materials

Edwin Kroke*, Ya-Li Li, Christoph Konetschny, Emmanuel Lecomte, Claudia Fasel,
Ralf Riedel
Technische UniversitaÈt Darmstadt, Fachbereich Materialwissenschaft, Fachgebiet Disperse Feststoffe, Petersenstr. 23,
D-64287 Darmstadt, Germany
Accepted 14 January 2000
Abstract
This review highlights the synthesis, processing and properties of non-oxide silicon-based ceramic materials
derived from silazanes and polysilazanes. A comprehensive summary of the preparation of precursor compounds
containing Si±N±Si units, including commercially available materials, is followed by the discussion of various
processing techniques. The fabrication of dense bulk ceramics in the Si/E/C/N systems is reported which involves
cross-linking of the polymeric ceramic precursor followed by a polymer-to-ceramic transformation step. The cross-
linked precursor can be milled, compacted and pyrolysed to form dense, additive-free, amorphous silicon
carbonitride monoliths or polycrystalline composites which withstand oxidation in air at 16008C. Furthermore, an
overview is given on the fabrication of silazane derived powders and coatings involving chemical vapour deposition
(CVD) methods utilising volatile precursors. Fibre spinning and fibre properties, as well as other processing
techniques like infiltration of preforms, the preparation of porous ceramics and joining are briefly discussed. A state
of the art of the mechanical properties of polymer derived amorphous Si/C/N and Si/B/C/N ceramics with respect to
hardness as well as high-temperature creep and oxidation resistance is summarised. Finally, some important aspects
of industrial applications will be considered. The review is in part based on our own work related to the polysilazane
derived ceramics, but will also cover a comprehensive state of the art including the published literature in this
field. # 2000 Elsevier Science S.A. All rights reserved.
Keywords: Silazane; Polysilazanes; Cross-linking; Silazane derived ceramics; Chemical vapour deposition
1. Introduction
Many of the important industrial objectives for future developments such as an increased
efficiency in energy generation (turbines), storage and utilisation (Na±S batteries) or increased
environmental compatibility in transport systems (ceramic motors) depend upon the successful
design of high-temperature inorganic components. Simple inorganic materials possess important
limitations in applications where mechanical performance at high temperature is required. In
principle, advanced ceramics meet these requirements, i.e. high chemical resistance and enhanced
mechanical properties at elevated temperatures. However, ceramics are difficult to apply with
confidence owing to their intrinsically brittle behaviour, while metals are limited by corrosion
problems and by reduced performance at temperatures approaching their melting point.
Consequently, the need for ceramics capable of performing under demanding conditions is evident.
There are a variety of ceramic systems which will play a major role in the field of the
*
Corresponding author. Tel.: 49-6151-16-6345; fax: 49-6151-16-6346.
E-mail address: kroke@hrzpub.tu-darmstadt.de (E. Kroke)
0927-796X/00/$ ± see front matter # 2000 Elsevier Science S.A. All rights reserved.
PII: S 0 9 2 7 - 7 9 6 X ( 0 0 ) 0 0 0 0 8 - 5
98 E. Kroke et al. / Materials Science and Engineering 26 (2000) 97±199
aforementioned applications where exceptional mechanical reliability under severe conditions is

required. To provide this reliability, the material must have a well-defined organisation of its
microstructure with respect to quantity, distribution and character of the constituent phases involved.
Therefore, novel ceramic processing techniques have to be developed in order to overcome the
limitations considered.
Many researchers all over the world have sought to establish the basic sciences and technology
for promising methods of reliable synthesis for ceramic materials that reveal tailor-made
microstructures. The techniques considered in the course of this review are based on the thermally
induced conversion of organoelement precursors to inorganic components, sometimes denoted as
hybrid processing (Fig. 1) [1]. The polymer-to-ceramic transformation process involves a complex
sequence of chemical and physical changes. The choice of the organoelement compound, the design
of the intermediate processing steps and the use of annealing treatments must all be done with full
understanding of the basic science. This requires contributions from many disciplines, namely
inorganic/organic chemistry, mineralogy, materials science and computer assisted modelling. It is
therefore important to develop a fundamental interdisciplinary basis so that effective exploitation of
the method can take place in industrial technology.
Fig. 1. Flow chart representing the hybrid processing for the fabrication of silicon nitride/silicon carbide based composites.
E. Kroke et al. / Materials Science and Engineering 26 (2000) 97±199 99
Polysilazane derived silicon nitride (Si3N4) and silicon carbide (SiC) and its composites have
been intensively studied in recent years, particularly in Europe, Japan and the United States. Silicon
nitride/silicon carbide-composites are candidate materials for high-temperature application as engine
and turbine parts and as cutting tools, as well as having importance for microelectronic devices. In
comparison to monolithic Si3N4 materials, the composites have been found to exhibit (i) enhanced
fracture strength values, not only at room temperature but also at elevated temperatures and (ii)
superplasticity at 16008C [2,3]. However, the conventional fabrication of dense Si3N4/SiC-
composites is difficult, owing to the different sintering behaviour of the Si3N4 and SiC powder
particles used as the starting materials [4]. One promising alternative route is to start from silicon
carbonitrides, Si3xN4Cxy, which are single phase materials with a homogeneous distribution of
silicon, carbon and nitrogen at an atomic level [5,6]. It is obvious that the ratio of these three
elements may be modified and controlled by the selection and design of appropriate Si/C/N
precursor compounds. Usually, silazanes containing Si±N±Si units have been used because they are
relatively easy to obtain from chlorosilanes and ammonia or amines. However, other classes of
molecular and polymeric precursors may be used as well, e.g. silylcarbodiimides [7], but these will
not be considered in this review article.
Polymeric silazanes, as well as cross-linked precursor materials, possess amorphous networks
which transform to amorphous, hydrogen free silicon carbonitride ceramics around 10008C. At this
point it should be mentioned that recently the first crystalline compounds in the ternary Si/C/N
system were synthesised below 10008C. They have the compositions SiC2N4 and Si2CN4 but
were not derived from silazane precursors [8]. In particular for silazane derived Si/C/N materials,
the crystallisation behaviour at temperatures >12008C, the phase transformations and the thermal
stability in different atmospheres have been studied and will be discussed in Sections 3 and 4.
The state of the art of the mechanical properties of amorphous Si/C/N ceramics with respect to
hardness, as well as high-temperature creep behaviour and oxidation resistance, is also summarised
[9,10].
The polysilazane derived ceramics of the general composition Si3xCxyN4, remain metastable
at T<14408C and give off nitrogen at T>14408C when the system reaches the tie line composition
Si3xCxN4, and finally crystallises to give Si3N4 and SiC (Eqs. (1)±(3)).
Si3x Cxy N4 ! Si3 N4 xSiC yC (1)

Si3 N4 yC ! 1 ÿ y=3Si3 N4 ySiC 2y=3N2 (2)
Si3x Cxy N4 ! 1 ÿ y=3Si3 N4 x ySiC 2y=3N2 (3)
Consequently, the thermal stability of polysilazane derived Si/C/N ceramic materials during
degradation can be significantly enhanced by avoiding nitrogen evolution in the high-temperature
region. This, however, requires the direct formation of Si3xCxN4 compositions. Therefore, the
carbon content may be adjusted by the reactive heat treatment of poly(hydridomethyl)silazanes in
mixed NH3/Ar atmospheres.
The generation of dense bulk ceramics in the Si/C/N system has been successfully achieved by
the hybrid processing route [1,11]. It involves the steps described above, i.e. cross-linking of the
polymeric precursor followed by a thermal induced polymer-to-ceramic transformation. On the one
hand, the cross-linked precursor can be milled, compacted and pyrolysed to form dense, additive-
free, amorphous silicon carbonitride monoliths or polycrystalline Si3N4/SiC/C micro/nano
composites which withstand oxidation up to 16008C in air [12]. On the other hand, the cross-
linked polymer can be pyrolysed to amorphous silicon carbonitride powders, which can be
subsequently processed to dense polycrystalline Si3N4/SiC micro/nano-composites by hot-pressing

in the presence of oxidic sintering additives. Due to the low additive content, enhanced
thermomechanical properties and oxidation resistance of the polymer derived Si3N4/SiC-composites
can be expected relative to conventionally processed silicon nitride-based ceramics [12].
A completely different way of processing has to be applied for the fabrication of powders,
fibres, composites or coatings by CVD, spin- or dip-coating, fibre-spinning, or infiltration of porous
substrates. These techniques are very important for many industrial applications of silazane derived
ceramics but will be considered only briefly in this review. In principle they can also be described by
the scheme in Fig. 1. CVD processing starts with molecular precursors, which are directly
transformed to amorphous or crystalline ceramic powders or coatings, depending on experimental
parameters like deposition temperature, pressure, kind of substrate etc. For spin-, dip- and spray-
coating liquid oligomeric/polymeric silazanes or solutions may be used. Similarly, fibres may be
drawn from melts or solutions. A very crucial step is the curing or cross-linking of the green fibres in
order to obtain the desired ceramic fibres. The infiltration of porous materials may be accomplished
by gaseous chemical vapour infiltration (CVI) or liquid precursors.
It is very important to note that at least some molecular, oligomeric and polymeric silazanes are
commercially available. Consequently, silicon carbonitrides are easily accessible and are, therefore,
suitable precursors for the fabrication of Si3N4/SiC ceramic bulk composites. Polyorganosilazanes,
as well as molecular silazanes, can be produced in large quantities and can be considered as highly
efficient Si3N4/SiC precursors. The conversion mechanism of different commercially available
polysilazanes has been studied in order to correlate the microstructure and the properties of the
polymer derived ceramics with the molecular structure of the starting preceramic polymer.
The polysilazanes have been chemically modified in order to produce quaternary systems like
Si/B/C/N and Si/P/C/N [13]. It has been shown that the polymer modification significantly
influenced (i) the crystallisation behaviour, (ii) the thermal stability and (iii) the ceramic
microstructure of the resulting advanced ceramic material [14].
The production of Si3N4/SiC ceramics by hybrid processing has now been proven to be a
commercial success for a number of products of small dimensions, e.g. for fibres and coatings. The
use of this processing route to fabricate bulk ceramics has problems, chiefly caused by a nanoscale
porosity owing to gas evolution during pyrolysis and subsequent ceramisation. Currently this
porosity must be reduced by costly and time-consuming post-processing techniques such as hot-
pressing or HIP treatments. In order to overcome this problem, sintering aids have been introduced
into the ceramic by chemically modifying the polysilazanes with Y and Al containing side groups,
using metal alkoxides as reagents.
The hybrid processing has been investigated by two methods:
1. Forming the polysilazane green body followed by the ceramisation of the compact, without any
sintering promoting additives. Application of different pyrolysis atmospheres like Ar and NH3 has
revealed that the carbon content can be adjusted over a broad range. Furthermore, the ammonia
treatment can be utilised to form functionally gradient materials with low carbon content in the
shell, and with high carbon content in the core of the ceramic part.
2. Hot-pressing of polysilazane derived amorphous Si/C/N powders, containing low amounts of
yttria, introduced either by the chemical modification of the precursor or by adding Y2O3 powder
to the as-produced silicon carbonitride. The densification has resulted in high density (99% of
T.D.) Si3N4-based ceramics and Si3N4/SiC-composites. The type of starting polysilazane
precursors (NCP 200 and perhydropolysilazane (PHPS)) and the hot-pressing conditions
influence the final microstructure, density, hardness, fracture toughness and crack formation of
the composite ceramics containing sintering aids.
In parallel with the experimental work, the sintering and grain growth has been simulated by Monte
Carlo methods in two-dimensional triangle lattices to optimise the design of the microstructure in the
precursor derived ceramics. The modelling clearly predicts an enhanced rate of densification of the
polysilazane derived silicon nitride and silicon carbonitride powders due to their amorphous state.
The microstructures of pyrolysed and densified ceramic bulk materials have been fully
characterised by comprehensive TEM investigations. The amorphous state as well as the
crystallisation behaviour of pure and modified silicon carbonitride ceramics (Si/(M)/C/N with
MB, P) has been studied. Polysilazane related polyborosilazanes synthesised from single source
precursors can be transformed into silicoboron carbonitrides (SiBxCyNz), exhibiting enhanced
thermal stability with respect to crystallisation and decomposition, in the temperature range between
1400 and 20008C [14]. The ultra-high temperature stability of the silicoboron carbonitrides can be
discussed in terms of the Si/B- and C/N-ratios.
Using the electron spectroscopic imaging (ESI) technique, two-dimensional maps of the
distribution of various elements in large specimen areas (several square micrometers) have been
obtained at high resolution (2±3 nm). High resolution electron microscopy and electron diffraction
investigations have been performed to characterise the order of crystallinity and the type of
segregations at the grain boundaries in the ceramic materials synthesised from different
polysilazanes heat treated at temperatures ranging from 1000 to 18008C.
The overall progress in the field of silazane derived ceramics provides important results and
information, regarding the basic science and technology of the hybrid process in which organosilicon
polymers are converted to inorganic components. The hybrid processing methods have been
precisely characterised to understand the complex sequence of the structural changes which occur
during the polymer-to-ceramic transformation. Novel techniques have been used to show the
characteristic properties of the materials prepared by this method. Transmission electron microscopy
has, in particular, revealed the unique and complex microstructures of these polymer derived
advanced ceramic materials.
Finally, the experimental and theoretical findings of the international research activities
unambiguously indicate that the microstructure of silicon nitride/silicon carbide-based ceramics can
be precisely controlled over a wide span of possibilities. This microstructural adjustment is attained
by chemically modifying appropriate molecular precursors and by the use of distinct conditions
during the course of the hybrid processing of the organic±inorganic polymers. The following
sections summarise the scientific findings in detail, with respect to the synthesis and modification of
silicon carbonitride forming silazanes, as well as to the densification (sintering), microstructural
characterisation, properties and applications of the resulting non-oxide ceramic materials.
2. Synthesis of polysilazanes
In order to obtain any kind of ceramic material from organometallic starting compounds the
precursors themselves have to be synthesised. It is the objective of this chapter to give an overview of
the formation and modification of silicon compounds containing Si±N±Si units, i.e. (poly)silazanes.
It has to be stressed that large volumes of Gmelin's handbook [15] and chapters in other monographs
[16,17] as well as some review articles [18±20] are devoted to related topics. Besides, a large and
increasing number of original papers and proceedings contributions are published each year. Here we
will discuss those molecular and polymeric compounds which are relevant to ceramic precursor
technology. This does not decrease the body of literature reviewed too much since a very large
fraction of all silazane compounds has been pyrolysed to form silicon containing ceramics.
Historically, the first efforts in silazane chemistry were simply directed towards the preparation,
classification and characterisation of silazanes and polysilazanes [21±24]. These and the following
studies resulted in a large number of (mostly cyclic) oligosilazanes known in the sixties and
summarised in review articles [25,26]. The successful commercialisation of silicones in the 1950s
and 1960s initiated the investigation of oligo- and polysilazanes as potential siloxane analogues as
well as early attempts to prepare silicon (carbo)nitrides and related materials [27±30]. Today, the
interest in silazanes primarily derives from their applications as silylating agents in synthetic
chemistry and as single source precursors for the preparation of ceramic materials by vapour, liquid
and solid phase pyrolysis.
2.1. Ammonolysis of halogenosilanes
Chlorosilanes are produced by the MuÈller±Rochow-Synthesis on a large scale. They are used for
the fabrication of siloxanes and silicones which in turn are the most important inorganic or
organometallic polymers [31]. Siloxanes are generally synthesised by hydrolysis of Si±Cl-bonds
followed by subsequent condensation of the Si±OH intermediates to form Si±O±Si groups. If the
water is replaced by ammonia, a similar ammonolysis reaction takes place and silylamines are
formed, which can condense to give (poly)silazanes. Thus, Eq. (4) gives a simplified general
description of this synthesis. It is assumed that two hydrogen atoms per ammonia molecule are
substituted and the chlorine atoms are replaced completely:
2nR4ÿx SiClx 3xnNH3 ! NHx=2 ÿ SiR4ÿx n 2xn;

x 1; 2; 3; 4 for x 1 follows n 1 (4)
It is obvious that for x1, small molecules like disilazanes R3Si±NH±SiR3 (or aminosilanes R3Si±
NH2) are obtained. These are useful single source precursors for MOCVD methods. In almost all
other cases, i.e. for x>1 linear or cyclic oligomers as well as linear or cross-linked polymers are
formed if RH or CH3.
The two-step reaction mechanism of silazane formation by ammonolysis, involving substitution
of chlorine atoms and condensation of the resulting aminosilanes, can be described by Eqs. (5) and
(6):
BSiÿCl 2NH3 ! BSiÿNH2 NH4 Cl (5)

2BSiÿNH2 ! BSiÿNHÿSiB NH3 (6)
This reaction scheme is completely analogous to the above-mentioned hydrolysis of

chlorosilanes. However, the remaining N±H-group has to be considered as a reactive functional
group which is in contrast to oxygen bridges. Further reactions are possible, e.g. formation of a
second equivalent of ammonia (7) or dehydrocoupling (8) (see the following sections):
BSiÿNH2 BSiÿNHÿSiB ! BSiÿ3 N NH3 (7)

BSiÿH BSiÿNHÿSiB ! BSiÿ3 N H2 (8)
In the following paragraphs we try to give an overview of the numerous silazanes prepared by
ammonolysis reactions.
2.1.1. Ammonolysis of monochlorosilanes R3SiCl

Hexamethyldisilazane or HMDS, (H3C)3Si±NH±Si(CH3)3, is the most important single source
precursor for the preparation of Si/C/(N) from the gas phase (see Eq. (9)). This is because of its
availability, its relatively low price of US$ 10/kg [32], and its volatility. HMDS is synthesised on
an industrial scale by ammonolysis of chlorotrimethylsilane [33].
2H3 C3 SiCl 3NH3 ! H3 C3 SiÿNHÿSiCH3 3 2NH4 Cl (9)
The exothermic reaction can be carried out in solvents such as hexane, toluene, or isododecane,
but is best performed in HMDS itself. Coarse-grained ammonium chloride is formed which can be
readily filtered off to give high turnovers and yields of HMDS.
Chlorosilanes with small substituents comparable to methylgroups like (H3C)(H5C2)2SiCl,
(H3C)2(H5C2)SiCl [23], or (H3C)2(H3CO)SiCl [34] react with ammonia similar to chlorotrimethyl-
silane to form disilazanes. Chlorotrialkylsilanes with larger substituents like the triethyl-,
triisopropyl- or tributyl-derivatives or chlorotriarylsilanes tend to form primary silylamines
R3SiNH2, see e.g. [35±37]. On the other hand, it is very difficult to prepare pure trimethylsilylamine
[16,38]. With one or more hydrogen atoms and small substituents at the silicon atom tertiary amines
(trisilylamines) are formed (Eq. (10)) [39,40]:
3RR0 HSiCl 4NH3 ! RR0 HSi3 N 3NH4 Cl (10)
The ammonolysis of chlorosilanes besides, a number of other routes to disilazanes have been
found [16]. Most of these syntheses are too expensive to be relevant for large scale production of
ceramic materials. However, some important principles will be mentioned in Section 2.6.
2.1.2. Ammonolysis of dichlorosilanes R2SiCl2

It is well known from siloxane chemistry that a variety of linear and cyclic oligomers are formed
by hydrolysis of dichlorosilanes [31]. Methyl- and phenyl-substituted dichlorosilanes are the main
components for the preparation of industrially important silicones. Also, the difunctional monomer
can bear cross-linkable functional groups and mono- as well as tri- and tetrafunctional (SiCl4) silane
units are used to design the properties of these polymers.
Silazanes are obtained by reactions with ammonia, similar to the formation of siloxane
polymers. Depending on the size of the substituents and on the reaction conditions different products
are observed. If the substituents are sufficiently large and low temperatures are applied,
diaminosilanes are formed in good yields (Eq. (11)) [41±44]:
R2 SiCl2 4NH3 ! R2 SiNH2 2 2NH4 Cl R e:g: mesityl; benzyl; t-butyl; i-propyl etc:
(11)
Upon heating, condensation takes place to form ammonia and oligomeric silazanes. For smaller
substituents (Rmethyl, ethyl, phenyl, H) it is much more difficult to prevent condensation.
In principle, two kinds of dimers can be formed by condensation of diaminosilazanes, linear
disilazanes and cyclodisilazanes (1,3,2,4-diazadisiletidines). However, due to the ring strain of
four-membered cyclosilazanes the formation of chains or cyclic tri- or tetramers is usually favoured
(Eq. (12)):
R2 SiNH2 2NH3 2R2 SiNH2 2 ! H2 NÿSiR2 ÿNHÿSiR2 ÿNH2 NH3 (12)
cyclodisilazanes linear trisilazane
Table 1
Selected list of cyclotrisilazanes prepared by ammonolysis reactions of dichlorosilanes
Dichlorosilane Formula of cyclotrisilazane References
Dichloro-diethylsilane [(CH3CH2)2SiNH]3 [23,54]
Ethyl-dichloro-methyl-silane [(CH3)(C2H5)SiNH]3 [55]
Dichloro-ethyl-silane [H(C2H5)SiNH]3 [52]
Dichloro-isopropyl-silane [H(C3H7)SiNH]3 [56]
Dichloro-methyl-phenyl-silane [(CH3)(C6H5)SiNH]3 [57,58]
Dichloro-diphenylsilane [(C6H5)2SiNH]3 [59,60]
Dibutyl-dichlorosilane [(C4H9)2SiNH]3 [42]
Dichloro-isopropyl-methyl-silane [(CH3)(C3H7)SiNH]3 [61]
Dichloro-hexyl-methyl-silane [(CH3)(C6H13)SiNH]3 [55]
Butyl-dichloro-methyl-silane [(CH3)(C4H9)SiNH]3 [55]
Dichloro-isobutyl-methyl-silane [(CH3)(C4H9)SiNH]3 [62]
Dichloro-cyclohexyl-methyl-silane [(CH3)(C6H13)SiNH]3 [61]
Dichloro-methyl-octyl-silane [(CH3)(C8H17)SiNH]3 [55]
Dichloro-methyl-nonyl-silane [(CH3)(C9H19)SiNH]3 [62]
Dichloro-methyl-vinyl-silane [(CH3)(CH2=CH)SiNH]3 [63]
Dichloro-methyl-phenethyl-silane [(CH3){C6H5(CH2)2}SiNH]3 [64]
1,1-Dichlorosilacyclobutane [(CH2)3SiNH]3 [65]
Bromomethyl-dichloro-methyl-silane [(CH3)(CBrH2)SiNH]3 [66]
4-(Dichloro-methyl-silanyl)-butyronitrile [(CH3){NC(CH2)3}SiNH]3 [64]
Cyclodisilazanes of the type [R2Si(NH)]2 are known, but they are prepared by methods other
than amminolysis reactions of dichlorosilanes [45±48]. At this point it has to be mentioned that
cyclotrisilazanes, important ammonolysis products of dichlorosilanes (see later and Table 1), can
react with further equivalents of dichlorosilane to form cyclodisilazanes by ring contraction [49,50]
(Eq. (13)):
(13)
Linear disilazanes, H2N±SiR2±NH±SiR2±NH2, have been found to form together with

diaminosilanes or with higher oligomers. Some of these compounds, e.g. RCH3 [51], n-butyl
and n-hexyl [41] have been isolated and characterised [52,53] as have silazanes derived from
dichlorosilanes with two different substituents. Cyclotrisilazanes have also been isolated from
ammonolysis reactions of many other dichlorosilanes. Further examples are listed in Table 1.
Most frequently, cyclotrisilazanes are formed together with higher oligomers. The reaction of
dichlorodimethylsilane, (CH3)2SiCl2, with NH3 gives 2,2,4,4,6,6-hexamethyl-cyclotrisilazane
(HMCTS) and approximately the same amount of 2,2,4,4,6,6,8,8-octamethyl-cyclotetrasilazane
(OMCTS), together with small amounts of other oligomers [23,24,67,68]. It was shown that the
exact product composition, i.e. the relative amount of each oligomer, strongly depends on the
experimental parameters like reaction time and temperature [26]. The product mixture can either be
separated by distillation, or used for pyrolysis to ceramic materials. According to recent reports [69],
the product mixture consists of a liquid-waxy to solid, colourless to yellowish substance, melting
around 55±908C, while the pure cyclotrisilazane is a liquid boiling at 1868C and the
cyclotetrasilazane is a solid with a melting point of 978C. Although of low volatility compared to
HMDS, the mixture, as well as the isolated cyclic compounds, have been used for gas phase
pyrolysis methods such as MOCVD [69±71].
As already mentioned, cyclotetrasilazanes are usually formed together with cyclotrisilazanes as

well as other oligomers. Disilazanes and some trisilazanes [72,73] are volatile and suitable for CVD
applications. Nevertheless, in many cases they are part of oligomeric and polymeric mixtures which
can be cross-linked and pyrolysed to form Si/C/N ceramics.
The ammonolysis of dichlorosilane, SiCl2H2, was first described by Stock and Somieski in 1921
[22]. In the gas phase the reaction gave only intractable solids, in benzene a soluble viscous oil was
obtained which was not stable at room temperature. Later, Seyferth and coworkers examined the
reaction using more polar solvents like diethylether or dichloromethane [74±76]. At 08C non-volatile
soluble oils were obtained. The viscosity of these polymers increased gradually at room temperature
to give glass-like solids after 1±5 days. Pyrolysis in argon at 11508C for 5 h leads to the formation of
silicon nitride together with elemental silicon. Nevertheless, this approach was used to prepare
ceramic composites and coatings [77]. Furthermore, the pyridine adduct of dichlorosilane was used
to prepare a PHPS with Mn800±1000, r1.1±1.2 g/cm3 and Z40±80 mPa s at 308C. Thermal
cross-linking occurs at 100±3008C to give glass-like hard solids [78]. PHPS polymers of this type
were prepared by Tonen Corp., Japan [79,80]. They may be cured, e.g. by treating it with pyridine
and NH3 at 80±1208C to obtain a solid polymer (Mn2400) that is soluble in pyridine or xylene and
can be drawn into fibres [81,82].
Dichloromethylsilane, CH3SiHCl2, is less volatile and less reactive than dichlorosilane and
relatively cheap. The reaction with ammonia, carried out in an organic solvent, gives principally a
mixture of cyclic oligomers, [CH3(H)SiNH]n with n2, 3, 4, 5, . . . [23]. If these cyclosilazanes are
not separated immediately, rearrangements occur that result in the formation of ±SiH2CH3 groups
[83]. These might arise from ring contraction processes described in literature [84,85]. The volatility
of the obtained low molecular weight oligosilazanes is usually too low to serve as pyrolytic
precursors for silicon nitride or silicon carbonitride, except when special gas-phase pyrolysis
methods or autoclave techniques are applied. Therefore, several methods to increase the molecular
weight have been developed (see Sections 2.6 and 3.2.1). Treatment with basic catalysts such as KH
[86,87] is the method most frequently used.
Analogous to dichloromethylsilane the ethyl-derivative, CH3CH2SiHCl2, was ammonolysed in
diethylether solution at 08C. After removing the NH4Cl and the solvent, an oily residue was isolated,
which was used directly for CVD experiments (vide infra) [88].
Dichlorodiphenylsilane was ammonolysed to prepare a poly(diphenyl)silazane (PDPS) oil,
which was modified by NH4Br and additional NH3 treatment. This material was used to produce
ceramic coatings [89]. A disadvantage of the phenyl substituents is the high carbon content which
either causes the formation of free carbon or Ð if volatile phenyl derivatives are formed Ð leads to a
low ceramic yield.
1,1,-Dichloro-1-silacylobutane, (CH2)3SiCl2, was used to obtain a polysilacyclobutasilazane by
a standard ammonolysis reaction in CH2Cl2 at ÿ20 to ÿ408C. The product was transformed by
equilibration/ammonolysis treatments to give precursors for the preparation of sintered silicon
carbide monoliths [90].
In order to achieve an appropriate molecular weight and degree of cross-linking vinyl-
functionalised silazanes were frequently produced. The C=C-double bond can either react with
Si±H±, N±H±, or with other vinyl groups. These additional reactions can be induced thermally or
catalytically, causing additional cross-linking.
Dichlorovinylsilane, (CH2=CH)HSiCl2, can be transformed to an oligomeric precursor by
ammonolysis in toluene, giving a mixture of cyclic oligomers [91±95]. The average molecular
weight measured by vapour pressure osmometry was 380, corresponding to a mean degree of
polymerisation of 5.4 [93]. The material may be cross-linked thermally at 1108C in toluene, or
catalytically in the presence of some wt.% H2PtCl6 at temperatures around 708C. Polymeric samples
ranging from viscous liquids to insoluble solids were obtained. The major cross-linking reactions are
hydrosilylation of the vinyl groups in a- or b-position [93,96] (Eq. (14)):
(14)
Transamination reactions which lead to trisilylated nitrogen atoms, together with the release of
ammonia, were reported to occur only to a small extent (Eq. (15)) [93]:
(15)
Polymerisation of the vinyl groups may also contribute to three-dimensional cross-linking [96]
(Eq. (16)).
(16)
Other cross-linking mechanisms which were already mentioned above are dehydrocoupling
reactions involving Si±H and N±H groups [92], as described by Eqs. (17) and (18):
(17)
(18)
It is possible to control which of the cross-linking reaction occurs simply by choosing different
chlorosilanes and using amines instead of ammonia to prepare the polymer (see Section 2.2). For
example, a vinyl-functionalised oligomer free of Si±H groups is obtained from ammonolysis reaction
of dichloromethylvinylsilane, (CH2=CH)CH3SiCl2 [92,97]. The synthesis can be performed
similarly to the above described preparation in toluene at 08C, giving slightly lower molecular
masses (see Table 2). Due to the absence of Si±H units the cross-linking is much slower and less
effective. At 1208C under argon it takes more than 8 days until the material solidifies, compared to
2 h for the dichlorovinylsilane derived oligomer.
To summarise the general results of the ammonolysis reactions of pure dichlorosilanes, it has to
be stressed that many authors applied this method for synthesising ceramic precursors. This is due to
the fact that effective curing via hydrosilylation is achieved if Si±H and vinyl groups are present. The
experimental techniques and parameters used by the authors are not always comparable. Even if the
same laboratory repeats a synthesis, the variation of the composition and structure of the obtained
silazane may be quite large [26,69]. This should be kept in mind, especially when the final ceramic
products of different authors are compared (see Table 2).
E. Kroke et al. / Materials Science and Engineering 26 (2000) 97±199
Table 2
Typical experimental conditions and product characteristics of ammonolysis reactions according to the equation: RR0 SiCl23NH3!(RR0 Si±NH±)n2NH4Cl
R R0 Reaction conditions Product properties and yield References
CH3 CH3 Toluene, T<608C, 3 h, NH3(g) y50%, solid, M (cryoscopy)238, Si:N-ratio1:1, O: <0.5%, Cl: <0.1% [97]
CH3 CH3 CH2Cl2, cooling, NH3(g) y80%, solid, mp55±908C, EAa: C: 37.5%, H: 6.5%, N: 17.5%, Si: 37.5% [69]
R,R0 =(CH2)3 CH2Cl2, ÿ20 to ÿ408C, NH3(g) y91%; Mw2015; Mn142; TMAb <208C; EA: C: 44.0%, H: 9.6%, N: 15.8%, Si: 30.7% [90]
H CH3 Toluene, T<608C, 3 h, NH3(g) y65%, Z0.012 Pa s, M (cryoscopy)383, Si:N-ratio1:1, O: <0.5%, Cl: <0.1% [97]
H CH3 Toluene, T08C, 5 h, NH3(g) y>80%, Mw335, Mn295 [92]
H CH3 Et2O or THF, 08C, NH3(g), 2 h M (cryoscopy)298 [87,93]
H CH3CH2 Et2O, 08C, NH3(g), 3 h Oily residue, vapour pressure at 373 K2.4 kPa [88]
CH3 CH=CH2 Toluene, T08C, 5 h, NH3(g) y>80%, Mw350, Mn310 [92]
H CH=CH2 Toluene, T<608C, 3 h, NH3(g) y91%, Z0.004 Pa s, M (cryoscopy)234, Si:N-ratio1:1, O: <0.5%, Cl: <0.1% [97]
H CH=CH2 Toluene, T08C, 5 h, NH3(g) y>80%, Mw 450, Mn390 [92]
a
Average values.
b
TMA softening point.
107
2.1.3. Ammonolysis of trichlorosilanes RSiCl3

Similar to the reactions of mono- and dichlorosilanes with ammonia, trichlorosilanes form
monomeric, oligomeric or polymeric silazane compounds. The degree of polymerisation and thus the
volatility of the products can be controlled by the reaction conditions and the stoichiometry.
Careful ammonolysis of trichlorosilane, HSiCl3, using CaCl2(NH3)8 as an ammonia source
leads to the formation of 1,1,3,3-tetrachlorodisilazane [99] (Eq. (19)):
3NH3 2HSiCl3 ! HCl2 SiÿNHÿSiCl2 H 2NH4 Cl (19)

The structure, chemical reactivity and thermal stability of this compound have been
investigated. The chlorine atoms can by replaced in order to obtain a molecular Si/C/N/(H)-
precursor.
Usually the reaction of trichlorosilane with excess ammonia leads to the formation of
polysilazanes [100] (Eq. (20)). Thermal treatment of these products gives `silicohydrocyanic acid'
(HSiN)n [101] (Eq. (21)):
4:5nNH3 nHSiCl3 ! HSiNH1:5 n 3nNH4 Cl (20)

D
HSiNH1:5n !0:5nNH3 HSiNn (21)
Oligomeric and polymeric alkylsilsesquiazanes have been prepared since the sixties. A standard
ammonolysis reaction was performed using a 15±20% solution of a trichlorosilane in benzene and
dry gaseous ammonia at 15±208C [55] (Eq. (22)):
RSiCl3 4:5NH3 ! 1=nRSiNH1:5 n 3NH4 Cl R ÿCH3 ; ÿC2 H5 ; ÿC6 H13 ;

ÿC7 H15 ; ÿC8 H17 ; ÿC9 H19 (22)
Oligomeric products with n6 and 8 were isolated by distillation and identified by elemental
analysis as well as molecular mass determination [55]. Yields ranged from 11±56%. The results of
the elemental analysis agree very well with the calculated values, indicating quite pure products, but
no structural or spectroscopic studies of the compounds were published.
Ammonolysis of methyltrichlorosilane, CH3SiCl3, at ÿ788C in the presence of sodium gives a
cage molecule, (CH3Si)6(NH)9, which was isolated and structurally characterised by single crystal
X-ray diffraction [102] (Eq. (23)):
6HSiCl3 9NH3 18Na ! CH3 Si6 NH9 18NaCl 9H2 (23)
In most cases trichlorosilanes were used together with dichlorosilanes to produce silazane
precursors for Si/C/N-ceramics, because an increased degree of cross-linking can be achieved compared
to pure dichlorosilanes. The ammonolysis of chlorosilane mixtures is discussed in Section 2.2.5.
2.1.4. Ammonolysis of tetrachlorosilane SiCl4

First attempts to prepare Si±N-compounds from SiCl4 and ammonia were reported as early as
the last century [21,103]. More systematic studies were published in the 1950s [104±109]. The
reaction leads to a polymeric solid, silicon diimide [Si(NH)2]n, if a sufficient amount of ammonia is
used (Eq. (24)):
nSiCl4 6nNH3 ! SiNH2 n 4nNH4 Cl or nSiCl4 18nNH3
! SiNH2 n 4nNH4 Cl 3NH3 (24)
The reaction can be performed in many variations using gaseous or liquid ammonia as well as
gaseous, liquid or in an organic solvent diluted SiCl4 [110,111]. After pyrolysis at 800±11008C
amorphous Si3N4 is formed. Diffusion-controlled crystallisation at 1200±15008C in nitrogen results
in a-Si3N4 powders. Particle morphology, size, and phase composition of the products are influenced
by annealing time, temperature and impurities. Compared to silicon nitride powders prepared by
direct nitridation or by carbothermal reduction, the products are characterised by high chemical
purity and high sintering activity, respectively. Today this so-called `diimide process' is one of the
synthetic routes used to produce silicon nitride on an industrial scale [110,111].
The partial ammonolysis of tetrachlorosilane has also been examined [40,112]. In diethyl ether
below ÿ608C hexachlorodisilazane Cl3Si±NH±SiCl3 (yield: 40%), hexachlorocyclotrisilazane
[SiCl2±NH]3 (yield: 3±5%), and polylchlorosilazanes, mainly [SiCl2±NH]x are formed [112]. Due
to the chlorine content these compounds are not suitable precursors for ceramic materials. However,
they can be modified by treatment with amines or related reagents to obtain volatile chlorine-free
molecules for CVD processes or polymers for condensed phase pyrolysis.
2.1.5. Ammonolysis of chlorosilane mixtures Ð copolysilazanes

Several different silane mixtures were used to prepare polysilazanes for pyrolysis experiments.
This is due to the possibility controlling the properties of the produced polymers (viscosity, thermal
stability etc.) simply by varying the stoichiometry. Monochlorosilanes, R3SiCl, may be used as chain
stoppers, dichlorosilanes to form silazane rings and chains, and trichlorosilanes as well as
tetrachlorosilane to introduce branching.
A mixture of dichlorosilane, H2SiCl2, and trichlorosilane, HSiCl3, was used to prepare a
perhydridopolysilazane, which was transformed into nearly stoichiometric silicon nitride by
pyrolysis under an ammonia atmosphere [113]. The reaction was carried out in ether, with an excess
of ammonia at 08C, to give an opaque, viscous liquid after separating the product from NH4Cl and
solvent.
The commercial polymer NCP 200 (Nichimen Corp., Japan, see also Section 2.8), a
polyhydridomethylsilazane (PHMS) is a copolymer prepared by ammonolysis of dichloromethyl-
and dichlorodimethylsilane in a 2:1 ratio as calculated from the corresponding peak areas in the 29Si
NMR spectrum [114].
A mixture of trichlorosilane, HSiCl3, and methylchlorosilane in a 1:3 ratio was reacted with
ammonia to obtain a polysilazane with an average molecular weight of 850 which was used for spray
pyrolysis [115]. The original synthesis conditions were described by Seyferth et al. [116,117].
Similar soluble polysilazanes were prepared analogously from methyltrichlorosilane, CH3SiCl3, and
dichloromethylsilane, (CH3)HSiCl3 [86] and used to produce preceramic polymer/metal powder
composites for pyrolysis to CMCs [118].
Ammonolysis of an equimolar mixture of dichloromethyl- and dichlorovinylsilane in toluene at
08C gave a colourless oil [92]. This precursor was thermally cross-linked at 1208C for 15 h to give a
white gummy-like solid, which was selectively crystallised to SiC or Si3N4 depending upon
processing conditions [95,119±121].
Mixtures of pyridine adducts of chlorosilanes have been used to prepare polysilazane polymers.
Co-ammonolysis of dichlorosilane-pyridine, H2SiCl2py, and dichloromethylsilane-pyridine,
CH3(H)SiCl2py, produced transparent liquids with low viscosity (40±80 mPa s at 308C),
Mn800±1000 and r1.1±1.2 g/cm3. Heating the materials to 100±3008C causes cross-linking
and glass-like hard solids are obtained [122].
Different degrees of branching in the Si±N backbone were realised using varying ratios of
trichlorovinylsilane, CH2=CH±SiCl3, and dichloromethylsilane (CH3)HSiCl3. Ratios of 1:10, 1:5,
1:2, 1:1, and 1:0.5 were studied [97]. The polymer yields varied between 73 and 78%, while the
chemical composition changed from SiC1.19N0.81H4.68 to SiC1.76N1.27H4.60. As expected, this was
accompanied by an increase of molecular weight and viscosity from 498 to 806 g/mol and 0.25 to
23.7 Pa s respectively. Correspondingly, the ceramic yield after pyrolysis (10008C, 5 K/min, Ar)
changed from 28 to 77%.
The chlorosilanes used for the syntheses described above are rather expensive compared to
ammonia or organic monomers used to produce organic polymers. Therefore, distillation residues
obtained from the above-mentioned industrial Mueller±Rochow synthesis were used to prepare
polysilazanes by ammonolysis [123,124]. The mixtures consist of differently substituted
methylchlorodisilanes, which were reacted with gaseous ammonia in aprotic organic solvents. The
products were cured and pyrolysed at 12008C to form amorphous Si/C/N-ceramics with N:C-ratios
of 1:1 to 1:1.5. The ceramic yield varied from 60 to 70%.
2.1.6. Miscellaneous ammonolysis reactions

