Synthesis of hydrocarbon-soluble, methyl-substituted highly branched polysilanes via the Wurtz-type reductive coupling of trifunctional trisilanes and their pyrolysis to silicon carbide

Abstract

A hydrocarbon-soluble, methyl-substituted highly branched polysilane was synthesized by subjecting a new type of monomer to the common Wurtz-type reductive coupling reaction conditions. For this purpose, the trichloro-substituted trisilane (ClMe₂Si)₂SiMeCl (1) was used as a monomer. Polymerization was conducted at elevated temperature using dispersed sodium in toluene. Termination was initiated after 4 h by the addition of MeMgBr and precipitation in MeOH. The branched polysilane was analyzed by GPC, NMR and IR spectroscopy after quenching with MeMgBr and additionally after precipitation in MeOH. Opto-electronic properties were tested by UV/Vis and PL spectroscopy. Thermal properties were examined via TGA and DSC. The synthesized polysilane is usable as a soluble precursor for the preparation of SiC, not least due to its high silicon content. The chemical process during pyrolysis was examined at different temperatures by means of ¹H NMR, IR and UV/Vis spectroscopy. XRPD analysis, IR spectroscopy, HR-SEM-EDS and elemental analysis confirmed the successful formation of β-SiC.

---

Introduction

Polysilanes, with an all silicon backbone and two organic substituents at each silicon atom, have been known since the early 1920s,^1–5 when Kipping et al. synthesized polydiphenylsilane [Ph₂Si]_n for the first time.^6,7 These early polymers were mainly insoluble and intractable substances, difficult to characterize and thus hardly industrially applicable. The scientific interest in polysilanes only increased when the first hydrocarbon-soluble and therefore easily analyzable high molar mass linear polysilanes (homo- and copolymers) were synthesized in the late 1970s and beginning of the 1980s.^3,8–19 In general, R¹R²SiCl₂ (R¹/R² = alkyl or aryl) was reacted with an alkali metal dispersion (usually sodium) at elevated temperatures in a high-boiling inert solvent (usually toluene), named the Wurtz-type reductive coupling reaction.^2–4,20 Subsequently, the thereby achieved accessibility enabled the development of various industrial applications. The extensive σ-conjugation of electrons along their silicon main chain grants this class of polymers unique optical, electronic and photophysical properties.^3,4,21,22 Consequently, polysilanes became suitable as basic material for semiconductors (when doped with AsF₅),^1,18,23 photoinitiators,^4,23,24 photoconductors,^4,23,25,26 non-linear optical materials^4,23 or photoresists^1,4,23 for microlithography.

Besides the aforementioned possible fields of applications, Yajima et al. discovered in 1975 that methyl-substituted polysilanes are suitable precursors for the pyrolytic preparation of β-silicon carbide (β-SiC).^27–29 According to Yajima, the conversion of hydrocarbon-insoluble polydimethylsilane [Me₂Si]_n towards β-SiC takes place in a two-step process (Scheme 1).^3,28–31

Scheme 1 Simplified illustration of the pyrolytic conversion of PDMS into β-SiC.³

Within the first low-temperature thermolysis (∼450 °C), the intractable polydimethylsilane (PDMS) is converted into a soluble and moldable polycarbosilane (PCS) to prevent weight loss at higher temperatures. Initially it was assumed, that during thermolysis α-carbon radicals are generated, which insert into adjacent Si–Si bonds (Kumada³² rearrangement). Subsequent hydrogen abstraction of the resulting silyl radicals from methyl groups lead to the reformation of α-carbon radicals. Ideally, repetition of these steps result in the formation of [MeHSiCH₂]_n. More recently, the involvement of silylenes and silenes during rearrangement is discussed as well. The actual structure of the gained PCS is much more complex than depicted within Scheme 1, since during rearrangement additional cross-linking and cyclization reactions might take place.^3,28–33

Within the next steps the PCS is surface oxidized by heating in air (∼350 °C) for stabilization reasons and is afterwards pyrolized in a final step at 1300 °C in an inert atmosphere to gain polycrystalline β-SiC.^3,28–33

SiC is of great interest for industrial applications, due to its high thermal stability, convenient mechanical properties, outstanding oxidation resistance up to 1400 °C, good thermal conductivity and its semiconductive nature.^34,35

However, Yajima's preparation method and all other common synthetic routes to SiC existing so far (Acheson process³⁵ or chemical vapor deposition^35,36) exhibit a big drawback: the preparation of structurally detailed components or fine ceramic structures on substrates is procedurally difficult due to the nature of the SiC preparation.¹ Such detailed components and fine structures can be of interest in the field of electronic and opto-electronic devices, especially for use at high temperatures or under environmental extremes.³⁰

An adequate basic material for this kind of application would meet the requirements for an idealized ceramic precursor stated by Dunoguès et al., i.e. especially a high silicon content and furthermore an adequate ratio of Si to C to grant a high ceramic yield, viscoelastic properties (e.g. solubility) to facilitate procedural handling and a sufficiently high molar mass to avoid oligomeric volatilization.^30,37

With regard to the silicon content and Si/C ratio, methyl-substituted polysilanes are favorable when compared to alternative organic side groups. Using methyl-side groups also helps to keep the amount of free carbon produced during and after pyrolysis comparatively low.^30,37 However, methyl-substituted linear polysilanes [Me₂Si]_n are not soluble in common organic solvents.^2,8,9,11,38 To address the requirement for solubility, complex polysilane structures (network or branched) have to be taken into account. Over the years, a large number of soluble, high molar mass network polysilanes^39–43 and (hyper)branched copolysilanes^41,44–47 with various organic side chain substituents – excluding pure methyl-substitution – have been prepared using the traditional Wurtz-type reductive dehalogenation reaction conditions. Network polysilanes were synthesized by polymerization of trifunctional monomers R¹SiCl₃ (R¹ = alkyl or aryl) and randomly branched polysilanes by the copolymerization of R¹SiCl₃ (R¹ = alkyl or aryl) and R²R³SiCl₂ (R²/R³ = alkyl or aryl).^39–47 Methyl-substituted network polysilanes [MeSi]_n (ref. 41 and 48) are also literature-known, but are, similarly to the linear ones, insoluble in common organic solvents. Thus, branched structures remain the only thinkable alternative to produce soluble polymethylsilanes. To our knowledge, the copolymerization of trichloro(methyl)silane MeSiCl₃ and dichloro(dimethyl)silane Me₂SiCl₂ resulting in [(MeSi)_n(Me₂Si)_m] has never been mentioned in literature. Although this procedure, commonly accepted for other organic side chains, might seem to be the obvious next step to the formation of branched polymethylsilanes, we chose an alternative synthesis route.