Ammonolysis of special chlorosilanes may by used to prepare defined silazane compounds
which are useful to study precursor chemistry and the polymer-to-ceramic conversion. An example is
the synthesis of pure cyclotetrasilazanes using 1,3-dihalodisilazanes [125]. A number of methods is
known for the synthesis of the latter [126]. This approach was applied to examine the cross-linking
mechanism of dichloromethylsilane derived silazane mixtures [127] (Eq. (25)).
(25)
Together with similar aminolysis reactions and the cross-linking behaviour of the corresponding
products it was possible to draw conclusions about the structure of the KH cured polysilazanes [127].
2.2. Aminolysis of chlorosilanes
In general, the ammonia used in the above described reactions can be replaced by amines in
order to obtain precursors for Si/C/N-ceramics. In the case of secondary amines R2NH volatile
molecular silylamines are formed (Eq. (26)):
Rx SiCl4ÿx 2nR02 NH ! Rx R02 Nn SiCl4ÿxÿn nR02 NH2 Cl (26)
The products are chlorine-free if sufficient amounts of the amine are present and the substituents
R and R0 are not too large. Due to their volatility these compounds can be used as MOCVD
precursors. However, this review article focuses on compounds containing Si±N±Si units, i.e.
silazanes, so a detailed discussion of silylamines is beyond the scope of this paper.
Tertiary amines do not react with chlorosilanes to form silazanes. If primary amines are used
oligomeric and polymeric products can be synthesised. The resulting products contain nitrogen
atoms bonded to one carbon and usually two silicon atoms as described in Eq. (27):
2nR4ÿx SiClx 3xnR0 NH2 ! NR0 x=2 ÿSiR4ÿx n 2xnR0 NH3 Cl

x 1; 2; 3; 4 for x 1 follows n 1 (27)
Again, it is obvious that for x1 small molecules, i.e. disilazanes are obtained. In all other cases
oligomers, as well as linear or cross-linked polymers, are formed. For ceramic precursor technology
small substituents R and R0 like methyl should be selected. Larger groups lead to an excess of carbon
in the precursor and in turn to low ceramic yields after pyrolysis and/or to the formation of free carbon.
Numerous aminolysis reactions of chlorosilanes with primary amines have been reported in
literature. A few examples will be described here. The reaction of methylamine with
chlorotrimethylsilane in benzene gives heptamethyldisilazane analogous to the synthesis of
hexamethylsilazane [128] (Eq. (28)):
CH3 3 SiCl 2CH3 NH2 ! CH3 3 SiÿNCH3 ÿSiCH3 3 CH3 NH3 Cl (28)
The chlorotrimethylsilane can be replaced by other monochlorosilanes, e.g. chlorodimethylsi-
lane (CH3)2HSiCl [129] or chloromethylsilane (CH3)H2SiCl [39] to form 1,1,2,3,3-pentamethyldi-
silazane and 1,2,3-trimethyldisilazane, respectively.
Dichlorosilanes react with primary amines to give linear or cyclic oligomers. Dichloromethyl-
silane for example reacts with methylamine to form octamethylcyclotetrasilazane [(CH3)HSi±
N(CH3)]4, hexamethylcyclotrisilazane [(CH3)HSi±N(CH3)]3 and pentamethyltrisilazane (CH3)HN±
SiH(CH3)±N(CH3)±SiH(CH3)±N(CH3)H together with minor amounts of other oligomers [130]. If
the same reaction is carried out at ÿ308C in THF an intermediate, the partially aminolysed 1,3-
dichloro-1,2,3-trimethyldisilazane, can be isolated in 43% yield [131]. Similar results were obtained
with dichlorodimethylsilane [125,132].
Trichlorosilanes do not necessarily form cross-linked oligomers or cage-like molecules as it might
be expected. An example is the aminolysis of alkyltrichlorosilanes using five equivalents of methylamine
in pentane. A cyclotrisilazane with three methylamine substituents was isolated [133] (Eq. (29)):
(29)
If the stoichiometry is carefully controlled using an amine:chlorosilane ratio of 3:1 the

corresponding chlorine substituted cyclotrisilazanes are obtained. This is even possible for
tetrachlorosilane. These products were isolated in yields of 27±43% [133] (Eq. (30)):
(30)
Surprisingly, reactions of three or five equivalents of isopropyltrichlorosilane or phenyltri-

chlorosilane with methylamine gave no cyclic oligomers, polymers were obtained instead [133]. It is
also remarkable that aminolysis as well as ammonolysis reactions of trichlorosilanes do not give
adamantane-analogous structures [133]. However, there are also no siloxanes with adamantan-like
structure described in literature. For siloxanes the formation of octamers (cubane-like
silsesquioxanes), hexamers (prismane-like tetracyclo<5.5.1.13.11.15.9>hexasiloxanes), and other cage
structures are known [134]. This has not been reported for the Si±N-system, i.e. ammonolysis or
aminolysis reactions of trichlorosilanes do not give analogous cage or polyhedral structures. An
exception is the above-mentioned amminolysis of methyltrichlorosilane [102] (see Section 1.2.3).

Furthermore, it has been reported that no nitridosilicates or silicon nitrides containing one of the
mentioned cage structures are known [135].
The aminolysis of dichlorosilane H2SiCl2 with methylamine gives oligomers and the
corresponding cyclotetrasilazane [136] (Eq. (31)):
(31)
These silazanes were pyrolysed to give a-Si3N4 with a ceramic yield of 38%. However, by catalytic
cross-linking higher yields are obtainable [136]. This method was applied to strengthen glass by
polymer derived ceramic coatings [137] and to compare the pyrolytic properties of two polysilazanes
having the same elemental composition but having different chemical structures, namely
polymethylsilazane [(CH3)HSi±NH]n and poly(N-methylsilazane) [H2Si±N(CH3)]n [138].
Dichloromethylsilane, (CH3)HSiCl2, was treated with methylamine in toluene at 08C to give an
oligosilazane mixture with Mn250 and Mw260 [92]. This mixture most likely contains the above-
mentioned penta-, hexa- and octamers [130]. It was not possible to cure this mixture thermally at
1208C within 8 days [92]. This was attributed to the fact that dehydrocoupling reactions between
Si±H groups may lead to cross-linking. An analogous reaction with dichlorovinylsilane gave a
similar product with Mn300 and Mw310. Due to the fact that this material allows hydrosilylation
reactions, curing was observed after 1 day at 1208C [92]. Structure and pyrolysis behaviour of the
obtained vinyl-N-methyl silazane have been studied [95].
The aminolysis of 1,3-dichlorodisilazanes can be used to prepare pure cyclotetrasilazanes [127].
An example is shown in Eq. (32):
(32)
These cyclotetrasilazanes were compared with the corresponding ammonolysis products with
the aim of understanding KH catalysed cross-linking (see also Sections 2.2.6 and 2.7) [127].
Further aminolysis reactions have been reported, e.g. reactions of boron-containing
chlorosilanes in order to obtain Si/B/C/N-ceramics which will be discussed later.
2.3. Hydrazinolysis of chlorosilanes
Another approach to producing Si/C/N precursors from chlorosilanes is a hydrolysis-,

ammonolysis- or aminolysis-analogous hydrazinolysis reaction. One reason for the use of hydrazine
was to achieve nitrogen-rich precursors for the preparation of silicon nitride [139]. A second reason
was the idea that UV curable precursors may be obtainable, since siladiazanes, such as
[(CH3)2Si]2N±N[Si(CH3)3]2 or [(CH3)2Si]2N±NH[Si(CH3)3], have been used to prepare silylaminyl
radicals [139].
Hydrazine, H2N±NH2, is a strong base and its chemical behaviour, e.g. in substitution reactions
like the replacement of halogen atoms from silicon, is similar to ammonia. Hence, chlorosilanes
react with hydrazine to form HCl and Si±N bonds. It is possible to replace one, two, three or even all
four hydrogen atoms of the hydrazine molecule with silyl groups. The degree of substitution depends
on several factors, the most important being the stoichiometry, the reaction conditions (especially
temperature) and the size of the substituents on the silicon atoms. The replacement of one hydrogen
atom of the hydrazine molecule by a silyl group leads to small volatile Si/C/N/H-molecules.
Examples are described in the literature [140], but no attempts to use these compounds for ceramic
precursor technology have been reported. If two hydrogen atoms are involved in the hydrazinolysis,
the following basic reaction describes the process (Eq. (33)):
2nR4ÿx SiClx 3xnN2 H4 ! N2 H2 x=2 ÿSiR4ÿx n 2xnN2 H5 Cl

x 1; 2; 3; 4 for x 1 follows n 1 (33)
In the case x1 small molecules of the type (R3Si)HN±NH(SiR3) and/or (R3Si)2N±NH2 are
obtained. For Rmethyl it was found that a 1:1 mixture of both isomers is formed [141]. Similar results
gave the reaction of ethyldimehylchlorosilane with hydrazine. t-Butyldimethylchlorosilane, on the other
hand, furnished only the 1,2-isomer which was attributed to steric effects [141]. For x2 different kinds
of rings and chains may be formed. 1,2,4,5-Tetraaza-3,6-disilacyclohexanes R2Si(NH±NH)2SiR2 have
been isolated from reactions with dichlorodimethyl- and dichlorodiphenylsilane, for example [142].
Preceramic polymers have been synthesised using dichloromethylsilanes R(CH3)SiCl2 with
RH, CH3, CH=CH2 [94]. The reactions were performed in toluene at 20±708C in the presence of
triethylamine, in order to reduce the formation of hydrazine hydrochloride. Oils of low molecular
weight were obtained for RCH3 and CH=CH2. When RH only partially soluble solids are
formed. From elemental analysis and 1H NMR spectroscopy it was concluded that hydrazine is
trifunctional toward R(CH3)SiCl2 for RRCH3 and CH=CH2. This was in contrast to the expected
polymer structure (SiH(CH3)±NH±NH)n [94]. The following structure was suggested for the
products [94,139] (Eq. (34)):
(34)
For RH partial tetrafunctionality was assumed to explain the higher molecular weight and the
results of the elemental analysis. These results correspond well with the observation that
chlorotrimethylsilane results in the trisubstituted product [(CH3)2Si]2N±NH[Si(CH3)3] in >50%
yield [143] and the fact that (CH3)2Si(NHNH)2Si(CH3)2 reacts with two equivalents of (CH3)HSiCl2
to give a ladder polymer [139] (Eq. (35)):
(35)
Based on the above described synthesis of a Si/C/N-precursor from dichloromethylvinylsilane

and hydrazine, commercial products called Pyrofine PV and Pyrofine PM have been developed
[144,145] (see section 2.8.2.). The polymer is a highly viscous liquid at room temperature,
containing 3.2 wt.% oxygen which originates from the hydrazine used for the synthesis. It can be
thermally cross-linked at 3008C to form an infusible and insoluble solid. After pyrolysis at 10008C a
ceramic material containing 7.4 wt.% oxygen is obtained in 60±65% yield [144]. The changes of
viscosity during heating at different temperatures and the pyrolysis behaviour were examined [145].
The material was used to fabricate monolithic ceramic bodies which were characterised with respect
to porosity, crystallinity, mechanical properties and oxidation resistance (see later) [144].
2.4. Silazanolysis of chlorosilanes
A problem which always arises when chlorosilanes are transformed into oligo- or polysilazanes,
using one of the above described methods, is the formation of HCl or the corresponding ammonium
chloride, (R1R2R3NH)Cl. This side product has to be separated from the silazane, usually by
filtration. An elegant way to overcome this problem is to use redistribution reactions of the type
(Eq. (36)):
(36)
The advantage of this approach is that the chlorine-containing side product chlorotrimethylsi-
lane is very volatile and can be easily removed by evaporation. In most cases HMDS and a
chlorosilane are used as starting materials. HMDS is a liquid which can be handled safely and like
ammonia, amines or hydrazine it does not cause problems such as the contamination of the products
with undesirable heteroelements.
In principle, the following simplified general equation describes the syntheses of oligosilazanes
by silazanolysis of chlorosilanes with HMDS, if both trimethylsilyl-groups react (Eq. (37)):
nR4ÿx SiClx xnHNSiCH3 2 ! NHx=2 ÿSiR4ÿx n 2xnClSiCH3 3

x 1; 2; 3; 4 for x 1 follows n 1 (37)
It is also possible to replace only one trimethylsilyl group of the HMDS molecule. For example,
tetrachlorosilane reacts with one equivalent of HMDS to form 1,1,1-trichloro-3,3,3-trimethyl-
disilazane, Cl3Si±NH±Si(CH3)3 [146] (Eq. (38)).
SiCl4 CH3 3 SiÿNHÿSiCH3 3 ! Cl3 SiÿNHÿSiCH3 3 ClSiCH3 3 (38)

1,1,1-Trichloro-3,3,3-trimethyl-disilazane has been used to prepare especially designed
precursors, e.g. for boron containing silazanes (see later). The redistribution reaction involving
HMDS may also be used to prepare many different symmetrically substituted molecular silazanes,
e.g. [126] (Eq. (39)):
2CH3 2 HSiCl CH3 3 SiÿNHÿSiCH3 3 ! CH3 2 HSiÿNHÿSiHCH3 2 2ClSiCH3 3

(39)
Preceramic polymers have been prepared from chlorosilane and HMDS. It was noted that the
basic reactions leading to polymer formation and cross-linking are completely analogous to
hydrolysis, aminolysis or ammonolysis reactions. The first step of the process consists of a
substitution of the chlorine atom, followed by two possible condensation reactions [147,148] (Eq.
(40a±c)):
substitution : SiÿCl CH3 2 SiNH ! SiÿNHSiCH3 3 CH3 3 SiCl (40a)
condensation a : SiÿNHSiCH3 3 ClÿSi ! SiÿNHÿSi CH3 3 SiCl (40b)
condensation b : SiÿNHSiCH3 3 ! SiÿNHÿSi CH3 2 SiNH (40c)
In order to develop a stable, melt-spinable polymer which produces a ceramic material after
pyrolysis, approaching the composition of Si3N4, trichlorosilane was treated with excess HMDS
[149]. In a typical reaction 1 mol of trichlorosilane was mixed with >3 mol of HMDS and stirred.
The exothermic reaction raised the temperature to 758C. Chlorotrimethylsilane (bp 588C) and excess
HMDS were removed by distillation and the temperature was increased up to 200±2308C. A
transparent, colourless solid was obtained in a yield exceeding 95%. This product was characterised
by 1H NMR- and IR-spectroscopy, as well as elemental analysis, gel permeation chromatography,
rheometry, TGA and TMA. It was used to draw fibres which were cured with trichlorosilane and
pyrolysed to obtain almost pure silicon nitride fibres [149].
Mixtures of chlorodisilanes, (CH3)xCl3ÿxSi±SiCly(CH3)3ÿy, which are by-products of the
Mueller±Rochow synthesis of chloromethylsilanes, were also used to prepare polysilazanes. It was
assumed that the incorporation of Si±Si units contributes to enhanced ceramic yields and still
maintains tractability [150]. The crude mixtures of disilanes with the empirical formula
Si2(CH3)2.6Cl3.4 were purified by distillation and the fraction boiling between 145 and 1528C was
combined with HMDS. The mixture was heated to 2758C to yield a clear, colourless and brittle
preceramic polymer. Its formula, (CH3)2.6(Si2)1.0(NH)1.5[NHSi(CH3)3]0.4 was deduced from
elemental analysis and detailed examination of the reaction of pure 1,2-dimethyltetrachlorosilane,
(CH3)Cl2Si±SiCl2(CH3), with HMDS which was examined by GC and 29Si NMR spectroscopy. The
precursor was pyrolysed giving Si/C/N ceramics in yields around 60% [150].
2.5. Miscellaneous routes
In addition to the preparative methods described in Sections 2.1±2.5, a number of syntheses to

produce silazanes, i.e. Si±N±Si containing compounds, are known. Most of them are described in the
review articles mentioned in Section 1 [16,25,26]. A few of these methods have been used for the
synthesis of ceramic precursors [17,18,32]. These include dehalogenation reactions with alkali
metals, i.e. Wurtz-type coupling reactions (Eq. (41)):
Lx EÿCl L0y E0 ÿCl 2Na ! Lx EÿE0 L0y 2NaCl (41)
If difunctional chlorine compounds are used oligomeric chain and ring molecules are formed.
This principle has been used to prepare the so-called polysilasilazanes (PSSZs) which were
transformed thermally to polycarbosilazanes (PCSZs) [151±153]. Equimolar amounts of
dichlorodimethylsilane, (CH3)2SiCl2, and 1,2-dichloro-1,2-dimethyldisilazane, [(CH3)SiHCl]2NH,
were reacted with sodium to form a liquid PSSZ copolymer in 72% yield [151] (Eq. (42)):
(42)
The reactions were performed by adding the chlorosilane mixture to a boiling suspension of
sodium in toluene followed by refluxing for 24 h. Anhydrous gaseous ammonia was bubbled through
the solution after removing the sodium chloride in order to eliminate the remaining Si±Cl bonds. The
product was isolated by filtration and devolatilisation under high vacuum. Different yellowish and
solid PCSZs were obtained by thermal cross-linking at temperatures from 335 to 4708C [151]. An
advantage of this process is that the nitrogen content of the precursor can be continuously modified
by changing the ratio of the two monomers. The products were fusible, soluble in several organic
solvents and melt-spinable. The pyrolysis behaviour of this precursor was studied and fibres were
drawn which were cured by oxygen treatment [154±158].
Another approach to prepare Si/C/N precursor polymers from monomers which already contain
Si±N±Si units is to polymerise the vinyl substituents of molecular vinylsilazane. This idea was
studied using 1,3,5-trimethyl-1,3,5-trivinylcyclotrisilazane as a starting material [159]. This silazane
is a commercially available, relatively inexpensive liquid which is normally used for surface
derivatisation in chromatography. Several common free-radical initiators were tested, but only
dicumyl peroxide was found to effectively polymerise the starting material at 120±1508C. Bulk
polymerisation with 0.5% of the initiator at 1358C for 12 h gave a clear, brittle, insoluble and
infusible polymer in 100% yield. A soluble product was also produced by performing the reaction
in toluene. After removing the solvent the solid cannot be redissolved in toluene. The products
were characterised by size exclusion chromatography and IR spectroscopy. They were pyrolysed
and the ceramic products examined by several IR, X-ray diffraction (XRD), TEM, X-ray
photoelectron spectrum (XPS), elemental analysis, as well as helium pycnometry and surface
area analysis.
Another alternative route to different types of oligo- and polysilazanes is based on
dehydrocondensation reactions involving N±H and Si±H groups [160,161]. This transition metal
catalysed process is similar to the alkali hydride dehydrogenation used to cross-link silazanes (see
later), e.g. (Eq. (43)):
(43)
Various silanes such as (C6H5)2SiH2, (C6H5)SiH3 and (C2H5)3SiH etc. have been reacted with
amines and ammonia. It has been pointed out that any kind of polysilazane can be prepared by this
method which has the advantage to a priori avoid the products being contaminated with any traces of
chlorine. On the other hand, chlorosilanes are usually cheaper than the corresponding hydridosilanes,
and besides, it is more dangerous to handle compounds like (CH3)SiH3 compared to (CH3)SiCl3.
Therefore, it is easier to prepare an oligomer via ammonolysis or aminolysis of a chlorosilane and
modify it in order to obtain a cross-linked ceramic precursor.
To convert cyclosilazane oligomers to materials of higher molecular weight either ring-opening
polymerisation or ring-fusion polymerisation may be used. For oligosilazanes, which in general are
thermolabile, both reactions may occur at the same time, especially at higher temperatures. Some
techniques described in the literature as ring-opening processes will be discussed in this section,
ring-opening and modification with urea as well as ring-fusion related reactions are discussed in
Section 2.6.
Ring-opening reactions of cyclotrisiloxanes using strong nucleophilic reagents like MOR, MOH
or MR are well known and can be applied to prepare linear polysiloxanes [162]. Similar reactions for
cyclosilazanes have been examined, but are more difficult to perform [19]. Stoichiometric and
catalytic ring-opening of hexaalkylcyclodisilazanes by organoalkali reagents has been reported [163]
(Eq. (44)):
(44)
Ring-opening polymerisation of organocyclosilazanes was also initiated by trifluoromethan-

sulfonicacid-methylester [164±168]. High molecular weight linear polysilazanes were obtained
using cyclodisilazanes with small R groups, e.g. (Eq. (45)):
(45)
It is remarkable that these products showed molecular masses up to 100,000, were soluble in
organic solvents and stable to air oxidation.
Linear oligodimethylsilazanes were also reported by Redl et al. to form upon heating
cyclotrisilazanes under ammonia (Eq. (46)):
CH3 2 SiNH3 NH3 ! H2 NÿCH3 2 SiNHx ÿH (46)
However, the rather low conversions around 10% and molecular weights of Mn1200 were
obtained.
Polymerisation of silazanes can also be promoted by acid catalysts like ammonium halides, at
temperatures of 1608C [169,170]. After 6±8 h a condensation polymerisation process leads to the
formation of ammonia and waxy polysilazane products. The polymers were found to have very high
molecular weights of Mn10.000. For these materials a structure consisting of rings linked by silyl
bridges was suggested (Eq. (47)):
(47)
Another approach is to promote ring opening polymerisation by transition metal catalysts. The
process seems to involve hydrogenolysis of Si±N bonds [160,171] (Eq. (48)):
(48)
Because the catalyst in this reaction cleaves Si±N bonds in both the tetracyclic starting material,
as well as the product oligomers, equilibration prevents the growth of high molecular weight species
[19].
Very recently, lithium nitride dissolved in diglyme, H3CO±(CH2)2±O±(CH2)2±OCH3, and

cyclohexyltrichlorosilane have been employed to synthesise organosilazanes [172]. After refluxing
for 2 h a soluble but highly branched polymeric silazane was obtained (Eq. (49)). A high molecular
mass of 26,000 Da was reported.
(49)
The product was characterised by nuclear magnetic resonance (NMR), IR, elemental analysis
and solubility examinations. It contained residues of diglyme and 3.2 wt.% chlorine which certainly
is a problem for its future applications as a precursor for Si/C/N-ceramics.
Transamination and deamination reactions may also by used to prepare preceramic
polysilazanes. In fact, the first example of an Si/C/N ceramic produced from a polysilazane
precursor described in literature, is based on pyrolysis induced deamination/condensation of
aminosilanes [173,174], (50a) and (50b):
(50a)
520 C
RSiNHCH3 3 ÿÿÿÿÿÿÿÿÿ
ÿ! polymer CH3 NH2 (50b)
Rmethyl or phenyl
These processes were optimised and the pyrolysis behaviour of the polymers were examined
[19,175,176]. The polymeric materials were used to prepare amorphous Si/C/N fibres.
2.6. Modification of silazanes
Many of the syntheses described above Ð especially the ammonolysis- and aminolysis-based
methods Ð give mixtures of oligomers which are too volatile to obtain high ceramic yields by
pyrolysis. The reason is simply the evaporation of low molecular weight fractions of the mixture.
Several techniques have been developed to overcome this problem. One approach is the application
of high pressure pyrolysis in an autoclave. This is not necessary, if the oligomers contain reactive
groups like Si±H or vinyl units. Prolonged heat treatment of these mixtures, e.g. by refluxing has
been used to increase the mean molecular weight (see above).
More sophisticated techniques utilise catalysts which enhance the rates of cross-linking
reactions like hydrosilylation, Si±HN±H or Si±HSi±H dehydrocoupling, transamination or vinyl
polymerisation. A well-known process is the treatment of low molecular weight silazanes, containing
±SiH±NH± units with basic catalysts such as KH, which leads to the elimination of hydrogen and the
formation of N±Si bonds and four-membered cyclodisilazane rings [86,177]. This reaction is similar
to the dehydrocoupling of hydridosilanes with ammonia or amines which was used to prepare
silazanes, as described in Section 2.5, except the fact that Si±H and N±H are vicinal groups in the
same molecule [127] (Eq. (51)):
(51)
Later, this process was further modified using methyl iodide, in order to quench the `living'
polymer which contains potassium amide functions [87]. The same dehydrocoupling mechanism
takes place thermally, although slowly, at 1208C [178].
A process based on transition metal catalysed dehydrocoupling has been used to accelerate the
cross-linking [19,136,138,179±181]. Poly(N-methylsilazane) and poly(Si-methylsilazane) have been
treated with 0.2 mol% Ru3(CO)12 for several hours, increasing the molecular weight significantly,
e.g. (Eq. (52)):
(52)
Dehydrocoupling reactions may also be accelerated by the catalytic action of pyridine. It was
reported that perhydropolysilazanes prepared from dichlorosilane in a pyridine solution were treated
with an excess of ammonia, causing dehydrocondensation [80]. However, no experimental proof was
given for this mechanism. Dehydrogenative condensation of cyclosilazanes, induced by heating in
the presence of pyridine, has been substantiated by other authors using 29Si solid state NMR
spectroscopy [83,182]. Alternatively, cross-linking can be explained by the disproportionation of
difunctional silanes in the presence of pyridine to give mono- and trifunctional silanes [183,184].
Alcoholate catalysts were used to prepare preceramic polymers from silazane oligomers which
contain N±H but no Si±H groups like OMCTS. The reactions were carried out using 1 wt.% KOCH3
as a catalyst under reflux conditions at 200±3608C for 6 h. The evolution of ammonia and methane
was detected while NH3 predominantly forms at lower temperatures, indicating transamination
reactions (see above). CH4 was the main product at higher temperatures which was explained by
coupling reactions between methyl and N±H groups [185,186] (Eq. (53)):
(53)
Another strategy to increase the molecular weight and the viscosity of oligosilazanes is to use
redistribution or substitution reactions involving dichlorosilanes. Oligohydridoalkylsilazanes,
(RSiH±NH)n, have been reacted with alkyldichlorosilanes in order to obtain polymeric
chlorohydridosilazanes [187]. Subsequent treatment with ammonia gives soluble polysilazanes
suitable for fibre drawing.
Mixtures of cyclosilazane oligomers obtained in the ammonolysis of dichloromethylsilane have
been converted to materials of higher molecular weight by the reaction with urea [182,188]. This
method was developed on the basis of an older report about the reaction of HMDS with urea giving
N,N-bis(trimethylsilyl)urea [189] (Eq. (54)):
(54)
The synthesis of poly(ureidosilazanes) (PUSZ) was performed in pyridine at 858C, varying the
Si(CH3)H±NH-unit:urea ratio between 4:1 and 12:1. Depending on this ratio, white solids, gums or
oils were obtained. IR, TGA-MS, TGA-FTIR, as well as solid-state and solution-state NMR were
applied in combination with selective isotopic labelling using 13C- and 15N-labelled urea, to examine
the constitution of the poly(ureidosilazanes). It was found that the following simplified polymer
structure was formed, which contained clearly detectable amounts of isocyanate and SiH2CH3 units
originating from side reactions (Eq. (55)):
(55)
The pyrolysis behaviour of these polysilazanes was examined, and it was suggested that they
should find useful applications in polymer derived ceramic fabrication.
Polysilazanes with a tailorable viscosity have been prepared by reacting poly(methylvinyl)si-
lazanes, obtained from co-ammonolysis of dichloromethyl- and dichloromethylvinylsilane, with
mono- and multifunctional isocyanates [190±193].
A commercial product called `CerasetTM SN' based on this synthesis was developed by Dupont
Lanxide Composites (DLC), a company now belonging to AlliedSignal (see Sections 2.8 and 3.2).
The following simplified structure was published by the manufacturer [194] (Eq. (56)):
(56)
In Ref. [194] the nature of R0 is not specified. This substituent originates from the isocyanate
used to modify the silazane. It has been known since the 1960s that silylamines and silazanes react
with isocyanates by insertion reactions to form urea derivatives in high yield [195] (Eq. (57)):
(57)
However, carbonyl groups were not detectable in the `CerasetTM SN' polymer by FTIR, 1H or
13
C NMR spectroscopy, indicating that only very small amounts of isocyanate are used to prepare
this product [196]. The low viscous liquid polyureasilazane is recommended to be cured through the
addition of free radical initiators such as organic peroxides, but it may also by cross-linked thermally
at 250±4508C.
2.7. Introduction of heteroelements (B, Al, Ti, . . .)
Multinary ceramics and composite materials are of interest due to their enhanced properties. For
example, the addition of boron nitride leads to a better thermoshock resistance of hot-pressed silicon
nitride ceramics [197]. Precursor derived ceramics may have a tendency to show better high-
temperature properties, i.e. crystallisation onset temperatures, going from binary through secondary
to ternary phases [32].
Precursor derived Si/B/C/N ceramics have a number of advantages compared to the Si/C/N
system, including higher thermal stabilities and oxidation resistance, as well as higher crystallisation
temperatures and ceramic yields. They have even been reported to function as self-sealing barriers
for the oxidation of carbon/carbon composite materials [198]. Therefore, the fabrication of boron
containing Si/C/N ceramics, including fibres and coatings, is currently under development, with the
target to produce composites for aircraft and spacecraft components as well as other applications
[198±200]. The synthesis and properties of some Si/B/C/N materials have been summarised in recent
reviews [4,13,32,202,203].
The first Si/B/C/N-polymers which have been pyrolysed to form corresponding ceramic
materials were prepared in 1984 [204,205]. Mixtures of polydimethylsilanes [(CH3)2Si]n with B-
trimethyl-N-triphenylborazine [(CH3)B-N(C6H5)]3 were either cross-linked in an inert atmosphere
using an irradiation method, or in a second variant oxidatively treated at 250±5008C. Subsequently,
fibres were drawn from the obtained polyorganoborosilazanes. An enhanced tensile strength
compared to pure SiC fibres was reported. Furthermore, the material was utilised to produce SiC-
based ceramic composites [206].
Based on earlier work on the reaction of diborane with HMDS by NoÈth [207], cyclic silazanes
derived from dichloromethylsilane were reacted with diborane or borane adducts in toluene solution
to form boron containing polysilazanes [208,209]. The products were viscous oils or solids,
depending on the borane to silazane ratio. 11B NMR spectroscopic investigations showed that
borazine rings are present, together with small amounts of species with BN3 connectivity. Eq. (58)
gives a simplified description of the reaction principle which can be described as a
dehydrocondensation between N±H and B±H groups, followed by rearrangements to borazine rings
and SiH2(CH3) groups (Eq. (58)):
(58)
The silazanyl side-chains in the product can react further with H3BS(CH3)2 forming a boron-
containing polymeric silazane network.
Boron-doped silazane polymers have also been synthesised by thermal condensation reactions
of polyhydridopolysilazanes, derived from silazanolysis of trichlorosilane (see Section 2.4 [149])
with liquid borazine at 60±1008C [203,210,211]. The polymers can be prepared with molecular
weights of Mn5000±24000 and boron contents of 2±17 wt.%. Spectroscopic investigations show
that the silazane backbone is retained and modified with pendant borazine rings (Eq. (59)):
(59)
Borazine-silazane copolymers have been synthesised similarly by thermal dehydro-condensa-

tion of borazine with HMCTS or tris(trimethylsilylamino)silane [203,212,213]. The spectroscopic
analysis of both series of copolymers revealed complex, branched structures, as indicated in the
following reaction scheme (Eq. (60)):
(60)
A similar dehydrocoupling reaction of tetrakis(methylamino)silane with a borane adduct was

applied to synthesise another kind of polyborosilazanes [200]. The ratio of silicon to boron was
varied from 10:1 to 1:1. It was reported that processing properties such as solubility, viscosity or
meltability are tuneable over a wide range (Eq. (61)):
(61)
A hydridomethylsilazane co-polymer of the form [(CH3)2Si±NH±]m[CH3SiH±NH±]n has been

treated with tris(dimethylamino)borane B[N(CH3)2]3 to obtain a boron-containing polysilazanes
[214±216]. Transamination and cross-linking reactions occurred upon refluxing the reaction mixture
for 340 h. Pyrolysis at 10008C under argon provided a Si/B/C/N ceramic in 74% yield containing
1.1 wt.% boron.
Similarly, the treatment of perhydropolysilazane (H2Si±NH)n with trimethoxyborane (CH3O)3B
results in the formation of boron doped precursors [82,217,218]. The reaction proceeds primarily by
the evolution of methanol and the formation of B±N bonds. However, the products are not free of
oxygen. An O:B atomic ratio of 1.09 was determined. The thermally induced polymer-to ceramic
transformation was studied and an increased high-temperature stability compared to boron-free
silicon nitride ceramics was reported.
Hydroboration reactions of unsaturated alkylsilanes and the subsequent pyrolysis in ammonia,
gives ceramic products in the quaternary Si/B/C/N system [219,220]. If dichloromethylvinylsilane
and dimethylsulfide-borane are used, a boron-containing chlorosilane is formed in high yield [221].
This Si/B/C/N monomer was ammonolysed to give a polyorganoborosilazane with 4 wt.% boron
content [11,222,223]. The ceramic materials obtained after pyrolysis are stable against
decomposition up to 20008C and start to crystallise above 17008C [14,222] (Eq. (62)):
(62)
The polymer was isolated in 80% yield. It consists of a white solid powder soluble in THF and
toluene. FTIR and NMR examinations of the product supported a trigonal planar co-ordination for boron.
It was concluded that cyclic oligosilazanes are formed, which are two-dimensionally cross-linked via
Si±C±C±B±C- and Si±C±B±C-bridges [14]. In order to decrease the boron content of these polymers,
co-ammonolysis reactions with dichlorodimethylsilane were performed [222,223] (Eq. (63)):
(63)
An increased boron content of 14.2 and 16.3 wt.% was realised by treating the
organopolyborosilazane with dimethylsulfide-borane. However, this method does not necessarily
give a homogenous distribution of boron in the polymer. Therefore, the hydroboration of vinylsilanes
was also applied to the dimethylsulfide adducts of chloroboranes like ClBH2 or Cl2BH [224]. For
example, dichloromethylvinylsilane or trichlorovinylsilane react with ClBH2S(CH3)2 to give the
corresponding a-hydroboration product (Eq. (64)):
(64)
The result of the replacement of borane by mono- and dichloroborane is monomers with
decreased Si:B ratios of 2:1 and 1:1, compared to 3:1 for BH3. Furthermore, the products were
modified by reactions with HMDS to replace the chlorine atoms bonded to boron by ±NH±Si(CH3)3
groups, if the corresponding stoichiometric amounts are used. This behaviour is analogous to the
silazanolysis reactions discussed in Section 2.4. Accordingly, the B±Cl groups may also be replaced
by B±NH±B units if two equivalents of chlorine atoms are reacted with one equivalent of HMDS.
The following range of different monomers was obtained and ammonolysed to give the
corresponding polyborosilazanes (Eq. (65)):
(65)
Upon ammonolysis, both the chlorine atoms bonded to silicon as well as those bonded to boron
react to form the polyborosilazane network. This network contains cyclosilazane units together with
boron containing heterocycles. Borazin rings can form if ±BCl2 or ±NH±BCl± groups are present in
the monomer.
Other polyborosilazane syntheses were developed based on the reaction of HMDS with
tetrachlorosilane to give 1,1,1,3,3,3-trichlorotrimethyldisilazane (TTDS, see also Section 2.4) [146].
TTDS forms (trichlorsilylamino)dichloroborane (TADB) upon treatment with BCl3 (Eq. (66)):
SiCl4 BCl3
H3 C3 Si2 NH ! Cl3 SiÿNHÿSiCH3 3 ! Cl3 SiÿNHÿBCl2 (66)
HMDS ÿH3 C3 SiCl TTDS ÿH3 C3 SiCl TADB
Ammonolysis with liquid NH3 at ÿ788C provides an insoluble solid, which forms Si3B3N7 after
pyrolysis. A yellow solid with a softening point of 1308C is obtained by aminolysis using
methylamine instead of ammonia. This soluble N-methylpolyborosilazane is isolated at 80% yield
with an average molecular weight of 2500 g/mol. It can be spun into fibres by melt- or dry-spinning
(see later). Reaction with dimethylamine provides a molecular Si/B/C/N compound, suitable for
CVD and CVI processing [32,198,199,201] (Eq. (67)):
(67)
These Si/B/C/N-precursors as well as their application for fibre drawing and the production of
coatings and composites etc. have been developed by the Bayer AG (Germany) to become
commercial products.
Besides boron, other elements have been introduced into silazanes, in order to obtain precursors
for quaternary Si/E/C/N ceramics. A promising candidate is aluminium since AlN, Al2O3, SiAlONes
and other Al-containing materials show interesting properties such as high strength, high hardness or
resistance to thermal shock and oxidation. Polysilazane derived ceramic materials in the Si/Al/O/N-
system have been found to show suppressed grain growth compared to aluminium-free analogues
[82]. Recently, high-strength alkali-resistant Si/Al/C/O-fibres have been produced from a
polyaluminocarbosilane [225]. Oxygen-free precursors for Si/Al/C/N-materials have been
synthesised by the pyrolysis of polyaluminosilazanes with varied Si:Al ratios, at varied pyrolysis
conditions [200,226±228].
Single-source precursors have been obtained from the reaction of ethylaluminum compounds
with 1,3,5-trimethyl-1,3,5-trivinylcyclotrisilazane [226]. The polymers are formed by a known
condensation process which occurs between the N±H groups of the silazane and the
organoaluminum reagent to form Al±N bonds and evolve ethane [229]. The same reaction type
takes place when HMCTS and triethylaluminum (TEA) are mixed at room temperature either neat or
with both compounds dissolved in dry hexane [228] (Eq. (68)):
(68)
The Si:Al ratios were varied from 1 to 5 for Rmethyl and from 1 to 9 for Rvinyl. The air-
sensitive products were examined by thermogravimetric analysis, FTIR, 29Si, 13C, and 27Al MAS
NMR spectroscopy. The polymers were pyrolysed and crystallised at temperatures up to 19008C to
give SiC/AlN-composite materials.
Aluminium-doped silazane polymers were also obtained, analogous to the above described
reactions of borane with tetrakis(methylamino)silane [200]. The alane Lewis-base adduct
AlH3N(CH3)3 dissolved in toluene was mixed with Si(NHCH3)4 at ÿ788C. Hydrogen evolution
indicates a dehydrocoupling reaction to give a colourless solid polymer containing 12.4 wt.%
aluminium (Eq. (69)):
(69)
A method similar to the above-described preparation of Si/B/C/N precursors by the treatment of

commercially available polysilazanes with tris(dialkylamino)borane B[NR2]3 has been utilised to
obtain phosphorous containing silazanes [216]. Tris(dimethylamino)phosphane P[N(CH3)2]3 was
mixed with mesitylen and polyhydridomethylsilazane derived from dichloromethylsilane and
dichlorodimethylsilane as described in Section 2.1.5. Thermal cross-linking applying 4008C for 4 h
gave a colourless solid which was pyrolysed at 10008C. A very low phosphorous content of 0.5 wt.%
was found. This value decreased further to 0.4 and 0.0 wt.% upon heating to 1400 and 15008C for
50 h, respectively. Besides, it was found that Si/P/C/N ceramics tend to crystallise faster than Si/C/N
materials which is opposite to the behaviour of analogous Si/B/C/N systems.
Besides boron, phosphorous or aluminium transition metals like titanium or zirconium may be
incorporated in silazane precursors. This is a useful task since cross-linking and curing may be
positively influenced by the presence of these metals, for example giving higher ceramic yields. TiN/
Si3N4-composites have been suggested as structural materials because of their mechanical properties
[230]. Polytitanocarbosilanes are already used for the production of the Si/Ti/C/O fibre `Tyranno' by
UBE Ltd. in Japan [231]. However, only a few molecular and polymeric Si/Ti/C/N precursors have
been prepared and thermally transformed into ceramic materials [232±234]. As was discussed for the
preparation of the Si/B/C/N-precursor TADB an analogous titanium compound is provided by the
reaction of TiCl4 with HMDS [232] (Eq. (70)):
TiCl4 CH3 3 SiÿNHÿSiCH3 3 ! CH3 3 SiÿNHÿTiCl3 ClSiCH3 3 (70)
The red-orange coloured crystalline solid is isolated at 58% yield. It was pyrolysed in nitrogen
atmosphere at 600, 800 and 10008C to give TiN.
Polytitanosilazanes were synthesised by treating different oligosilazane mixtures with
tetrakis(dimethylamino)titanium and tetrakis(diethylamino)titanium [216,234]. A homogenous
distribution of Si and Ti results in a statistical attachment of Ti[N(CH3)2]3 groups to the silazane
backbone (Eq. (71)):
(71)
For R0 methyl, the silazane mixture consists mainly of ring systems with m0 and n3±4, as
described in Section 2.1.2. It reacts only very slowly with tetrakis(dimethylamino)titanium in
refluxing toluene. For R0 vinyl or hydrogen m1 and n<15, these silazane mixtures react more
readily with the titanium compounds to give brown viscous resins or black solids respectively. The
products were characterised by IR and NMR spectroscopy, as well as thermogravimetric analysis.
The correlation between ceramic yield and Si:Ti ratio in the precursor starting mixture was
examined. A maximum of 72% was found for a Si:Ti ratio of 5.
2.8. Commercially available Si/(E)/C/N-precursors
Sections 2.1±2.7 clearly show that a large number of studies concerning the synthesis and
modification of molecular, oligomeric and polymeric compounds containing Si±N±Si units have
been conducted. Many laboratories have developed their own precursor systems and may even
synthesise the precursors on a semi-technical scale. However, the properties of the ceramic products
not only depend on the composition and structure of the starting material, but also are strongly
influenced by the processing parameters. Therefore, it should always be kept in mind, that
commercially available Si/C/N- and Si/E/C/N-precursors exist. In order to make the results of
different laboratories comparable, the use of commercial products is recommended.
2.8.1. Molecular compounds

Molecular Si/C/N compounds are usually volatile and suitable for gas phase pyrolysis methods
like CVD or CVI. By far the most important exponent of this group of ceramic precursors is HMDS
(see Section 3.3). A number of other molecular compounds are available, examples are listed in
Tables 3 and 4.
2.8.2. Polymeric compounds

Compared to molecular silazanes significantly less silazane oligomers and polymers are
available commercially. Some systems have been developed but are not produced anymore. An
overview is given in Table 5.
3. Processing from Si/(E)/C/N precursors to ceramic materials
3.1. Historical sketch and overview
3.1.1. Historical sketch

The availability of a wide variety of polysilazane precursors with selected physical and
chemical properties offers multiple ways to perform polymer-to-ceramic conversion and allows the
formation of different geometries of Si-based ceramics such as fibres, coatings, films and bulk
materials. The polysilazane precursors can also be applied as binders for conventional ceramic
processing or as a liquid medium for infiltration of porous substrate materials. Depending on the
processing conditions the structure of the ceramics can either be amorphous, nanocrystalline or
crystalline. Their compositions change from binary (Si/N), over ternary (Si/C/N) ceramics to
quaternary systems (Si/E/C/N), where E refers to an additional element, like B, Al, Ti, P or Zr. Since
the pioneering work of Yajima and co-workers who synthesised SiC fibres from polycarbosilazanes,
various polysilazanes were developed and were processed to generate Si-based ceramics. Table 6
lists some of the related work based on the Si/(B)/C/N system in chronological order.
Table 3
A selection of commercially available linear molecular silazanes
Name Structure CAS-number Properties Supplier
1,1,3,3-Tetramethyldisilazane 15933-59-2 Flammable liquid; bp: 99±1008C Acros, Fisher, Gelest, Lancaster,
Sigma (Aldrich)

1,1,1,3,3,3-Hexamethyldisilazane (HMDS) 999-97-3 Flammable liquid; bp: 1258C Acros, Alfa, Fisher, Gelest,
Lancaster, Sigma etc.
1,3-Diphenyl-1,1,3,3-tetramethyldisilazane 3449-26-1 bp: 96±1008C/0.1 Torr Sigma, ABCR

(DPTMDS)
1,3-Divinyl-1,1,3,3-tetramethyldisilazane 7691-02-3 Combustible liquid; bp: 161±1638C Gelest (ABCR), Lancaster, Sigma
1,3-Dimethyl-1,1,3,3-tetraphenyldisilazane 7453-26-1 mp: 88±908C; bp: 218±2208C/1.5 Torr Acros, Gelest (ABCR), Sigma
1,3-Bis(chloromethyl)-1,1,3,3-tetramethyldisilazane 14579-91-0 Combustible liquid; bp: 103±1058C Lancaster, Sigma
1,3-Di-n-butyl-1,1,3,3-tetramethyldisilazane 82356-80-7 bp: 818C/2 Torr Gelest (ABCR), Lancaster
1,3-Dimethyl-1,3-diphenyl-1,3-divinyldisilazane 23038-10-0 bp: 1188C/1 Torr Gelest (ABCR)
Heptamethyldisilazane 920-68-3 bp: 148±1508C Gelest (ABCR)
Tris(trimethylsilyl)amine or nonamethyltrisilazane 1586-73-8 mp: 67±698C; bp: 728C/10 Torr Gelest (ABCR), Lancaster, Sigma
or nitrilotris(trimethylsilane)
127
128
Table 3 (Continued )
4-Bromo-N,N-bis(trimethylsilyl)aniline 5089-33-8 bp: 156±1588C/23 Torr Sigma±Aldrich
Germanium(II)-bis[bis(trimethylsilyl)amide] 55290-25-0 mp: 32±338C; bp: 608C/0.4 Torr Gelest (ABCR)
Tin(II)-bis[bis(trimethylsilyl)amide] 59863-13-7 mp: 37±388C; bp: 1128C/0.05 Torr Sigma, Gelest (ABCR)
2,2-Thiobis(1,1,1,3,3,3-hexamethyldisilazane) 18243-89-5 mp: 65±678C; bp: 1428C/15 Torr Sigma
Tris(dimethylaminosilyl)aminobis 155633-91-3 bp: 2698C Bayer AG

(dimethylamino)borane
Table 4
Selected commercially available cyclic molecular silazanes
2,2,5,5-Tetramethyl-2,5-disila-1-aza-cyclopentane 7418-19-1 bp: 928C/185 Torr Gelest (ABCR)

1,1,3,3,5,5-Hexamethyl-cyclotrisilazane (HMCTS) 1009-93-4 mp: ÿ108C; bp: 186±1888C Acros, Fisher, Fluka, Gelest,
Lancaster, Sigma (Aldrich)
1,2,3,4,5,6-Hexamethyl-cyclotrisilazane 2587-46-4 bp: 568C/5 Torr Gelest (ABCR)
1,3,5-Triethyl-2,4,6-trimethylcyclotrisilazane 17921-78-7 bp: 85±868C/12 Torr Gelest (ABCR)
1,3,5-Trimethyl-1,3,5-trivinylcyclotrisilazane 5505-72-6 bp: 106±1098C/11 Torr Gelest (ABCR)
1,1,3,3,5,5,7,7-Octamethyl-cyclotetrasilane (OMCTS) 1020-84-4 mp: 978C; bp: 2258C Gelest (ABCR), Lancaster
1,3,5,7-Tetraethyl-2,4,6,8-tetramethyl-cyclotetrasilazane 211935-45-4 bp: 1458C/1 Torr Gelest (ABCR)
1,3,5,7-Tetramethyl-1,3,5,7-tetravinyl-cyclotetrasilazane 5162-63-0 bp: 104±1078C/4 Torr Gelest (ABCR)
129
130
Table 5
Commercially available oligomeric and polymeric silazanes
Supplier Product name Composition/structure Properties Availability References
Allied Signal Composites Inc. CERASETTM SZ Colourless liquid: mp <ÿ788C, Available [194,236]
bp >1508C, d0.96 g/ml,
Z(258C)0.05 Pa s,
Z(1008C)15 mPa s

Bayer AG (Germany) Oligoborosilazane [(NMe)1.5Si±NH±B(NMe)]n M500 g/mol, d1.1 g/ml, Not available [235]
Z(258C)0.08 Pa s, any more
Z(1608C)4 mPa s
Bayer AG (Germany) Polyborosilazane Same as oligoborosilazane, M30 000 g/mol, Not available [235]
but thermally cross-linked d1.2 g/ml, Z(1608C)100 Pa s any more
Elf Attochem (France) Pyrofine PV d1 g/ml, Z(508C)1 Pa s, Not available [94,139,237]
Z(1508C)0.02 Pa s any more
Gelest Inc. (USA) PSN-2M01: d1.04 g/ml, Z>0.05 Pa s, Available [238]

poly(1,1-dimethyl- M500±900 g/mol,
silazane) telomer Tgÿ828C
Gelest Inc. (USA) PSN-2M02: Z>1 Pa s Available [238]

poly(1,1-dimethyl-
silazane) cross-linked
Gelest Inc. (USA) PSN-2M11: d0.99 g/ml, Z0.5±0.8 Pa s Available [238]

poly(1,2-dimethyl-
silazane)
HoÈchst AG (Germany) ET 70 ±[CH3CH2Si(NH)1.5]n± M1600 g/mol, 70 mol% Not available [216,239]

solution in THF any more
HoÈchst AG (Germany) VB 40 ± Not available [216,239]
any more
HoÈchst AG (Germany) VT 50 ±[CH2=CH±Si(NH)1.5]n± M2000 g/mol, 50 mol% Not available [216,239]

solution in THF any more
Nichimen Inc. (Chisso, Japan) NCP 100 [(Me2Si)2NH]x[MeSiHNH]y[MeSiN]z M4000 g/mol, meltable Not available [216,240]

solid any more
Nichimen Inc. (Chisso) NCP 200 [Me2SiNH]x[MeSiHNH]y[MeSiN]z M1100 g/mol, meltable Not available [216,240]
with xyz0.5 solid any more
Tonen Corp. (Japan) NN 110 ±(SiH2±NH)n±, pure PHPS Mn700 g/mol, 20 wt.%, Available [278]
(perhydropolysilazane) soluble in xylene
Tonen Corp. (Japan) NN 310 ±(SiH2±NH)n± with N(SiMe)2 Mn900 g/mol, 20 wt.%, Available [278]
groups soluble in xylene
Tonen Corp. (Japan) NN 710 (SNC) ±(SiHMe±NH)x±(SiH2±NH)y± Mn900 g/mol, 20 wt.%, Available [278]
soluble in xylene
University of Bayreuth, HVNG M620 g/mol, Available [422]

IMA1 (Germany) Z(208C)29 Pa s
University of Bayreuth, HPS M440 g/mol, Available [422]

IMA1 (Germany) Z(208C)0.05 Pa s
University of Bayreuth, ABSE M1100 g/mol, Available [422]

IMA1 (Germany) solid, mp1008C
131
Table 6
Selected chronological list of some (poly)silazanes and polyborosilazanes used for producing Si/(B)/C/N (Si3N4±SiC)
ceramics
Precursors Ceramics Year References
Polycarbosilazane Si/C/N fibres 1974 [173]
Polycarbosilazane Si3N4±SiC fibres 1975 [241,242]
Poly(silsesqui-N-methylsilazane) Si/C/N fibres 1982 [243]
Polyvinylsilazane Si/C/N 1982 [74,117,177]
Poly(N-methylsilazane) Si3N4±SiC fibres 1984 [245]
Polymethylsilazane Si3N4±SiC fibres 1986 [244]
Polymethylsilazane Si3N4C 1986 [136,138,246]
Polysilazane Si/C/N fibres 1987 [247]
Polyhydridodisilazane Si3N4±SiC fibres 1987 [149]
Polyethylenesilazanes Si/C/N 1988 [248]
Polymethyldisilazane Si3N4±SiC fibres 1988 [249]
Polysiladiazanes Si/C/N powders 1989 [94]
Polysilacyclobutasilazane Si/C/N powders 1990 [250]
Octamethylcyclotetrasilazane Si/C/N 1990 [186]
Polyborsilazane Si/B/C/N 1990 [208]
Polyvinylsilazane Si/C/N powders 1991 [93]
Polydihydrosilazane Si/C/N bulk 1992 [1]
Polymethylsilazane Si3N4±SiC 1992 [252]
Cyclodisilazane Si/C/N 1992 [164,253]
Polyborosilazane Si/B/C/N 1992 [199]
Polysilacyclobutasilazane Si/C/N/(O) 1992 [90]
Polycarbosilazane Si/C/N 1993 [151]
Polyvinylmethylsilazane Si/C/N powders 1993 [119]
Poly(hydrazinomethylsilane) Si/N/(C0.5H0.5) 1995 [254]
Poly(ureidosilazanes) Si/C/N 1995 [188]
Polyvinylsilazane Si/C/N bulk 1996 [251]
Polyborosilazane Si/B/C/N bulk 1996 [14]
Poly(disilazane) Si/C/N 1998 [124]
The study on polymer derived ceramics from polysilazanes was started by Verbeek [173] in
Germany in 1974 who synthesised N-methyl polysilazanes by reacting trichloromethylsilane with
methylamine and successfully prepared Si/C/N fibres by melt-spinning from the polycarbosilazanes.
The Si/C/N fibres derived from the polysilazane reveal superior strength and elastic modulus
comparable to the conventionally SiC fibres. Additionally, an excellent oxidation resistance was
observed up to 12008C. Therefore, the synthesis of ceramic fibres became a research field of great
interest and various kinds of polysilazanes were synthesised and transformed into Si/C/N (or Si3N4±
SiC) fibres (see Table 6) Before the mid 1980s, most research work on processing of polysilazanes
was focused on developing Si3N4±SiC fibres, including the improvement of fibre drawing
techniques. An important precursor developed in this period is polyhydridomethylsilazane which
was produced by Seyferth and co-workers by cross-linking cyclic silazane oligmers using strong
base catalyst in 1984 [74,117]. The cross-linked polysilazane gives a high ceramic yield (85 wt.%)
which is important for an economic polymer-to-ceramic conversion.
In the last 10 years, the synthesis of amorphous Si/C/N powders from polysilazane precursors
has attracted increasing scientific interest [94,186,250]. The major motivation for the synthesis of
Si/C/N powders was to develop nano/nano or nano/micro Si3N4±SiC ceramic composites which
are of great interest because of their high mechanical performance which was demonstrated by
Niiharra et al. in 1991 [2,255,256]. Various investigations show that the polysilazane derived Si/C/N
powders can be fully densified in the presence of sintering aids under hot-pressing conditions.
The obtained ceramics exhibit high fracture toughness and high-temperature stability that is
superior to comparable monolithic coarse grained Si3N4 and SiC ceramic materials. Due to the large
effects of phase composition and microstructure on the material properties, the synthesis of Si/C/N
powders with controlled elemental and phase composition is the subject of current investigations
in this field.
Apart from the traditional powder sintering route to form Si3N4±SiC bulk ceramics, a
bulk conversion process via pyrolysis of preceramic polymer compacts was developed which
enables the formation of dense amorphous Si/C/N ceramics having various three-dimensional
geometries [1]. In this process, dense ceramics are generated directly through reaction bonding
during pyrolysis. The conversion into bulk ceramics, as well as the changes during pyrolysis, are
currently not completely understood and continue to be an interesting subject for material scientists
and chemists.
3.1.2. Overview: precursor properties and processing

The processing of precursor to ceramic conversion can be conducted as a solid, liquid or vapour
process. The different ways of ceramic processing or the applicability of a precursor for a specific
ceramic conversion greatly depend on the physical and chemical properties of the selected precursors
such as solubility, rheology, infusibility, the behaviour of cross-linking and pyrolysis. There are
general relationships between the properties of the precursor and the chosen processing technique
which are summarised in the following paragraphs [257±261]:
1. Silazanes with high volatility can be applied in CVD processes to synthesise ceramic powders or
to prepare coatings and films. CVI may be used to produce composite components starting from
porous preforms. A vapour phase of the precursor can be generated either by evaporating the
molecules above their boiling point at a certain pressure or by using a carrier gas bubbling
through the liquid precursor below its boiling point. It is necessary that the precursors do not
cross-link or decompose upon heating to the required temperature.
2. Liquid, meltable solid or soluble polymers can be applied in dry- or melt-spinning of fibres or in
the fabrication of coatings, films or laminates. Precursors with a low viscosity and a high ceramic
yield are needed for infiltration of porous materials or as binders for ceramic powder compaction.
The liquid polysilazanes also can be used to produce fine ceramic powders or films via liquid
processes like spray pyrolyses and spray coating.
3. Highly cross-linked and infusible precursors are suitable for the synthesis of ceramic powders or
bulk ceramics respectively. The infusibility is required for processing the polymeric powders and
for bulk pyrolysis. Additionally, a high degree of cross-linking is essential to ensure a high
ceramic yield and a high density of the pyrolysed bodies.
By design of reactant functionality, a polysilazane precursor with tailored chemical and physical
properties can be achieved according to the well-established chlorosilane chemistry described in
Section 2. Fig. 2 shows the general behaviour of viscosity versus temperature for polysilazanes,
(CH3SiH±NH)x(CH3SiN)y having different degrees of cross-linking [258]. For the meltable
silazanes, spinning and coating is possible in a temperature range where a suitable viscosity is
attained but no enhanced cross-linking and decomposition occurs. For the infusible polymer, shaping
of the precursor powders is carried out around the glass temperature where the polymer particles
show softening and allow plastic forming. Therefore, the viscosity of preceramic polysilazane
precursors can be tailored using the synthesis and cross-linking chemistry described in Section 2.
This, in turn can be applied for different processing techniques to obtain ceramic materials.
Fig. 2. Schematic diagram showing the viscosity vs. temperature for fusible and infusible polysilazanes [258].
3.2. Cross-linking and pyrolysis
3.2.1. Cross-linking
The as-formed silazane products obtained by ammonolysis, aminolysis or related reactions (see
Section 2) of chlorosilanes are generally of low molecular weight which gives a very low ceramic
yield upon pyrolysis. Cross-linking of the oligomeric silazanes into large molecules with highly
interlocked backbones is needed to avoid the evaporation of oligomers during pyrolysis [259].
Oligomeric silazanes can be cross-linked either thermally or chemically using radical initiators
like peroxides or catalysts. Seyferth et al. [75] successfully cross-linked the cyclic oligomeric
silazanes, produced from ammonolysis of methyldichlorosilane or trichlorosilanes using strong base
catalysis. The cross-linking occurred through dehydrocyclodimerisation via Si±H and H±N groups
without ring opening as described in Section 2.6 (see Eq. (51)). A high ceramic yield of 85 % was
obtained after pyrolysis of the cross-linked products in inert gases. They also processed cyclo-
(Me2SiNMe)2 and cyclo-(Me2SiNiPr)2 under chemical cross-linking using organoalkali reagents
(RLi, RLi/Me3CONa, RLi/Me3COK) and observed that addition of catalytic amounts of RLi, RLi/
Me3CONa could induce ring-opening polymerisation of the cyclo-(Me2SiNMe)2 (Section 2.6) [163].
The dehydrogenation condensation of silazane rings can also be induced under transition metal
catalysis, as used by Laine et al. [161,262,263], who processed a series of cyclic silazanes by this
method. One example is cross-linking of N-methyl silazane using Ru3(CO)12 which leads to the
transamination of the liquid polymer to gels or resins upon heating up to 908C (see Eq. (52)). An acid
catalyst was also used in thermal polymerisation of polysilazane precursors [264].
The oligomeric silazane can also be thermally cross-linked into non-volatile polysilazanes when
reactive functional groups such as vinyl units are present in the precursors (see Sections 2.2 and 2.5).
On consideration of the control of elemental and phase composition of the final ceramics,
thermal cross-linking is favourable because no additional elements are introduced to the polymer.
The major cross-linking reactions involved in the pyrolysis polysilazanes include hydrosilylation of
Table 7
Cross-linking conditions and reactions of several kinds of polysilazanes
Polymer Cross-linking conditions Major reactions References
PHMS (NCP200) 300±4008C in Ar Dehydrogenation via Si±H and N±H [114,265]
PVS (VT50) 200±3008C in Ar Hydrosilylation, transamination, vinyl-polymerisation [114]
PUMVS (Ceraset) 250±4008C in Ar Hydrosilylation and addition of vinyl groups [190]
90±1308C in Ar, Vinyl-polymerisation [196,266,267]
0.1 wt.% peroxide
vinyl groups (Eq. (14) in Section 2.1.2) in the case of vinylsilazanes, dehydrocoupling (Eqs. (17) and
(18)), transaminations (Eq. (15)) and vinyl polymerisation (Eq. (16)) [17,92,264,265]. Choong Kwet
Yive et al. [92] systematically investigated the effects of the different functional groups on the
thermal cross-linking behaviour of a series of oligomeric silazanes with successively decreasing
numbers of active groups (vinyl, Si±H and N±H) relative to the inactive methyl groups. They found
that the time for the transformation of the starting liquid to gel-like polymer under heating (1208C)
decreased significantly with an increasing number of active vinyl and Si±H groups. Based on their
experimental results, the cross-linking activity can be arranged in the following order:
hydrosilylation>dehydrocoupling>transaminations>vinyl group polymerisation. Further cross-link-
ing of polysilazane may be needed to attain the state of infusibility.
Several kinds of commercial polysilazanes including NCP 100 and NCP 200 (Chisso Corp.
Tokyo, Japan), VT 50 and ET 70 (Hoechst AG, Frankfurt, Germany) as well as Ceraset (Lanxide
Corp., USA) are cross-linkable under heating. The cross-linking conditions that are generally used in
ceramic conversion and the associated cross-linking reactions are summarised in Table 7. Cross-
linking of NCP 200 is conducted between 350±4508C for transforming the as-purchased low cross-
linked polysilazane into infusible products as required for bulk pyrolysis. Dehydrogenation via Si±H
and N±H groups is the major cross-linking mechanism under these conditions [1]. Thermal cross-
linking of Si±H free polyvinylsilazane occurred differently, predominately via the vinyl groups at
lower temperature [114]. In the case of Si±H functionalised polyvinylsilazanes, hydrosilylation of
vinyl groups occurred upon heating [93]. Similar reactions were reported for the poly(ureamethyl-
vinyl)silazane (Ceraset) precursor [190,266,267]. Besides the hydrosilylation, transamination was
also detected during the cross-linking but to a smaller extent [93]. The hydrosilylation of solutions of
polyhydridovinylsilazanes in toluene or hexane lead to transformation into glassy solids [93]. The
hydrosilylation rate is faster when a Pt catalyst is used: the liquid±solid transformation was complete
in 10 min at 2008C [93]. The presence of 0.1±1.0 wt.% of dicumyl peroxide also can induce cross-
linking of the polyvinylsilazane (trivinylcyclotrisilazane) at much lower temperature (1358C) [159].
Similarly, thermal setting of the Ceraset liquid took place between 90±1508C in the presence of small
amounts of peroxides [190,266,267]. Thus, compared with NCP 200 lower cross-linking
temperatures are used for VT50 (polyvinylsilazane), polyhydridovinylsilazane and Ceraset
precursors because they contain the active vinyl groups.
3.2.2. Conversion to Si/C/N ceramics

Polysilazanes consisting of Si, N and C backbone elements are ideal precursors for formation of
the ternary Si/C/N ceramics. A lot of work was devoted to the fabrication of ternary Si/C/N ceramics
by using different kinds of polysilazane precursors. The processing behaviour of the polysilazanes is
discussed in the following section including ceramic yield, composition as well as pyrolysis
mechanism.
Table 8
Ceramic yield of polysilazanes with varying structure and composition according to LuÈcke et al. [97]
(±SiR1R2±NH)x/(±NHR3)y Ratio (Si/N) Yield (wt.%)
R1 R2 R3 x/y
S1a CH3 CH3 ± ± 1:1 50
S1b H CH3 ± ± 1:1 65
S1c CH=CH2 CH3 ± ± 1:1 91
S1d CH=CH2 CH3 SiHCH3 1:1 1:1 76
S2a CH=CH2 N±H SiHCH3 1:10 1:1.5 73
S2b CH=CH2 N±H SiHCH3 1:5 1:1.08 77
S2c CH=CH2 N±H SiHCH3 1:2 1:1.17 75
S2d CH=CH2 N±H SiHCH3 1:1 1:1.25 78
S2e CH=CH2 N±H SiHCH3 1:.5 1:1.33 74
3.2.2.1. Ceramic yield. It is clear that in almost any kind of ceramic fabrication a high yield is
needed. The ceramic yield may determine the applicability of the precursors for the corresponding
processing and the quality of the resultant ceramics.
The ceramic yield of polysilazanes strongly depends on their chemistry including backbone
structure, the functionality as well as degree of cross-linking. Cross-linked polysilazanes of the type
((RSiHNH)n and (R2SiNH)m) gave a high ceramic yield of 80±85 wt.% upon pyrolysis at 14008C
under argon whereas the uncross-linked silazane gave only 20 wt.% ceramic residue [86,259,260].
Similarly, cyclotetrasilazanes gave a ceramic yield of 84 wt.% under argon at 10008C [127].
Boury et al. [268] observed that the ceramic yield of poly(carbosilane) increased with the vinyl
content in the precursors changing from 12 to 60 wt.% under conditions of 14008C in N2 atmosphere
for 2 h. Because of the significant hydrosilylation cross-linking of the oligomers around 2008C, the
increased ceramic yield was ascribed to the increased degree of cross-linking [268].
The effects of substituents and backbone structure of polysilazanes on their pyrolysis behaviour
were investigated systematically by LuÈcke et al. [97]. They prepared branched and unbranched
polysilazanes with different substitution patterns (Si±H, Si±CH3, Si±CH=CH2, Si(NH3) and Si(NH3))
that were synthesised by the well-known ammonolysis of chlorosilanes using different kinds of
chlorosilanes (Table 8).
TGA analysis performed under argon atmosphere utilising a heating rate of 58C/min showed
that the unbranched silazane (Fig. 3a) gave a very low ceramic yield <11 wt.% at 10008C, while the
branched silazanes (Fig. 3b) showed a higher ceramic yield in range of 51±77 wt.%. The lower yield
for the unbranched silazanes was caused by the evaporation of oligomers at T<3008C. The weight
loss at this stage was reduced from >80 wt.% for the unbranched precursor to 28 wt.% for the
branched one, with the lowest degree of branching in the group of the branched precursors. For the
branched precursor, the weight loss decreases with the decreasing degree of branching and the major
difference between them is the evaporation of oligomers around 3008C [97].
Besides the precursor chemistry and pyrolysis temperatures, the ceramic yield is also affected
by pyrolysis conditions including, heating rate and pyrolysis atmosphere. Bahloul et al. [119]
observed that the ceramic residue of vinylsilazanes increases from 15 to 30 wt.% when the heating
rate is changed from 1 to 608C/min.
Table 9 lists the ceramic yields of several polysilazanes under typical pyrolysis conditions. The
ceramic yield of PHMS (NCP 200) is in the range of 60±70 wt.% which is lower than that of 70±
80 wt.% for the vinyl-containing PVS and PUMVS polymers. The higher yield for the PVS and
PUMVS is due to the enhanced low temperature cross-linking via vinyl groups.
Fig. 3. Thermogravimetric analysis (TGA) of (a) unbranched precursors (b) branched precursors (5 K/min, Ar) [97].
The weight loss during pyrolysis was caused predominately by the evolution of oligomers at
lower temperatures and by-product gases at higher temperatures. A three-stage weight loss was
generally observed for the polysilazanes, as demonstrated by Choong Kwet Yive et al. who examined
the pyrolysis behaviour of oligomeric silazanes with different substitutions (OVS: (±SiViH±NH±)n;
Table 9
Ceramic yield of polysilazanes pyrolzed under the indicated conditions
Silazanes Ceramic yield (wt.%) Pyrolysis conditions References
PHMS (Poly(hydridomethylsilazane)) 80±85 10008C, Ar, 108C/min [86]
PHMS (NCP 200) 60±70 10008C, Ar [216]
PUMVS (Ceraset) 70±80 1000±11008C, Ar [196]
PVS (Poly(vinylsilazane)) 85 11508C, Ar, 108C/min [93]
PVS (VT 50) 70±80 10008C, Ar [216]
Poly(carbosilazane) (PCS) 50±80 10008C, Ar [151]
Poly(phenylsilazane) (PPS) 50±70 8008C, N2, 0.58C/min [161]
Poly(hexylsilazane) (PHS) 25±40 8008C, N2, 0.58C/min [161]
Poly(silacyclobutasilazane) 66±73 12008C, Ar [90]
Poly(ureidosilazanes) 61±73 1000±14008C (Ar, NH3) [188]
Fig. 4. TGA-MS analysis of PUMVS (Ceraset) cross-linked precursor at 260 and 4008C [196].
OMS: (±SiMeH±NH±)n and OMNMS: (±SiMeH±NMe±)n) by TG/MS investigations [92]. The first
weight loss occurred between 150 and 2208C and was caused by the distillation of light oligomers;
the second one between 220 and 4008C was associated with dehydrogenation and transamination and
the third one above 4008C was caused mainly by evolution of methane and hydrogen gases. A
similar three-stage weight loss was found in other kinds of polysilazanes [186,196] implying that this
may be a general pyrolysis behaviour for polysilazanes.
The pyrolysis behaviour changes with cross-linking temperatures as observed for the PUMVS
precursors [196]. Two samples differently cross-linked at 260 and at 4008C were analysed by TGA-
MS (Fig. 4).
The sample with a lower degree of cross-linking exhibits a three stage weight loss curve. The
first step of 6 wt.% occurs around 2008C and the second one of further 11 wt.% at around 4008C
while the high cross-linked sample has negligible mass loss below 4008C and only exhibits the third
weight loss step. Evolution of ammonia gas was detected corresponding to the second weight loss,
indicating the occurrence of transamination reactions in this temperature range. However, the
evolution of ammonia can not explain the large weight loss of 11 wt.%. Therefore, the large weight
Table 10
Ceramic yield and loss of silicon for the cured oligomeric silazanes with different substituents after pyrolysis at 10008C for
5 h [92]
Starting Formula Ceramic DSi/Si0 First weight
silazane yield (wt.%) (%) loss (wt.%)
OVS (±SiViH±NH±)n 83 0.5 3.5
OVNMS (±SiViH±Nme±)n 66 13 13.5
OMS (±SiMeH±NH±)n 46 46 47
O(VS/MMS) (±SiViH±NH±)x(±SiMe2±NH±)y 63 17 21
O(VS/MS) (±SiViH±NH±)x(±SiMe± NH±)y 77 8 10.5
loss at the second stage was probably due to the escape of structural units from the polysilazane
network because transamination involves significant rearrangement of backbone of the Si±N bonds.
The major difference in the weight loss for different kinds of polysilazanes is associated with
the evolution of oligomers. This in turn causes a significant Si loss. Estimation of the Si loss (DSi)
after pyrolysis is useful for evaluating the suitability of a precursor for the formation of ceramics.
This has been done by calculating the Si content before and after pyrolysis (Si0 and Si, respectively)
from the ceramic yield (Y) using the simple relation [92]:
DSi Si0 ÿ YSic (72)

Table 10 gives the relative Si loss (DSi/Si0) for polysilazanes with different substituents as
reported by Choong Kwet Yive et al. [92]. For the oligomeric vinylsilazanes (OVA), the DSi/Si0
value is very low being 0.5% due to the enhanced cross-linking compared with the other silazanes,
for instance the oligomeric methylsilazane (OMS) loosing almost half of its Si content (DSi/
Si046%) after pyrolysis. Analysis of the DSi/Si0 values listed in the last column of Table 10 (which
corresponds to the first stage of weight loss) suggests that the first weight loss step is most important.
Thus, most significant effects for raising ceramic yields are gained by preventing the evolution of Si-
containing oligomeric products at lower temperatures. On the other hand, enhanced weight loss at
the higher temperatures which correspond to the third stage is due to the formation of hydrocarbon
gases and other non-silicon-containing compounds.
3.2.2.2. Elemental and phase composition. Pyrolysis of polysilazanes between 800 and 12008C in
inert gases generally yields amorphous Si/C/N ceramics. The chemical compositions of the ceramics
derived from several kinds of polysilazanes under typical pyrolysis conditions are listed in Table 11.
Table 11
Chemical composition of the Si/C/N ceramics derived from several kinds of polysilazanes under typical pyrolysis
conditions
Polysilazane Chemical composition of Pyrolysis References
the derived ceramics conditions
Polyhydridomethylsilazane, PHMSZ (NCP 200) SiC0.58N0.90 10008C, Ar [1]
Polyvinylsilazane (PVSZ) SiC1.47N0.98O0.06H0.06 14008C, N2 [93]
Polyureavinylmethylsilazane, PUMVSZ (Ceraset) SiC0.86N0.82O0.02 10008C, Ar [196]
Polymethylvinylsilazane (PMVSZ) SiC1.08N0.78O0.16H0.66 8508C, N2 [119]
Polymethylvinylsilazane (PMVSZ) SiC0.06N0.01O2.08H0.18 8008C, O2 [119]
Polycarbosilazane (PCSZ) SiC0.78N0.48O0.05H0.18 11008C, Ar [151]
Fig. 5. Tenary Si±C±N phase diagram showing the compositions of the Si/C/N ceramics obtained from several commercial
polysilazanes after high temperature annealing at T>14408C [196]. PVSZ: polyvinylsilazane (VT5O, Hoechst AG,
Germany) derived ceramic obtained at 10508C/Ar/4h: PHMSZ: polyhydridomethylsilazane (NCP200, Nichimen Corp.,
Japan) derived ceramic obtained at 10508C/Ar/4h; A: polyureavinylmethylsilazane (Ceraset, Allied Signal, USA) as-
received; B: Ceraset cross-linked at 4508C/A/2h; C: Ceraset cross-linked and pyrolysed at 10008C/Ar/2h, D: Ceraset
pyrolysed and annealed at 17008C/N2/2h.
The composition and phases of the silazane derived ceramics can be evaluated using the Si/C/N
phase diagram. The amorphous Si/C/N ceramics will transform to Si3N4 and SiC above 14408C and
excess carbon reacts with Si3N4 to form SiC by loss of nitrogen (see Sections 3.5.1 and 3.5.2). The
compositions of the Si/C/N ceramics derived from several kinds of polysilazanes are indicated in the
isothermal section of the Si/C/N diagram in Fig. 5, as are the crystalline phases formed therefrom at
high temperatures (>14408C).
Polyvinylsilazane (PVSZ) derived Si/C/N has a composition located on the carbon-rich side of
the tie line between N2 and SiC in Fig. 5, indicating the presence of excess C. The composition
of Polyhydridomethylsilazane (PHMSZ) derived Si/C/N is located on the silicon rich side of the
N2±SiC tie line, indicating that there is no excess carbon in the final Si3N4 and SiC ceramics. Ceraset
(PUMVSZ) derived Si/C/N is found close to the N2±SiC tie line and transforms to SiC with small
amounts of Si3N4 and without excess carbon. Annealing experiments showed that 88 wt.% SiC was
formed from the Si/C/N under Ar at 17008C without graphite phases as checked by XRD [196]
which is in good agreement with that predicted from the Si/C/N phase diagram. This result suggests
that PUMVS is a precursor suitable for producing SiC-based ceramics.
The composition of the amorphous ceramics is determined by precursor chemistry and is
affected by pyrolysis conditions. Choong Kwet Yive et al. [92] found that Si±CH2=CH, Si±CH3,
N±H or N±CH3 groups in the cyclic polysilazanes had a large influence on the final Si/N ratio in the
pyrolysed products (Table 12) despite of the same nitrogen to silicon ratio N/Si1 in the starting
polysilazanes. It can be seen from Table 12 that the N/Si ratios in the ceramics derived from OVA,
O(VS/MMS) and O (VS/MS) are significantly lower than 1. The reduction of nitrogen after pyrolysis
was caused mainly by the transamination reactions occurring between 200 and 5008C. This result
suggested that (1) the backbone elements of the polysilazane may also escape during pyrolysis and
(2) the transamination is detrimental to the retention of the backbone in the pyrolysed products.
Table 12
Elemental composition of some oligosilazanes (theoretical) and ceramic composition (pyrolysed under Ar, 5 h at 10008C)
[92]
Starting Formular Composition of Composition of the
silazane starting silazane final ceramics
OVS (±SiViH±NH±)n SiC2NH5 SiC1.48N0.92H0.06O0.04
OVNMS (±SiViH±NMe±)n SiC3NH7 SiC1.58N1.00H0.05O0.14
OMS (±SiMeH±NH±)n SiCNH5 SiC0.64N0.81H0.05O0.05
O(VS/MMS) (±SiViH±NH±)x(±SiMe2±NH±)y SiC2NH6 SiC1.04N0.78H0.04O0.03
O(VS/MS) (±SiViH±NH±)x(±SiMe±NH±)y SiC1.5NH5 SiC1.01N0.81H0.13O0.16
The final crystalline phases are mainly determined by the composition of the amorphous Si/C/N
materials obtained at lower temperatures. Analysis of the ternary Si/C/N phase diagrams provides a
very useful way to evaluate the composition of a amorphous Si/C/N ceramics or to predict the final
crystalline phases.
Because most polysilazanes contain alkyl groups, excess carbon is readily formed in the
pyrolysed products. Burns et al. [270] prepared various polysilazanes by ammonolysis of RSiCl3
with different alkyl groups (Rmethyl, ethyl, isopropyl, . . .) and unsaturated R groups and examined
their pyrolysis behaviour. They found that the carbon residue increases with the number of carbon
atoms in the unsaturated substituents, whereas different alkyl groups gave the same amount of
carbon in the final pyrolysed ceramics.
The carbon content in the polysilazane derived ceramics can be adjusted to a certain extent,
through pyrolysis or isothermal annealing in active gases like ammonia. This treatment reduces the
carbon content by the formation of methane as described in Eq. (73) which is similar to the base
catalysed cross-linking mentioned in Section 2 (Eq. (53)):
BSiÿCH3 NH3 ! BSiÿNH2 CH4 (73)