Instead of using the mentioned chloro-substituted monosilanes as educts (MeSiCl₃ and Me₂SiCl₂) we used the trichloro-substituted oligomer (ClMe₂Si)₂SiMeCl (1) as starting material. This trisilane already contains the desired ratio of substituents, i.e. the tertiary silicon branching unit and the linear silicon units are already pre-built.⁴⁹ In contrast, when using monosilanes, it is likely that their different chemical affinities to be incorporated into the growing polymer chain lead to the formation of irregular, possibly even linear or network, and consequently insoluble structures.⁴¹ Using trisilanes as starting material should ease the formation of highly branched polymers and hence assure the solubility in common organic solvents, although methyl-groups are applied as side chains. Highly branched polysilanes with a regular substitution pattern, a controlled degree of branching and high silicon content are expected to be generated.

This kind of polymers attracted our attention not only with regard to their possible suitability as soluble precursor for the pyrolytic preparation of SiC,^28–30 but also with regard to their wide range of applications due to their exceptional optical, electronic and photophysical properties.^1,4,22,23 Using this type of polysilane, previous common applications might be reformed or extended due to its combination of solubility, methyl-substitution and branching. Defined structures and components composed of polysilane or SiC may become feasible.

Experimental section

Experimental methods and materials

All experiments were performed under an atmosphere of argon (purity: 99.998%, Westfalen AG) using standard Schlenk and glovebox techniques unless otherwise stated. For the additional removal of oxygen and moisture in the argon, a gas clean filter (OT3-4 trap, Agilent) was directly installed in front of the Schlenk line. All reactions were conducted using glassware (including a glass coated magnetic stirring bar), which was heat dried under vacuum prior to use. The monomer 1,2,3-trichloro-1,1,2,3,3-pentamethyltrisilane (1) was synthesized using a method described elsewhere.⁵⁰ All chemicals were purchased from common suppliers and were used as received without further purification unless otherwise stated. Sodium and MeMgBr (3.0 M in diethyl ether) were purchased from Sigma-Aldrich. Toluene (>99.8%, Fisher Sci. GmbH) and n-hexane (>97%, Sigma-Aldrich) were dried over sodium, benzophenone (99%, Sigma-Aldrich) and tetraethylene glycol dimethyl ether (≥98%, Fluka) and were distilled, degassed (freeze–pump–thaw technique, three times) and stored under argon atmosphere prior to use.⁵¹ THF (99.8%, Fisher Sci. GmbH) was dried over sodium and benzophenone and was distilled, degassed and stored under argon atmosphere prior to use, too.⁵¹ Dry methanol (>99.8%, Fisher Sci. GmbH) was obtained by the addition of sodium and subsequent distillation, degassing and storage under argon atmosphere.⁵¹ Chlorotrimethylsilane (99%, ABCR) was distilled for purification reasons and stored under argon atmosphere.

Analytical methods and equipment used

¹H (500.36 MHz), ²⁹Si{¹H} INEPT (99.41 MHz) and ¹H/²⁹Si{¹H} HMBC NMR spectra were recorded on a Bruker AV-500cryo spectrometer at ambient temperature in [D₈]THF or C₆D₆ and referenced versus internal tetramethylsilane (¹H NMR: δ_H = 0 ppm, ²⁹Si{¹H} INEPT: δ_Si = 0 ppm). Spectroscopic chemical shifts δ are reported in ppm. [D₈]THF and C₆D₆ were obtained from Euriso-top® GmbH and were each stirred over NaK alloy before they were transferred into another flask via condensation and stored in a glove box (LABmaster sp, MBraun). In order to perform the ¹H/²⁹Si{¹H} HMBC NMR spectra, a Bruker standard HMBC NMR parameter file was modified detecting ²⁹Si nuclei instead of ¹³C nuclei. Furthermore, a gradient was implemented to reduce the measurement time (GPZ [%]: 1 = 60, 2 = 20, 3 = 55.89379). Gel permeation chromatography (GPC) was carried out using a Varian PL-GPC-50 Plus chromatograph with two PLgel 5 μm MIXED-C columns and a differential refractive index (DRI) detector for the determination of the molar masses and the polydispersity (M_w/M_n) of the branched polymethylsilanes. Measurements were conducted with THF as eluent (2 g L⁻¹) at 35 °C and a flow rate of 1 mL min⁻¹. Linear polystyrene standards were used to calibrate the GPC system. Attenuated total reflectance infrared spectroscopy (ATR-IR) was performed on a Bruker Vertex 70 FT-IR spectrometer (range: 3500–400 cm⁻¹) with a Platinum ATR unit (Bruker) at room temperature. The following abbreviations were used for the evaluation of the absorption bands: s – strong, m – medium, w – weak, v – very and sh – shoulder. Ultraviolet-visible (UV/Vis) spectra were recorded on a Varian Cary-50 Scan UV/Vis spectrophotometer at ambient temperature in a quartz glass cuvette (path length = 10 mm) using n-hexane as solvent (0.005 g L⁻¹). Photoluminescence (PL) spectroscopy was conducted on an AvaSpec-2048 fiber optic spectrometer (Avantes GmbH) with a Mic-LED-365 (Prizmatix) as light source for excitation (adjusted excitation wavelength λ_ex = 365 nm). Measurements were performed using a quartz glass cuvette (cuvette diameter = 10 mm; 90° cuvette holder) and THF (1.5 g L⁻¹) as solvent at ambient temperature. Thermogravimetric analysis (TGA) was carried out on the thermobalance TG 209 F1 Libra (Netzsch) with a heating rate of 5 K min⁻¹ in an argon flow of 20 mL min⁻¹ (purity: 99.998%, Westfalen AG) in platinum pans (∼4 mg). Differential scanning calorimetry (DSC) measurements were performed on the differential scanning calorimeter Q2000 (TA Instruments) using hermetic aluminum pans sealed under argon atmosphere (∼5 mg). A heating/cooling rate of 10 K min⁻¹ was used, whereas the data of the 2^nd heating cycle were utilized for analysis. Polymer samples were heat-treated using a Nabertherm RD 30/200/13 compact tube furnace. X-ray powder diffraction (XRPD) was carried out on a STOE Stadi P diffractometer (CuK_α1 radiation; Ge(111) monochromator) equipped with a Dectris Mythen microstrip solid-state detector. The sample was thoroughly ground in an agate mortar before it was sealed inside a glass capillary (0.3 mm). The measurement was conducted applying the Debye–Scherrer method (2Θ-range: 15–90°) for 16 hours. High resolution scanning electron microscopy (HR-SEM) and energy-dispersive X-ray spectroscopy (EDS) analyses were performed on a JEOL JSM-7500F scanning electron microscope equipped with a X-Max detector (Oxford Instruments, size: 50 mm²) using an accelerating voltage of 1 kV for secondary electron observation and 7 kV for EDS analysis. Elemental analyses were conducted at the Chair of Inorganic Chemistry, Technische Universität München, using the Euro EA Elemental Analyzer (HEKAtech GmbH) for the elements C, H, N and S. The silicon content was determined photometrically (UV-160, Shimadzu) after an alkaline digestion of the silicon containing samples with Na₂O₂ and the subsequent conversion with ammonium heptamolybdate in an acid setting in order to form molybdenum blue. The latter was then photometrically reviewed.