The retention of C in the pyrolysed products derived from NCP 200 was examined [269] based
on the experimental results that the major gaseous products upon pyrolysis are H2 and CH4.
Obviously, only Si±N and Si±C bonds are formed by reactions among N±H, Si±H and ±Si±CH3
groups [269]. The pyrolysis reactions forming Si±N bonds can be expressed by Eq. (75), which is a
summary of Eqs. (8) and (74). Eqs. (76) and (77) represent the Si±C bond forming reactions. They
can be combined to Eq. (78). All these transformations may be summarised by the overall Eq. (79):
NÿH HÿSiB ! NÿSiB H2 (74)

2xNÿH 2xHCH3 Si ! 2xNÿSiB xH2 2xCH4 (75)
SiCH3 2 ! BSiÿC CH4 H2 (76)
SiCH3 2 SiHCH3 ! 2BSiÿCB CH4 3H2 (77)
2ySiCH3 2 ySiHCH3 ! 3yBSiÿCB 2yCH4 4yH2 (78)
2xNÿH 2x ySiHCH3 2ySiCH3 2 ! 2xNÿSiB 3yBSiÿCB
2x 2yCH4 x 4yH2 (79)
According to reaction (79) and the empirical formula of the starting precursor (Si1C1.3N0.9O0.1H4.4),
the composition of the pyrolysed products was expressed as Si1.00C1.31ÿ2(xy)N0.91H4.47ÿ(10x16y) where
Fig. 6. Weight fractions of carbon in powders obtained by pyrolysis of poly(hydridomethyl)silazane in Ar. The diagram is
based on the individual reactions given by Eqs. (75) and (78) and on the overall reaction (79) for xy0.28100%. The
experimentally analysed carbon contents of the pyrolysed products at 1100 and 5258C are also shown [269].
10x16y4.47. The carbon content in the pyrolysis product can vary from 10.9 wt.% when reaction
(75) takes place exclusively (y0, x0.447), up to 18.1 wt.% if only reaction (78) is taken into
account. This is illustrated in Fig. 6 which also shows the result of two pyrolysis experiments. For the
sample Ar525 pyrolysed in argon at 5258C, the reaction parameters are estimated to be x0.14
(31%) and y0.19 (69%). The corresponding values for sample Ar1000, which was obtained from a
pyrolysis at 10008C, are x0.31 (69%) and y0.086 (31%). This implies that annealing the
polysilazane at a transient temperature around 500±5508C favours the decomposition according to
Eq. (78) over Eq. (75), a fact which may be used to control the carbon content in the pyrolysed
products. This can be combined with the above mentioned effects of ammonia containing
atmospheres reducing excess carbon as described in [269].
3.2.2.3. Pyrolysis behaviour and mechanism. The pyrolysis mechanism involved in the polymer-to-
ceramic transformation is a very complex process and is not completely understood. This is partly
due to the poorly defined structure of the highly cross-linked polysilazanes and the resulting
amorphous and crystalline composite ceramics. By applying solid state NMR, TG/MS, TG/FTIR,
and other modern methods it is possible to investigate the pyrolysis reactions which occur during the
pyrolysis of polysilazanes.
Although polyhydridosilazanes were used to synthesise Si/C/N ceramics for many years, the
pyrolysis mechanism involved in the process of ceramic conversion has not been fully understood.
Bill and Seitz et al. characterised in detail the commercial NCP 200 polymers and their pyrolysis
[114,271] by spectral characterisation of the pyrolysed products obtained at different temperatures.
Based on FTIR and NMR results, they proposed the pyrolysis mechanism shown in Fig. 7.
Fig. 7. Pyrolysis mechanism proposed for PHMS precursors (NCP 200) [114].
The 13C NMR spectrum indicates the presence of Si±CH2±Si units around 5508C. TG/MS
proved the evolution of methane at this temperature suggesting reactions among Si±H and Si±CH3.
The pyrolysis reactions involving N±H groups proceeded between 550 and 6258C to form SiN4 by
successive replacement of methyl groups reacting with H±N groups. Above 6258C, major pyrolysis
Table 13
Various possible pyrolysis reactions among N±H, Si±CH3, Si±H groups and the calculated reaction enthalpy DE (kJ/mol)
[271]
Pyrolysis reactions DE (kJ/mol)
(a) BSiCH2ÿHHÿSiB!BSiCH2±SiBH2 ÿ2
(b) BSi±CH3H±CH2SiB!BSi±CH2SiBCH4 0
(c) BSi±HH±CH(SiB)2!BSi±CH(SiB)2H2 ÿ2
(d) BSi±CH3H±CH(SiB)2!BSi±CH(SiB)2CH4 0
(e) BSi±HH±C(SiB)3!BSi±C(SiB)3H2 ÿ2
(f) BSi±CH3H±C(SiB)3!BSi±C(SiB)3CH4 0
(g) (BSi)2N±HH±SiB!(BSi)2N±SiBH2 ÿ55
(h) BSi±CH3H±N(SiB)2!BSiÿN(SiB)2CH4 ÿ54
(i) BSiCH2ÿHH±CH2SiB!BSiCH2±CH2SiBH2 51
(j) BSiCH2ÿHH±N=!BSiCH2ÿN=H2 66
(k) BSiCH3H±CH2SiB!BSiCH=CHSiB2H2 177
(l) BSiCH3H3CSiB!BSiCBCSiB3H2 296
reactions again involve elimination of Si±CH3 and Si±H groups. This is in accordance with the FTIR
spectra. A successive increase of the number of Si±N or Si±C bonds with pyrolysis temperatures was
supported by the 13C and 29Si NMR data.
Seitz et al. [271] discussed various possible pyrolysis reactions between Si±CH3, Si2N±H
and Si±H groups in NCP 200 by estimating the reaction enthalpies DE which were calculated
based on bond dissociation energies and the assumption of methane and hydrogen formation
(Table 13).
The results show that the formation of new Si±C (reactions (a)±(f)) and Si±N (reactions (f)
and (g)) is favoured, while formation of C±N and new C±C double or triple carbon bonds (reactions
(i)±(l)) is less possible due to the large DE values. This result is in agreement with the spectral
analysis showing favourable formation of Si±C and Si±N bonds in the pyrolysed products. However,
the formation of Si±N bonds was detected only at T>6008C by 29Si NMR spectra which contradicts
somewhat to the low DE values of the related reactions (reactions (f)±(g)).
For vinyl-containing polysilazanes, the pyrolysis behaviour is completely different. Bill et al.
and Seitz et al. [114,271] also investigated in detail, the pyrolysis behaviour of polyvinylsilazanes
(VT 50, Hoechst AG, Germany) based on structural characterisation of pyrolysed products and
proposed a pyrolysis mechanism as shown in Fig. 8.
Below 3508C, cross-linking among the vinyl groups takes place to form carbon chains. At
higher temperatures, the carbon chains transformed into sp2 carbon while the number of Si±N bonds
increase by pyrolysis reactions between Si±H or Si±CH3 and N±H groups. The sp2 carbon was
detected by 13C NMR and the formation of Si±H around 5508C as result of cleavage of Si±C bonds
was detected by FTIR.
The pyrolysis behaviour of the Si±H containing vinyl silazane, (ViSiHNH)x±(MeSiHNH)y was
investigated by Bahloul et al. [119]. The decomposition of the polymer network took place above
4008C with the evolution of mainly methane and hydrogen and to a lesser extent ethylene and other
hydrocarbons like acetylene and propylene. Also, the evolution of ammonia was detected by MS
(Fig. 9) indicating that transamination reactions are also involved in the cross-linking process. The
formation of methane was explained by radical cleavage of SiCH3 and SiCH(CH3)Si units followed
by H abstraction and the ethylene by-products were explained by degradation of Si±CH2±CH2±Si
bridge via a two-step rearrangement.
Fig. 8. Pyrolysis mechanism suggested for the ceramisation of the polyvinylsilazane PVS [114].
3.2.3. The binary Si/N system

Si3N4 or amorphous SiNx ceramics can be obtained from silazane precursors with or without a
low amount of carbon-contamination (Table 14). Early work in this area dates back to the last
century [21,103]. Systematic studies were publised since 1950 [104±109,272±273]. Silicon diimide
[Si(NH2)n] is a well-known precursor for production of pure Si3N4 and gives a-Si3N4 upon heating to
12508C. However, this precursor is insoluble, hence limiting its utility for infiltrations or coatings via
liquid processing technologies.
The larger molecular carbon-free silazane perhydropolysilazane, (H2SiNH)x was obtained as
low viscosity oil products by ammonolysis of H2SiCl2 (see Section 2.1.2). Heating the oily precursor
Fig. 9. TG/MS analysis of cross-linked polyvinylsilazane pyrolysed under He [119]. The m/z values correspond to the
following species: 2 Ð H2; 15 Ð CH3 (NH); 17 Ð NH3 (OH); 26 Ð C2H2 (CN); 27 Ð C2H3 (HCN); 28 Ð
N2 (CO) (C2H4); 41 Ð C3H5 (C2H3N).
to 2008C in argon led to solidification of the samples. Pyrolysis of the perhydropolysilazane at

11508C yields Si3N4 ceramics with a ceramic yield of 70±80 wt.% [75,275]. Contrary to the
amorphous Si/C/N phases derived from most of the carbon-containing polysilazanes, crystalline a-
and b-Si3N4 (and elemental Si) were detected by XRD in the products obtained after pyrolysis at
11508C in argon from (H2SiNH)x (Eq. (80)) [75]. X-ray diffraction analysis indicated that the grain-
size of the pyrolysed products (a- and b-Si3N4 and elemental Si) is in the mm-range. The products
were reported to be partially crystalline utilising pyrolysis temperatures up to 14008C [113].
4H2 SiNHx ! Si3 N4 Si 6H2 (80)

One disadvantage of this precursor is the formation of free Si in the pyrolysed products
[75,275]. Quantitative phase analysis of the pyrolysed products under different pyrolysis conditions
by X-ray diffraction revealed that the free Si decreased with increasing pyrolysis temperature under
Table 14
List of (poly)silazanes for producing Si3N4 (SiNx) ceramics
Silazanes/Pyrolysis atmosphere/Temperature Ceramics References
(H2SiNH)x/N2/11508C Si3N4 powders (free Si) [75]
(H2SiN(CH3))x/N2/850±16008C Si3N4 powders [274]
(H2SiNH)x/N2 or NH3/1100±13008C Si3N4 powders & fibres (free Si) [275,276]
(RSi(NH)1.5)x/NH3/400±12008C a-Si3N4 powders [277]
(H2SiNH)x/NH3/400±10008C a-Si3N4 powders [79,82]
((CH3)HSiNH)x/NH3/600±8008C a-Si/N/(H) films (CVD) [71]
(CH3CH2SiNH)x/NH3 or H2/600±8008C a-Si/(C)/N/H films (CVD) [88]
((CH3)2SiNH)4/NH3/900±15008C Si2N2(NH) & Si3N4 powders [279]
nitrogen [113]. Furthermore, when pyrolysis was conducted under ammonia the free Si could be
eliminated and near stoichiometric Si3N4 ceramics could be produced [79,82,113].
Monolithic Si3N4 was also synthesised from various carbon-containing polysilazanes by
ammonolysis [17,94,277]. Pyrolysis of ±(SiH2NMe)x± gave a a-Si3N4 product free of SiC and C at
1000 and 12008C as checked by XRD; however, the ceramic yield was much lower (38 wt.%) due to
the lower molecular weight of the precursors [177]. After cross-linking of the precursors using
transition metal catalysts such as Ru3(CO)12, the ceramic yield was raised to 64±75 wt.%; the
ceramics contain >80 wt.% a-Si3N4 but contain a certain amount of carbon (2.5±18 wt.%) which was
found to be dependent on processing conditions [136]. Low-carbon amorphous Si3N4 ceramics with
carbon contents below 0.1 wt.% were obtained by pyrolysis of a hydridopolysilazane (HPZ) in
ammonia between 800 and 12008C [277]. Significant carbon removal from the starting HPZ
precursors occurred between 400 and 6008C which was found to be independent of polymer
structure or functionality but partially dependent on the initial cross-linking temperatures of the
precursors [277]. The precursor HPZ was also used to deposit amorphous Si±N films on single
crystal Si by CVD and a film having a carbon content of 2.7 at.% was formed at 600±8008C
in presence of ammonia [71]. A vinyl-containing silazane was also pyrolysed under ammonia
and crystalline Si3N4 with 2 wt.% of carbon was obtained [94]. In this case, the nitridation
temperatures were much lower, they were found to be between 300 and 6508C [94]. In addition,
Si3N4 was also synthesised using a carbon-containing ethylsilazane-precursor and ammonia in a
CVD process [88].
3.2.4. The quaternary Si/(E)/C/N system (EB, Ti,. . .)

Multi-element composite ceramics such as quaternary Si/(E)/C/N with EAl, B, (P), Ti or Zr
have been found to show enhanced thermal stability and oxidation resistance. Hence, these materials
have drawn great attention from researchers all over the world in recent years [14,213]. Such
composite ceramics are generally difficult to synthesise by traditional methods, owing to the
thermodynamic instability of their solid solution. An important approach to multi-element composite
ceramics is the polysilazane route. Various kinds of metals can be incorporated as pendent groups or
into the polysilazanes backbone. For instance, various boron-containing polysilazanes have been
synthesised by reactions of boron or borosilicon compounds with silazanes or amines (see Section
2.7) [212,280±282]. The polyborosilazane, synthesised by hydroboration of dichlormethylvinylsi-
lazane gave a pyrolytic ceramic Si3.0B1.0C4.3N2.0 with 58% conversion after pyrolysis under argon at
10008C [223]. Another route to the boron-containing silazanes is modification of polysilazane by
various boron sources through dehydrocoupling or hydroboration to attach a pendent boron-
containing species to the Si±N backbone [205,208,217,283]. This strategy is particularly useful for
the formation of Si/B/C/N materials containing low amounts of boron. Pinacolboranzine-modified
hydridopolysilazane (PIN-HPZ) gave a Si/B/C/N ceramic containing 1±3% boron after pyrolysis at
14008C [284]. Compared with the HPZ derived ceramics containing only 4% nitrogen at 18008C, the
PIN-HPZ derived Si/B/C/N contained 23% nitrogen, indicating the enhanced thermal stability of the
B-containing composites.
Besides, other metals can also be incorporated into polysilazanes, for instance using the
transamination reaction of polysilazanes with Al(NR2)3, B(NR2)3 or P(NR2)3 or Zr(NR2)4 for
producing the corresponding quaternary ceramics [258].
Synthesis of Ti-containing ceramics from precursor route to Si/Ti/C/N system has been reported
in recent years. The work is aimed to produce SiC/Si3N4±TiC/TiN composite ceramics that were
produced predominately by mixing of the powders of respective phases and pyrolysis under inert or
active atmosphere. Molecular incorporation of titanium into silazanes has been realised by reactions
with tetrakis(dimethylamino)titanium [232,234,285]. In comparison with the carbon-containing

polymers derived from the other polysilanes and alkoxides, the poly(titanosilazanes) contains no
oxygen. The liquid precursor was suitable for producing TiC (and/or TiN)-containing ceramics
[232]. TGA analysis of the poly(titanosilazanes) gave a yield in the range of 30±70 wt.%. The
ceramic yield was found to be strongly dependent on the substituents of the starting polysilazane
precursors [(CH3)3SiNH]m[Si(CH3)3 R±NH]n[Si(CH3)]m with (R±CH3, ±CH=CH2 and ±H). For
R±H the highest yield found was of 72 wt.% [232]. Chemical analysis of the pyrolysed products
obtained at 10008C revealed a larger amount of Ti in the range of 15.7±19.6 wt.% [232].
Incorporation of Al into polysilanes or polysilazanes was noted by several authors with respect
to the pyrolysis conversion to Si/Al/N, SiC/AlN composites. Si/Al/N precursor was synthesised from
polyhydrosilazanes and organoaluminiumamides [286]. SiC/AlN ceramics were also produced by
pyrolysis of polyalkylaminoalanes and polysilanes [287] or by pyrolysis of trimethylsilyl -substituted
aluminium compounds [288]. Reaction of tetrakis(methylamino)silane with a Lewis-base adduct of
alane AlH3NMe3 was recently reported to produce polyalumosilazanes. The polyalumosilazanes
synthesised by this route have tailored physical properties such as solubility and viscosity and
ceramic compositions by controlling synthesis reaction and cross-linking. Pyrolysis of the
polyalumosilazanes yielded aluminium nitride/silicon nitride nanocomposite ceramic powders with
a yield of 60±70 wt.% [200].
Another interesting aspect of aluminium-containing silazanes is the production of Si/Al/O/N
ceramics, a ceramic system with high thermal stability and mechanical properties. Schmidt et al.
[289] synthesised aluminium- and oxygen-containing polysilazanes by reaction of 1,3,5-trimethyl-
1,3,5-trivinylcyclotrisilazane with Et3Al or Et2AlOEt. They obtained Si3N4/AlN and Si/Al/O/N
composites by pyrolysis of polyaluminosilazanes in NH3. The composite ceramics consisted of
crystalline phases a-, b-Si3N4/AlN, b-Si3N4, AlN, Al2O3 and mixed Al/O/N amorphous phases
rather than a single phase SiAlON. Yu and Mah [290] used (CH3SiHNH)n/CH3Si(H)O)n hybrid
polymers blended with aluminium hydroxides. Pyrolysis of the products under ammonia yields
mixtures of mullite and alumina. Formation of fine grained b0 -SiAlON was demonstrated by Soraru
et al. [291] by ammonolysis of aluminium alkoxide modified polycarbosilanes. Finally, synthesis of
polyaluminosilazanes using common perhydropolysilazane (PHPS) with addition of (ethylaceto-
acetate)aluminium diisopropoxide (ALCH) was performed by Funayama et al. [292]. Pyrolysis
of the polyaluminosilazanes under ammonia at 10008C yields amorphous products SiA-
l0.12O0.64N1.1C0H2,4. Subsequent annealing of the amorphous products under nitrogen at 15008C
gave partially crystalline phases consisting of b0 -SiAlON and mullite.
3.3. Gas phase and liquid phase processes
3.3.1. Chemical vapour deposition (CVD)

Chemical vapour deposition (CVD) is an active branch in the field of precursor derived
ceramics because unconventional forms of ceramics such as nano powders, thin films or thick
coatings can be produced by this process. Gaseous or volatile liquid precursors are required and the
ceramics are formed by decomposition of the vapour under heating, either by electric resistance or
by plasma or laser. Organosilazanes are unique single source CVD-precursors for Si/C/N or (Si3N4-
SiC) ceramics. This ternary system has been investigated extensively in recent years to produce Si-
based nano ceramic powders [69,98,293±305] and structural or functional films [71,306±315]. Some
results in this field are reviewed briefly in this section.
Si-based nano ceramic powders, including Si3N4, SiC or Si3N4±SiC, can be produced from
organosilazanes by vapour phase reactions. In the CVD process, the evaporated silazane precursor is
heated by hot-wall, plasma or laser and the solid particles nucleate and grow rapidly in the hot zone
and lead to the formation of fine particles in the range of several nanometers up micro-meters. The
particle size of the product depends on the heating method as well as other parameters like
temperature and flow rate. Hexamethyldisilazane (HMDS) is a common precursor used in CVD
processing for the fabrication of Si-based ceramics because of its low boiling point. It has also strong
absorption to CO2 laser photons, hence HMDS is suitable as a precursor in laser pyrolysis.
CVD synthesis of fine ceramic powders using hexamethyldisilazane (HMDS) was started 10
years ago by Japanese researchers [303,304]. Vapour phase HMDS mixed with ammonia was heated
in Al2O3 reaction tubes under heating at temperatures of around 10008C. Si/C/N powders with
diameters in the range of 40±200 nm were produced by this process [303±305]. Additionally, the
annealing of these powders at higher temperature was investigated [316]. The significance of the Si/
C/N powders synthesised by this CVD process has been demonstrated several years later by Niiharra
et al. who developed Si3N4±SiC nano composite ceramics from such powders [2,255,256].
Laser synthesis, developed 10 years ago by Haggerty et al. [317], is considered to be a superior
process for producing ceramic powders. The process is especially useful for producing high purity
nano powders with small diameters (particle size <50 nm) which are useful for producing
nanostructured monolithic or composite ceramics. Additionally their use as reinforcements for
ceramics or metal-based composites was under investigation. Si3N4, SiC and their composite
powders were produced predominately from the expensive SiH4 with additional gases such as C2H4
and/or NH3 or (CH3)2NH [318,319]. SiH4 has the advantage of strong absorption to CO2 laser light
at the corresponding wavenumber of 10.06 mm. Rice and Woodin [296,297] first introduced
hexamethyldisilazane into the laser process in 1984 and obtained Si/C/N powders with a mean
diameter of 0.1 mm. However, the ceramic powders were strongly agglomerated and contained
excess carbon. With improvement of the process and variation of synthesis parameters, Si/C/N
powders with a wide range of compositions (Si, N and C) [298] and nanosized crystalline SiC were
synthesised [299]. With addition of ammonia in the HMDS precursors, amorphous Si3N4 powders
were also obtained [299]. The particle diameter was found to be in the range of 10±35 nm which is in
the same order of the materials produced from the SiH4NH3 system. A two-step nucleation
mechanism of Si3N4 was proposed in the case of SiH4NH3 precursor. It involves decomposition of
SiH4 followed by nucleation and nitridation under ammonia at high temperatures [317]. In the case
of HMDS the nucleation of solid particles is accompanied by the formation of hydrocarbon gases
[300]. The final composition of the powders is governed by a competition between the cross-linking
rate and the diffusion of volatile hydrocarbons. Again, this strongly depends on synthesis parameters
such as the laser flux energy and the flow rate of the vapour phase. Compared with the use of silane
gas the organosilazane reactant offers advantages of (1) lower reaction temperatures; (2) absence of
free Si in the powder products and (3) a wide range of Si, N and C compositions [299].
Vapour phase pyrolysis was also applied to oligomeric silazanes Si(CH3)2NH)n n3±4
synthesised from dimethyldichlorosilane and ammonia [69]. The liquid oligomeric precursor was
evaporated and pyrolysed by means of hot tube heating between 600 and 10008C. Si/C/N particles
with diameters of 100±600 nm and BET surfaces of 10±350 m2/g were obtained. The particle
diameter and the BET value, as well as the chemical composition, depended on the conditions for the
synthesis of the starting materials and the pyrolysis process. The as-produced powders contained
organic groups on the particle surface which could be removed by the subsequent annealing. After
crystallisation at 14008C, the amorphous powders transformed to crystalline Si3N4 and SiC phases.
Furthermore, the amorphous Si/C/N powders were hot pressed after addition of sintering aids at
18008C applying 30 MPa for 1 h. Fully dense Si3N4±SiC composite sintered bodies with nanosize
SiC dispersed in the Si3N4 grains were obtained and the mechanical properties were measured.
3.3.2. Liquid phase pyrolysis

Liquid polysilazanes are used to produce coatings, ceramic infiltrations, binders and to generate
fine particles.
Spray pyrolysis of liquid polysilazanes is unique technique for the generation of fine Si3N4 or
Si/C/N powders [98,115]. The process involves evaporation of the polysilazane liquid by means of
ultrasonication and then pyrolyzing the spray by means of hot-tube [115,320], plasma or laser [98]
heating. Liu [115] reported synthesis of spherical Si3N4 particles with a diameter of 38 nm by
pyrolysis of polysilazanes prepared from reactions between trichlorosilane and methyldichlorosilane
according to Seyferth's method. The polysilazanes were first dissolved in benzene and then diluted
with acetonitrile. A micro-tube pump was used to supply the solution to an atomiser. Amorphous
Si3N4 particles were formed at a pyrolysis temperature of 7008C in nitrogen atmosphere. SEM
investigations showed clearly the spherical shapes of the particles. The spherical shapes retained
after crystallisation by annealing at 14008C. The smooth surface of the Si3N4 particles was ascribed
to the unique liquid±solid transformation from silazane droplets during pyrolysis. This unique shape
can not be obtained by traditional solid phase pyrolysis or CVD methods because Si3N4 itself has no
liquid phase.
The liquid oligomer (CH3SiHNH)n was also pyrolysed using laser heating to produce fine
particles [319]. In this process the liquid precursor was forced through a nozzle which was irradiated
by a CW CO2 laser beam. The strong coupling between the laser radiation and the injected aerosol
droplets resulted in their rapid evaporation and the formation of a plasma plume. The precursor
decomposed and the decomposition products were rapidly quenched as they emerged from the
plume. TEM observation showed that the powder particles were nearly spherical. However, the
particle size presented a bimodal distribution with smaller particles around 45 nm and larger ones of
120 nm. The larger particles were explained to be formed by agglomeration of the small particles. In
this process Si/C/N powders were obtained with a production rate of 100 g/h. 70% of the liquid
phase was found to be converted into the nanosize powder particles.
3.4. Processing to dense Si/C/N bulk materials
Before the 1990s, polysilazane-derived ceramics were mainly prepared in the form of materials
of low dimension such as fibres, coatings or powders. Conventional powder sintering is still the
major route to larger bulk-materials.
Attempts to produce bulk ceramics by pyrolysis of liquid or cross-linked polysilazanes were not
successful because the evolution of oligomers during pyrolysis resulted in considerable pores or
cracks of the pyrolysed bodies [93,151,261]. A glassy solid polyvinylsilazane obtained by cross-
linking of liquid oligomers was pyrolysed up to 14008C in Argon [93]. SEM investigations revealed
the presence of extensive cracks and pores which were found to be due to the generation of stress
induced by the volume decrease during pyrolysis. However, the pyrolysed porous bodies had locally
dense regions and maintained their shapes without bloating after pyrolysis, even at higher heating
rates. This suggests the possibility of obtaining dense Si/C/N bulk ceramics through the pyrolysis of
polysilazanes. However, to reduce the porosity and to avoid cracking an optimised processing is
needed.
In 1991, Riedel et al. developed a new process which produces dense Si/C/N bulk ceramics
[1,5]. This method (Fig. 10) involves cross-linking of polysilazanes into infusible solids, milling the
cross-linked silazanes into fine powders followed by shaping green bodies and pyrolyzing the
compacts around 10008C. The unique feature of this novel process is the direct transformation from
highly cross-linked polymer networks to dense monolithic Si/C/N bodies by pyrolysis bonding
Fig. 10. Schematic illustration of the powder processing route and bulk pyrolysis of perhydropolysilazane [321].
without sintering aids [1]. This shortened route from polymers to ceramics promises to become a
new process compared with the traditional liquid sintering of ceramic powders and is also a great
challenge to material scientists and ceramists in academia as well as in industry. At present, this new
process is still in the earlier stages of development and only three kinds of polysilazanes have been
applied in such a processing. They include a polydihydromethylsilazane (PHMS, NCP 200)
[1,5,322] a polyureamethylvinylsilazane (PUMVS, Ceraset) [323] and polyvinylsilazanes (VT 50)
[251] all of which are commercial products (see also Section 2.8).
In the following sections, the processing characteristics of polysilazane bulk pyrolysis to obtain
monolithic specimens are discussed.
3.4.1. Curing
The general requirements for a precursor suitable for bulk pyrolysis processing include (1)
cross-linkablility to form infusible polymers; (2) compactability into dense green compacts [258].
The starting polymers have to be transformed to a cross-linking-state which is characterised by the
presence of a small amount of oligomers. This is necessary for the bonding between the particles
during compaction and ceramisation. The success in application of polysilazanes for bulk ceramic
transformation is ascribed in large to their suitable cross-linking properties which have been already
discussed above.
The degree of cross-linking was checked by thermo-mechanical analysis (TMA) observing the
shrinkage upon heating [322]. PHMS (NCP 200) was cross-linked into infusible polysilazanes at
>3508C in inert gases [1,5]. The cross-linked polysilazanes were compacted and subjected to TMA
measurements (Fig. 11).
Fig. 11. TMA analysis of the warm-pressed powder-compacts. The amorphous Si/C/N was obtained from PHMS cross-
linked at different temperatures [321].
A rapid shrinkage above 1508C was observed for the 3508C-cross-linked specimens which was
attributed to a softening of the silazane-bodies, indicating a low degree of cross-linking. In
comparison, the 3808C and 4008C-cross-linked samples exhibited a continuous shrinkage. The onset
temperature for the shrinkage increases with the cross-linking temperature. Pyrolysis results showed
that 3808C is the optimal temperature of cross-linking for achieving high density in the ceramic
product after pyrolysis [321].
3.4.2. Compaction
Dense and crack-free preceramic powder compacts of the cross-linked polysilazanes are
required for obtaining crack-free dense bulk ceramics after pyrolysis. Early attempts to prepare poly-
mer compacts for bulk pyrolysis used isostatic or uniaxial cold pressing leading to densities of the
compacts of 84% [1]. The final bulk density of the Si/C/N ceramics after pyrolysis at 10008C was 93%.
Warm-pressing methods have recently been used by Seitz and Bill to compact cross-linked
polyvinylsilazane powders for bulk pyrolysis [251]. The pressing temperature should be in the range
of the softening point of the cross-linked polysilazanes since plastic forming can be realised under
these conditions. TMA analysis of the cold compacted cross-linked specimens was used to detect the
softening region of the precursor material between 100 and 2758C (Fig. 12). Thus, a pressing
temperature of 2508C was used to compact the polyvinylsilazane powders (28 mm) under an acting
pressure of 47.75 MPa. Dense green bodies with a porosity of 7.5% were obtained by this
compaction technique. After pyrolysis, Si/C/N ceramics with a relative density of 97% were
obtained. SEM investigation of the cold-pressed and warm-pressed samples before and after
pyrolysis provided evidence of the enhanced density in the case of the warm-pressing process
[251,321,322].
Fig. 12. TMA analysis of the cold-pressed polyvinylsilazane compacts [251]. The oscillation of the curve between 100 and
2508C indicates a softening of the sample.
The warm-pressing technique was recently applied to process the PHMS polymers [321,322].
Pyrolysed Si/C/N ceramic materials exhibiting an open porosity in the range of 1.3±5 vol.% were
produced. This is in contrast to the cold-pressed samples containing 10±15 vol.% of open porosity
after pyrolysis. The polymer powders pressed at 3008C (cross-linking temperature: 4208C) yielded
dense translucent compacts with an open porosity of 1 vol.% and an average pore diameter below
1.0 mm (Fig. 13).
SEM observation of the fracture surface of the green bodies reveals that the original particles are
still present in the cold-pressed samples (Fig. 14 (a)) while for the warm-pressed samples no
particulate structure can be observed (Fig. 14 (b)). This indicates the high degree of densification
which occurs during the plastic deformation in the process of warm-pressing. After pyrolysis there
are appreciable amount of pores in the cold-pressed samples (Fig. 14 (c)), whereas most of the pores
are eliminated in the warm-pressed samples (Fig. 14 (d)).
The quality and density of the warm-pressed green bodies strongly depends on the cross-linking
state of the polymer and the warm-pressing parameters, as well as the pressing technique. Fig. 15
shows the variation of green density with pressing temperatures for the PHMS precursor cross-linked
at 3808C [322]. The green density increases with warm-pressing temperatures above 1008C. Similar
behaviour was observed for the thermally cross-linked PUMVS precursors [323]. It is noted that the
temperature for the onset of increasing density is nearly the same, i.e. around 1008C, and seems to be
independent of precursor structure and cross-linking temperature. Although the green density
increases with warm-pressing temperatures, the warm-pressing temperature should always be lower
than the cross-linking temperature [251,323]. Otherwise cracks will be generated due to the
evolution of oligomers. From Fig. 15 it can be seen that the final pyrolysis density also increases
distinctly with the warm-pressing temperatures as a result of the higher density of the green bodies.
The unique feature of the warm-pressing technique is that it allows a plastic forming of green
bodies. The plastic deformation and the formation of strong (covalent) bonding between the particles
Fig. 13. Pore size distribution of cross-linked infusible PHMS: (a) isostatically compacted at 640 MPa and (b) pyrolysed at
10008C in argon [1].
results in machinability of the green bodies. One example is shown in Fig. 16. The samples are
obtained from warm-pressing of cross-linked PUMVS polysilazanes.
3.4.3. Bulk pyrolysis

Crack-free Si/C/N ceramics were formed after pyrolysis of the polymer compacts made from
cross-linked PHMS at 10008C and 58C/min in flowing argon atmosphere [1]. The pyrolysed bodies
were X-ray amorphous with a composition of Si1.73N1.56C1.0. The density of the pyrolysed samples
measured by mercury pressure porosimetry was 2.15 g/cm3, which corresponds to a 93% relative
density on the basis of the solid phase density of 2.30 g/cm3 for the Si/C/N ceramics. The mercury
porosimetry revealed the presence of large pores with diameters up to 110 mm in the green bodies.
This was attributed to the cold pressing which was used to prepare the green bodies and to the low
green density (relative density84%). However, warm-pressed PHMS green compacts with a
relative density as high as 98% were obtainable recently. This shows that dense ceramic materials
with porosity below 3% can be prepared [222,223].
Polyvinylsilazane (VT 50) was also processed using the bulk pyrolysis route after cross-linking
at 3008C/90 min. Compaction was accomplished by cold-pressing (at room temperature) or warm-
pressing applying 48 MPa/2508C [251]. Pyrolysis of the warm-pressed compacts at 10008C for 4 h
under argon yields crack free-dense ceramics with relative density >97%, as measured by mercury
pressure porosimetry. In contrast, the cold pressed specimens still have 7.5% residual open porosity
in the pyrolysed state.
Fig. 14. SEM pictures of the fracture surface of (a) cold pressed green bodies; (b) warm-pressed green bodies; (c) after
pyrolysis of the cold-pressed samples and (d) after pyrolysis of the warm-pressed samples. Polysilazanes: PHMS; cross-
linking temperature: 4008C [322].
Recently, the PUMVS precursors have been processed along the route of bulk pyrolysis. Si/C/N
ceramics were obtained by pyrolysis of the warm-pressed compacts. These were made from
precursor powders cross-linked between 300 and 4008C. No crack formation was observed on the
surface and inner part of the pyrolysed bodies (Fig. 17).
After pyrolysis the sample revealed a shrinkage of about 60% by volume. The shrinkage along
the axial and radial direction of the cylindrical samples was found to be different, with a higher
radial shrinkage (28±30%) compared with the axial shrinkage (15±24%). This anistropic shrinkage
was related to the textured structure of the warm-pressed bodies due to the uniaxial pressing.
The pyrolysis of the warm-pressed PUMVS-compacts was investigated by TG/MS and TMA
analysis which reveals that the shrinkage occurred concurrently with the weight loss between 450
and 7008C (Fig. 18). Also, the evolution of H2 and CH4 at this temperature implies the presence of
macro pores larger than micrometers. These pores are not intrinsically related to the evolution of
gaseous by-products, but are due to compacting faults. Below 4808C an expansion is observed with
raising temperature which is associated with the thermal expansion and/or continued cross-linking
reactions of the compacted polysilazanes.
Bulk pyrolysis was also applied to a liquid polysilazane containing N±N bonds (Pyrofine PM,
ELF Atochem) [94,144]. Besides pyrolysis of compacts of the cross-linked and pyrolysed
polysilazane powders, mixtures of the liquid precursor with 10008C-pyrolysed Si/C/N ceramic
powders and 3008C-cross-linked powders were compacted and pyrolysed, respectively. It was found
Fig. 15. Influence of the warm-pressing temperature on the green density and on the density of the pyrolysed Si/C/N
ceramics [321].
that the monolithic pyrolysis processing gave a much lower bulk density (open porosity>24%).
Whereas, pyrolysis of the mixture of the cross-linked powders and the liquid precursor gave a higher
density (open porosity>12%). In addition, the porosity of the compacts pyrolysed from the mixture
system increases to 10±11% after annealing the pyrolysed bodies at 14008C. Although the mixture
system gave a higher density in this case, this process is less successful compared with the monolith
bulk pyrolysis using the other three kinds of precursors considered above. The mixture strategy was
recently also applied to polyvinylsilazane precursors of the type ([±(NH)1.5±SiCH=CH2]x-
[±NHSiHCH3]y) [324]. Pyrolysis of compacts made from a mixture containing 70 vol.% 3008C
cross-linked polysilazane and the liquid polysilazane results in pyrolysed ceramics with a residual
Fig. 16. Warm-pressed samples made from PUMVS (Ceraset): (a) as-pressed sample; (b) machined sample [323].
Fig. 17. Appearance of a warm-pressed, PUMVS-derived sample before and after pyrolysis [323].
Fig. 18. TMA and TG/MS examination of warm-pressed PUMVS-derived compacts [323].
open porosity of 10±15%. This is in contrast to the PVS (polyvinylsilazane VT 50) derived bulk
ceramics with an open porosity as low as 3% [251].
3.4.4. Bulk pyrolysis of filler systems