Polymerizations and work-up procedures

Two different batches of branched polysilanes are described within the present work, PS1 and PS2. Both batches were prepared using exactly the same polymerization conditions. PS2 is described within this work since samples for characterization purposes were retrieved after work-up procedure I (PS2-I) and after work-up procedure II (PS2-II).

Details on polymerization entry PS1. Polysilane PS1 was prepared in a round bottom flask (50 mL) equipped with a glass coated magnetic stirring bar and an overpressure safety valve on top of a reflux condenser. Initially, freshly cut sodium metal (650 mg, 28.2 mmol, 3 equiv.) was added in a glove-box to a solution of dry toluene (20 mL) and 1,2,3-trichloro-1,1,2,3,3-pentamethyltrisilane (1) (2.5 g, 9.4 mmol, 1 equiv.). The reaction mixture was immediately brought to 110 °C and was stirred for 4 h. Afterwards work-up procedure I and II were performed resulting in polysilane PS1-II. No isolation and characterization of the polymer occurred before its precipitation in MeOH took place.
Work-up procedure I. Upon completion of the reaction time, the reaction mixture (PS1) was allowed to cool to room temperature. Toluene and other volatile compounds were removed under reduced pressure. Subsequent, the branched polysilane was extracted from the residue by dissolving in dry n-hexane (50 mL). The resulting cloudy solution was filtrated to remove unreacted sodium metal and precipitated sodium salts generated during the polymerization reaction. The solid residue was washed once again with n-hexane (50 mL), filtrated and combined with the polysilane containing n-hexane filtrate. Afterwards, the volume of the polysilane containing solution was reduced to a volume of 10 mL in vacuo. The residue was diluted with THF (10 mL). In order to quench unreacted Si–Cl groups, a solution of methyl magnesium bromide in diethyl ether (3 M) was added until a hydrolyzed aliquot of the reaction mixture tested a neutral pH value. After stirring for 1 h, a small amount of chlorotrimethylsilane was added as terminating agent for silyl anionic chain ends until a hydrolyzed aliquot of the reaction mixture tested a slightly acidic pH value after rigorous stirring (30 min). The slight excess of chlorotrimethylsilane was again quenched via the addition of an ethereal solution of MeMgBr to the stirred reaction mixture. All volatile compounds and solvents were removed under reduced pressure.
Work-up procedure II. Immediately after the removal of all volatile compounds and solvents under reduced pressure, the polysilane was dissolved in THF (10 mL) and added to stirred methanol (100 mL) at room temperature. Isolation of the resulting white precipitate from solution by filtration and drying in vacuo yielded polysilane PS1-II as white, hydrocarbon-soluble solid (176.5 mg, 1.11 mmol, 12%).
PS1-II. ¹H NMR (500 MHz, [D₈]THF): δ_H = 3.39 (br, Si–OCH₃), 0.04–0.58 ppm (br, Si–CH₃); ²⁹Si{¹H} INEPT NMR (99 MHz, [D₈]THF): δ_Si = 19.0 to 22.0 (br, Si–OCH₃), −10.0 to −15.0 (br, Si(CH₃)₃), −20 to −45 (br, Si(CH₃)₂), −60 to −85 ppm (br, Si(CH₃)); ATR-IR (neat): [small nu, Greek, tilde] = 2951 (s, ν_as(C–H)), 2891 (m, ν_s(C–H)), 1402 (m, δ_as(Si–CH₃)), 1243 (s, δ_s(Si–CH₃)), 1072 (sh, ν(Si–OCH₃)), 1019 cm⁻¹ (m, ν_as(Si–O–Si)); GPC: M_n = 1700 g mol⁻¹, PDI = 2.0; UV/Vis: λ_max(n-hexane) = 300 nm; PL: λ_max(THF) = 449 nm; TGA: T_o = 320 °C; elemental analysis: found: C > 34.81; H, 9.42; Si, 45.61. Calc. for [Me₅Si₃]_n: C, 37.67; H, 9.48; Si, 52.85%.

Details on polymerization entry PS2. Polymerization of PS2 was conducted according to polymerization entry PS1, but using a slightly modified version of work-up procedure I and II due to the sample taking of the non-precipitated polymer. Initial weights of the educts: sodium metal (260 mg, 11.29 mmol, 3 equiv.), 1,2,3-trichloro-1,1,2,3,3-pentamethyltrisilane (1) (1 g, 3.76 mmol, 1 equiv.) and toluene (8 mL). Upon completion of the reaction time and cooling the reaction mixture to room temperature, THF (8 mL) was added at first in order to be able to perform the following quenching reactions in an ethereal solution. Reactive Si–Cl groups of the polysilane were quenched initiating a Grignard reaction with an ethereal solution of MeMgBr (3 M). Silyl anionic chain ends were quenched using Me₃SiCl (see work-up procedure I for further details). Afterwards, all volatile compounds and solvents were removed under reduced pressure. The branched polysilane was extracted from the residue by dissolving in dry n-hexane (2 × 50 mL) and filtration of the cloudy solution. After evaporation of the solvent, the gained polymer was thoroughly dried in vacuo prior to characterization. The methyl-substituted highly branched polymer PS2-I, was obtained as viscous colorless oil.
PS2-I. ¹H NMR (500 MHz, [D₈]THF): δ_H = 0.03–0.49 ppm (br, Si–CH₃); ATR-IR (neat): [small nu, Greek, tilde] = 2950 (s, ν_as(C–H)), 2891 (m, ν_s(C–H)), 1403 (m, δ_as(Si–CH₃)), 1244 (s, δ_s(Si–CH₃)), 1030 cm⁻¹ (m, ν_as(Si–O–Si)); GPC: M_n = 1000 g mol⁻¹, PDI = 2.5.