The bulk pyrolysis of polymer powder compacts to fabricate dense Si/C/N ceramics provides
the opportunity to produce multi component materials. Instead of using a pure single source
polysilazane precursor active or passive powders (fillers) may be added to the precursor. The active
fillers may react with the silazane, its decomposition products or active atmospheres during pyrolysis
to form carbide, nitride, carbonitride, oxide etc. phases, which in turn can reduce the shrinkage and
the amount of volatile side products. A similar effect may be gained by passive fillers which do not
react with the silazane matrix, but decrease shrinkage and formation of volatiles due to a reduction of
the total polymer amount [325,326].
Loading of polymers with various metallic fillers was conducted to produce bulk amorphous
ceramics [325,327,328]. In this process liquid or solid polymer precursors were mixed with active
filler powders, followed by shaping and pyrolysis. Suitable active fillers are elements or compounds
forming carbide, nitride or oxide reaction products such as Al, B, Si, Ti, CrSi2, MoSi2 which exhibit
a high specific volume increase upon reaction [325]. By this process, zero shrinkage after pyrolysis
could be achieved allowing a near net shape forming of ceramic composite materials. However, the
pyrolysed bodies were not fully densified with an appreciable amount of residual open porosity. This
process was previously mainly applied to polysiloxane precursors to produce oxygen-containing
ceramic composites such as Si/O/C±TiC, Si/O/C±TiN, Si/O/C±SiC. However, it may also be used to
process polysilanes or polysilazanes to produce nonoxide ceramic composites [118,326,329]. Some
studies on ceramic conversion via the filler strategy have been summarised recently by Greil [325].
3.5. Densification of silazane derived Si/C/N powders with sintering aids
It has been established in recent studies that Si3N4/SiC composites exhibit improved high-
temperature mechanical properties combined with good oxidation resistance in comparison to
monolithic materials. An advantageous method for growing homogeneously distributed ultrafine SiC
grains, both at Si3N4 grain boundaries and within Si3N4 grains, is to use amorphous
polycarbosilazane derived Si/C/N powders as starting materials. These amorphous powders can
either be used as single starting materials or mixed with a crystalline Si3N4 powder. A variety of
processes have been reported to produce amorphous polycarbosilazane derived powders with the
purpose of using them as starting materials for the fabrication of Si3N4/SiC composites.
Niihara et al. [2] were the first research group to investigate in detail the sintering of
polycarbosilazane derived Si3N4/SiC composites. The amorphous Si/C/N powder used in their
research was prepared via CVD (see Section 3.3.1). Other studies were conducted with Si/C/N
amorphous powders produced by plasma chemical synthesis [330]. Another route is based on
amorphous Si/C/N powders obtained by pyrolysis of a polycarbosilazane precursor as starting
materials for the fabrication of Si3N4/SiC composites [11,330±332].
The densification of amorphous Si/C/N powders or mixtures of amorphous Si/C/N powder and
crystalline Si3N4-powder to Si3N4/SiC composite materials can only be achieved by liquid phase
sintering. For this purpose the starting powder has to be homogenised with sintering aids in a
conventional powder metallurgical way.
One advantage of the chemical processing route makes it possible to incorporate [M(OR)x,
MAl, Y] oxide groups in the polymer chain of the precursor, so that the sintering aids are
homogeneously distributed in the precursor [333,334]. The pyrolysis of the modified precursors
leads to [Si/(Al,Y)/O/C/N] amorphous products. Similar powders can be obtained by other routes,
like laser spray pyrolysis. In this case [M(OR)x] compounds are mixed to the Si/C/N precursor [335].
3.5.1. Thermodynamic aspects

One major consideration for the understanding of the sintering behaviour of amorphous Si/C/N-
powder derived Si3N4/SiC composites is the thermodynamic stability of phases in contact during
sintering. The crystallisation of amorphous Si/C/N-powders, which is described in detail in the
following section in the case of a polyhydridomethylsilazane derived Si/C/N-powder, leads to the in
situ formation of a-Si3N4 and a mixture of a- and b-SiC. Additionally, the formation of free carbon
containing inclusions was observed. Considering the thermodynamic stability of this three phase
system, three reactions (1, 2 and 3) are to be taken into account in a first approach. They depend
strongly on the nitrogen pressure and on the temperature as it is shown in Fig. 19.
Under a given nitrogen pressure, the sintering temperature of the Si3N4/SiC composites is
limited by the decomposition temperature of Si3N4 according to Eq. (81) which defines the lower
bound of the stability range in Fig. 19.
Si3 N4 s ! 3Si l 2N2 g (81)

Since Si3N4 and C are in contact during sintering as mentioned before, reaction (82) occurs at
temperatures above 14408C under a nitrogen pressure of 0.1 MPa [339]. This reaction leads to the
formation of nitrogen and is consequently accompanied by a weight loss. Since the influence of
carbon in Si3N4/SiC composites is very detrimental for thermomechanical properties, it is of benefit
to shift reaction (82) to the right side. This has a major consequence for the optimisation of gas
pressure sintering cycles. To promote reaction (82), sintering experiments have to be carried out in a
Fig. 19. Thermodynamic stability range of Si3N4 and SiC, relative to temperature and nitrogen pressure. Literature data
taken from [330]. Curves 1 and 5 represent data from Sundman et al. [336], curves 2 and 6 from represent data from Barin
[364], curve 3 represents data from Panek [337] and curves 4 and 7 represent data from Johnson [338].
first stage using a nitrogen pressure of 0.1 MPa up to temperatures of 17008C [330]. The upper
bound of the stability range of Si3N4 and SiC in Fig. 19 is given by reaction (82).
>1440 C=0:1MPa=N2
Si3 N4 s 3C sÿÿÿÿÿÿÿÿÿÿÿ!3SiC s 2N2 g (82)
Under gas pressure sintering conditions i.e. using high nitrogen pressures the nitridation of SiC
according to reaction Eq. (83) (reverse of reaction (82)) may occur. Reaction (83) implies
incorporation of nitrogen from the atmosphere and, therefore, leads to a weight gain.
3SiC s 2N2 g ! Si3 N4 s 3C s (83)

Prior to liquid phase sintering the crystallisation of the amorphous powder was induced by a heat
treatment in many studies [330,340,341]. As amorphous Si/C/N powders generally contain an excess
of carbon, this intermediate step is carried out in order to remove carbon according to Eq. (82). By
this procedure the evolution of nitrogen due to reaction (2) during liquid phase sintering can be
avoided. The evolution of nitrogen and the presence of carbon in the powder mixture have to be
considered as two major factors influencing the sintering behaviour (see Section 2.5.3).
Crystallisation tests were carried out in the range of coexistence of Si3N4 and SiC [340] so that
the crystallite size remains in the same range as the particle size of the amorphous powder.
In addition to the interaction between nitrogen and the solids Si3N4, SiC and carbon (Eqs. (81)±
(83)), the reactions involving oxygen containing compounds like CO (g), SiO (g) or SiO2 (s, l) have
to be considered. Oxygen is usually present as an impurity in amorphous Si/C/N powders. It was
reported that CO (g) like SiO (g) is likely to be formed in thermodynamically unstable amorphous
powders [342]. Moreover, the crystallisation of SiO2 in amorphous Si/C/N/(O)-powders can also be
envisaged. SiO2 is present on the surface of crystalline powders, which are in some studies mixed
with amorphous powders to prepare starting materials for the sintering of Si3N4/SiC composites.
Since an excess of carbon is commonly present in amorphous Si/C/N-powders, two cases have
to be considered:
If no crystallisation of the amorphous powder has been carried out prior to liquid phase
sintering, in situ crystallisation and reaction (82) take place during sintering. It follows that reactions
involving carbon and CO (g), SiO (g) or SiO2 (s, l) have to be considered. The possible chemical
interactions are described by the Eqs. (84)±(87).
SiO2 s; l 3C s ! SiC s 2CO g (84)

and
SiO g 2C s ! SiC s CO g (840 )

3SiO2 s; l 6C s 2N2 g ! Si3 N4 s 6CO g (85)
and
3SiO g 3C s 2N2 g ! Si3 N4 s 3CO g (850 )
2SiO2 s N2 g 3C s ! Si2 N2 O s 3CO g (86)
and
2SiO g N2 g C s ! Si2 N2 O s CO g (860 )

2SiO g CO g N2 g ! Si2 N2 O s 2CO2 g (87)

2Si3 N4 s 3CO g ! 3SiC s 3SiO g 4N2 g (88)
Among the above mentioned reactions, the equilibrium constants at 16278C of reactions (840 ),
(85 ), (860 ) and (87) have the highest values [343].
0
In amorphous powders which were crystallised before sintering, excess carbon has been
removed, so that the possible chemical interactions involve SiO2 (l) contained in the intergranular
glassy phase, and the SiC and Si3N4 phases [330] (Eqs. (88)±(900 )). Since CO (g) is produced by
reactions (84)±(860 ), (89) and (900 ) a further reaction between Si3N4 and CO (g) has to be considered
(Eq. (88)).
2SiC s SiO2 l 2N2 g ! Si3 N4 s 2CO g (89)
Si3 N4 s 3SiO2 l ! 6SiO g 2N2 g (90)
SiC s 2SiO2 l ! 3SiO g CO g (900 )
It has to be pointed out that reactions (84)±(850 ), and (88)±(900 ) lead to the consumption of SiO2 (l),
which causes reduction of the amount of liquid phase during sintering. This may result in an
incomplete densification. Since the formation of a liquid phase can be prevented locally, these areas
in the specimens have to be considered as defects in the sintered Si3N4/SiC composites.
Taking into consideration the partial pressures of CO (g) and SiO (g) during sintering in
nitrogen atmosphere, Herrmann et al. [330] deduced an optimised nitrogen pressure±temperature
curve within the Si3N4/SiC stability range in Fig. 19, which if observed during the sintering cycle
minimises the decomposition of the liquid phase.
The stability of sintering aids like Al2O3, Y2O3 or Yb2O3 is high in comparison to SiO2, SiO (g)
or CO (g), so that their chemical interactions with other compounds will be neglected.
3.5.2. Evolution of crystalline phases during in situ crystallisation

The knowledge of phase evolution during the in situ crystallisation of amorphous Si/C/N
powders is of special interest in the understanding of the sintering behaviour, as well as the
microstructure evolution of the sintered composite materials. Such investigations are reported in
several studies [332,341,343].
The in situ crystallisation of amorphous Si/C/N powders will be discussed in this section taking
into consideration an amorphous Si/C/N powder obtained by pyrolysis of a commercially available
polyhydridomethylsilazane (NCP 200) [332]. This amorphous powder has a chemical composition
SiN0.8C0.57O0.11 which means that a carbon excess of 6 wt.% is present. The evolution of the
crystalline phases was studied in a series of hot pressing experiments with temperatures between
1500 and 17508C and an annealing time of 30 min. These experiments were carried out in a graphite
furnace under a nitrogen atmosphere of 0.1 MPa. The results of XRD investigations are shown in
Fig. 20.
The crystallisation of the amorphous Si/C/N powder starts at 15008C and leads simultaneously
to the formation of a-Si3N4 and a-SiC. The intensity of the peaks related to the a-Si3N4 phase
increases rapidly above 15008C. A very small amount of b-Si3N4 could be detected above 16008C.
The crystallisation of SiC occurs progressively between 1550 and 16508C and accelerates above
16508C. The crystallised SiC is a mixture of the b-phase (3C-type) and the a-phase (2H- and 6H-
types). Raman-investigations (Fig. 21) were carried out in order to get insight into the correlation
between the evolution of free carbon content and the in situ crystallisation of SiC. It is obvious in
Fig. 20. X-ray diffraction patterns of amorphous Si/C/N powder derived materials sintered between 1500 and 17508C for
30 min in 0.1 MPa N2 [332].
Fig. 21 that the decrease in the intensity of the bands which characterise free carbon is accompanied
by an increase in the SiC signals intensity. After annealing at 17508C for 30 min, no free carbon
could be detected, which implies that reaction (2) is completed.
3.5.3. Densification behaviour

Powder mixtures consisting of amorphous Si/C/N particles, eventually crystalline Si3N4
particles and sintering aids are commonly densified by liquid phase sintering under hot pressing or
gas pressure sintering conditions in nitrogen, resulting in Si3N4/SiC composites. The sintering
behaviour depends on the sintering aids used since their amount, the viscosity of the liquid phase and
its wetting properties strongly influence this mechanism. Owing to the very low green density of
amorphous powder compacts, high amounts of sintering aids are required to proceed to full
densification by pressureless liquid phase sintering [344]. With the purpose of identifying the
different densification steps, shrinkage and temperature were recorded continuously during the
densification tests under hot-press conditions of the amorphous polyhydridomethylsilazane derived
Si/C/N powder. The sintering behaviour is investigated in dependence on various sintering aids
additions (Table 15). The densification curves are shown in Fig. 22.
Fig. 21. Raman spectra of amorphous Si/C/N powder derived materials sintered between 1500 and 17508C for 30 min in
0.1 MPa N2 [332].
Three densification steps can be distinguished in Fig. 22. In the first densification step, the
dependence of shrinkage on temperature is quite similar for the three considered powder mixtures.
Two processes are to be taken into consideration to explain this shrinkage. First, crystallisation of
amorphous Si/C/N powder particles occurs. Since the density of the amorphous Si/C/N powder
(2.35 g/cm3) is lower than that of Si3N4 or SiC crystallites (both 3.2 g/cm3), the crystallisation is
accompanied by crack formation in the Si/C/N powder particles. It follows that the original Si/C/N
amorphous powder particles will decompose to smaller ones, which are partly amorphous and
Table 15
Composition and content of sintering aids in the Si3N4/SiC composites 5Y/3Al, 2Y/6Yb and 5Y/2S, as well as lowest
eutectic temperature of the corresponding sintering aids systems
Si3N4/SiC composite 5Y/3Al 2Y/6Yb 5Y/2S
Sintering aids composition 5 wt.% Y2O33 wt.% 2 wt.% Y2O36 wt.% 5 wt.% Y2O32 wt.%
Al2O3 Yb2O3 SiO2
Sintering aids content in 4.3 2.6 4.3
starting-powder (vol.%)
Lowest eutectic temperature (8C) 1387 ca. 1550 1650
Fig. 22. Comparison of the hot pressing densification behaviour vs. temperature and time of an amorphous Si/C/N powder
mixed with various sintering aids compositions (see Table 15) [332].
crystalline, respectively. Second, a liquid phase is formed in the powder mixture in the temperature
range of the first densification step. The liquid phase causes a decrease of the friction forces between
the powder particles and leads to a decrease of the distance between the powder particles. These
processes are schematically described in Fig. 23A±C. Since a liquid phase will form above 13808C
in the sample 5Y/3Al (Table 15), the viscosity of the liquid phase is lower in comparison to the other
samples, thus allowing a faster particle movement. Nevertheless, the samples reached the same
shrinkage at 17508C.
The main densification takes place in step II (Fig. 23D). Three overlapping processes occur
during this step:
1. Further crystallisation of the powder particles. This is accompanied by a further decomposition of
the powder particles and by the reaction of Si3N4 with carbon, leading to the evolution of nitrogen
(see Section 3.5.2).
2. Particle rearrangement in the liquid phase.
3. Beginning of the solution-precipitation process, which is characterised by the a- to b-Si3N4 phase
transformation.
Although the volume of sintering aids and the viscosity of the liquid phases are varying
(Table 15), the shrinkage of the powder compacts in step II is found to be identical. It can be deduced
from these observations that the densification is governed and controlled by the evolution of nitrogen
according to reaction (82). The rearrangement of the powder particles in the liquid phase appears to
be the main contributing factor to the densification of the amorphous Si/C/N powder.
The densification occurring during step III (Fig. 23E) is attributed to the solution-precipitation
mechanism. The sintering behaviour in this step is clearly dependant on the amount and viscosity of
the liquid phase. The sample 5Y/3Al attained its final shrinkage rapidly while densification of the
sample 5Y/2S slowly continued during the whole holding time. The lower viscosity of the liquid
phase in the sample 5Y/3Al caused the solution precipitation mechanism to be accelerated.
Fig. 23. Schematic representation of the sintering process: (A) crystallisation of a-Si3N4 and formation of a graphite
dispersion; (B) crystallisation of a- and b-SiC, formation of a liquid phase, beginning of the reaction 2 between carbon and
Si3N4; (C) powder particle decomposition; (D) particle rearrangement in the liquid phase; (E) solution±precipitation
process.
In all cases fully dense Si3N4/SiC-composites were obtained after 2 h hot pressing at 18008C.
X-ray diffraction analysis of the hot-pressed Si3N4/SiC composites revealed that the a- to b-Si3N4
phase transformation is completed in the 5Y/3Al composite. In materials 5Y/2S and 2Y/6Yb,
a-Si3N4 could be detected. From the very low intensity of the b-Si3N4 reflexes in the 2Y/6Yb
composite it can be deduced that the a-Si3N4-phase is stabilised. Moreover, the X-ray diffraction
analysis shows that the grain boundary partly crystallised to Yb4Si2N2O7. The presence of the
Fig. 24. Weight-loss occurring during hot pressing densification vs. oxygen content of the starting powder and hot pressing
temperature [332].
a-Si3N4- and Yb4Si2N2O7-phases in the 2Y/6Yb composite has major consequences for the
mechanical properties (see Section 4.2).
Supposing that the 6 wt.% excess of carbon present in this amorphous Si/C/N-powder reacts
according to Eq. (82), a theoretical weight loss of 9.5 wt.% should be observed. The weight loss
measured after hot-press experiments is substantially higher and is found to be in the range of 11±
15 wt.%. This is due to the carbothermal reduction of SiO2 or SiO (g) by carbon. These chemical
interactions lead to a weight-loss depending on the oxygen content of the starting powder and the
hot-pressing temperature (Fig. 24).
3.6. Dense Si/C/N materials of small dimensions
3.6.1. Moulding of miniaturised structures

Freimuth et al. [345] reported that complex ceramic parts of very small dimension can be
prepared using a combination of lithography, casting of a polyvinyldisilazane containing solution
followed by a pyrolysis process. This route leads to nearly perfect, complementary bodies of the
original miniaturised moulds. It can be explained by the high smoothness of the surface of the
amorphous ceramics after pyrolysis, which is characteristic for pyrolysed warm-pressed bodies. In
comparison, the slurry casting of ceramic powders has limitations because of the presence of
agglomerates. Due to the fact that the ceramic powder is constituted from grains, the roughness of
the surface of the miniaturised structure is much higher than that of the amorphous material. A
similar process was carried out with a polyhydridomethylsilazane [346].
3.6.2. Fibres
Compared to the conventional powder processing methods, the hybrid-processing route has
several advantages. One of the major benefits is the possibility to draw fibres from precursor
materials. Yajima et al. [241] pioneered this process in the seventies in succeeding to form
continuous SiC-fibres from a polycarbosilane precursor. SiC- and Si3N4-fibres have attracted
considerable interest in the last two decades because they reveal a combination of excellent
mechanical properties, combining high tensile rupture modulus with good oxidation resistance
[176,347±349].
Since the ceramic precursors are melt spun on a filament device into fibres, the viscosity of the
polymers is one of the main parameters of the process. The drawn fibres have to be carefully cross-
linked to make the precursor infusible, enabling shape stability and high ceramic yield. The
precursor to ceramic transformation is performed by a controlled pyrolysis process.
In 1973, Verbeek [273] described the fabrication of Si3N4/SiC fibres which were prepared by
pyrolysing a precursor fibre made by melt spinning a polycarbosilazane. The melt-spun green fibre
was exposed to moist air at 1108C and then pyrolysed at 15008C in an inert atmosphere. High purity
perhydropolysilazane derived amorphous Si3N4 fibres were developed by Tonen Corp. [81]. The
crystallisation of these fibres under nitrogen seems to proceed by the vapour phase growth
mechanism. Commercially amorphous Si3N4/SiC fibres are drawn from HPZ (Dow Corning Corp.)
[149] or methylpolydisilazane (MPDZ) (Hoechst Celanese) [249]. After curing in the presence of
trichlorosilane and pyrolysing at 12008C, the surface of the fibres is coated with a microporous silica
rich layer of approximately 0.2 mm thickness [350].
It was established that the introduction of the heteroatom boron into the Si/C/N-fibres is a
suitable way to improve their temperature resistance [351]. One processing route involves modifying
the surface of Si/C/N-fibres drawn from a polysilazane precursor by subsequently curing in an
atmosphere containing boron trichloride and ammonia. SEM investigations on cross-sections of
these fibres after annealing at 18008C reveal that the grains formed during crystallisation are smaller
in boron-containing fibres compared to boron free SiC-fibres. These results are consistent with the
above mentioned improved resistance to crystallisation of polymer derived Si/B/C/N-materials
[224].
Other methods comprise an exposure during the infusibilisation stage to an HCl containing
atmosphere, possibly followed by a treatment in a water or an ammonia gas atmosphere [352].
In other studies, the surface of hydridopolysilazane derived Si3N4/SiC fibres was coated with a
BN, BN/SiC, or BN/Si3N4 layer by a chemical vapour deposition process. This additional step was
found to influence both tensile strength and Weibull modulus of the fibres [350]. A comparison of
characteristics and properties of some precursor derived SiC, Si/C/N and Si/B/C/N-fibres is
presented in [20,32,223,224].
3.6.3. Coatings
High temperature applications of oxygen-sensitive materials in oxygen-containing atmospheres
can only be realised if they are protected by a pore free and adhesive coating. Ceramic coatings, in
particular, are of potential value in industry as a means of providing wear and corrosion resistance in
materials used in adverse environments. The requirements for the generation of coatings are
complex. To realise a good adhesion of the ceramic coatings on the substrates, the mismatch of the
coefficients of thermal expansion between the coating and the substrate should be minimised. The
coating should act as an oxygen barrier and the oxygen diffusion to the substrate has to be prevented
as does the diffusion of oxygen sensitive compounds from the substrate to the atmosphere. The
conditions for a long resistance of the coating include low partial pressure and good temperature
interaction behaviour. The expected properties of a coating are summarised in Fig. 25.
A variety of methods can be used for the fabrication of ceramic coatings. Using the CVD
method, the thickness of the coating can be adjusted between 0.1 and 100 mm. Ceramic coatings with
a thickness between 0.01 and 20 mm can be realised using the physical vapour deposition technique
Fig. 25. Overview of the requirements to be fulfilled in order to achieve an efficient oxidation protective coating.
(PVD). Coating with polymeric precursors utilising simple techniques such as spin- or dip-coating,
followed by pyrolysis at relatively low-temperatures offers a versatile and potentially more
economical route. The thickness of polymer coatings can be varied between 0.3 and 20 mm. The
fabrication of ceramic coatings by polymer pyrolysis was first described by Coblenz et al. [77]. They
showed that an improvement in the oxidation resistance of graphite-based materials can be reached
by coating and infiltration with Si/N or Si/B/C/O precursors. More recently, Mucalo et al. [89] reported
the fabrication of approximately 2 mm thick crack-free ceramic coatings on Al2O3 substrates. The
large shrinkage associated with the polymer-to-ceramic transformation leads to delamination and to
formation of cracks if the thickness of the coating exceeds a critical value, typically in the region of
1±2 mm. These problems can be overcome by dispersing appropriate fillers within the precursor
material. This reduces the thermal coefficient mismatch between the coating and the substrate, which
is Ð in addition to the shrinkage during pyrolysis Ð responsible for the formation of cracks
[321,353] and the fabrication of adherent crack-free ceramic coatings can then be achieved.
Langlois and Lum [354,355] spread a polysilazane solution containing AlN particles as a filler
on metals such as Mg, Ti or Al. The coatings have been shown to contribute to an improvement in
oxidation resistance. A second possibility is adding Al2O3 as filler particles to the polysilazane.
Depending on the Al2O3 content, this method is appropriate for realising dense, adhesive and crack-
free coatings [321]. Fig. 26 shows a SEM microgragh of a dense layer utilising a
polyhydridomethylsilazane/Al2O3 slurry after pyrolysis on a Si±C±O substrate.
Metals are also commonly used as fillers. Heimann et al. [356] produced coatings by using a Si-
powder/polysilazane suspension to protect carbon fibre composites from oxidation. A significant
improvement of the oxidation resistance in the temperature range between 1160 and 12608C was
achieved. Recently, promising results were achieved by using polysilazanes, which were modified by
incorporating metal-containing compounds [357]. In this case, no filler was added to the polymer
slurry. It was reported that transparent and adhesive coatings can be realised with these modified
polysilazanes on a variety of surfaces including metals, glasses or plastics.
In addition to the fabrication of Si/C/N amorphous coatings, polysilazanes have been recently
employed to produce silica films at low temperatures, mainly for applications in electronics. A
solution containing a polysilazane and catalysts like amine compounds, acid compounds and
Fig. 26. Scanning electron micrographs of a layer obtained by dip-coating using a polyhydridomethylsilazane/Al2O3 slurry
followed by 11008C pyrolysis in Ar: (a) lower magnification and (b) higher magnification demonstrating Al2O3 (bright)
particles homogeneously distributed in a pore-free Si/C/N layer (dark) [322].
peroxides is deposited upon a substrate such as plastic films or substrates used in electronic products
[358]. The coated substrates are then heated to very low temperatures in the range of 50±1008C
under oxidising atmosphere to form amorphous SiO2 films. This processing has to be considered as a
simple way to achieve highly pure amorphous SiO2 films. Such films find suitable applications as
layers on microelectronic components, optical materials as well as chemical inert coatings and
barrier coatings for a variety of substrate materials. By dispersing UV-absorbing fine metal oxide
particles in the polymer solution, UV-shielding SiO2 coatings can be realised using this technique
[359].
Ceramic coatings are frequently used as oxidation protective materials. They may also find
application as a means to improve wear resistance. For example moulds can be efficiently protected
from wear in this way [360].
An alternative route to produce ceramic coatings consists of depositing thin polymeric films
with a CVD technique and subsequent pyrolysis. This process may be applied to produce oxidation
protective coatings on carbon fibres. A Si/B/C/N ceramic coating of 1±5 mm thickness on the surface
of carbon fibres was achieved using a CVD process with a boron containing polysilazane as
precursor [361]. The deposited polymeric film was subsequently pyrolysed. This process leads to the
formation of ceramic coatings which have strong adhesion and offer an excellent oxidation
protection.
Ceramic coatings on powder particles can act as a binder phase to produce dense composites.
Nearly dense Si/C/N-green bodies containing 30 wt.% diamond particles after pyrolysis could be
achieved by warm-pressing of diamond particles previously coated with a cross-linked
polyhydridomethylsilazane precursor [321]. The subsequent pyrolysis of the green bodies leads to
dense Si/C/N-diamond composites.
In a similar way, ZrB2- or Si3N4-powder particles can be bound together after being coated first
with a polysilazane precursor [362,363]. The green bodies are then ceramised and densified by a
sintering process.
3.6.4. Joining
Ceramic precursors may be used to realise precise, high temperature resistant joins with good
strength. Joining methods have become an active area of research in the industry as a means to
realising complex ceramic structures by assembling several Si3N4-, SiC-parts [365±369] or
composite materials [370]. This method also finds application in the electronics industry for the
adhesion of functional components to substrates [371]. Since the requirements for joinings are
similar to those for coatings, polysilazanes used as preceramic polymers are mixed with filler
particles like SiC-, Si3N4- or Si-powders to prevent shrinkage during the polymer-to-ceramic
transformation. In the case of joinings between ceramic parts, sintering aids may be added. The
fracture resistance of a joining produced by this process is sufficient for a static use of the combined
ceramic parts in various fields.
3.7. Microporous ceramics and membranes
The synthesis of microporous silazane-derived ceramics exhibiting pores in the nano- to micro-
meter range attracted interest recently since these materials may find applications as membranes or
filters with better thermal resistance and stability in reducing environments than oxide materials.
Low molecular weight polysilazanes are suitable precursors for microporous ceramics if the
pyrolysis is carried out under an ammonia containing reactive atmosphere at low average heating
rates (<28C/min). These porous materials are stable over a pyrolysis temperature range of 500±
10008C. Low heating rates appear to be the most important factor for the synthesis of microporous
materials. In comparison, pyrolysis of the same precursor under inert atmosphere results in
microporous materials, only if the pyrolysis temperature is in a limited range of 400±6008C. One
other route to achieve stable microporous materials consists of producing nano- or microcolloids by
adding ceramic or metal particles to the polysilazane. The pyrolysis of the colloid can be carried out
either under inert atmosphere (Ar, N2) or ammonia. Moreover, microcellular foams can be
manufactured by replicating a dissolvable particle preform with a polysilazane solution and
subsequently leaching away the preform and pyrolysing the polymeric polysilazane skeleton [372].
If polysilazanes are commonly employed as Si/C/N precursors, they may also be used to realise
SiO2 coatings (see Section 3.6.3.) and SiO2 membranes. Solutions consisting of a polysilazane, metal
alkoxides and a ceramic formation accelerator were applied on porous inorganic supports [373].
After drying and firing at 500±7008C in oxidising atmosphere, these materials lead to (micro)porous
SiO2 membranes which may find applications as filters, catalyst and enzyme supports, respectively.
3.8. Infiltration of porous preforms
The solubility or low viscosity of ceramic precursors suggest potential processing routes for the
infiltration of preforms. The polymer infiltration/pyrolysis technique offers distinct advantages over
the CVI technique since the first one implies a more simple and economical processing. Both
methods are commonly used to decrease the porosity of preforms consisting of whiskers and/or
fibres with the purpose of enhancing the mechanical properties. The processing has been revealed to
be a viable method of manufacturing continuous fibre-reinforced ceramic-matrix composites in
complex shapes.
Similar to the coating techniques, special attention has to be focused on the volume shrinkage
occurring during pyrolysis. The formation of internal stresses is governed by the thermal mismatch
between the fibres and the matrix and strongly influence the fracture strength of the composites
[374]. The shrinkage can be diminished or prevented by adding fillers to the ceramic precursor. The
properties of the preform, and the viscosity and ceramic yield of the polymer are relevant processing
parameters. The relationships between these properties and the molecular structure of the
polysilazanes were investigated systematically [97]. It was established that a processing without
solvent can be carried out with polymers showing a low molecular weight and a low viscosity. High
ceramic yields up to 80% were reached by branching the oligomers, thus reducing the gas evolution
during pyrolysis.
Three-dimensional preforms consisting of carbon-, SiC-, Si3N4-, Al2O3-fibres or a combination
thereof were infiltrated with a polysilazane containing slurry in order to improve the density of the
matrix [236,356,372,374±376]. Simultaneously this processing results in the formation of an
oxidation protective coating of the carbon fibres. The shrinkage of the precursor can be compensated
for by the volume extension occurring during reaction of a filler like silicon with nitrogen to Si3N4.
Nitrogen may evolve as a pyrolysis product or be present in the furnace atmosphere. Subsequent to
the impregnation of the preforms with a slurry containing Si-powder, the nitridation of silicon was
achieved by annealing the samples in nitrogen atmosphere at 14008C [356,377]. An alternative
approach to reduce the thermal mismatch between fibres and matrix is the use of inactive fillers like
Si3N4 [378]. To obtain dense composites, impregnation, curing and firing processes were repeated.
Combining CVI and polymer infiltration/pyrolysis cycles results in a higher interface strength
between the fibres and the matrix constituted by the pyrolysed polymer [379].
Si/C/N monoliths obtained after pyrolysis of cold uniaxial or isostatically pressed cross-linked
powder compacts are porous (see Section 3.4.2) because of the gas evolution taking place during the
pyrolysis. The volume of the pore channel system can be minimised by infiltrating the monoliths
with a polysilazane solution and subsequent pyrolysis [380].
4. Structure and properties of silazane derived bulk ceramic materials
4.1. Amorphous Si/C/N materials
4.1.1. Structural characteristics

In most ceramic processing techniques such as the fabrication of ceramic fibres, ceramic
powders or bulk specimens, pyrolysis of polysilazane precursors is conducted at around 10008C
since the organic±inorganic transformation is completed up to this temperature. The pyrolysed
amorphous ceramics are thermodynamically metastable Si/C/N or Si/C/N-containing materials.
X-ray diffraction (XRD) analysis of the polysilazane derived Si/C/N products normally gives a
featureless diffraction pattern. However, under careful TEM and SEM examination the presence of
microsize or nanosize crystalline SiC in the amorphous matrix was frequently observed, both in the
as-pyrolysed ceramics, as well as in samples subsequently annealed below 11008C. This effect was
found for a variety of SiC precursors, [381±383]. Similarly, carbon-free polysilazanes have a strong
tendency to give crystalline products. Pyrolysis of perhydropolysilazane between 1100 and 11508C
yielded crystalline a- and b-Si3N4 that could be clearly detected by XRD [61,274]. Thus, the
structure of the polymer derived ceramics strongly depends on the chemistry of the precursor. The
formation of amorphous Si/C/N is favoured when polysilazanes are used as precursors while
formation of crystalline products is increased in the case of C or N-free Si-polymer precursors.
Even for the X-ray amorphous Si/C/N ceramics derived from polysilazane precursors, the
structural configuration may be significantly different, changing with precursor structure and with
pyrolysis conditions. Recently, an increasing number of studies on the understanding of the micro
structure of X-ray amorphous Si/C/N ceramics has been published. Various structural characterisa-
tion techniques were performed to provide information of the amorphous ceramics on an atomic
level including NMR [114,271,282,384], FTIR and Raman [385] neutron diffraction [138] and XRD
in terms of radial distribution function analysis [386,387] and electron spin resonance (ESR) [388] as
well as high resolution transmission electron microscopy (HRTEM) [9,12,324,420].
Solid state NMR and FTIR spectra were performed on the X-ray amorphous pyrolysis products
derived from poly(hydridomethyl)silazane (PHMS, NCP200) and polyvinylsilazane (PVS, VT 50)
around 10008C revealing two different bonding configurations of the amorphous materials
[114,271,282,383]: the PHMS derived ceramics obtained around 10008C have a short range order
of the elements consisting mainly of tetrahedral SiCxNy (xy4) units with some SiN2CH
tetrahedra. In contrast, the PVS derived ceramics consisted of SiN4 tetrahedral units and sp2-
hybridised carbon. These results indicate that even minor changes in the functionality of the
precursors can have a large impact on the structure of the pyrolysed ceramics. Interestingly, a similar
structure consisting of SiN4 units mixed with amorphous carbon was deduced from a poly(N-
methylsilazane), [MeHSiNMe]x. The ceramic materials were obtained at 14008C and characterised
using FTIR, NMR and Raman analysis [138]. However, neutron diffraction analysis of the PVS
derived ceramics gave a average SiN3C tetrahedron unit [385]. Besides, the perhydropolysilazane
derived silicon nitride ceramic fibres were characterised by XRD in terms of radial distribution
functions [386]. Results showed that either a a- or b-Si3N4-like structure or a mixture of them exists
in the fibres [386]. X-ray and neutron diffraction analysis of the pyrolysed products of NCP 200
obtained at 10508C under argon showed the existence of amorphous Si3N4 and excess C in the form
of graphite-like amorphous carbon [387]. This configuration is different from the a- and/or b-Si3N4-
like structure of the Si/C/N fibres [386]. It is also in contrast to the tetrahedral SiCxNy (xy4) units
of the products derived from the same kind of precursors [114,271,282]. The diversity of structures
might not only be caused by different processing conditions but might also be due to slight
differences in precursor compositions and structures.
Structural characterisation of the Si/C/N films deposited from TMS-NH3-H2 at 1000±12008C by
XPS revealed that the products consist of a silicon carbonitride rather than a mixture of Si3N4 and
SiC [313]. Gheorghiu et al. [389] and Tenegal et al. [390] investigated the structure of nanosize Si/C/
N powders by Si K-edge extended X-ray absorption fine structure (EXAFS) and proposed a different
atomic structure. The former group found a local SiCN3 arrangement around Si while the latter
researchers concluded that the Si atoms are randomly linked to C and N in a C±N network. In
another work, magic angle spinning nuclear magnetic resonance (MAS-NMR) and ESR were
performed on CO2 laser pyrolysed ceramic particles obtained from a SiH4C2H4NH3 mixture. It
was found that the structure of the particles strongly depends on the nitrogen content. The lower N
containing particles consisted of a SiC phase with dissolved N. The N atoms were partly present in
SiC microcrystal and partly in the grain boundary layer between the particles. For the high N content
powders, they consisted of crystalline SiC, Si3N4 and residual amorphous phases [391]. The wide
variation of the structure and composition was partly due to the pyrolysis conditions during the CVD
process being far from the equilibrium. Besides, the films deposited on single crystal Si from a
mixture of SiH4CH4N2 using an electron cyclotron resonance plasma were found to have a
totally amorphous Si/C/N structure [392].
In the Si/C/N system it is well known that only two thermodynamically stable phases exist:
Si3N4 and SiC. No ternary crystalline phases in the Si/C/N system was observed in conventional
inorganic synthesis or reactions involving this system. Recently, work on poly(organosilylcarbo-
diimides) showed that new ternary phases Ð SiC2N4 and Si2CN4 Ð can be synthesised [8,393±395].
On the Si/C/N phase diagram these two phases are located on the tie line of the hypothetical C3N4
phase and the Si3N4 phase. In situ X-ray powder diffraction revealed that the amorphous SiC2N4
transformed to crystalline phases above 4008C. The crystalline SiC2N4 undergoes a reversible phase
transition at 1508C. The high temperature phase of SiC2N4 has cubic structure with anticuprite
structure topology. The silicon carbodiimide SiC2N4 is thermally stable under Argon up to 9008C. It
transforms to crystalline Si2CN4 above 9508C. HRTEM revealed the crystalline Si2CN4 phase with
nanosised grains in the range of 20±70 nm. SiC2N4 and Si2CN4 are the first crystalline phases in the
ternary Si/C/N system. However, it is unlikely that these phases are formed during the pyrolysis of
silazane precursors since both structures contain the unsaturated N=C=N or carbodiimide unit. Upon
annealing at T>1400±15008C, the Si2CN4 phase decomposed into amorphous Si/C/N and then
transformed to the thermodynamically stable phases Si3N4 and SiC at T>15008C.
New crystalline ternary Si/C/N phases may also be formed in CVD deposited materials, as
reported recently by Chen et al. [396]. They claimed that crystalline Si/C/N films were grown by
microwave-enhanced chemical vapour deposition using H2CH4N2SiH4 mixtures and deposi-
tion temperatures between 400 and 12008C. XRD showed that the films were a mixture consisting of
a-Si3N4 and a new hexagonal Si/C/N phase with lattice parameters of a0.54 and c0.67 nm [397].
The N content of the new ternary phase is in the region of 35±40 at.%, whereas the Si content
amounts only to 10 at.% [396]. Nitrogen K edge XANES spectra showed that the crystalline Si/C/N
had a similar local environment around the nitrogen atoms compared to a-Si3N4, and silicon K edge
XANES and EXAFS spectra confirmed the existence of a local tetrahedral SiCN3 arrangement with
a long range ordered crystalline like structure [398]. Electrical and optical measurements revealed
that the new Si/C/N phase had a direct band gap around 3.8 eV at room temperature which makes the
material a potential candidate for blue and UV emitting optoelectronic devices [396].
4.1.2. Crystallisation and thermal stability

The thermal stability of the polymer derived amorphous ceramic materials includes (1)
chemical stability that is related to the decomposition of the materials and (2) structural stability
which is associated with crystallisation and phase transition processes. The thermal decomposition
and crystallisation may induce extensive porosity and stress hence degrading the mechanical,
physical and chemical performance of the ceramics. Thus, stabilisation of the structure of the
materials is necessary. For a comparison and discussion of the crystallisation behaviour of polymer
derived ceramics it is important to note that there is no clear definition for the onset point of
crystallisation. On the one hand, it strongly depends on the heating rate and holding time as well as
other experimental parameters like the composition of the atmosphere or the crucible material. On
the other hand, the method of identification of crystalline phases and the crystallite size is crucial: by
TEM smaller crystallites are detectable compared to XRD. Similarly, thermal stability measured by
Fig. 27. Weight loss vs. heat-treatment temperature for various ceramics submitted to standard TGA conditions [399].
weight-loss has to be considered in combination with the experimental set-up, i.e. heating rates,
annealing times, atmospheres, etc.
Polysilazane derived amorphous Si/C/N ceramics were found to have a structural and chemical
stability comparable to the binary SiC or SiNX amorphous ceramics, derived from N or C free
polysilane, oxygen-containing ternary or quaternary phases like Si/C/O and Si/C/N/O, derived from
oxygen-containing polysilane precursors [151,154,155,157,399±402]. Thermal stability investiga-
tions showed that the polycarbosilane derived SiC has a higher chemical stability, decomposing
above 16008C compared to the polysilazane derived Si/C/N systems (Fig. 27). However, amorphous
SiC tends to crystallise at relatively low temperatures around 1000±11008C compared to the Si/C/N
ceramics which withstand crystallisation up to 14008C.
For the oxygen containing systems Ð namely Si/C/O and Si/C/N/O Ð the decomposition
temperature is slightly decreased compared to the Si/C/N system. Additionally, an increased degree
of decomposition was observed (Fig. 27), due to a significant evolution of oxygen or oxygen
containing gases like CO and CO2 at elevated temperatures. If there is oxygen in polysilazane or
polysilane derived Si/C/N (or SiC) fibres, generally introduced during the process of fibre curing in
air, the mechanical performance of the fibre material was found to be greatly decreased compared to
oxygen free fibres. However, the presence of oxygen impurities had positive effects enhancing the
oxidation resistance at high temperatures [158]. The crystallisation behaviour of the polysilazane
derived Si/C/N or Si/C/N containing ceramics is closely related to the molecular architecture of the
precursors. Most kinds of polysilazane derived amorphous Si/C/N ceramics remain amorphous up to
14008C under inert atmospheres. This crystallisation temperature is higher than that of CVD-
produced amorphous Si3N4 which starts to crystallise between 1200 and 14008C [403,404].
However, the Si/C/N ceramics derived from polycarbosilazanes began to crystallise between 1250
and 14008C [151]. The crystalline phases were found to be b-SiC and free carbon [151].
Also, the presence of other elements in the amorphous Si-based ceramics has a large influence
on the chemical and structural stability of the materials. Amorphous silicon nitride derived from
chlorine containing reactants was found to show a crystallisation temperature 2008C higher than that
for chlorine free polysilazanes which started to crystallise around 11008C under nitrogen [405].
However, the activation energy of crystallisation obtained from a modified Avrami-Erofe'ev
Equation was almost constant in the two cases: The amorphous Si3N4 derived from the chlorine free
and chlorine containing polysilazanes was determined to be 31845 kJ/mol [405,406]. The kinetics
of the phase transformation agreed with the laws of diffusion controlled reactions with one-
dimensional grain growth and spontaneous nucleation. The retardation of the crystallisation was
explained by aggregation of the chlorine atoms between the amorphous and crystalline phases
inhibiting grain growth. This is probably associated with the higher bond energy of Si±Cl bonds (398
kJ/mol) in comparison to Si±N bonds (335 kJ/mol) which decreases the nucleation rate and the
diffusion. As a result, the a/b transformation was also retarded as observed previously [407].
Therefore, the chlorine-doped Si3N4 or Si/C/N may show an enhanced thermal stability compared to
the undoped materials. In this respect, incorporation of a small amount of boron can have dramatic
effects on the structure, raising the crystallisation temperature up to 18008C [14,212].
However, incorporation of hetero elements does not always cause a structural stabilisation. It
was shown that the presence of about 1 wt.% phosphorus in the Si/C/N system decreases the
crystallisation temperature from 1400 to 15008C to 13508C [282,383]. Similarly, Si/N or Si/C/N
materials containing oxygen may show enhanced crystallisation and grain growth. These effects of
oxygen are due to various vapour phase reactions involving SiO and/or CO2/CO gases producing
Si3N4 or SiC in the form of needles or whiskers, as frequently observed after thermal annealing of
the amorphous materials.
The crystallisation of amorphous Si/C/N was also found to be related to the microstructure of
the materials such as porosity and particulate interface. Kleebe et al. [324] recently observed that the
bulk-pyrolysis derived amorphous Si/C/N ceramics from a mixture of cross-linked powders and a
liquid polyvinylsilazane did not crystallise up to 15408C for 6 h in nitrogen atmosphere, while the
ceramics produced from a mixture of powders pyrolysed at 10008C with the same liquid
polyvinylsilazane crystallised to form SiC and Si3N4 grains which were predominately located on
the surface of the specimens. The chemical stability of the Si/C/N ceramics is also closely related to
the elemental Si:C:N ratio. Most kinds of silazane derived ceramics are C rich with respect to
stoichiometric Si3N4±SiC phases. The excess of carbon greatly degrades the stability of the Si/C/N
materials by inducing decomposition and reaction with Si3N4 to form SiC according to Eq. (82).
Also, the elemental carbon tends to crystallise to form graphite phases which greatly impairs the
mechanical properties of the materials. Therefore, adjusting the Si/C/N composition on the Si3N4±
SiC tie line in the Si/C/N ternary phase diagram (Fig. 5) would enhance the chemical and structural
stability of the materials [257].
However, formation of free carbon always seems to be the first crystalline phase, even for the
Si±C system containing low amounts of carbon (C/Si0.7) as observed in the different systems Si/C,
Si/C/N, Si/C/O and Si/C/N/O produced from a variety of Si-polymer precursors [399]. HRTEM
revealed that the nucleated carbon was in the form of two to three polyaromatic layers of 1 nm in
lateral extension piled up as plates in turbostratic order. The stability of the free carbon is in part due
to the presence of small amounts of hydrogen which saturate the peripheral carbon atoms. Retention
of a certain amount of hydrogen in the pyrolysis products supports the assumption that hydrogen
stabilises the observed turbostratic graphite. Based on these microstructural characteristics, it was
argued that an interconnected carbon network may be formed by controlling the amount of carbon in
the starting materials. Improved mechanical properties were expected for these amorphous materials,
with simultaneously enhanced structural stability [399]. The latter was based on the assumption that
the presence of excess carbon could hinder the nucleation of SiC which is usually the second
crystalline phase following the graphite [399].
Incorporating boron into Si/C/N to form quaternary Si/(B)/C/N can greatly retard crystallisation
and enhance the thermal stability of the materials [14,212]. The amorphous Si3.0B1.0C4.3N2.0
Fig. 28. Thermal gravimetric analysis of powdered polycrystalline Si3N4, a silicon carbonitride (Si1.7C1.0N1.6) derived
from PHMS (NCP 200) and a silicoboron carbonitride (Si3.0B1.0C4.3N2.0) under stationary 0.1 MPa He in a graphite
crucible; heating rate: 5 K/min [248].
ceramics derived from polyborosilazane resist crystallisation below 17008C. This upper limit for the
crystallisation is higher by 300±4008C than for ternary Si/C/N ceramics. Annealing of the
Si3.0B1.0C4.3N2.0 at 17008C for 50 h in Ar atmosphere resulted in the formation of b-Si3N4 and b-
SiC. XRD did not show the presence of crystalline carbon and boron nitride. But FTIR revealed the
existence of amorphous boron nitride in the annealed products. Based on chemical analysis, XRD
and FTIR, the crystallisation of the silicon borocarbonitride can be expressed by Eq. (91).
>1700 C=0:1MPa=Ar
Si3:0 B1:0 C4:3 N2:0 ÿÿÿÿÿÿÿÿÿÿÿ! b-Si3 N4 9b-SiC 4BN 8:2C (91)
The Si3.0B1.0C4.3N2.0 ceramics were found to have extraordinary thermal stability compared
with pure silicon nitride and other silicon-nitride-based materials [14]. Fig. 28 shows the TGA curves
of the quaternary Si3.0B1.0C4.3N2.0, a ternary Si/C/N and the binary Si3N4 ceramics which were
analysed under He atmosphere. The Si3.0B1.0C4.3N2.0 withstands thermal degradation up to 20008C
while significant decomposition of the Si/C/N and Si3N4 occurs around 16008C. The weight loss
between 1400 and 20008C for the Si3.0B1.0C4.3N2.0 materials is below 6 wt.% which was associated
with the evolution of oxygen-containing gases like B2O3, SiO and/or CO. Chemical analysis showed
that the Si3.0B1.0C4.3N2.0 ceramics annealed at 20008C for 2 h show almost no change in
composition, indicating the ultra-high thermal stability of the materials.
When heating above 20008C the decomposition rate of the ceramics increases. Since
crystallisation completes below 20008C, the enhanced thermal degradation above 20008C was
correlated with the decomposition of Si3N4 (Eqs. (81) and (82)). The overall reactions at T>20008C
can, therefore, be expressed by Eq. (92).
>2000 C=0:1MPa=Ar
Si3:0 B1:0 C4:3 N2:0 ÿÿÿÿÿÿÿÿÿÿÿ! 3b-SiC 1BN 1:3C 0:5N2 (92)
4.1.3. Mechanical properties

Some mechanical properties of bulk silicon carbonitride amorphous ceramics derived from the
polyhydridomethylsilazane NCP 200 have been characterised. The mechanical strength was
375 MPa and the Vickers hardness was determined to be in the region of 9.5 GPa. The hardness
value is comparable to that of reaction bonded silicon nitride [408], but is significantly lower than
that of sintered and crystalline SiC (27±30 GPa) and Si3N4 (18±20 GPa), respectively. This may be
due to the presence of appreciable open porosity in the ceramic bodies (7%). Recent improvement of
the pyrolysis process led to the formation of dense Si/C/N ceramics with significantly enhanced
hardness attaining 16 GPa [322].
Due to the fact that the polymer derived Si/C/N ceramics contain no second phase in the
monolithic amorphous network, increased high temperature properties can be expected. Recent work
showed that the polysilazane derived amorphous Si/C/N (Si1.7C1.0N1.5) and polyborosilazane derived
Si/B/CN (Si2B1.0C3.4N2.3) exhibited an exceptionally low creep rate at high temperatures [9,10]
(Fig. 29). The viscosity of the silicon carbonitride and silicoboron carbonitride was three- and six-
orders of magnitude higher than that of fused silica. However, at T>15508C the viscosity is about
two-orders of magnitude below that of CVD derived [409] or reaction bonded SiC [410].
Fig. 29. Comparison of the high temperature viscosity of amorphous Si/B/C/N materials with that of amorphous Si/C/N
and silica as well as polycrystalline Si3N4 and SiC [10].
Fig. 30. Change of the strain rate in Si/(B)/C/N vs. time showing three different stages [10].
Both the amorphous materials exhibited a three-stage creep behaviour: (1) the creep rate declined
which was associated with the further densification; (2) the strain rate approached a steady state and
(3) it resumed a declining strain rate that ultimately decreased below the measuring limit of the
system (Fig. 30). The density of the tested Si/C/N ceramics increased from 2.06 at the beginning of
the experiment to 2.30 g/cm3 at the end of the stage (1) and then remained constant during stages (2)
and (3). The enhanced creep resistance in stage (3) was related to crystallisation in the samples (about 5%
crystalline Si3N4 was detected in the 12808C tested Si/C/N ceramics). The evolution of hydrogen in
particular was associated with the presence of small pores in the as-synthesised Si/C/N ceramics.
4.1.4. Oxidation resistance

The crystallisation of silicon carbonitride under argon or nitrogen atmosphere starts at
temperatures exceeding 14008C and is accompanied by a significant weight loss due to the
generation of nitrogen (see above) [12]. As a consequence, both the content of open porosity and the
mean pore radius increase.
This is related to the in situ crystallised Si3N4 and C formed during phase partitioning of silicon
carbonitride according to Eq. (1). Simultaneously, carbon reacts with Si3N4 to give SiC and N2 (Eq.
(2)) which are thermodynamically stable products at T>14408C. The overall solid state reaction
given in Eq. (3) can be verified quantitatively by chemical analysis.
It was found that this undesired degradation of silicon carbonitride in an inert environment
above 14008C can be circumvented by heating the material in air. During oxidation a cristobalite
layer of about 2 mm thickness builds up on the surface of the ceramic substrate [12]. The weight gain
during oxidation as measured by TGA indicates an initial reaction controlled mechanism (parabolic
curve) which is transformed into a diffusion-controlled mechanism (asymptotic curve), see Fig. 31.
Fig. 31. Thermogravimetric analysis of the oxidation behaviour of additive free Si/C/N ceramics with open porosity of 7%
and mean pore radii of 75 nm at different holding temperatures in air.
The oxidation resistance was remarkably enhanced in comparison to that of conventionally

fabricated Si3N4-based ceramics. Subjected to the same heat treatment additive doped Si3N4 was
completely oxidised. In this case, the surfacial SiO2 connects to the intergranular oxynitride
secondary phase which exhibits a significant higher diffusion coefficient for oxygen and thus
provides a proper diffusion path. Consequently the complete Si3N4 matrix internally was oxidised.
The improved thermal stability of bulk silicon carbonitride covered by a protective silica surface
layer can be explained by the low nitrogen diffusion coefficient through SiO2 which hinders the solid
state reaction (Eq. (2)). Thus, the equilibrium of the reaction given in Eq. (2) is shifted to the side of
the educts, enabling the formation of novel metastable bulk composite materials in the ternary
system Si/C/N, with Si3N4, SiC and C as discrete phases (Eq. (1)). Moreover, due to the low oxygen
diffusion coefficient of 10ÿ11±10ÿ6 cm2/s at 900 and 14008C, respectively, the passivating dense
SiO2 layer effectively protects the bulk silicon carbonitride from further oxidation [12].
4.1.5. Electrical properties

In general, ternary Si/C/N materials were characterised by semiconductive electrical behaviour
[151,155]. The electrical properties of this class of materials strongly depend on the starting
composition of the precursor material, the type of microstructure obtained and the utilised processing
technology in terms of pressing and pyrolysing conditions [411,412]. Depending on the above
mentioned parameters, the dc-conductivity of polysilazane (PHMS) derived amorphous Si/C/N
materials can be adjusted between 100 and 10ÿ15 Oÿ1 cmÿ1 at room temperature [413]. Decreasing
the carbon content of the Si/C/N ceramics by performing the pyrolysis in a reductive ammonia
atmosphere results in decreased dc-conductivity, whereas an increased conductivity is observed by
annealing the material in nitrogen atmosphere. Starting phase separation and beginning
crystallisation at annealing temperatures of 13008C results in high conductivity hence this is the
limiting temperature for amorphous Si/C/N semiconductors. Based on solid state physics, the dc-
conductivity of the polymer derived ceramics shows a T ÿ1/4 dependence at low temperatures which
indicates that electron transport is realised occupying midgap-states by variable-range-hopping
[414]. At higher temperatures a change of the charge transport mechanism is observed utilising tail-
states by polarone-tunneling [415]. Comparable results were obtained by the preparation of polymer
Fig. 32. Specific electrical resistance of Si/C/N ceramics derived from PHMS (NCP 200) and Si/B/C/N ceramics. Form
comparison some typical values for insulators, semiconductors and conductors are also shown [413].
derived amorphous Si/B/C/N materials (Fig. 32). Polymer derived amorphous Si/(B)/C/N semicon-
ductors have high potential for future applications in microelectronic devices due to their tailorable
wide range conductivity. Additionally, the comfortable and relatively inexpensive production is an
advantageous characteristic of this novel type of material.
4.2. Si3N4 /SiC composites derived from amorphous Si/C/N powders
The concept of the reinforcement of Si3N4 ceramics by nanosized SiC particles was presented at
the beginning of the nineties by Niihara [2]. The fabricated Si3N4/SiC composites revealed
outstanding mechanical properties at both ambient and elevated temperatures. More recently,
Herrmann et al. [330] and Rendtel et al. [416] correlated the type of starting powder used for the
sintering of Si3N4/SiC composites with the resulting mechanical properties. The best mechanical
properties at elevated temperatures were achieved with Si3N4/SiC composites sintered from a
mixture of crystalline Si3N4 powder and amorphous Si/C/N powder that was produced via plasma
chemical synthesis. Only a few studies in the literature reported data concerning mechanical
properties of amorphous Si/C/N powder derived Si3N4/SiC composites [2,191,331,332]. Other
authors reported about the mechanical properties of Si3N4/SiC composites obtained from a mixture
of crystalline Si3N4 powder and amorphous Si/C/N powder [330,332,369,416,417].
In the following sections, the mechanical properties at both ambient and elevated temperature of
amorphous Si/C/N powder derived Si3N4/SiC composites are presented. Especially, the effects
described in the literature relating to the use of an amorphous Si/C/N powder on the microstructure
development, fracture mechanics and chemistry of the grain boundary phase are summarised.
4.2.1. Microstructural investigations

Fig. 33 shows an SEM micrograph of the 5Y/5S Si3N4/SiC composite which was sintered from
an amorphous Si/C/N-powder obtained by pyrolysis of polyhydridomethylsilazane (Section 3.5).
5 wt.% Y2O3 and 5 wt.% SiO2 were added to the amorphous Si/C/N powder as sintering aids. These
Fig. 33. Scanning electron micrograph of the microstructure of the 5Y/5S Si3N4/SiC composite sample [332].
powder mixtures were fully densified by hot pressing at 18008C for 2 h. The composite was plasma
etched after polishing, in order to make its microstructure apparent in SEM. This technique revealed
the Si3N4 grains to be the darker areas and the SiC grains to be the grey areas. Both elongated and
globular shaped Si3N4 grains can be identified. A mean Si3N4 grain size (d50) of 0.2 mm was
measured.
The TEM micrograph of this composite in Fig. 34 revealed that SiC particles are incorporated
within the Si3N4 grains. These SiC inclusions have a size of 20±110 nm. Fine globular shaped SiC
grains can be distinguished between the Si3N4 grains and are concentrated in domains. The mean
SiC grain size was measured as 0.15 mm and was found to be virtually independent of the sintering
aids used.
The microstructural changes observed when changing the sintering aids system or/and the
volume concern the mean size and the aspect ratio of the Si3N4 grains. The composites 5Y/2S and
5Y/3Al are composed of globular shaped Si3N4 and SiC grains. The composite 2Y/6Yb showed an
unique microstructure consisting of a matrix composed of globular shaped SiC and a-Si3N4 grains.
In this matrix elongated b-Si3N4 grains revealing a very high aspect ratio of 9 were detected [332].
4.2.2. Room temperature mechanical properties
4.2.2.1. Hardness. Hardness measurements were performed on amorphous Si/C/N powder derived
Si3N4/SiC composites, which were sintered as discussed in Section 3.5. Fig. 35 shows the
dependence of hardness on increasing SiC content. If the a- to b-Si3N4 phase transformation during
liquid phase, sintering of the Si3N4/SiC composites has been completed, Si3N4 is only present in the
Fig. 34. Transmission electron micrograph of the microstructure of the 5Y/5S Si3N4/SiC composite sample [332].
b-modification. The hardness of Si3N4/SiC composites, which contain Si3N4 only in the b-
modification, will be considered first. In this case, an increase of the SiC content in the Si3N4/SiC
composite results in a linear increase in the hardness. The line in Fig. 35 characterises this linear
dependence.
Fig. 35. The variation of Vickers hardness with SiC content for Si3N4/SiC composites [332].
The SiC content of the samples which are derived from the amorphous Si/C/N powder obtained by
pyrolysis of polyhydridomethylsilazane is 45 wt.%. The average hardness of these Si3N4/SiC com-
posites after complete a- to b-Si3N4 phase transformation was measured as 18.1 GPa (Sample 5Y/3Al).
At a given SiC content and mean SiC particle size, the rate of the a- to b-Si3N4 phase transformation
is dependent on the volume and on the viscosity of the liquid phase. The higher viscosity of the liquid
phase of the samples 5Y/2S and 2Y/6Yb leads to an incomplete a- to b-Si3N4 phase transformation.
Due to the presence of a-Si3N4, the hardness of the Si3N4/SiC composites differs strongly from the
linear dependence. Since the SiC content remains constant, this effect is due to the greater hardness
of a-Si3N4.. The hardness of the composite sample 2Y/6Yb reaches a value higher than 22 GPa.
4.2.2.2. Fracture toughness. Fracture toughness data of amorphous Si/C/N powder derived Si3N4/
SiC composites reported in the literature are in the range of 4.4±7.5 MPa m1/2. The amorphous Si/C/
N powder derived Si3N4/SiC composites obtained by the methods described in Section 3.5 exhibited
a fracture toughness between 4.6 and 6.2 MPa m1/2. The correlation between these measurements
and the microstructural examinations reveals that fracture toughness of the materials exhibiting
elongated b-Si3N4 grains in their microstructure reaches 5.5±6.2 MPa m1/2.
Herrmann et al. [330] determined that Si3N4/SiC composites sintered from a powder mixture
consisting of crystalline Si3N4 powder and amorphous Si/C/N plasma derived powder showed a
slightly improved fracture toughness compared to materials sintered from commercial crystalline
powders. This difference was attributed to the higher content of impurities in the commercial
powders used in this study.
Since no influence of an increasing SiC volume fraction on the fracture toughness could be
clearly established in the literature, it must be concluded that the evidence of a toughening
mechanism associated with the presence of SiC particles in a Si3N4 matrix remains doubtful.
4.2.2.3. Fracture strength. The influence of the use of an amorphous Si/C/N powder as starting
material for the sintering of Si3N4/SiC composites on the room temperature strength remains a
subject of discussion. Herrmann et al. [330] concluded that the type of starting powder did not affect
either the room temperature strength or the flaw size. Contrarily, SÏajgalõÂk et al. [417] demonstrated a
positive influence of the use of an amorphous Si/C/N powder on the room temperature strength. The
presence of amorphous Si/C/N powder in the starting powder was found to have a positive effect on
the microstructure since a high density of SiC inclusions in Si3N4 grains could be achieved. The
mean size of SiC inclusions was measured to be 50 nm, which is extremely fine. Because SiC has
a higher coefficient of thermal expansion (a4.510ÿ6 Kÿ1) than Si3N4 (a3.610ÿ6 Kÿ1),
complex microstructures characterised by an inter- and intragranular SiC dispersion in a-Si3N4
matrix lead to residual stresses. The presence of SiC inclusions in Si3N4 grains influences the
strength of the boundaries between these grains [417]. This analysis takes into account two types of
SiC inclusions. Si3N4 grains with SiC inclusions surrounded by a glassy layer are supposed to have
no residual stresses because of the relaxation of thermal stresses within the glassy layer. The
presence of Si3N4 grains containing SiC inclusions which are not coated by a glassy phase in the
microstructure are considered to result in micro-stresses in the grains and cause tensions at their
grain boundary. This statement implies that the intergranular fracture mode will be promoted, thus
leading to an increase in the fracture toughness. The SiC dispersion induces a refinement as well as a
homogeneity of the microstructure of Si3N4/SiC composites. As a consequence, the flaw size in
Si3N4/SiC composites is reduced in comparison to monolithic Si3N4 materials. The combination of
these two influences i.e. increase in the fracture toughness and decrease of the flaw size could
explain the higher room temperature strength of Si3N4/SiC composites reported in the literature.
4.2.3. Mechanical properties at elevated temperatures

Si3N4/SiC composites are attractive materials for structural applications because of their
outstanding high temperature mechanical properties such as creep resistance and fracture strength.
4.2.3.1. Fracture strength. The retained strength of Si3N4/SiC composites up to temperatures of

14008C was demonstrated in several studies [416]. In a systematic study, Rendtel et al. [416]
correlated the source of the starting powder used for the fabrication of Si3N4/SiC composites with
the resulting mechanical properties. The authors measured the highest absolute strength at both
ambient and elevated temperatures for Si3N4/SiC composites fabricated from a Si/C/N plasma
powder. These results are attributed to the reduction of high temperature crack growth by the
intergranular SiC particles [2].
4.2.3.2. Creep properties. The improvement of the creep behaviour of monolithic Si3N4 materials
by the incorporation of SiC particles was demonstrated on nano-composites which were fabricated
from a powder mixture consisting of crystalline Si3N4 powder and amorphous Si/C/N plasma derived
powder [416]. The results of creep tests performed on amorphous Si/C/N powder derived Si3N4/SiC
composites found in the literature are shown in Fig. 36.
The improvement of the creep properties of Si3N4/SiC composites can be attributed to three
types of SiC particle induced influences, as suggested by Rendtel et al. [416]. These influences
overlap microstructural, micromechanical and chemical effects:
4.2.3.2.1. Microstructural effects. Keeping the SiC content constant causes the mean Si3N4 grain
size to become finer if the mean SiC grain size is decreased [418]. This refinement of the
Fig. 36. Literature creep rate data for amorphous Si/C/N powder derived Si3N4/SiC composites. Curve 1: (8 wt.% Y2O3±
30 wt.% SiC) Si3N4/SiC composite from Niihara et al. [419]; curve 2: (8 wt.% Y2O3±30 wt.% SiC) Si3N4/SiC composites
from Rendtel et al. [416]; curve 3: ((5 wt.% Y2O35 wt.% SiO2)±45 wt.% SiC) Si3N4/SiC composite from Lecomte
[332]). Creep experiments were carried out under 100 MPa (curves 1 and 2) and under 200 MPa (curve 3).
microstructure implies an increased interlocking of the elongated b-Si3N4 grains. Creep of liquid
phase sintered materials is characterised by dilatational deformation. An increased interlocking of
the elongated b-Si3N4 grains leads to a better creep resistance. Using polysilazane derived
amorphous Si/C/N powder in the starting material as a precursor for SiC results in the formation of
very fine SiC grains (typically 0.1 mm or even smaller) in the microstructure of the Si3N4/SiC
composites. A refinement of the elongated b-Si3N4 grains can be achieved in this case at low SiC
contents because of their fine size. This has to be considered as an advantage since Si3N4/SiC
composites with high SiC contents are difficult to densify.
4.2.3.2.2. Micromechanical effects. Niihara et al. [419] suggested that the intergranular SiC
particles inhibited grain-boundary sliding, which governs the dilatational deformation mechanism of
the Si3N4 matrix during creep experiments. SiC particles located at the grain boundaries act as
obstacles since they increase the roughness of the grain boundary and exert an additional resistance
against the grain movement. TEM studies of Si3N4/SiC interfaces revealed that a direct bonding
between the two phases can be developed without the formation of an amorphous layer, if the size of
the SiC particle is 0.1 mm or smaller [2,420]. However, this statement was put in doubt by other
authors [417]. Furthermore, the grain boundary diffusivity may be decreased by intergranular
particles because the cross-section that is available for diffusive mass transport is reduced. Only a
limited number of SiC particles can be placed at Si3N4 grain boundaries. That is why the creep
properties are thought to be increased up to a certain SiC content which is in turn dependant on the
mean SiC grain size.
4.2.3.2.3. Chemical effects. As described in Section 3.5, free carbon inclusions will form if the
amorphous Si/C/N powder is sintered without prior crystallisation. This has a major consequence on
the chemistry of the grain boundary phase since SiO2, which is dissolved in the liquid phase, will be
reduced by the carbon. For illustration of this effect, the Si3N4/SiC composite sample 2Y/6Yb will be
considered. By removing SiO2 from the liquid phase, the composition of the remaining liquid
changes toward higher Y2O3/SiO2 and Yb2O3/SiO2 ratios and, hence, shifts away from the glass-
forming region of the Yb2O3±Y2O3±SiO2 system. An increase in the viscosity occurs, which leads to
an increase in creep resistance. By this mechanism the use of amorphous Si/C/N powders as starting
materials is thought to promote the devitrification of the glassy phase in Si3N4/SiC composites.
The partial crystallisation of the grain boundary results in a reduction of the volume content of the
glassy phase. The crystallisation of the grain boundary phase of the Si3N4/SiC composite sample
2Y/6Yb to Yb4Si2N2O7 is supposed to have a positive influence, both on mechanical properties such
as hardness and fracture toughness, as well as on the oxidation behaviour. Klemm et al. [421] pointed
out that the improved creep properties of Si3N4/SiC composites have to be mainly related to the
as-described modification of the grain boundary phase composition in comparison to monolithic
Si3N4 materials, rather than to micromechanical interactions in the liquid phase, which will be
reduced by the carbon.
The main impurity in the commercially available crystalline SiC powders is fluorine. The
concentrations were found to be much higher compared to Si/C/N powders that were obtained from
plasma process or derived from polymeric precursors. The increasing impurity level of the glassy
phase causes the viscosity to decrease, which leads to a loss of creep resistance. The devitrification
of the glassy phase as well as the lower amount of impurities are considered to be the main reasons
for the lower creep rates that have been measured in different amorphous Si/C/N powder derived
Si3N4/SiC materials [416].
4.2.3.3. Oxidation resistance. Klemm et al. [421] reported an altered oxidation mechanism for
Si3N4/SiC composites in comparison to monolithic Si3N4 materials. They observed an increased
formation of Si2N2O in the bulk region instead of SiO2 during oxidation of the Si3N4/SiC
composites. In comparison, only SiO2 was formed in the bulk of monolithic Si3N4 materials. The
formation of SiO2 is accompanied by the evolution of nitrogen and induces thermal mismatches in
the bulk as a result of the different coefficients of thermal expansion of SiO2 and Si3N4/SiC.
Consequently, the formation of SiO2 results in microstructural degradations such as pores, cracks or
grain boundary inhomogeneities. These effects are minimised in the case of Si3N4/SiC composites
due to the presence of Si2N2O crystallites in the bulk of the material.
In the preceding section, the use of amorphous Si/C/N powders for the sintering of Si3N4/SiC
composites was correlated with the shift of the composition of the grain boundary phase towards
compositions containing low amounts of SiO2. This effect has a major influence on the oxidation
behaviour of the composites since it leads to an increase of the viscosity of the grain boundary phase
which in turn corresponds to a reduction of oxygen diffusion. The oxidation resistance can be
additionally improved if a crystallisation of the grain boundary to oxidation stable phases is
achieved. The in situ crystallisation of an amorphous Si/C/N powder used as starting powder for the
fabrication of Si3N4/SiC composites is accompanied by the formation of free carbon, which is known
to promote the devitrification of the grain boundary phase. This point can be illustrated by
considering the Si3N4/SiC composites which were fabricated by liquid phase sintering of a
polysilazane derived Si/C/N powder in the sintering aids system SiO2/Y2O3/Yb2O3. The
crystallisation of Yb4Si2N2O7 at the grain boundary of the composites contributes to improving
the oxidation resistance, since the amount of amorphous grain boundary through which oxygen
diffuses to the bulk of the material is reduced. The use of polysilazane derived Si/C/N powders for
the fabrication of Si3N4/SiC composites combines a decrease of the SiO2 content in the grain
boundary phase and its devitrification, which both contribute to significantly improve the oxidation
resistance.
5. Industrial applications of silazanes and silazane derived materials
It is not easy to give an overview concerning the field of applications of silazanes in industry.
The reason for this, is that a large fraction of the research activities and technical developments of
industrial laboratories are not published at all. In many cases patents are the only source of
information available. Therefore, the objective of this section is to give a very brief impression of
industrial applications based on the patent literature.
Fig. 37 gives the number of patents filed since the beginning of research work on silazane
derived ceramics. It is obvious that industrial researchers got interested in this field during the 1980s.
This correlates with the general euphoria about any kind of ceramic material in this period. There
was a maximum number of patents filed around 1992, declining after that. However, it seems that the
activities are increasing again since 1997.
In Table 16 a list of recent patents concerning the industrial application of silazanes and silazane
derived ceramics is given. It is obvious that a number of companies from Japan, the United States
and Europe are or intend to be active in this field. A wide range of different topics are covered, which
shows that there is not only one small market for these materials. The most promising areas for the
application of silazane derived ceramics seem to be protective or hard coatings, ceramic fibres and
composites. Other applications like insulating films or joining techniques might also become
important in the future.
Fig. 37. Number of patents filed per year according to the hits of an online data-bank search in chemical abstracts using the
command ``?ceramic? and ?silazan? and patent/dt and . . ./py'' on 9 April 1999.
6. Conclusions and outlook
The properties of Si3N4- and SiC-based ceramics, for example, resistance to thermal shock,
corrosion, creep and oxidation, continue to generate interest in their application as high temperature
resistant structural materials. However, conventionally processed nonoxide ceramics contain larger
amounts of secondary phases in order to enable densification. These oxides are located at the grain
boundaries of the sintered material and impair high temperature properties like oxidation and creep
resistance.
Considerable attention has been directed to the use of organometallic polymers as precursors to
high performance ceramics such as Si3N4 (derived from polysilazanes or polycarbosilazanes) and
Si3N4/SiC Ð composites as well as amorphous Si/(E)/C/N Ð materials. Some of the main
advantages of the preparation of ceramic materials by this hybrid route include the following points:
1. no necessity to use densification aids;
2. better compositional homogeneity in the product due to the elemental distribution on the
molecular level within the polymer;
3. high purity amorphous or crystalline ceramic products with tailorable microstructure;
4. preparation of refractory ceramics at relatively low temperatures;
5. application of processing techniques well established and used for polymers (e.g. fibre drawing,
coating, infiltration);
188
Table 16
Recently filed patents concerning silazane derived ceramics and their applications
Patents Country Date Topic Patent number
Association for Fine Ceramics Technology Japan 04-11-1998 SiC powders, chemically modified JP 10291864 A2
Research, Agency of Industrial Sciences
and Technology
Bayer AG Germany 15-10-1998 Boron-containing carbosilazanes, Si/B/C/N-ceramics WO 9845303 A1
Bayer AG Germany 15-10-1998 Silylalkylboranes, carbosilazanes, Si/B/C/N-ceramics WO 9845302 A1

Bayer AG Germany 21-01-1998 Polymeric silaborocarboazanes, Si/B/C/N-ceramics EP 819720 A1
Bayer AG Germany 04-06-1997 Si/B/C/N, fibres EP 759414 A3
Bayer AG Germany 20-02-1997 Infusible polyborosilazanes DE 19530390 A1
Bosch, Robert, GmbH Germany 25-02-1999 Soft magnetic, deformable composite WO 9909565 A1
Bosch, Robert, GmbH Germany 20-11-1997 Controlling the porosity of ceramic composites WO 9743229 A1
Bosch, Robert, GmbH Germany 25-04-1996 Electric insulators DE 19538695 A1
D.R. Bujalski; G.A. Zank; T.D. Barnard USA 13-05-1998 Crystalline SiC fibres from borosilazanes GB 2319022
K.K. Canon Japan 23-06-1998 Surface treatment of ink-jet recording head US 5770271
Catalysts and Chemicals Industries Co. Ltd. Japan 08-04-1997 Coatings, polysilazane compositions JP 09095646 A2
Daimler-Benz AG Germany 04-02-1999 Brake units, heat-resistant, tribological material DE 19727587 A1
Daimler-Benz AG Germany 04-03-1999 Brake units DE 19719634 C1
Daimler-Benz AG Germany 01-10-1998 High-temperature-resistant, fibre-reinforced ceramics WO 9842635 A1
Dow Corning Corp. USA 26-01-1999 Crystalline SiC fibres, from borosilazanes US 5863848 A
Dow Corning Corp. USA 30-09-1998 Secondary lithium batteries, Si/C/O anode EP 867958 A1
Dow Corning Corp. USA 15-09-1998 Tamper-proof ceramic coatings, electronic devices US 5807611 A
Dow Corning Corp. USA 13-05-1998 Polycrystalline silicon carbide fibres GB 2319022 A1
Dow Corning Corp. USA 06-05-1998 Tamper-proof coatings on electronic devices JP 10113610 A2
Dow Corning Corp. USA 29-04-1998 Cross-linking silazanes, densifying porous ceramics EP 808863 A3
Dow Corning Corp. USA 08-04-1998 Thick silica-containing coating on electronic devices EP 834489 A1
Dow Corning Corp. USA 08-04-1998 Cross-linking silazane polymers EP 808862 A3
Dow Corning Corp. USA 17-12-1997 Electrodes for Li-ion batteries EP 813259 A1
Dow Corning Corp. USA 20-05-1997 Cathodes for Li-ion batteries US 5631106 A
Dow Corning Corp. USA 26-03-1997 Adhering electronic components to substrates EP 764704 A1
Dow Corning Corp. USA 27-12-1996 Coatings EP 750337 A2
Dow Corning Corp. USA 18-12-1996 Coatings EP 749155 A2
Duncan, Barnard USA 15-06-1996 Fibres CA 2023377 AA
Exxon Research and Engineering Co. USA 13-01-1999 Amorphous ceramics, microporous EP 889862 A1
Exxon Research and Engineering Co. USA 10-07-1997 Polymer-stabilised metal colloids, microporous ceramics WO 9724297 A1
Fine Ceramics Center Japan 05-03-1996 Porous ceramics JP 08059364 A2
Fraunhofer-Gesellschaft zur FoÈrderung der Germany 11-03-1999 Carbon composites WO 9911581 A1
Angewandten Forschung e.V.
Gosudarstvenny Nauchny Tsentr Rossyskoi Russia 09-07-1998 Organosilicon polymers with metal clusters WO 9829472 A1
Federatsii, Gosudarstvenny Nauchno-
Issledovatelsky Institut Khimii i Tekhnologii
Elementoorganicheskikh Soedin
Heat System Research & Industry, Inc. Japan 02-07-1997 Water-soluble pyrolytic paint EP 781817 A1
Hitachi, Ltd. Japan 26-05-1998 Composites, complex shape at low cost JP 10139558 A2
Honda Giken Kogyo Kabushiki Kaisha Japan 05-01-1999 Fibre-reinforced ceramics, slip casting US 5856253 A
Isuzu Ceramics Res. Inst. Japan 09-04-1996 Filters JP 08093445 A2
Kuraray Co. Ltd. Japan 03-06-1997 Molds, lens sheets JP 09141663 A2
Kyocera Corp. Japan 13-10-1998 Alumina-base sintered bodies JP 10273359 A2
Kyocera Corp. Japan 18-08-1998 Porous silica membranes JP 10218690 A2
Kyocera Corp. Japan 23-06-1998 Alumina-based sintered membranes JP 10167808 A2
Kyocera Corp. Japan 23-06-1998 Low-loss dielectric material, high-frequency application JP 10167809 A2
Lanxide Technology Co., L.P. USA 10-04-1997 Thin composite ceramics WO 9712844 A1