Subsequent to work-up procedure I, the gained branched polysilane PS2-I was dissolved in THF (5 mL) and was added to stirred methanol (100 mL) at room temperature (work-up procedure II). The resulting white precipitate was isolated from solution by filtration and was thoroughly dried in vacuo. The methyl-substituted highly branched polymer PS2-II was obtained as white, hydrocarbon-soluble solid.

PS2-II. ¹H NMR (500 MHz, [D₈]THF): δ_H = 3.38 (br, Si–OCH₃), 0.07–0.55 ppm (br, Si–CH₃); ²⁹Si{¹H} INEPT NMR (99 MHz, [D₈]THF): δ_Si = 22.0 (br, Si–OCH₃), −10.0 to −15.0 (br, Si(CH₃)₃), −20 to −45 (br, Si(CH₃)₂), −60 to −85 ppm (br, Si(CH₃)); ATR-IR (neat): [small nu, Greek, tilde] = 2951 (s, ν_as(C–H)), 2892 (m, ν_s(C–H)), 1403 (m, δ_as(Si–CH₃)), 1244 (s, δ_s(Si–CH₃)), 1080 (sh, ν(Si–OCH₃)), 1025 cm⁻¹ (m, ν_as(Si–O–Si)); GPC: M_n = 2000 g mol⁻¹, PDI = 2.1, UV/Vis: λ_max(n-hexane) = 302 nm; PL: λ_max(THF) = 449 nm.

Thermal treatment of PS1-II

Thermal rearrangement of PS1-II to polycarbosilane PS1–PCS-330 and PS1–PCS-365. Two samples of polysilane PS1-II (each ∼15 mg) were sealed in glass ampoules under vacuum and were separately transferred in a screwed steel ampoule (burst protection) into a compact tube furnace. The respective samples were heated from ambient to either 330 °C or 365 °C using a heating rate of 20 K min⁻¹ and maintained at the operating temperature for 45 min. The products PS1–PCS-330 (yellowish sticky substance) and PS1–PCS-365 (yellowish-brown solid) were analyzed by means of NMR, IR and UV/Vis spectroscopy.
PS1–PCS-330. ¹H NMR (500 MHz, C₆D₆): δ_H = 4.41 (br, Si–H), 0.06–0.85 ppm (br, Si–CH₃); ATR-IR (neat): [small nu, Greek, tilde] = 2950 (s, ν_as(C–H)), 2891 (m, ν_s(C–H)), 2092 (s, ν(Si–H)), 1404 (m, δ_as(Si–CH₃)), 1352 (vw, δ(Si–CH₂)), 1245 (s, δ_s(Si–CH₃)), 1021 cm⁻¹ (s, ω(Si–CH₂–Si)).
PS1–PCS-365. ¹H NMR (500 MHz, C₆D₆): δ_H = 4.41 (br, Si–H), 0.06–0.83 ppm (br, Si–CH₃); ATR-IR (neat): [small nu, Greek, tilde] = 2953 (m, ν_as(C–H)), 2892 (w, ν_s(C–H)), 2098 (s, ν(Si–H)), 1406 (w, δ_as(Si–CH₃)), 1354 (vw, δ(Si–CH₂)), 1248 (s, δ_s(Si–CH₃)), 1018 cm⁻¹ (s, ω(Si–CH₂–Si)).

Pyrolysis of PS1-II to silicon carbide PS1–SiC-1200. For the performance of pyrolysis experiments, polysilane PS1-II (∼25 mg) was placed in a quartz tube as thin wall coating by dissolving in n-hexane, evaporation of the solvent over several hours and sealing the tube under vacuum. The heating program for the pyrolysis of polysilane PS1-II to PS1–SiC-1200 comprised several temperature levels and holding times:^27,52

Special care must be taken on opening the quartz tubes as volatile substances were generated during pyrolysis and pressurize the tube. For this reason a sufficiently large volume of the sealed quartz tube should be chosen. The black product PS1–SiC-1200 (observed residue: 5.6 mg; ceramic yield: 29.1%) was analyzed by means of IR spectroscopy, XRPD, HR-SEM-EDS analysis and elemental analysis.

PS1–SiC-1200. ATR-IR (neat): [small nu, Greek, tilde] = 825 cm⁻¹ (s, ν(Si–C)); elemental analysis: found: C, nd; Si, 68.6. Calc. for SiC: C, 29.96; Si, 70.04%.

Results and discussion

In order to pursue our goal to synthesize methyl-substituted, hydrocarbon-soluble highly branched polysilanes of high silicon content, we chose the oligomer 1,2,3-trichloro-1,1,2,3,3-pentamethyltrisilane (1) to be the monomer for the present polymerization study. Monomer 1 was prepared via the chlorodephenylation of the corresponding phenyl-substituted trisilane (PhMe₂Si)₂SiMePh using Friedel–Crafts acylation conditions: acetyl chloride and anhydrous AlCl₃ in n-hexane at 0 °C (Scheme 2).⁵⁰

Scheme 2 Synthesis of the trichloro-substituted monomer (ClMe₂Si)₂SiMeCl (1) via chlorodephenylation of (PhMe₂Si)₂SiMePh.⁵⁰

Polymerizations were conducted by subjecting the trichloro-substituted trisilane 1 to common Wurtz-type reductive coupling reaction conditions by varying the kind of alkali metal (sodium or lithium), the sort of solvent (toluene or THF), as well as the reaction temperature (reflux or r.t.), the reaction time (4 h or 7 d) and the work-up procedure (quenching with methyl-magnesium bromide (MeMgBr) or MeLi/precipitation in MeOH) in order to find promising polymerization conditions. All reaction conditions and work-up procedures were derived from literature-known synthetic routes and were newly assigned to the trifunctional monomer 1.^{2,3,10,11,39,49,53}

Within this polymerization study it was found that reacting (ClMe₂Si)₂SiMeCl (1) with molten sodium in boiling toluene for 4 h and termination by the addition of MeMgBr/Me₃SiCl (work-up procedure I) and precipitation in MeOH (work-up procedure II) was one of the most promising reaction conditions applied with regard to satisfactory number-average molar mass (M_n) and relatively narrow molar mass distributions of the polysilanes gained (Scheme 3). In addition, the obtained polymeric products were soluble in common organic solvents at ambient temperature and therefore well manageable as desired.