Logic Tools L.L.C. USA 18-07-1996 Textile-reinforced molds WO 9621547 A1
McDonnell Douglas Corp. USA 17-11-1998 Composites, fibre-reinforced, low-dielectric, high-strength US 5837318 A
McDonnell Douglas Corp. USA 17-11-1998 Composites, fibre-reinforced, low dielectric constant US 5837318 A
Micropyretics Heaters International USA 07-03-1996 Filters WO 9606814 A2
Mitsui Toatsu Chemicals, Inc. Japan 31-03-1997 Transparent resin electrode plates JP 09085877 A2
N.E. Chemcat Corp. Japan 22-09-1998 Coating, polysilazanes, on large substrates JP 10249276 A2
Neos Co. Ltd. Japan 10-03-1998 Sealing thermal spray coated films, corrosion resistance JP 10068086 A2
Nikken Toso Kogyo K.K. Japan 17-06-1997 Heat-resistant films on metal surfaces JP 09155285 A2
Nippon Chemi-Con Corp. Japan 02-05-1997 Electrolytic capacitor, amorphous ceramic layer JP 09115785 A2
Nippon Oil Co. Ltd. Japan 13-01-1998 Fibre-reinforced ceramics, boron-containing silazanes JP 1007472 A2
Optrex K.K. Japan 13-10-1998 Ceramic gas-barrier films, liquid-crystal display JP 10272736 A2
Optrex K.K. Japan 08-09-1998 Ceramic-coated plastics JP 10235801 A2
Optrex K.K. Japan 28-04-1998 Plastic base plates for liquid crystal displays, oxygen JP 10111499 A2
barrier properties
Optrex K.K. Japan 14-10-1997 Colour liquid crystal display element JP 09269481 A2
Optrex K.K. Japan 05-09-1997 Colour liquid crystal displays, ceramic protective films JP 09230314 A2
Pall Corp. USA 11-03-1999 Porous ceramic structures WO 9911582 A1
Petroleum Energy Center; Tonen K.K. Japan 17-06-1997 Coatings for reinforcement fibres JP 09157051 A2
Politechnika Warszawska Poland 30-06-1998 Sintered silicon nitride ceramic PL 174126 B1
Raytheon Ti Systems, Inc. USA 25-06-1998 Low-carbon ceramics from polysilazanes WO 9827025 A1
Shinetsu Chem Ind. Co. Japan 16-07-1996 Silicon nitride powders for injection molding JP 08183606 A2
Shinko Electric Industries Co. Ltd. Japan 26-08-1997 Ceramic substrates having implanted conductive pins JP 09223872 A2
Showa Electric Wire & Cable Co. Japan 14-01-1997 Fibres, insulated wires, heat resistance JP 09013278 A2
Showa Electric Wire & Cable Co. Ltd. Japan 10-03-1998 Coatings, on electric wires and coils JP 10069819 A2
I. Sigalas; B.A. Roberts; R. Riedel S. Africa 22-07-1997 Manufacturing abrasives US 5649984 A
Southwest Research Institute USA 27-02-1996 Fibre-reinforced silicon nitride composites US 5494867 A
Sumitomo Electric Industries, Ltd. Japan 09-09-1997 Porous ceramic membranes, filters JP 09234319 A2
Suzuki Co. Ltd. Japan 14-07-1998 Coating, on aluminum cylinder bores JP 10183369 A2
Suzuki Sogyo Co. Ltd. Japan 09-10-1998 Base plates for display panels JP 10269938 A2
Teijin Ltd. Japan 20-01-1998 Gas-impermeable laminated films JP 10016150 A2
Teijin Ltd. Japan 20-01-1998 Transparent gas-barrier laminates JP 10016142 A2
Teijin Ltd. Japan 08-07-1997 Coating, laminates JP 09174782 A2
Texas Instruments Inc. USA 14-04-1998 Composites, reinforced with ceramic woven cloth US 5738750 A
Tokyo Ohka Kogyo Co. Ltd.; Tonen K.K. Japan 13-05-1997 Coatings JP 09125006 A2
189
190
Table 16 (Continued )
Patents Country Date Topic Patent number
Tokyo Shibaura Electric Co. Japan 09-07-1996 Composite material JP 08175879 A2
Tonen Co. Japan 10-07-1997 Polysilazane WO 9724391 A1
Tonen Co. Ltd. Japan 11-08-1998 Silica-based ceramic films JP 10212114 A2
Tonen Co. Ltd. Japan 28-07-1998 Formation of Si3N4 thin films JP 10194873 A2
Tonen Co. Ltd. Japan 28-07-1998 Silica-based ceramic material, low dielectric constant JP 10194826 A2
Tonen Co. Ltd. Japan 28-07-1998 Alkali passivation coatings JP 10194781 A2

Tonen Co. Ltd. Japan 07-07-1998 Phenylsilyl-cross-linked polysilazanes, ceramic, coatings JP 10182835 A2
Tonen Co. Ltd. Japan 20-10-1998 SiO2 ceramic coatings from silazanes JP 10279362 A2
Tonen Co. Ltd. Japan 07-04-1998 Gas-separation JP 10085568 A2
Tonen Co. Ltd. Japan 31-03-1998 Coating, chemical apparatus JP 10080630 A2
Tonen Co. Ltd. Japan 15-07-1997 Hard coatings JP 09183949 A2
Tonen Co. Ltd. Japan 15-07-1997 Coating, optical devices JP 09183665 A2
Tonen Corp. Japan 13-01-1999 Heat-resistant, low dielectric constant EP 890599 A1
Tonen Corp. Japan 28-07-1998 Ceramic coatings JP 10194753 A2
Tonen Corp. Japan 10-07-1997 Producing polysilazanes WO 9724391 A1
Tonen Corp. Japan 25-06-1997 UV-shield transparent plates JP 08165145 A2
Tonen Corp. Japan 18-02-1997 Coloured coating, coloured glass JP 09048951 A2
Tonen Corp. Japan 04-02-1997 Coating, polysilazane JP 09031202 A2
Tonen Corp. Japan 30-01-1997 Silazane polymer compositions WO 9703131 A1
Tonen Corp. Japan 21-01-1997 Coating, silica ceramics JP 09019658 A2
Tonen Corp. Japan 14-01-1997 Ceramic-coated plastics JP 09010687 A2
Tonen Corp. Japan 12-11-1996 Liquid crystal display JP 08297274 A2
Tonen Corp. Japan 13-09-1996 Electroluminescent device JP 08236274 A2
Tonen Corp. Japan 10-09-1996 Paints JP 08231899 A2
Tonen Corp. Japan 11-06-1996 Display electrode JP 08152610 A2
Tonen K.K. Japan 08-07-1997 Polycarbonate product, hard coating JP 09175868 A2
Tonen K.K. Japan 17-06-1997 Coating, surface-ceramised plastics JP 09157528 A2
Tonen K.K. Japan 08-04-1997 UV-resistant wood products JP 09094525 A2
Toshiba Corp. Japan 28-03-1997 Copper-clad ceramic circuit board JP 09082843 A2
Toshiba Corp.; Toshiba Electronic Engineering Japan 31-03-1997 Fibre-reinforced ceramic composites JP 09087049 A2
Group
Toyo Polymer K.K.; Nitto Shoji K.K. Japan 28-04-1998 Gas-barrier ceramic films, laminates JP 10110280 A2
Toyo Tanso Co. Japan 14-04-1998 Carbon composites, high oxidation resistance JP 10095685 A2
U.S. NASA USA 29-09-1998 Waterproofing ceramics, coating US 5814397 A
UniversitaÈt Bayreuth, Lehrstuhl fuÈr Keramik Germany 24-07-1997 Nanocomposites DE 19545346 A1
u. Verbundwerkstoffe
University of Florida USA 22-12-1998 Boron-doped SiC fibres US 5851942 A
Wisconsin Alumni Research Foundation USA 18-02-1997 Plasma-cross-linked polymers, thin layer US 5604038 A
6. preparation of novel composite materials and multinary ceramics not available by conventional
technologies.
The results discussed in this review article show that a number of laboratories successfully
contributed to the evolution of elaborate ways to utilise especially designed and synthesised
silazanes, as well as commercial Si-based polymers, as precursors for the hybrid route to amorphous
Si/C/N and polycrystalline Si3N4 and Si3N4/SiC Ð composite ceramics. It was demonstrated that
silazanes can be processed to ceramic materials by (i) the classical approach i.e. by liquid phase
sintering of silazane-derived amorphous Si/C/N powders with additives, and (ii) by the non-classical
approach which involves the direct transformation of molecular and polymeric silazanes to silicon
nitride/carbide-based ceramics.
The strength of the polymer-to-ceramic transformation technique is to control stoichiometry and
thus phase composition, microstructure, degree of crystallinity, and porosity in the amorphous and
crystalline states.
For further development of the silazane derived silicon-based non-oxide ceramics the formation
and amount of volatile side products during the polymer-to-ceramic transformation has to be
controlled and minimised, respectively. Other very important factors are the costs and availability of
appropriate silazane starting materials. In some areas, like the use of HMDS for vapour-phase
synthesis of ceramic products, the processes have shown to be competitive. In the area of silazane
polymers some commercial products have been produced, but are not available any more. Other
materials are still produced or are currently under development.
Acknowledgements
The authors thank the Deutsche Forschungsgemeinschaft, Bonn (Germany), for financial
support. R.R. acknowledges the support of Fonds der Chemischen Industrie, Frankfurt (Germany), of
Bayer AG, Leverkusen and the KSB foundation, Stuttgart. Y.-L.L. acknowledges the Humboldt
Foundation for granting a research fellowship, and E.L. acknowledges the European Union for
granting a Marie Curie fellowship. The authors thank Y. Iwamoto and O. Funayama who provided
useful information about commercially available silazane polymers of Tonen Corp. (Japan).
References
[1] R. Riedel, G. Passing, H. SchoÈnfelder, R.J. Brook, Nature 355 (1992) 714.
[2] K. Niihara, J. Ceram. Soc. Jpn. 99 (1991) 974.
[3] F. Wakai, Y. Kodoma, S. Sakaguchi, N. Murayama, K. Izaki, K. Niihara, Nature 344 (1990) 421.
[4] W. Dressler, R. Riedel, Int. J. Refractory Met. Hard Mater. 15 (1997) 13.
[5] R. Riedel, M. Seher, J. Mayer, D.-V. SzaboÂ, J. Eur. Ceram. Soc. 15 (1995) 703.
[6] R. Riedel, in: R.W. Cahn, P. Haasen, E.J. Kramer, R.J. Brook (Eds.), Materials Science and Technology, Vol. 17 B,
1996, VCH, Weinheim, p. 1.
[7] R. Riedel, E. Kroke, A. Greiner, A.O. Gabriel, L. Ruwisch, J. Nicolich, Chem. Mater. 10 (1998) 2964.
[8] R. Riedel, A. Greiner, G. Miehe, W. Dressler, H. Fuess, J. Bill, F. Aldinger, Angew. Chem. Int. Ed. Engl. 36 (1997)
603.
[9] L. An, R. Riedel, C. Konetschny, H.-J. Kleebe, R. Raj, J. Am. Ceram. Soc. 81 (1998) 1349.
[10] R. Riedel, L.M. Ruwisch, L. An, R. Raj, J. Am. Ceram. Soc. 81 (1998) 3341.
[11] R. Riedel, Naturwissenschaften 82 (1995) 12.
[12] R. Riedel, H.-J. Kleebe, H. SchoÈnfelder, F. Aldinger, Nature 374 (1995) 526.
[13] R. Riedel, J. Bill, A. Kienzle, Appl. Organomet. Chem. 10 (1996) 241.
[14] R. Riedel, A. Kienzle, W. Dreûler, L. Ruwisch, J. Bill, F. Aldinger, Nature 382 (1996) 796.
[15] P. Eigen, H. Keller-Rudek, W. Kurtz, P. Merlet, H. Schaefer, F. Schroeder, Gmelin Handbook of Inorganic
Chemistry, 8th Edition, Vol. B4, Silicon Suppl., Springer, Berlin, 1989, Chapter 8, p. 226.
[16] S. Pawlenko, in: O. Bayer, E. MuÈller (Eds.), Methoden der organischen Chemie (Houben-Weyl), 4th Edition, Vol.
13, Part 5, Metallorganische Verbindungen Silicon, Georg Thieme, Stuttgart, 1980, p. 226.
[17] C.K. Narula, Ceramic Precursor Technology and its Applications, Marcel Dekker, New York, 1995, Chapter 3.4, p.
83.
[18] M. Birot, J.-P. Pillot, J. Dunogues, Chem. Rev. 95 (1995) 1443.
[19] R.M. Laine, Y.D. Blum, D. Tse, R. Glaser, in: M. Zeldin, K.J. Wynne, H.R. Allcock (Eds.), Inorganic and
Organometallic Polymers, ACS Symposium Series 360, American Chemical Society, Washington, DC, 1988,
Chapter 10, p. 124.
[20] R.C.P. Cubbon, Polymeric Precursors for Ceramic Materials, Vol. 7, Rapra Review Report No. 76, 1994.
[21] P. Schutzenberger, H. Colsen, C. R. Acad. Sci. 92 (1885) 1508.
[22] A. Stock, K. Somieski, Chem. Ber. 54 (1921) 740.
[23] S.D. Brever, C.P. Haber, J. Am. Chem. Soc. 70 (1948) 3888.
[24] R.C. Osthoff, S.W. Kantor, Inorg. Synth. 5 (1957) 61.
[25] W. Fink, Angew. Chem. 78 (1966) 803.
[26] E.G. Rochow, Pure Appl. Chem. 13 (1966) 247.
[27] B.J. Aylett, G.M. Burnett, L.K. Peterson, N. Ross, SCI Monogr. 13 (1961) 5.
[28] B.J. Aylett, in: E.P. Popper (Ed.), Special Ceramics, Academic Press, New York, 1964, p. 105.
[29] P.G. Chantrell, E.P. Popper, in: E.P. Popper (Ed.), Special Ceramics, Academic Press, New York, 1964, p. 87.
[30] O. Glemser, K. Beltz, P. Naumann, Z. Anorg. Allg. Chem. 291 (1957) 51.
[31] H.-H. Moretto, M. Schulze, G. Wagner, in: Ullmann's Encyclopedia of Industrial Chemistry, Vol. A24, VCH,
Weinheim, 1993, p. 57.
[32] H.-P. Baldus, M. Jansen, Angew. Chem. 109 (1997) 338.
[33] L. RoÈsch, P. John, R. Reitmeier, in: Ullmann's Encyclopedia of Industrial Chemistry, Vol. A24, VCH, Weinheim,
1993, p. 35.
[34] W.E. Weibrecht, E.G. Rochow, J. Organomet. Chem. 5 (1966) 520.
[35] D.L. Bailey, L.H. Sommer, F.C. Whitmore, J. Am. Chem. Soc. 70 (1948) 435.
[36] U. Wannagat, F. Brandmair, Z. Anorg. Allg. Chem. 280 (1955) 223.
[37] H.-J. Goetze, B. Bartylla, M. Ismeier, Spectrochim. Acta Part A 49 (1993) 497.
[38] N. Wiberg, W. Uhlenbrock, Chem. Ber. 104 (1971) 2643.
[39] E.A.V. Ebsworth, H.J. Emeleus, J. Chem. Soc., London (1958) 2150.
[40] U. Wannagat, Fortschr. Chem. Forsch. 9 (1967) 102.
[41] K. Ruhlandt-Senge, R.A. Bartlett, M.M. Olmstead, P.P. Power, Angew. Chem. 105 (1993) 459.
[42] L.H. Sommer, L.J. Tyler, J. Am. Chem. Soc. 76 (1954) 1030.
[43] H. Niederpruem, Angew. Chem. 75 (1963) 165.
[44] F. Metras, Bull. Soc. Chim. Fr. (1966) 2155.
[45] K.A. Andrianov, N.V. Delazari, Bull. Acad. Sci. USSR Div. Chem. Sci. 21 (1972) 2679 (Engl. Transl.).
[46] T. Kottke, U. Klingebiel, M. Noltemeyer, U. Pieper, S. Walter, D. Stalke, Chem. Ber. 124 (1991) 1941.
[47] U. Klingebiel, N. Vater, Angew. Chem. 94 (1982) 870.
[48] U. Kliebisch, U. Klingebiel, N. Vater, Chem. Ber. 118 (1985) 4561.
[49] U. Wannagat, E. Bogusch, Monatsh. Chem. 102 (1971) 1806.
[50] L.W. Breed, H.L. Budde, R.L. Elliot, J. Organomet. Chem. 6 (1966) 676.
[51] B.D. Lavrukhin, B.A. Astapov, A.V. Kisin, A.A. Zhdanov, Bull. Acad. Sci. USSR Div. Chem. Sci. 32 (1983) 959
(Engl. Transl.).
[52] M.G. Katsitadze, V.V. Kazakova, L.I. Nakaidze, A.A. Zhdanov, L.M. Khananashvili, Bull. Acad. Sci. USSR Div.
Chem. Sci. 37 (1988) 562 (Engl. Transl.).
[53] D.Ya. Zhinkin, G.N. Mal'nova, J. Gen. Chem. USSR 34 (1964) 1482 (Engl. Transl.).
[54] W. de Man, J. van Steenis, H.I. Waterman, Recl. Trav. Chim. Pays-Bas 67 (1948) 865.
[55] K.A. Andrianov, G.V. Kotrelev, J. Organomet. Chem. 7 (1967) 217.
[56] K. Ruehlmann, G. Tuchtenhagen, Z. Chem. 5 (1965) 107.
[57] K. Hizawa, E. Nojimoto, Kogyo Kagaku Zasshi 59 (1956) 1445; Chem. Abstr. (1959) 4176.
[58] Zh.V. Gorislavskaya, J. Gen. Chem. USSR 42 (1972) 2200 (Engl. Transl.).
[59] Am. Cyanamid Co., US Patent 2,885,370 (1957).
[60] E. Larsson, L. Bjellerup, J. Am. Chem. Soc. 75 (1953) 995.
[61] K.A. Andrianov, J. Gen. Chem. USSR 42 (1972) 844 (Engl. Transl.).
[62] K.A. Andrianov, Chem. Heterocycl. Compd. 5 (1969) 585 (Engl. Transl.).
[63] K.A. Andrianov, Bull. Acad. Sci. USSR Div. Chem. Sci. (1963) 860 (Engl. Transl.).
[64] E.P. Lebedev, V.O. Reikhsfel'd, J. Gen. Chem. USSR 40 (1970) 1070 (Engl. Transl.).
[65] N.S. Nametkin, Chem. Heterocycl. Compd. 1 (1965) 305 (Engl. Transl.).
[66] W. Simmler, Chem. Ber. 96 (1963) 1495.
[67] J.-P. Pometan, C. Dumas, J.-B. Fourtillan, P. Courtois, B. Maillard, Eur. J. Med. Chem. Chim. Ther. 16 (1981)
425.
[68] M.M. Morgunova, Yu.M. Varezhkin, A.E. Chernyshev, V.N. Bochkarev, D.Ya. Zhinkin, J. Gen. Chem. USSR 52
(1982) 1633 (Engl. Transl.).
[69] G. Boden, A. Neumann, T. Breuning, E. Tschernikova, W. Hermel, J. Eur. Ceram. Soc. 18 (1998) 1461.
[70] H. Du, B. Gallois, K.E. Gonsalves, Chem. Mater. 1 (1989) 569.
[71] H. Du, B. Gallois, K.E. Gonsalves, J. Am. Ceram. Soc. 73 (1990) 764.
[72] T.P. Smirnova, L.V. Yakovkina, Yu.I. Amosov, O.V. Danilovich, Inorg. Mater. 32 (1996) 615 (Transl. of Neorg.
Mater.).
[73] F. Maury, Z. Hatim, A. Reynes, R. Morancho, J. Anal. Appl. Pyrolysis 17 (1989) 67.
[74] D. Seyferth, G.H. Wiseman, C. Prud'homme, Mater. Sci. Res. 17 (1982) 263.
[75] D. Seyferth, G.H. Wiseman, C. Prud'homme, J. Am. Ceram. Soc. 66 (1983) C13.
[76] D. Seyferth, G.H. Wiseman, Polym. Prepr. 25 (1984) 10.
[77] W.S. Coblenz, G.H. Wiseman, P.B. Davis, R.W. Rice, Mater. Sci. Res. 17 (1982) 271.
[78] T. Morozumi, K. Sato, A. Tezuka, H. Kaya, T. Isoda, in: P. Vincenzini (Ed.), Proceeding of the Topical Symposium
on Advanced Structural Fibre Composites of the 8th CIMTEC-World Ceramic Congress and Forum on New
Materials, Florence, Italy, 28 June±04 July, 1994, Faenza, Italy, 1995, p. 403.
[79] O. Funayama, M. Arai, Y. Tashiro, H. Aoki, T. Suzuki, K. Tamura, H. Kaya, H. Nishii, T. Isoda, Seramikusu
Ronbunshi 98 (1990) 104.
[80] T. Isoda, in: H. Ishida (Ed.), Proceedings of the 3rd International Conference on Composite Interfaces (ICCI-III):
Controlled Interfaces in Composite Materials, Elsevier, New York, 1990, p. 255.
[81] H. Matsuo, O. Funayama, T. Kato, H. Kaya, T. Isoda, J. Ceram. Soc. Jpn. 102 (1994) 409.
[82] O. Funayama, Y. Tashiro, A. Kamo, M. Okumura, T. Isoda, J. Mater. Sci. 29 (1994) 4883.
[83] N.R. Dando, A.J. Perrotta, C. Strohmann, R.M. Stewart, D. Seyferth, Chem. Mater. 5 (1993) 1624.
[84] U. Klingebiel, Phosphorus, Sulfur, Silicon Relat. Elem. 41 (1989) 361.
[85] K. Dippel, E. Werner, U. Klingebiel, Phosphorus, Sulfur, Silicon Relat. Elem. 44 (1992) 15.
[86] D. Seyferth, G.H. Wiseman, J. Am. Ceram. Soc. 67 (1984) C132.
[87] H.N. Han, D.A. Lindquist, J.S. Haggerty, D. Seyferth, Chem. Mater. 4 (1992) 705.
[88] Y.W. Bae, H. Du, B. Gallois, K.E. Gonsalves, B.J. Wilkens, Chem. Mater. 4 (1992) 478.
[89] M.R. Mucalo, N.B. Milestone, I.C. Vickridge, M.V. Swain, J. Mater. Sci. 29 (1994) 4487.
[90] G.T. Burns, C.K. Saha, G.A. Zank, H.A. Freeman, J. Mater. Sci 27 (1992) 2131.
[91] N.S. Choong Kwet Yive, R.J.P. Corriu, D. Leclerq, P.H. Mutin, A. Vioux, New J. Chem. 15 (1991) 85.
[92] N.S. Choong Kwet Yive, R.J.P. Corriu, D. Leclerq, P.H. Mutin, A. Vioux, Chem. Mater. 4 (1992) 141.
[93] A. Lavedrine, D. Bahloul, P. Goursat, N. Choong Kwet Yive, R. Corriu, D. Leclerq, H. Mutin, A. Vioux, J. Eur.
Ceram. Soc. 8 (1991) 221.
[94] C. Colombier, in: G. de With, R.A. Terspstra, R. Metselaar (Eds.), Euro-Ceramics Ð Proceedings of the
1st European Ceramic Society Conference, Vol. 1, Maastricht, 18±23 June, 1989, Elsevier, Amsterdam, 1989,
p. 143.
[95] D. Bahloul, M. Pereira, C. Gerardin, J. Mater. Chem. 7 (1997) 109.
[96] C.M. Gerardin, F. Taulelle, J. Livage, Mater. Res. Soc. Symp. Proc. 287 (1993) 233.
[97] J. LuÈcke, J. Hacker, D. Suttor, G. Ziegler, Appl. Organomet. Chem. 11 (1997) 181.
[98] K.E. Gonsalves, P.R. Strutt, T.D. Xiao, P.G. Klemens, J. Mater. Sci. 27 (1992) 3231.
[99] H. Fleischer, K. Hensen, D. Burgdorf, E. Flindt, U. Wannagat, H. BuÈrger, G. Pawelke, Z. Anorg. Allg. Chem. 621
(1995) 239.
[100] A. Stock, F. Zeidler, Chem. Ber. 56 (1923) 986.
[101] O. Ruff, K. Albert, Chem. Ber. 38 (1905) 2235.
[102] B. RaÈke, H.W. Roesky, I. Uson, P. MuÈller, Angew. Chem. 110 (1998) 1508.
[103] M. Persoz, Ann. Chim. Phys. 44 (1839) 315.
[104] W.C. Schumb, L.H. Towle, J. Am. Chem. Soc. 75 (1953) 6085.
[105] M. Billy, C. R. Acad. Sci. 246C (1958) 433.
[106] M. Billy, C. R. Acad. Sci. 248C (1959) 807.
[107] M. Billy, Ann. Chim. (Paris) 4 (1959) 795.
[108] M. Billy, Bull. Soc. Chim. Fr. (1962) 183.
[109] O. Glemser, P. Naumann, Z. Anorg. Allg. Chem. 298 (1959) 134.
[110] G.M. Crosbie, in: A.W. Weimer (Ed), Carbide, Nitride and Boride Materials Synthesis and Processing, Chapman &
Hall, London, 1997, Chapter 15, p. 391.
[111] H. Lange, G. WoÈtting, G. Winter, Angew. Chem. 103 (1993) 1606.
[112] U. Wannagat, H. Moretto, P. Schimdt, M. Schulze, Z. Anorg. Allg. Chem. 381 (1971) 288.
[113] C.R. Blanchard, S.T. Schwab, J. Am. Chem. Soc. 77 (1994) 1729.
[114] J. Bill, J. Seitz, G. Thurn, J. DuÈrr, J. Canel, B.Z. Janos, A. Jalowiecki, A. Sauter, S. Schempp, H.P. Lamparter, J.
Mayer, F. Aldinger, Phys. Stat. Solidi (a) 166 (1998) 269.
[115] T. Liu, O. Sakurai, N. Kieda, M. Kato, N. Mizutani, Nippon Seramikkusu Kyokai Gakujutsu Ronbunshi 97 (1989)
1321.
[116] D. Seyferth, G.H. Wisman, J.M. Schwark, Y.F. Fu, C.A. Poutasse Inorganic organometallic polymer, ACS Symp.
Ser. 360 (1988) 143.
[117] D. Seyferth, G.H. Wisman, C.A. Poutasse, J.M. Schwark, Y.F. Fu, Polym. Prepr., Am. Chem. Soc. Div. Polym.
Chem. 28 (1987) 389.
[118] D. Seyferth, N. Bryson, D.P. Workman, C.A. Sobon, J. Am. Ceram. Soc. 74 (1991) 2687.
[119] D. Bahloul, M. Pereira, P. Goursat, N.S. Choong Kwet Yive, R.J.P. Corriu, J. Am. Ceram. Soc. 76 (1993) 1156.
[120] D. Bahloul, M. Pereira, P. Goursat, J. Am. Ceram. Soc. 76 (1993) 1163.
[121] H. Hu, Z. Chen, C. Feng, C. Zhang, Y. Song, Cailiao Yanjiu Xuebao 12 (1998) 64; Chem. Abstr. 129:139513.
[122] T. Aoki, O. Funayama, Y. Shimizu, H. Nakahara, T. Isoda, in: Proceedings of the 12th Meeting of the Society of
Polymer Science, Japan, 1993, p. 78.
[123] G. Boden, Key Eng. Mater. 132 (1997) 2256.
[124] G. Boden, T. Breuning, CFI, Ceram. Forum Int. 75 (1998) 24.
[125] U. Wannagat, G. Schreiner, Monatsh. Chem. 96 (1965) 1895.
[126] E. Bacque, J.-P. Pillot, M. Birot, J. Dunogues, G. Bourgeois, M. Petraud, J. Organomet. Chem. 481 (1994) 167.
[127] D. Seyferth, R.M. Stewart, Appl. Organomet. Chem. 11 (1997) 813.
[128] R.C. Osthoff, S.W. Kantor, Inorg. Synth. 5 (1957) 58.
[129] K.A. Andrianov, Bull. Acad. Sci. USSR Div. Chem. Sci. (1968) 340 (Engl. Transl.).
[130] E.A. Semenova, Bull. Acad. Sci. USSR Div. Chem. Sci. (1962) 1945 (Engl. Transl.).
[131] H. Fleischer, D.C. McKean, C.R. Pulham, M. Buehl, J. Chem. Soc., Dalton Trans. 4 (1998) 585.
[132] U. Wannagat, R. Braun, Monatsh. Chem. 100 (1969) 1910.
[133] P. Jutzi, F.W. Schroeder, Z. Naturforsch. B 24 (1969) 649.
[134] R.H. Baney, M. Itoh, A. Sakakibara, T. Suzuki, Chem. Rev. 95 (1995) 1409.
[135] W. Schnick, H. Huppertz, Chem. Eur. J. 3 (1997) 679.
[136] Y.D. Blum, R.M. Laine, K.B. Schwartz, D.J. Rowcliffe, R.C. Bening, D.B. Cotts, Mater. Res. Soc. Proc. Symp. 73
(1986) 389.
[137] Y.D. Blum, R.M. Platz, E.J. Crawford, J. Am. Ceram. Soc. 73 (1990) 170.
[138] R.M. Laine, F. Babonneau, K.Y. Blowhowiak, R.A. Kennish, J.A. Rahn, G.J. Exarhos, K. Waldner, J. Am. Ceram.
Soc. 78 (1995) 137.
[139] J. He, M. Scarlete, J.F. Harrod, J. Am. Ceram. Soc. 78 (1995) 3009.
[140] U. Wannagat, U. Liehr, Z. Anorg. Allg. Chem. 297 (1958) 129.
[141] R. West, M. Ishikawa, R.E. Bailey, J. Am. Chem. Soc. 88 (1966) 4648.
[142] J. He, J.F. Harrod, Phosphorus, Sulfur Silicon Relat. Elem. 94 (1994) 349.
[143] N. Wiberg, M. Veith, Chem. Ber. 104 (1971) 3176.
[144] M.F. Gonon, G. Fantozzi, M. Murat, J.P. Disson, J. Eur. Ceram. Soc. 15 (1995) 591.
[145] M.F. Gonon, S. Hampshire, J.P. Disson, G. Fantozzi, J. Eur. Ceram. Soc. 15 (1995) 683.
[146] H. Grosse-Ruyken, K. Schaarschmidt, Chem. Tech. 11 (1959) 451.
[147] R.H. Baney, J.H. Gaul, T.K. Hilti, in: Proceedings of the Conference on Emergent Process Methods for High-
Technology Ceramics, Vol. 17, Mater. Sci. Res., Raleigh, NC, USA, 8±10 November, 1982, p. 253.
[148] E. Kroke, in: P. Vincenzini (Ed.), Proceedings of the 9th CIMTEC Ð World Ceramics Congress on Ceramics:
Getting into the 2000s, Part C, Techna Srl., 1999, p. 123.
[149] G.E. Legrow, T.F. Lim, J. Lipowitz, R.S. Reaoch, Am. Ceram. Soc. Bull. 66 (1987) 363.
[150] R.H. Baney, Polym. Prep. (Am. Chem. Soc., Div. Polym. Chem.) 25 (1984) 2.
[151] D. Mocaer, R. Pailler, R. Naslain, C. Richard, J.P. Pillot, J. Dunogues, C. Gerardin, F. Taulelle, J. Mater. Sci. 28
(1993) 2615.
[152] E. Bacque, J. Donugues, C. Biran, P. Olry, J.P. Pillot, French Patent 2589037 (1986).
[153] E. Bacque, J.P. Pillot, J. Dunogues, P. Olry, Eur. Patent 296028 (1988).
[154] D. Mocaer, R. Pailler, R. Naslain, C. Richard, J.P. Pillot, J. Dunogues, J. Mater. Sci. 28 (1993) 2632.
[155] D. Mocaer, R. Pailler, R. Naslain, C. Richard, J.P. Pillot, J. Dunogues, O. Delverdier, M. Monthioux, J. Mater. Sci.
28 (1993) 2639.
[156] O. Delverdier, M. Monthioux, D. Mocaer, R. Pailler, in: A.R. Bunsell, J. Jarnet, A. Massiah (Eds.), Proceedings of
the 5th European Conference on Composite Materials, Bordeaux, France, 1992, p. 691.
[157] O. Delverdier, M. Monthioux, D. Mocaer, R. Pailler, J. Eur. Ceram. Soc. 14 (1994) 313.
[158] R. Pailler, I. Jaymes, R. Naslain, C. Richard, J.P. Pillot, J. Dunogues, P. Olry, D. Mocaer, E. Chassagneux, in: A.R.
Bunsell, J. Jarnet, A. Massiah (Eds.), Proceedings of the 5th European Conference on Composite Materials,
Bordeaux, France, 1992, p. 631.
[159] W. Toreki, C.D. Batich, M.D. Sacks, A.A. Morrone, Ceram. Eng. Sci. Proc. 11 (1990) 1371.
[160] Y.D. Blum, R.M. Laine, Organometallics 5 (1986) 2081.
[161] Y.D. Blum, K.B. Schwartz, R.M. Laine, J. Mater. Sci. 24 (1989) 1707.
[162] W. Noll, Chemistry and Technology of Silicones, Academic Press, New York, 1968 (Chapter 5).
[163] D. Seyferth, J.M. Schwark, R.M. Stewart, Organometallics 8 (1989) 1980.

[164] E. Duguet, M. Schappacher, A. Soum, Macromolecules 25 (1992) 4835.
[165] S. Bruzaud, A. Soum, Macromol. Chem. Phys. 197 (1996) 2379.
[166] H. Tang, R.E. Prud'homme, A.-F. Mingotaud, M. Schappacher, A. Soum, Macromolecules 30 (1997) 1400.
[167] S. Bruzaud, A.-F. Mingozaud, A. Soum, Macromol. Chem. Phys. 198 (1997) 1873.
[168] C. Cazalis, A.-F. Mingotaud, A. Soum, Macromol. Chem. Phys. 198 (1997) 3441.
[169] E.G. Rochow, Mon. Chem. 95 (1964) 750.
[170] C.R. Kruger, E.G. Rochow, J. Polym. Sci. A 2 (1964) 3179.
[171] M.T. Zoeckler, R.M. Laine, J. Org. Chem. 48 (1983) 2539.
[172] J.E. Bender IV, W. Chin, M.M.B. Holl, Chem. Mater. 11 (1999) 154.
[173] W. Verbeek, US Patent 3,853,567 (1974).
[174] G. Winter, W. Verbeek, M. Mansmann, US Patent 3,892,583 (1975).
[175] B.G. Penn, F.E. Ledbetter III, J.M. Clemons, Ind. Eng. Chem. Process. Des. Dev. 23 (1984) 217.
[176] B.G. Penn, J.G. Daniels, F.E. Ledbetter III, J.M. Clemons, Polym. Eng. Sci. 26 (1986) 1191.
[177] D. Seyferth, G.H Wisman, US Patent 4,482,669 (1984).
[178] N.S. Choong Kwet Yive, R.J.P. Corriu, D. Leclerq, P.H. Mutin, A. Vioux, Chem. Mater. 4 (1992) 1263.
[179] K.A. Youngdahl, R.M. Laine, R.A. Kennish, T.R. Cronin, G.A. Balavoine, Mater. Res. Soc. Symp. Proc. 121 (1988)
489.
[180] R.M. Laine, Platinum Met. Rev. 32 (1988) 64.
[181] Y.D. Blum, G.A. McDermott, A.S. Hirschon, Inorg. Organomet. Polym. Proc. IUPAC Symp. Macromol. 33 (1991)
161.
[182] D. Seyferth, C. Strohmann, N.R. Dando, A.J. Perrotta, J.P. Garnder, Mater. Res. Soc. Symp. Proc. 327 (1994) 191.
[183] H.J. Campbell-Ferguson, E.A.V. Ebsworth, J. Chem. Soc. (A) (1966) 1508.
[184] D.S. Kim, E. Kroke, R. Riedel, A.O. Gabriel, S.C. Shim, Appl. Organomet. Chem. 13 (1999) 495.
[185] D. Bahloul, M. Pereira, P. Goursat, Ceram. Int. 18 (1992) 1.
[186] F. Sirieix, P. Goursat, A. Lecomte, A. Dauger, Composites Sci. Technol. 37 (1990) 7.
[187] T. Gerdau, H.-J. Kleiner, M. Peuckert, M. Brueck, F. Aldinger, German Patent 373,372,7 A1 (1989).
[188] D. Seyferth, C. Strohmann, N.R. Dando, A.J. Perrotta, Chem. Mater. 7 (1995) 2058.
[189] U. Wannagat, H. BuÈrger, C. KruÈger, J. Pump, Z. Anorg. Allg. Chem. 321 (1963) 208.
[190] J.M. Schwark, US Patent 4,929,704 (1990).
[193] J.M. Schwark, M.J. Sullivan, Mater. Res. Soc. Symp. Proc. 271 (1992) 807.
[194] CERASETTM SN Inorganic Polymer Technical Bulletin, Lanxide Performance Materials, Inc., Newark, DE, USA.
[195] J.F. Klebe, J.B. Bush, J.E. Lyons, J. Am. Chem. Soc. 86 (1964) 4400.
[196] Y.-L. Li, E. Kroke, R. Riedel, C. Fasel, C. Gervais, F. Babonnea, submitted for publication.
[197] K.S. Mazdiyasni, R. Ruh, J. Am. Ceram. Soc. 64 (1981) 415.
[198] H.-P. Baldus, G. Passing, Mater. Res. Soc. Symp. Proc. 346 (1994) 617.
[199] H.-P. Baldus, O. Wagner, M. Jansen, Mater. Res. Soc. Symp. Proc. 271 (1992) 821.
[200] J. LoÈffelholz, M. Jansen, Adv. Mater. 7 (1995) 289.
[201] H.-P. Baldus, J. LuÈcke, G. Passing, K. Vogt, Fortschrittsberichte der Deutschen Keramischen Gesellschaf, Vol. 12,
No. 1, 1997, ISSN 0177-6983, p. 117.
[202] J. Bill, F. Aldinger, Z. Metallkunde 11 (1996) 827.
[203] T. Wideman, P.J. Fazen, K. Su, E.E. Remsen, G.A. Zank, L.G. Sneddon, Appl. Organomet. Chem. 12 (1998) 681.
[204] M. Takamizawa, T. Kobayashi, A. Hayashida, Y. Takeda, N. Joetsu, German Patent 344,430,6 A1 (1985).
[205] M. Takamizawa, T. Kobayashi, A. Hayashida, Y. Takeda, US Patent 4,604,367 (1986).
[206] M. Takamizawa, T. Kobayashi, A. Hayashida, Y. Takeda, N. Joetsu, T. Hongu, German Patent 350,096,2 (1985).
[207] H. NoÈth, Z. Naturforsch. B 16 (1961) 618.
[208] D. Seyferth, H. Plenio, J. Am. Ceram. Soc. 73 (1990) 2131.
[209] D. Seyferth, H. Plenio, W.S. Rees, K. BuÈchner, in: A.R. Bassindale, P.P. Gaspar (Eds.), Frontiers of Organosilicon
Chemistry, The Royal Society of Chemistry, Cambridge, 1991, p. 15.
[210] K. Su, E.E. Remsen, G.A. Zank, L.G. Sneddon, Chem. Mater. 5 (1993) 547.
[211] K. Su, E.E. Remsen, G.A. Zank, L.G. Sneddon, Polym. Prepr. 34 (1993) 334.
[212] T. Wideman, K. Su, E.E. Remsen, G.A. Zank, L.G. Sneddon, Chem. Mater. 7 (1995) 2203.
[213] T. Wideman, E. Cortez, E.E. Remsen, G.A. Zank, P.J. Carroll, L.G. Sneddon, Chem. Mater. 9 (1997) 2218.
[214] R. Riedel, W. Dressler, in: P. Vincenzini (Ed.), Ceramics: Charting the Future, Vol. 3C, Techna Stl. Faenza, 1995, p.
1727.
[215] J. Bill, A. Kienzle, M. Sasaki, R. Riedel, F. Aldinger, in: P. Vincenzini (Ed.), Ceramics: Charting the Future, Vol. 3B,
Techna Stl. Faenza, 1995, p. 1291.
[216] M. Frieû, Ph.D. Thesis, MPI fuÈr Metallforschung, Stuttgart, 1994 (in German).
[217] O. Funayama, T. Kato, Y. Tashiro, T. Isoda, J. Am. Ceram. Soc. 76 (1993) 717.
[218] O. Funayama, H. Nakahara, M. Okoda, M. Okumura, T. Isoda, J. Mater. Sci. 30 (1995) 410.
[219] S.R. Riccitiello, M.T.S. Hsu, T.S. Chen, US Patent 4,987,201 (1991).
[220] M.T.S. Hsu, T.S. Chen, S.R. Riccitiello, J. Appl. Polym. Sci. 42 (1991) 851.
[221] R.P. Jones, J.K. Meyers, J. Organomet. Chem. 34 (1972) C9.
[222] R. Riedel, A. Kienzle, G. Petzow, M. BruÈck, T. Vaahs, German Patent 432,078,4-87 (1993).
[223] A. Kienzle, Ph.D. Thesis, University of Stuttgart, Stuttgart, 1994 (in German).
[224] L.M. Ruwisch, Ph.D. Thesis, Darmstadt University of Technology, Darmstadt, 1998, (in German).
[225] T. Ishikawa, Y. Kohtoku, K. Kumagawa, T. Yamamura, T. Nagasawa, Nature 391 (1998) 773.
[226] W.R. Schmidt, W.J. Hurley, R.H. Doremus, L.V. Interrante, P.S. Marchetti, in: M.D. Sacks (Ed.), Advanced
Composite Materials: Ceramic Transactions, Vol. 19, American Ceramic Society, Westerville, OH, 1991, p. 19.
[227] D. Seyferth, G. Brodt, B. Boury, J. Mater. Sci. Lett. 15 (1996) 348.
[228] G. Verdecia, K.L. O'Brien, W.R. Schmidt, T.M. Apple, Chem. Mater. 10 (1998) 1003.
[229] S.N. Borisov, M.G. Voronkov, E.Y. Lukevits, Organosilicon Heteropolymers, Heterocompounds, Plenum Press, New
York, 1970, p. 309.
[230] R.L. Allor, G.M. Crosbie, E.L. Cartwright, R.K. Govila, in: Proceedings of the 6th Annual ASM/ESD Advanced
Composites Conference, ASM International, Metals Park, OH, 1990, p. 555.
[231] S. Yajima, K. Okamura, S. Hayashi, M. Omori, J. Am. Ceram. Soc. 59 (1976) 324.
[232] C.K. Narula, B.G. Demczyk, P. Czubarow, D. Seyferth, J. Am. Ceram. Soc. 78 (1995) 1247.
[233] K.E. Gonsalves, R. Yazici, S. Han, J. Mater. Sci. Lett. 10 (1991) 834.
[234] J. Hapke, G. Ziegler, Adv. Mater. 7 (1995) 380.
[235] CFI GmbH & Co. KG, VorlaÈufige Produktinformation: Verbundwerkstoffe mit keramischer Matrix aus Borosilazan-
Poylmer, RoÈdental, Germany, January 1998.
[236] Y. Schimizu, Toso to Toryo 569 (1997) 27.
[237] Elf Attochem, personal communication, 1999.
[238] Gelest Inc., Catalog 2000, Silanes, Silicones, Metal-Organics.
[239] Hoechst AG, Product Information, Germany, 1994.
[240] Chisso Corp. personal communication, 1999 and data sheet ``Precursor for Advanced Ceramics-Polysilazame NCP-
200''.
[241] S. Yajima, J. Hayashi, M. Omori, Chem. Lett. (1975) 931.
[242] S. Yajima, K. Okamura, J. Hayashi, Chem. Lett. (1975) 1209.
[243] B.G. Penn, F.E. Ledbetter, J.M. Clemons, J.G. Daniels, J. Appl. Polym. Sci. 27 (1982) 3751.
[244] V.A. Mazaev, V.N. Talanov, M.M. Lin, A.K. Tsapuk, L.S. Miklin, T.M. Domashenko, Izv. Akad. Nauk SSSR,
Neorg. Mater. 19 (1986) 1796.
[245] H. Schmidt, Mater. Res. Sec. Symp. Proc. 32 (1984) 327.
[246] Y.D. Blum, R.M. Laine, Mater. Res. Soc. Symp. Proc. 73 (1986) 407.
[247] J. Lipowitz, J.A. Rabe, T.M. Carr, Polym. Prepr. 28 (1987) 411.
[248] P. Abrahams, Y.H. Mariam, Polym. Mater. Sci. Eng. 29 (1988) 404.
[249] R. Chaim, A.H. Heuer, J. Am. Ceram. Soc. 71 (1988) 960.
[250] G.T. Burns, US Patent 4, 929,742 (1990).
[251] J. Seitz, J. Bill, J. Mater. Sci. Lett. 15 (1996) 391.
[252] Y. Abe, T. Ozai, Y. Kuno, Y. Nagao, T. Misono, J. Inorg. Organomet. Polym. 2 (1992) 143.
[253] E. Duguet, M. Schappacher, A. Soum, Polym. Int. 33 (1994) 129.
[254] M. Scarlete, J. He, in: J.F. Harrod, R.M. Laine (Eds.), Application of Organometallic Chemistry in the Preparation
and Processing of Advanced Materials, Kluwer Academic Publishers, The Netherlands, 1995, p. 125.
[255] K. Niihara, T. Hirano, A. Nakahira, K. Ojima, K. Izaki, T. Kawakami, in: Proceedings of the MRS International
Meeting on Advanced Materials, Vol. 5, 1989, p. 107.
[256] K. Niihara, T. Hirano, A. Nakahira, K. Ojima, K. Izaki, T. Kawakami, J. Jpn. Soc. Powder Metal. 36 (1989) 243.
[257] R. Riedel, W. Dressler, Ceram. Int. 22 (1996) 233.
[258] R. Riedel, A. Kienzle, M. Friess, in: J.F. Harrod, R.M. Laine (Eds.), Application of organometallic Chemistry in the
Preparation and Processing of Advanced Materials, Kluwer Academic Publishers, The Netherlands, 1995, p. 155.
[259] D. Seyferth, in: G. Ziegler, H. Hausner (Eds.), Euro-Ceramics 2, Vol. 1, Deutsche Keramische Gesellschaft e. V.,
Augsburg, 1991, p. 567.
[260] D. Seyferth, C. Strohmann, H.J. Tracy, J.L. Robinson, Mater. Res. Soc. Symp. 249 (1992) 1.
[261] K.J. Wynne, R.W. Rice, Ann. Rev. 14 (1984) 297.
[262] R.M. Laine, Y.D. Blum, A. Chow, R. Hamlin, K.B. Schwartz, D.J. Rowcliffe, Polym. Preprints 28 (1987) 393.
[263] Y.D. Blum, R.M. Laine, K. Schwartz, D.J. Rowcliffe, R.C. Bening, D.B. Cotts, Better ceramics through chemistry,
in: C.J. Brinker, D.E. Clack, D.R. Ulrich (Eds.), MRS Proceedings, Vol. 32, 1984, p. 389.
[264] D. Seyferth, in: J.M. Ziegler, F.W.G. Feason (Eds.), Silicon Based Polymer Science Ð A Comprehensive Resource,
American Chemical Society, Washington, DC, 1990, p. 565.
[265] M. Reucket, T. Vaahs, M. BruÈck, Adv. Mater. 2 (1990) 398.
[267] Lanxide company, Technical Bulletin of CERASETTM SN Inorganic Polymer, 1997.

[268] B. Boury, R.J.P. Corriu, W.E. Douglas, Chem. Mater. 3 (1991) 487.
[269] D. Galusek, S. Reschke, R. Riedel, W. Dresser, P. Sajgalik, Z. Lences, J. Majling, J. Eur. Ceram. Soc. 19 (1999)
1911.
[270] G.T. Burns, T.P. Angelotti, L.F. Hanneman, G. Chandra, J.A. Moore, J. Mater. Sci. 22 (1987) 2690.
[271] J. Seitz, J. Bill, N. Egger, F. Aldinger, J. Eur. Ceram. Soc. 16 (1996) 885.
[272] K. Niedenzu, J.W. Dawson, Boron Nitrogen Compounds, Academic Publishers, New York, 1965.
[273] W. Verbeek, German Patent 221,896,0 (1973); 223,607,8 (1974); 224,352,7 (1975).
[274] K.B. Schwartz, D.J. Rowcliffe, Y.D. Blum, R.M. Laine, Mater. Proc. Symp. Proc. 73 (1983) 407.
[275] M. Arai, T. Isoda, Japanese Patent JP 61-89230.
[276] M. Arai, S. Sakurada, T. Isoda, T. Tomizawa, Polym. Preprints 28 (1987) 407.
[277] G.T. Burns, G. Chandra, J. Am. Ceram. Soc. 72 (1989) 333.
[278] O. Funayama, Tonen Corp., Japan, personnel information.
[279] S. Schaible, R. Riedel, E. Werner, U. Klingebiel, Appl. Organomet. Chem. 7 (1993) 53.
[280] T. Wideman, K. Su, E.E. Remson, G.A. Zank, L.G. Sneddon, Mater. Res. Soc. Symp. Proc. 410 (1996) 185.
[281] R. Riedel, A. Kienzle, V. Szabo, J. Mayer, J. Mater. Sci. 28 (1993) 3931.
[282] J. Bill, F. Aldinger, Adv. Mater. 7 (1995) 775.
[283] O. Funayama, H. Nakahara, A. Tezuka, T. Ishii, T. Isoda, J. Mater. Sci. 29 (1994) 2238.
[284] M. Kawaguchi, T. Kawashima, T. Nakajima, Chem. Mater. 8 (1996) 1197.
[285] J. Bill, M. Frieû, F. Aldiniger, R. Riedel, Mater. Res. Soc. Proc. 346 (1994) 605.
[286] M.L.J. Hackney, L.V. Interrante, G.A. Slack, P.J. Schields, in: J.D. Mackenzie, D.R. Ulrich (Eds.), Ultrastructure
Processing of Advanced Ceramics, Wiley, New York, 1988, p. 99.
[287] T.J. Clark, R.E. Johnson, US Patent 4,687,657 (1987).
[288] J.F. Janik, E.N. Duesler, R.T. Paine, Inorg. Chem. 26 (1987) 4341.
[289] W.R. Schmidt, W.J. Hurley Jr., R.H. Doremus, L.V. Interrante, P.S. Marchetti, Ceram. Trans. 19 (1991) 19.
[290] Y.F. Yu, T.I. Mah, in: J.D. Mackenzie, D.R. Ulrich (Eds.), Ultrastructure Processing of Advanced Ceramics, Wiley,
New York, 1988, p. 773.
[291] G.D. Soraru, A. Ravagni, R. Campostrini, F. Babonneau, J. Am. Ceram. Soc. 74 (1991) 2220.
[292] O. Funayama, Y. Tashiro, T. Isoda, J. Ceram. Soc. Jpn. 102 (1994) 908.
[293] K. Szulzewsky, C. Olschewski, I. Kosche, H.D. Klotz, R. Mach, Nanostr. Mater. 6 (1995) 325.
[294] S. Klein, M. Winterer, H. Hahn, Chem. Vapor. Dep. 4 (1998) 143.
[295] Z. Pan, H. Li, L. Zhang, J. Mater. Res. 13 (1998) 1996.
[296] G.W. Rice, J. Am. Ceram. Soc. 69 (1986) C183.
[297] G.W. Rice, R.L. Woodin, SPIE 458 (1984) 98.
[298] Y. Li, Y. Liang, F. Zheng, Z. Hu, Mater. Sci. Eng. A174 (1994) L23.
[299] Y. Li, Y. Liang, F. Zheng, Z. Hu, Mater. Lett. 22 (1995) 87.
[300] G.W. Rice, R.J. Woodin, J. Mater. Res. 4 (1989) 1538.
[301] M. Cauchetier, O. Croix, N. Herlin, M. Luce, J. Am. Ceram. Soc. 77 (1994) 993.
[302] Y. Li, Y. Liang, F. Zheng, Z. Hu, J. Am. Ceram. Soc. 77 (1994) 1662.
[303] T. Suzuki, T. Kawakami, T. Koyama, Jpn. J. Ceram. Soc. 95 (1987) 81.
[304] T. Suzuki, T. Kawakam, T. Koyama, H. Izaki, R. Nakano, T. Shitara, K. Hakkei, T. Hirai, K. Niihara, Yogyou-
kyokai-shi 95 (1987) 81.
[305] J. Hojo, H. Maeda, A. Kato, Jpn. J. Ceram. Soc. 95 (1987) 45.
[306] S.K. Ray, C.K. Maiti, N.B. Chakrabarti, J. Electon. Soc. 20 (1991) 907.
[307] W. Zhang, K. Zhang, B. Wang, Mater. Sci. Eng. B26 (1994) 133.
[308] A. Kienzle, J. Bill, F. Aldinger, R. Riedel, Nanostr. Mater. 6 (1995) 349.
[309] K.W. Gerstenberg, W. Beyer, J. Appl. Phys. 62 (1987) 1782.
[310] F. Maury, Z. Hatin, A. Reynes, R. Morancho, in: Proceedings of the 10th International Conference on CVD,
Electrochemical Society, Pennington, NJ, 1987, p. 1080.
[311] A.M. Wrobel, A. Walkiewiez-Pietrzykowska, Chem. Vapor Dep. 4 (1998) 133.
[312] A. Kienzle, A. Obermayer, A. Simon, R. Riedel, F. Aldinger, Chem. Ber. 126 (1993) 2569.
[313] M. Ducarroir, W. Zhang, R. Berjoan, J. de Phys. 3 (1993) 247.
[314] A. Bendeddouche, R. Berjoan, E. Beche, J. Appl. Phys. 81 (1997) 6147.
[315] W.F.A. Besling, A. Goossens, B. Meester, J. Schooman, J. Appl. Phys. 83 (1998) 544.
[316] T. Amano, T. Hira, K. Izaki, in: P. Vincenzini (Ed.), Ceramic Today Ð Tomorrow's Ceramics, Elsevier, Amsterdam,
1991, p. 785.
[317] J.S. Haggerty, S.C. Danforth, J.H. Flint, J.S. Haggerty, R.A. Marra, J. Am. Ceram. Soc. 65 (1982) 330.
[318] E. Borsella, S. Botti, R. Fantoni, R. Alexandrescu, I. Morjan, S. Endo, C. Popescu, J. Mater. Res. 7 (1992) 2257.
[319] R. Alexandrescu, I. Mojan, E. Borsella, S. Botti, R. Fantoni, R. Giogi, S. Endo, J. Mater. Res. 6 (1991) 2442.
[320] T. Liu, O. Sakurai, N. Kieda, N. Mizutani, M. Kato, Jpn. J. Ceram. Soc. 97 (1989) 935.
[321] C. Konetschny, D. Galusek, S. Reschke, C. Fasel, R. Riedel, J. Eur. Ceram. Soc. 19 (1999) 2789.
[322] C. Konetschny, Ph.D. Thesis, Darmstadt University of Technology, Darmstadt, 1998 (in German).
[323] Y.-L. Li, R. Riedel, unpublished work.
[324] H. Kleebe, D. Suttor, H. MuÈller, G. Ziegler, J. Am. Ceram. Soc. 81 (1998) 2971.
[325] P. Greil, J. Eur. Ceram. Soc. 18 (1998) 1905.
[326] T. Erny, M. Seibold, O. Jarchow, P. Greil, J. Am. Ceram. Soc. 76 (1993) 207.
[327] P. Greil, M. Modelling, J. Mater. Sci. 27 (1992) 1053.
[328] P. Greil, J. Am. Ceram. Soc. 78 (1995) 835.
[329] Z. Jiang, W.E. Rhine, Mater. Res. Soc. Symp. Proc. 286 (1993) 437.
[330] M. Herrmann, C. Schubert, A. Rendtel, H. HuÈbner, J. Am. Ceram. Soc. 81 (1998) 1095.
[331] R. Riedel, M. Seher, G. Becker, J. Eur. Ceram. Soc. 5 (1989) 113.
[332] E. Lecomte, Ph.D. Thesis, Darmstadt University of Technology, Darmstadt, 1999 (in German).
[333] B. Bauer, Ph.D. Thesis, Oxford University, 1998.
[334] Y. Iwamoto, K. Kikuta, S. Hirano, J. Mater. Res. 13-2 (1998) 353.
[335] Y. El Kortobi, H. Sfihi, A.P. Legrand, E. Musset, N. Herlin, M. Cauchetier, Colloids Surf. A 115 (1996) 319.
[336] B. Sundman, B. Jansson, J.O. Anderson, CALPHAD 9 (1985) 153.
[337] Z. Panek, J. Am. Ceram. Soc. 78 (1995) 1087.
[338] S. Johnson, The Si-Al-O-N System Ð Thermodynamic Properties and Phase Diagrams, Report of the Division of
Physical Metallurgy, Royal Institute of Technology, Stockholm, Sweden, 1991.
[339] J. Weiss, H.L. Lukas, J. Lorenz, G. Petzow, H. Krieg, CALPHAD 5 (1981) 125.
[340] Z. Lences, M. Haviar, Processing 25 (1996) 179.
[341] J.-J. Park, O. Komura, A. Yamakawa, K. Niihara, J. Am. Ceram. Soc. 81 (1998) 2253.
[342] A.W. Weimer, G.A. Eisman, D.W. Susnitzky, D.R. Beaman, J.W. McCoy, J. Am. Ceram. Soc. 80 (1997) 2853.
[343] Z. Lences, A. Bellosi, F. Monteverde, Mater. Chem. Phys. 41 (1995) 46.
[344] M. Seher, Ph.D. Thesis, University of Stuttgart, Stuttgart, 1993 (in German).
[345] H. Freimuth, V. Hessel, H. KoÈlle, M. Lacher, W. Ehrfeld, T. Vaahs, M. BruÈck, J. Am. Ceram. Soc. 79 (1996) 1457.
[346] T. Hanemann, J.H. Hausselt, Am. Chem. Soc. 39 (1998) 659.
[347] J. Lipowitz, H.A. Freeman, R.T. Chen, E.R. Prack, Adv. Ceram. Mater. 2 (1987) 121.
[348] R.E. Jones, J.A. Rabe, A.L. Peterson, Can. Chem. News 44 (1992) 9.
[349] T. Vaahs, M. BruÈck, W.D.G. BoÈcker, Adv. Mater. 4 (1992) 224.
[350] N.P. Bansal, R.M. Dickerson, Mater. Sci. Eng. A222 (1997) 149.
[351] L. Lu, J. Mater. Sci. Lett. 17-7 (1998) 599.
[352] D.R. Bujalski, G.A. Zank, T.D. Barnard, US Patent 91-772384 19910927.
[353] M. Labrousse, M. Nanot, P. Boch, E. Chassgneux, Ceram. Int. 19 (1993) 259.
[354] C.J. Langlois, G.S. Lum, US Patent 5,258,224 (1993).
[355] G.S. Lum, C.J. Langlois, US Patent 5,258,229 (1993).
[356] D. Heimann, J. Bill, F. Aldinger, P. Schanz, F. Gern, W. Krenkel, R. Kochendoerfer, Zeitschrift fuÈr
Flugwissenchaften und Weltraumforschung 19-3 (1995) 180.
[357] G. Motz, G. Ziegler, in: J. Heinrich, G. Ziegler, W. Hermel, H. Riedel (Eds.), Keramik, Werkstoffwoche 0 98, Band
VII, MuÈnchen, 12±15 October, 1998, Wiley-VCH, Weinheim, 1999, p. 667.
[358] Y. Schimizu, H. Matsuo, K. Yamada, Japanese Patent 96-JP1976 19960715.
[359] J. Tashiro, T. Oohayashi, H. Matsuo, S. Suzuki, Japanese Patent 94-308917 19941213.
[360] K. Inoe, Y. Komatsu, S. Suzuki, T. Isoda, Japanese Patent 90-413273 19901221.
[361] M.A. Rooke, P.M.A. Sherwood, Chem. Mater. 9 (1997) 285.
[362] G.A. Zank, US Patent 94-282300 19940729.
[363] S.T. Schwab, S.F. Timmons, C.R. Blanchard, M.D. Grimes, R.C. Graef, J.D. Katz, D.E. Rees, T.W. Hardek, Mater.
Res. Soc. Symp. Proc. 430 (1996) 619.
[364] I. Barin, Thermochemical Data of Pure Substances, VCH, Weinheim, 1993.
[365] N. Noaki, Japanese Patent 89-343396 19891229.
[366] M. Takeuchi, Japanese Patent 89-300324 19891117.
[367] M. Shinho, G. Yokobori, T. Katano, Japanese Patent 93-225092 19930818.
[368] G.I. Orlov, I.M. Sytova, S.Y. Vodyanitskij, Russian Patent RU 2047636 C1; RU 2034892 C1; RU 2034890 C1; RU
2032701 C1.
[369] G. Boden, S. Siegel, F. Maspero, in: J. Heinrich, G. Ziegler, W. Hermel, H. Riedel (Eds.), Keramik, Werkstoffwoche
0
98, Band VII, MuÈnchen, 12±15 October, 1998, Wiley-VCH, Weinheim, 1999, p. 31.
[370] R.C. Graef, D.G. Paquette, S.T. Schwab, US Patent 5,294,425.
[371] C.R. Bearinger, R.C. Camilletti, G. Chandra, L.A. Haluska, T.E. Gentle, US Patent 95-533217.
[372] T. Morozumi, K. Sato, Shinsozai 6 (1995) 41.
[373] Y. Tsumori, H. Oshima, Japanese Patent 97-23692 19970206.
[374] G. Ziegler, J. Luecke, I. Richter, D. Suttor, Key Eng. Mater. 132±136 (1997) 1870.
[375] R. Lundberg, R. Pompe, R. Carlsson, P. Goursat, Composites Sci. Technol. 37 (1990) 165.
[376] K. Sato, T. Suzuki, O. Funayama, T. Isoda, J. Ceram. Soc. Jpn. 100 (1992) 444.
[377] W.E. Rhine, A. Lighfoot, D. Seyferth, J.S. Haggerty, in: R.A. Lowden, M.K. Ferber, J.R. Hellman, K.K. Chawla,
S.G. DiPietro (Eds.), Proceedings of the Ceramic Matrix Composites Conference on Advanced High Temperature
Structural Materials, Boston, USA, 28 Novemeber±02 December 1994, Pittsburgh, USA, 1995, p. 101.
[378] K. Suzuki, S. Kume, J. Ceram. Soc. Jpn. 106 (1998) 348.
[379] M.F. Ganon, G. Fantozzi, M. Murat, J.P. Disson, J. Eur. Ceram. Soc. 15 (1995) 185.
[380] H. SchoÈnfelder, F. Aldinger, R. Riedel, in: Proceedings of the 3rd European Conference on Advanced Materials and
Processes, Vol. 2, J. Phys. IV, 1993, p. 1293.
[381] B.A. Bender, R.W. Rice, J.R. Spann, J. Am. Ceram. Soc. 70 (1987) C58.
[382] B.E. Walker, R.W. Rice, P.F. Becher, B.A. Bender, W.S. Cobkenz, Am. Ceram. Soc. Bull. 62 (1983) 916.
[383] S. Yajima, Am Ceram. Soc. Bull. 62 (1983) 893.
[384] J. Bill, F. Aldinger, Prog. Mater. Synth. 87 (1996) 11.
[385] J. Dixmier, R. Bellissent, D. Bahloul, P. Goursat, J. Eur. Ceram. Soc. 13 (1994) 293.
[386] Y. Yokoyama, T. Nanba, I. Yasui, H. Kaya, T. Maeshima, T. Isoda, J. Am. Ceram. Soc. 74 (1991) 654.
[387] S. Schempp, J. DuÈrr, P. Lamparter, J. Bill, F. Aldinger, Z. Naturforsch. 53a (1998) 127.
[388] M.R. Mucalo, D.G. Mcgavin, N.B. Milestone, J. Mater. Sci. 32 (1997) 3271.
[389] A. Gheorghiu, C. Senemaud, H. Roulet, G. Dufour, T. Moreno, S. Bodeur, C. Reynaud, M. Cauchetier, M. Luce, J.
Appl. Phys. 71 (1992) 4118.
[390] F. Tenegal, A.M. Flank, N. Herlin, Phys. Rev. B 54 (1996) 12029.
[391] M. Suzuki, Y. Hasegawa, M. Aizawa, Y. Nakata, T. Okutani, J. Am. Ceram. Soc. 78 (1995) 83.
[392] F.J. Gomez, P. Prieto, E. Elizalde, J. Piqueras, Appl. Phys. Lett. 69 (1996) 773.
[393] R. Franke, S. Bender, A.A. Pavlychev, P. Kroll, R. Riedel, A. Greiner, J. Electron Spectrosc. Relat. Phen. 96 (1998)
253.
[394] P. Kroll, A. Greiner, R. Riedel, Mater. Res. Soc. Symp. Proc. 437 (1996) 231.
[395] A. Obermeyer, A. Kienzle, J. Weidlein, R. Riedel, A. Simon, Z. Anorg. Allg. Chem. 620 (1994) 1357.
[396] L.C. Chen, C.K. Chen, S.L. Wie, Appl. Phys. Lett. 72 (1998) 2463.
[397] L.C. Chen, D.M. Bhusar, C.Y. Yang, Thin Solid Films 303 (1997) 66.
[398] Y.K. Chang, H.H. Hsieh, W.F. Pong, Phys. Rev. B 58 (1998) 9018.
[399] M. Monthioux, O. Delverdier, J. Eur. Ceram. Soc. 16 (1996) 721.
[400] D. Kurtenbach, H.P. Martin, E. MuÈller, G. Roewer, A. Hoell, J. Eur. Ceram. Soc. 18 (1998) 1885.
[401] O. Delverdier, M. Monthioux, D. Mocaer, R. Pailler, J. Eur. Ceram. Soc. 12 (1993) 27.
[402] D. Mocaer, R. Pailler, R. Naslain, C. Richard, J.P. Pillot, J. Dunogues, C. Darnez, M. Chambon, M. Lahaye, J.
Mater. Sci. 28 (1993) 3049.
[403] Y. Li, Y. Liang, Z. Hu, J. Mater. Sci. Lett. 13 (1994) 1588.
[404] H.T. Sawhill, J.S. Haggerty, J. Am. Ceram. Soc. 65 (1982) C131.
[405] M. Seher, J. Bill, F. Aldinger, R. Riedel, J. Crystal Growth 137 (1994) 452.
[406] R. Riedel, M. Seher, J. Eur. Ceram. Soc. 7 (1991) 21.
[407] D.R. Clark, J. Am. Ceram. Soc. 65 (1982) C21.
[408] G. Ziegler, J. Heinrich, G. WoÈtting, J. Mater. Sci. 22 (1987) 3041.
[409] C.H. Carter Jr., R.F. Davis, J. Bently, J. Am. Ceram. Soc. 67 (1984) 732.
[410] C.H. Carter Jr., R.F. Davis, J. Bently, J. Am. Ceram. Soc. 67 (1984) 409.
[411] M. Scarlete, J. He, J.F. Harrod, I.S. Butler, in: J.F. Harrod, R.M. Laine (Eds.), NATO ASI Series E: Applied
Sciences, Vol. 297, 1994, p. 125.
[412] A. Sawaguchi, K. Toda, J. Am. Ceram. Soc. 74 (1991) 1142.
[413] C. Haluschka, Ph.D. Thesis, Darmstadt University of Technology, Darmstadt, 1997 (in German).
[414] N.F. Mott, E.A. Davies, Electronic Processes in Non-crystalline Materials, Clarendon Press, Oxford, 1979.
[415] W. Heywang, Amorphe und polykristalline Halbleiter, Halbleiterelektronik Bd. 18, Springer, Berlin, 1984.
[416] A. Rendtel, H. HuÈbner, M. Herrmann, C. Schubert, J. Am. Ceram. Soc. 81 (1998) 1109.
[417] P. SÏajgalõÂk, K. Rajan, P. Warbichler, F. Hofer, J. Dusza, Key. Eng. Mater. 159±160 (1999) 405.
[418] T. Hirano, A. Nakahira, K. Niihara, J. Ceram. Soc. Jpn. 14 (1994) 1243.
[419] K. Niihara, T. Hirano, K. Izaki, F. Wakai, Silicon-base structural ceramics, in: B.W. Sheldon, S.C. Danforth (Eds.),
Ceramic Transactions, Vol. 42, American Ceramic Society, Westerville, OH, 1994, p. 207.
[420] X. Pan, J. Mayer, M. RuÈhle, J. Am. Ceram. Soc. 79 (1996) 585.
[421] H. Klemm, M. Herrmann, C. Schubert, G. Pezzotti, J. Am. Ceram. Soc., in press.
[422] G. Motz, Universitat Bayreuth, Germany, personal information.

Alumina Etching

Uploaded by

Copyright:

Available Formats

Alumina Etching

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Alumina Etching

Uploaded by

Copyright:

Available Formats

Materials Science and Engineering, 26 (2000) 97±199

Silazane derived ceramics and related materials

Accepted 14 January 2000

aforementioned applications where exceptional mechanical reliability under severe conditions is

Si3x Cxy N4 ! Si3 N4  xSiC  yC (1)

subsequently processed to dense polycrystalline Si3N4/SiC micro/nano-composites by hot-pressing

2.1. Ammonolysis of halogenosilanes

2nR4ÿx SiClx  3xnNH3 !  NHx=2 ÿ SiR4ÿx n  2xn;

BSiÿCl  2NH3 ! BSiÿNH2  NH4 Cl (5)

This reaction scheme is completely analogous to the above-mentioned hydrolysis of

BSiÿNH2  BSiÿNHÿSiB ! BSiÿ3 N  NH3 (7)

2.1.1. Ammonolysis of monochlorosilanes R3SiCl

2 H3 C3 SiCl  3NH3 ! H3 C3 SiÿNHÿSi CH3 3  2NH4 Cl (9)

3RR0 HSiCl  4NH3 ! RR0 HSi3 N  3NH4 Cl (10)

2.1.2. Ammonolysis of dichlorosilanes R2SiCl2

Linear disilazanes, H2N±SiR2±NH±SiR2±NH2, have been found to form together with

As already mentioned, cyclotetrasilazanes are usually formed together with cyclotrisilazanes as

2.1.3. Ammonolysis of trichlorosilanes RSiCl3

3NH3  2HSiCl3 ! HCl2 SiÿNHÿSiCl2 H  2NH4 Cl (19)

4:5nNH3  nHSiCl3 ! HSi NH1:5 n  3nNH4 Cl (20)

RSiCl3  4:5NH3 ! 1=nRSi NH1:5 n  3NH4 Cl R  ÿCH3 ; ÿC2 H5 ; ÿC6 H13 ;

6HSiCl3  9NH3  18Na ! CH3 Si6 NH9  18NaCl  9H2 (23)

2.1.4. Ammonolysis of tetrachlorosilane SiCl4

2.1.5. Ammonolysis of chlorosilane mixtures Ð copolysilazanes

2.1.6. Miscellaneous ammonolysis reactions

2.2. Aminolysis of chlorosilanes

Rx SiCl4ÿx  2nR02 NH ! Rx R02 Nn SiCl4ÿxÿn  nR02 NH2 Cl (26)

2nR4ÿx SiClx  3xnR0 NH2 !  NR0 x=2 ÿSiR4ÿx n  2xnR0 NH3 Cl

If the stoichiometry is carefully controlled using an amine:chlorosilane ratio of 3:1 the

Surprisingly, reactions of three or five equivalents of isopropyltrichlorosilane or phenyltri-

exception is the above-mentioned amminolysis of methyltrichlorosilane [102] (see Section 1.2.3).

2.3. Hydrazinolysis of chlorosilanes

Another approach to producing Si/C/N precursors from chlorosilanes is a hydrolysis-,

2nR4ÿx SiClx  3xnN2 H4 !  N2 H2 x=2 ÿSiR4ÿx n  2xnN2 H5 Cl

Based on the above described synthesis of a Si/C/N-precursor from dichloromethylvinylsilane

2.4. Silazanolysis of chlorosilanes

nR4ÿx SiClx  xnHNSi CH3 2 !  NHx=2 ÿSiR4ÿx n  2xnClSi CH3 3

SiCl4  CH3 3 SiÿNHÿSi CH3 3 ! Cl3 SiÿNHÿSi CH3 3  ClSi CH3 3 (38)

2 CH3 2 HSiCl  CH3 3 SiÿNHÿSi CH3 3 ! CH3 2 HSiÿNHÿSiH CH3 2  2ClSi CH3 3

2.5. Miscellaneous routes

In addition to the preparative methods described in Sections 2.1±2.5, a number of syntheses to

Lx EÿCl  L0y E0 ÿCl  2Na ! Lx EÿE0 L0y  2NaCl (41)

Ring-opening polymerisation of organocyclosilazanes was also initiated by trifluoromethan-

Very recently, lithium nitride dissolved in diglyme, H3CO±(CH2)2±O±(CH2)2±OCH3, and

2.6. Modification of silazanes

2.7. Introduction of heteroelements (B, Al, Ti, . . .)

Borazine-silazane copolymers have been synthesised similarly by thermal dehydro-condensa-

A similar dehydrocoupling reaction of tetrakis(methylamino)silane with a borane adduct was

A hydridomethylsilazane co-polymer of the form [(CH3)2Si±NH±]m[CH3SiH±NH±]n has been

A method similar to the above-described preparation of Si/B/C/N precursors by the treatment of

TiCl4  CH3 3 SiÿNHÿSi CH3 3 ! CH3 3 SiÿNHÿTiCl3  ClSi CH3 3 (70)

2.8. Commercially available Si/(E)/C/N-precursors

2.8.1. Molecular compounds

2.8.2. Polymeric compounds

3. Processing from Si/(E)/C/N precursors to ceramic materials

3.1. Historical sketch and overview

3.1.1. Historical sketch

E. Kroke et al. / Materials Science and Engineering 26 (2000) 97±199

1,3-Diphenyl-1,1,3,3-tetramethyldisilazane 3449-26-1 bp: 96±1008C/0.1 Torr Sigma, ABCR

Si3x Cxy N4 ! Si3 N4 xSiC yC (1)

2nR4ÿx SiClx 3xnNH3 ! NHx=2 ÿ SiR4ÿx n 2xn;

BSiÿCl 2NH3 ! BSiÿNH2 NH4 Cl (5)

BSiÿNH2 BSiÿNHÿSiB ! BSiÿ3 N NH3 (7)

2H3 C3 SiCl 3NH3 ! H3 C3 SiÿNHÿSiCH3 3 2NH4 Cl (9)

3RR0 HSiCl 4NH3 ! RR0 HSi3 N 3NH4 Cl (10)

3NH3 2HSiCl3 ! HCl2 SiÿNHÿSiCl2 H 2NH4 Cl (19)

4:5nNH3 nHSiCl3 ! HSiNH1:5 n 3nNH4 Cl (20)

RSiCl3 4:5NH3 ! 1=nRSiNH1:5 n 3NH4 Cl R ÿCH3 ; ÿC2 H5 ; ÿC6 H13 ;

6HSiCl3 9NH3 18Na ! CH3 Si6 NH9 18NaCl 9H2 (23)

Rx SiCl4ÿx 2nR02 NH ! Rx R02 Nn SiCl4ÿxÿn nR02 NH2 Cl (26)

2nR4ÿx SiClx 3xnR0 NH2 ! NR0 x=2 ÿSiR4ÿx n 2xnR0 NH3 Cl

2nR4ÿx SiClx 3xnN2 H4 ! N2 H2 x=2 ÿSiR4ÿx n 2xnN2 H5 Cl

nR4ÿx SiClx xnHNSiCH3 2 ! NHx=2 ÿSiR4ÿx n 2xnClSiCH3 3

SiCl4 CH3 3 SiÿNHÿSiCH3 3 ! Cl3 SiÿNHÿSiCH3 3 ClSiCH3 3 (38)

2CH3 2 HSiCl CH3 3 SiÿNHÿSiCH3 3 ! CH3 2 HSiÿNHÿSiHCH3 2 2ClSiCH3 3

Lx EÿCl L0y E0 ÿCl 2Na ! Lx EÿE0 L0y 2NaCl (41)

TiCl4 CH3 3 SiÿNHÿSiCH3 3 ! CH3 3 SiÿNHÿTiCl3 ClSiCH3 3 (70)

Gelest Inc. (USA) PSN-2M01: d1.04 g/ml, Z>0.05 Pa s, Available [238]

Gelest Inc. (USA) PSN-2M11: d0.99 g/ml, Z0.5±0.8 Pa s Available [238]

HoÈchst AG (Germany) ET 70 ±[CH3CH2Si(NH)1.5]n± M1600 g/mol, 70 mol% Not available [216,239]

HoÈchst AG (Germany) VT 50 ±[CH2=CH±Si(NH)1.5]n± M2000 g/mol, 50 mol% Not available [216,239]

University of Bayreuth, HVNG M620 g/mol, Available [422]

University of Bayreuth, HPS M440 g/mol, Available [422]

University of Bayreuth, ABSE M1100 g/mol, Available [422]

DSi Si0 ÿ YSic (72)

BSiÿCH3 NH3 ! BSiÿNH2 CH4 (73)

NÿH HÿSiB ! NÿSiB H2 (74)

4H2 SiNHx ! Si3 N4 Si 6H2 (80)

3.2.4. The quaternary Si/(E)/C/N system (EB, Ti,. . .)

Si3 N4 s ! 3Si l 2N2 g (81)

3SiC s 2N2 g ! Si3 N4 s 3C s (83)

SiO2 s; l 3C s ! SiC s 2CO g (84)

SiO g 2C s ! SiC s CO g (840 )

2SiO g N2 g C s ! Si2 N2 O s CO g (860 )

2SiO g CO g N2 g ! Si2 N2 O s 2CO2 g (87)