Scheme 3 Reductive coupling of monomer 1 gaining a methyl-substituted highly branched polysilane.

In Table 1 only the parameters used and the results obtained of the most effective polymerization out of the performed polymerization study are summarized. PS1-II is the designation of the polysilane obtained after quenching with MeMgBr/Me₃SiCl (work-up procedure I) and precipitation in MeOH (work-up procedure II). In addition, polymerization entry PS2 is listed, since within this experiment samples were retrieved after work-up procedure I (PS2-I) and after work-up procedure II (PS2-II) for the purpose of characterization. Both batches were prepared using exactly the same polymerization conditions.

Table 1 Parameters and analytical results of polymerizing (ClMe₂Si)₂SiMeCl (1) resulting in polysilane PS1-II, PS2-I and PS2-II

Entry	Metal^a	Reaction		Work-up procedure	Mn^d/(g mol^−1)	PDI^d	Product
		Temp	Time
a [Monomer 1] : [alkali metal] = 1 : 3, solvent: toluene, concentration of monomer 1 in toluene: ∼0.5 M.b Work-up procedure I – quenching agent: MeMgBr (3 M in Et2O) and Me3SiCl.c Work-up procedure II – precipitation in MeOH.d Determined by GPC in THF versus linear polystyrene standards (sample concentration: 2 g L^−1).
PS1	Na	Reflux	4 h	I^b + II^c: PS1-II	1700	2.0	White solid
PS2	Na	Reflux	4 h	I: PS2-I	1000	2.5	Viscous colorless oil
				I + II: PS2-II	2000	2.1	White solid

---

However, for entry PS2 the work-up procedures were adjusted due to the sample taking before precipitation in MeOH (see the Experimental section for further details). As polysilanes PS1-II and PS2-II differ only slightly in their molecular masses, their chemical and physical properties should be very similar and therefore comparable.

The successive performance of the different work-up procedures had the following purposes: within work-up procedure I, the addition of an ethereal solution of MeMgBr led to the quenching of remaining Si–Cl bonds.^39,40,49 Silyl anionic chain ends were supplementary quenched via the addition of chlorotrimethylsilane. Subsequent precipitation of the polysilanes in MeOH (work-up procedure II) removed low molar mass linear oligosilanes and possibly formed small cyclic oligomers (M_n ≈ 500 g mol⁻¹), as can be clearly seen from the GPC elution profiles of PS2-I and PS2-II (Fig. 1).^{3,38,39,41,53}

Fig. 1 Molar mass distribution of PS2 before (PS2-I, dashed line) and after (PS2-II, solid line) precipitation in MeOH. Molar masses are relative to polystyrene standards.

After the precipitation in MeOH, polysilanes of medium molar mass with M_n in the range of 1700 to 2000 g mol⁻¹ (≈30 to 35 Si atoms within in the branched polysilanes chains) and with relatively low polydispersity indices (PDI ≈ 2.0) were gained (Table 1, Fig. 1 and S1 (ESI†)). A narrow molar mass distribution is important to ensure the structure property relationship of the polymer. Moreover, the average molar mass and the PDIs are in the same order of magnitude as those published for comparable network polysilanes^49,54 or copolymers^41,46,49,55 synthesized via the polymerization of trichloro(organo)silanes R¹SiCl₃ (R¹ = alkyl or aryl) and dichloro(diorgano)silanes R²R³SiCl₂ (R²/R³ = alkyl or aryl). However, the real number-average molar mass of the polysilanes PS1-II, PS2-I and PS2-II may be higher than the values of M_n obtained from GPC, since the branched polysilanes feature a smaller hydrodynamic volume in comparison to the linear polystyrene standards, which were used for calibration of the GPC system.⁵⁶ In contrast to many references, precipitation of the polymers in water was not performed to avoid crosslinking of the polysilane chains via Si–O–Si bonds and the resulting falsification of the molar masses.^1,38–40,57

In order to unravel the structure of the gained polymers, one-dimensional ²⁹Si{¹H} INEPT NMR spectroscopy was conducted first. However, this method turned out to be unsuitable due to the broadening of the polymer signals (see Fig. S2, ESI†). Therefore, the microstructure was analyzed by ¹H/²⁹Si{¹H} HMBC NMR spectroscopy. This 2D NMR method detects silicon hydrogen couplings along several bonds (of about two to four).¹ However, the assignment of single Si atoms of the silicon backbone is not possible due to the high amount of detectable silicon hydrogen long-range couplings leading to overlapping resonance signals. It is only possible to assign individual building units of the polymeric species to a specific range of the chemical shifts within the spectra.

¹H/²⁹Si{¹H} HMBC NMR spectra of PS2 are illustrated in Fig. 2 before (A: PS2-I) and after (B: PS2-II) precipitation in MeOH. In general, three assignable main resonance peaks are visual within the 2D NMR spectra of polysilane PS2-I after quenching with MeMgBr and Me₃SiCl (Fig. 2A). These signals indicate the presence of trifunctional MeSi(Si)₃ branching units between −60 and −85 ppm, linear Me₂Si(Si)₂ units between −20 and −45 ppm and terminal Me₃Si–Si units around −10 ppm, as it is expected for branched polysilanes gained via the polymerization of (ClMe₂Si)₂SiMeCl (1).^1,58,59

Fig. 2 ¹H/²⁹Si{¹H} HMBC NMR (¹H: 500 MHz/²⁹Si: 99 MHz) spectra of PS2 (solvent: [D₈]THF). (A) NMR spectrum of PS2-I recorded after performing work-up procedure I (quenching with MeMgBr and Me₃SiCl). (B) NMR spectrum of PS2-II recorded after quenching with MeMgBr and Me₃SiCl and subsequent precipitation in MeOH (work-up procedure I + II). For the colored version of this figure see Fig. S3 in the ESI.†

The comparison of the 2D NMR spectra of polysilane PS2 before (PS2-I) and after (PS2-II) precipitation in MeOH clearly shows, that not all Si–Cl chlorine groups were quenched through the addition of MeMgBr within work-up procedure I.

As new peaks around 20 ppm appear within the ²⁹Si NMR spectra of polymer PS2-II, the formation of methoxy-substituted end groups, like (MeO)Me₂Si–Si or (MeO)MeSi(Si)₂, must have taken place (Fig. 2B). Thereby, the cross peaks at 3.41 ppm and 0.35 ppm on the ¹H axis of the ¹H/²⁹Si{¹H} HMBC NMR arise from the long-range coupling of methoxy hydrogen atoms to proximate silicon atoms and confirm the presence of Si–OMe groups.^58,60–63 The 2D NMR spectrum of polysilane PS1-II shows chemical shifts similar to those of PS2-II and is presented within the ESI (Fig. S4†). Altogether, the range of the chemical shifts within the ¹H/²⁹Si{¹H} HMBC NMR spectra suggest the successful formation of highly branched polysilanes gained via the polymerization of the trifunctional trisilane 1.

ATR-IR spectroscopy supports this assumption, since the positions of the infrared bands within the spectra for polysilane PS2-I correlate well with those expected for the characteristic organosilicon groups of the methyl-substituted polysilanes (Fig. 3).^64,65 Several intense signals around 2950, 2891, 1403 and 1244 cm⁻¹ occur within these IR spectra in the range of 3500 and 1000 cm⁻¹ and can be attributed to different kinds of Si–CH₃ vibrations, as they are reported for lower polysilanes as well (see Fig. 3 for structural assignments).^10,64–66 Thereby, it is irrelevant if the vibrations arise from MeSi(Si)₃, Me₂Si(Si)₂ or Me₃Si–Si structural units. Beyond, infrared bands in the region of 1030 cm⁻¹ are always present and indicate the existence of Si–O–Si bonds within the polysilane polymers, albeit not many since the intensity of the Si–O–Si band is very low compared to the strong absorption band of pure siloxanes compounds.^10,64,65 As those polysilanes, which were only quenched with MeMgBr and not precipitated in MeOH (PS2-I), still contain remaining Si–Cl bonds and the IR spectra were measured under exposure to air, the small degree of oxidation can easily be explained. Certainly, oxidation could also have occurred during synthesis or polymer isolation.

Fig. 3 ATR-IR spectra of the branched polymethylsilane PS2 after performing work-up procedure I (PS2-I) and after precipitation in MeOH (work-up procedure I + II, PS2-II). ν – stretching, δ – deformation, as – asymmetric, s – symmetric.

IR spectra of polysilanes precipitated in MeOH (PS1-II: Fig. S5 in the ESI† and PS2-II: Fig. 3) show two overlapping bands in the range of 1000 to 1100 cm⁻¹, of which the less intense one (shoulder) probably arises from the C–O stretching vibration mixed with contributions of Si–O stretching and confirms the additional existence of Si–OMe bonds.^64,65,67

Nevertheless, it can be assumed that during polymerization reactions long uninterrupted sequences of Si–Si bonds were predominantly formed, since polysilane PS1-II shows characteristic absorptions in the UV region ranging from 200 to 350 nm (see Fig. S6, ESI†). As the branched polysilane PS1-II is composed of linear Me₂Si(Si)₂ units and trifunctional MeSi(Si)₃ branching units, the UV/Vis spectra consist of several overlapping bands. Thereby, the long-wavelength UV absorption maximum λ_max in the 300 nm region is typical for the absorption of linear alkyl-substituted polysilanes originating from σ → σ* π-like transitions of the conjugated electrons within the Si–Si backbone.^1–3,68 The more intense and broad absorption UV band appearing at shorter wavelengths (≤250 nm) is, by contrast, characteristic of the electronic absorption of network or branched polysilanes [MeSi]_n. Its intensity decreases monotonically as the wavelength reaches the visible region.^{39–41,47,69}

Besides UV/Vis spectroscopy, solution PL spectroscopy of PS1-II verifies the formation of long uninterrupted sequences of Si–Si bonds within the polysilane, too (see Fig. S7, ESI†). A broad signal arises in the visible light region at 449 nm (λ_max) tailing down to 700 nm, which is characteristic of a branched conjugated Si–Si σ system extended into three dimensions. The broadness of the signal can as well be associated with the variety of polymer sizes comprised within the sample. In contrast, materials of low dimension, like linear polysilanes [R¹R²Si]_n (R¹/R² = alkyl or aryl), show sharp photoemissions (σ → σ* fluorescence) at shorter wavelength. In case of PS1-II, the fixed arrangement of linear and branching units due to the shape of the monomer prohibits the appearance of such sharp emission peaks at shorter wavelengths, as the length of the linear units is too little.^{44,45,47,48,69,70}

Altogether, NMR, IR, UV/Vis and PL spectroscopy of the polymerization entries PS1 and PS2 identify the respective products as branched methyl-substituted polysilanes, which in addition are soluble in common organic solvents.

After the performance of the common characterization methods of the branched methyl-substituted polysilane PS1-II and the examination of its optical and electronic properties, the thermal properties were investigated. To get an initial overview, TGA was conducted first (see Fig. S8, ESI†). The branched polysilane PS1-II is thermally stable up to temperatures of approximately 250 °C followed by a rapid weight loss starting at temperatures above 300 °C (extrapolated onset decomposition temperature T_o = 320 °C) and completed by 400 °C. During heat-treatment only one single mass-loss step appears. It is supposed that the Si–Si main chain is thermally cleaved (depolymerization) and the evaporation of volatile compounds (oligomers) leads to the almost complete decomposition of the polysilane and a weight residue of 4.76% at 800 °C.^28,71 The last named value does not reflect the potential ceramic yield of the precursor PS1-II, since during pyrolysis different conditions are applied. Thermal analysis of PS1-II by DSC does not show any characteristic transitions between −150 °C and 220 °C (see Fig. S9, ESI†).

As mentioned in the introduction part, polysilanes (e.g. insoluble polydimethylsilane) are suitable for the pyrolytic preparation of β-SiC at temperatures above 1000 °C in an inert atmosphere. It is thereby mandatory to apply PCS as intermediates to minimize weight loss. PCS are obtained by the thermal rearrangement of a polysilane at temperatures around 450 °C (Scheme 1).^27–29 For that reason we performed a study, in order to discover to what extent the hydrocarbon-soluble, highly branched polysilane PS1-II is initially suitable for the preparation of PCS and eventually of β-SiC.

In general, the branched polysilane PS1-II possesses a variety of advantages in contrast to the linear PDMS, which was used by Yajima et al. for the pyrolytic preparation of β-SiC.^3,28–31 Besides the already mentioned hydrocarbon-solubility of polymethylsilane PS1-II, its pyrolysis would lead to a theoretical SiC yield of 75.5 wt%. In contrast, the pyrolysis of Yajima's linear PDMS can only achieve a theoretical SiC yield of 69 wt%.⁷² Furthermore, the branching of the polymethylsilane PS1-II should minimize weight loss during thermal treatment. Therefore, the thermal rearrangement of polysilane PS1-II towards PCS structural units should no longer be mandatory. However, we initially tried to transfer the branched polymethylsilane PS1-II into PCS structural units in order to intensify the favorable nature of the branched polymer and to support the preforming of the Si–C network as well as in order to keep the conversion processes comparable.

The temperature, at which the rearrangement of the polysilane to the PCS possibly takes place, can be indirectly derived from the TG curve. In our case, considerable weight loss of the polysilane starts at 320 °C. Therefore, polysilane PS1-II was heat-treated from ambient to either 330 °C (PS1–PCS-330) or 365 °C (PS1–PCS-365). The respective products were analyzed by means of IR, UV/Vis and ¹H NMR spectroscopy in order to track the thermal rearrangement of the branched polysilane PS1-II to the assumed polycarbosilanes PS1–PCS-330 and PS1–PCS-365.

The IR spectra of PS1-II, PS1–PCS-330 and PS1–PCS-365 show several signals around 2951 (ν_as(C–H)), 2891 (ν_s(C–H)), 1402 (δ_as(Si–CH₃)) and 1243 cm⁻¹ (δ_s(Si–CH₃)), as expected for different kinds of Si–CH₃ vibrations of methyl-substituted polysilanes (Fig. 4).^10,64–66 By comparison of the IR spectra it is clearly visible that new absorption bands around 2100 cm⁻¹ and 1020 cm⁻¹ appear within the spectra of the heat-treated samples, indicating the presence of Si–H and Si–CH₂–Si structural units. Similar observations were reported by Yajima et al. during the rearrangement of the linear polydimethylsilane gaining PCS.^28,29 The observed absorption bands can be related to vibrations arising from stretching of Si–H and CH₂ wagging of the Si–CH₂–Si sequences formed.^28–30,64 The latter obviously obscures the IR absorptions originated from the excitation of Si–O–Si (1019 cm⁻¹) and Si–OMe (1072 cm⁻¹) bonds, but the disilylmethylene band is nevertheless sharp.^64,65,67 Additionally, a considerable increase of the absorption signal at 1020 cm⁻¹ can be observed with higher temperatures during heat treatment. Furthermore, a second distinctive band of the Si–CH₂–Si structural unit appears at 1352 cm⁻¹ which can be attributed to the CH₂ deformation vibrations.^30,64,73 Consequently, IR spectroscopy shows that thermolysis of PS1-II led to the rearrangement yielding PCS structural units with newly formed Si–CH₂ and Si–H bonds.

Fig. 4 Comparison of the ATR-IR spectra of the methyl-substituted branched polysilane PS1-II and of the polycarbosilanes PS1–PCS-330 and PS1–PCS-365. ν – stretching, δ – deformation, ω – wagging.

The transformation to PCS is further confirmed by UV/Vis spectroscopy since with proceeding thermal rearrangement the σ → σ* absorptions induced by linear Me₂Si(Si)₂ units and trifunctional MeSi(Si)₃ branching units decrease (Fig. 5). During thermal treatment an increasing number of Si–CH₂–Si units develop which do not possess the ability of UV absorption and are therefore UV/Vis inactive (compare dotted line).

Fig. 5 Comparison of the UV/Vis absorption spectrum of polysilane PS1-II (solid line) with those of the polycarbosilanes PS1–PCS-330 (dashed line) and PS1–PCS-365 (dotted line).

Within the ¹H NMR spectra of both PS1–PCS-330 and PS1–PCS-365 the distinct Si–H bond of the PCS can be detected at 4.41 ppm, confirming the successful rearrangement of the polysilane, too (see Fig. S10, ESI†). As mentioned above, the thermal rearrangement of PS1-II towards PCS structural units should not be mandatory in the present case. Therefore, the precise structure of PS1–PCS-330 and PS1–PCS-365 was not determined. However, it might include irregular cross-linking, rings and chains as already stated by Yajima et al. for PCS generated out of linear PDMS.²⁸

Based on these findings, we developed a heating program which comprises several temperature steps and holding times in order to enable a rearrangement from polysilane to PCS at first (see the Experimental section for further details). In the end the pyrolytic transformation to SiC should take place at 1200 °C and at best, crystallization to β-SiC.^27–29 The resulting pyrolytic residue PS1–SiC-1200 was all black and was analyzed by means of IR spectroscopy, XRPD, HR-SEM, EDS and elemental analysis.

Within the IR spectrum of PS1–SiC-1200 only a single major absorption band is visible, which is centered at 825 cm⁻¹ and can be attributed to the stretching vibration mode of the Si–C bond in the crystalline β-SiC phase (see Fig. S11, ESI†).^29,72 No additional absorption bands corresponding to vibrations of Si–CH_x (x = 2 or 3), Si–H or Si–O–Si bonds are present. Hence, conversion of PS1-II to an inorganic SiC network must have taken place. Presumably according to the suggested reactions occurring during the conversion of linear polydimethylsilane to SiC.^28–31 As the absorption band is rather broad, one can assume a compositional inhomogeneity of the sample. Absorption bands of amorphous SiC (a-SiC) would appear within the same range.³⁶

The morphology of the pyrolytic residue PS1–SiC-1200 was analyzed by means of XRPD (Fig. 6). The pattern indicates that besides crystalline products, amorphous substances must be existent due to the broadening of the background signal. The distinct diffraction peaks can be assigned to the (111), (220) and (311) reflections of cubic β-SiC. An extension of the annealing time of the sample at 1200 °C or the increase of the annealing temperature to a maximum of 1700 °C would have led to the sharpening of these peaks. In doing so, transition from a-SiC to crystalline β-SiC would have been completed and the fraction of perfect β-SiC crystallites would have been larger.^27,34,76 Furthermore, free graphite must be present, as the diffraction peak (002) can clearly be identified.⁷⁷ The latter is probably responsible for the black color of the pyrolysis residue.⁷² The amorphous phase might consist of a-SiC and of free amorphous carbon. The amount of free carbon could be evolved from the β-SiC phase with further heating under vacuum.^77,78

Fig. 6 X-ray (CuK_α1) diffraction pattern of PS1–SiC-1200 (black line). For the purpose of comparison, calculated XRPD patterns of β-SiC⁷⁴ (dark gray bars) and graphite⁷⁵ (light gray bars) are depicted as well.

The microstructure and the detailed composition of the heat-treated polymer sample PS1–SiC-1200 was characterized by HR-SEM and EDS analysis (Fig. 7 and 8). The HR-SEM images of a fragment of sample PS1–SiC-1200 show that the sample is homogeneous inside covered with a porous layer on top.

Fig. 7 HR-SEM image (magnification: ×30 000, 1 kV) of PS1–SiC-1200.

Fig. 8 HR-SEM image (magnification: ×20 000, 1 kV) and EDS analysis of PS1–SiC-1200.

EDS analysis reveals a high silicon content of around 69 wt% within the center of the sample (Fig. 8, area 1 and 2). This result is in good agreement with the expected theoretical values for pure SiC: Si 70 wt%, C 30 wt%. However, 3.6 wt% of oxygen were detected as well. This result is not surprising since Si–O bonds were detected within polysilane PS1-II by NMR and IR spectroscopy. According to the published ongoing processes during pyrolysis of polysilanes, Si–O moieties are not removed until temperatures above 1200 °C are reached.^31,78,79 The closer the measurements are conducted to the surface of the sample (Fig. 8, area 3 and 4), the higher is the content of carbon. The top layer of the ceramic consists of about 64 wt% carbon. It should be added that area 3 was measured on an area nearby the surface where a piece of the top layer was splintered due to the power of the electron beam. The oxygen content remains almost constant over the different regions of the ceramic.

Comparing the results of the XRPD and EDS measurements, it can be assumed that β-SiC, together with a-SiC, forms the main part of the particle with an additional top layer predominantly consisting of graphite or amorphous carbon. Individual crystals of β-SiC could not be detected within the matrix of the particle. The particle size of β-SiC crystallites at an annealing temperature of 1200 °C is expected to be in the range of a few nanometers.^{27,31,77–79}

The elemental analysis of PS1–SiC-1200 confirms the EDS results, as it offers a value of 68.6 wt% for the average Si content of the sample. Assuming a stoichiometric composition of the SiC within the pyrolytic residue and complete retention of all Si atoms of the polymer, the ceramic yield can be determined to be 29.1% of the theoretical SiC yield.⁷² Generally it was found, that the value of the average silicon content of PS1–SiC-1200 (68.6 wt%) is considerably higher than the silicon content gained for the SiC fibres manufactured from linear PDMS by Yajima (50.5 wt%).^80,81 Hence, the stoichiometric SiC content of PS1–SiC-1200 is likewise higher than in Yajima's SiC. Furthermore, the excess free carbon content of PS1–SiC-1200 is comparatively low, since excess free carbon and Si–O moieties comprise approximately 2 wt% of the pyrolytic residue PS1–SiC-1200 (assuming a stoichiometric composition of SiC and complete removal of hydrogen due to IR spectroscopic data). In contrast, the pyrolytic residue of PDMS exhibits levels of excess free carbon within the range of 15–20 wt%.^29,72,80,81

In summary, the branched polymethylsilane PS1-II shows a great potential to be applied as soluble precursor for the preparation of SiC. The branched structure of the polymethylsilane benefits a high SiC content and a low excess of free carbon in comparison to the linear structure of PDMS. However, in order to achieve maximum ceramic yields and best results in terms of crystal growth and low excess of free carbon via pyrolysis of PS1-II, the pyrolysis program would have to be further optimized.

Conclusions

The trifunctional trisilane (ClMe₂Si)₂SiMeCl (1) was successfully converted into a hydrocarbon-soluble, highly branched polymethylsilane using common Wurtz-type reductive coupling reaction conditions (dispersed sodium in boiling toluene). It is advantageous to choose this kind of monomer, as the tertiary silicon unit is already inherent and the formation of highly branched polysilanes with a regular substitution pattern is eased. The introduction of the linear silicon units assures the solubility of the branched polysilanes in common organic solvents, despite methyl-substitution.

Polymerization reactions were terminated after 4 h by the addition of MeMgBr and precipitation in MeOH yielding polymethylsilanes with Me₃Si–Si, (MeO)Me₂Si–Si and/or (MeO)MeSi(Si)₂ terminal units. Polymers with M_n in the range of 1700 to 2000 g mol⁻¹ and a PDI of approximately 2.0 were obtained. ¹H/²⁹Si{¹H} HMBC NMR spectra confirm the existence of trifunctional MeSi(Si)₃ branching units, linear Me₂Si(Si)₂ units and terminal Me₃Si–Si units. Methoxy-functionalized end groups were identified by NMR and IR spectroscopy. However, IR spectroscopy revealed the presence of Si–O–Si bonds, which were probably generated by oxidation reactions occurring during synthesis, polymer isolation or analysis. The formation of long uninterrupted sequences of Si–Si bonds within the polysilane was verified by the distinctive shape of its UV/Vis absorption spectra and the characteristic emission wavelength performing solution PL spectroscopy.

TGA revealed that the synthesized hydrocarbon-soluble, highly branched polysilane PS1-II is stable up to 320 °C and is therefore suitable as a soluble precursor for the preparation of SiC. In order to intensify the favorable nature of the branched polysilane and to support the preforming of the Si–C network, PS1-II was first rearranged at around 350 °C to gain PCS. IR, UV/Vis and ¹H NMR spectroscopic data indicate the successful rearrangement of polysilane PS1-II to the polycarbosilanes PS1–PCS-330 and PS1–PCS-365 at temperatures below 400 °C. A specifically edited pyrolysis program was developed to enable the initial formation of PCS, before the formation of SiC at 1200 °C under vacuum takes place. Results of IR spectroscopy, XRPD measurements, HR-SEM and EDS analysis confirm the successful conversion of polysilane PS1-II to SiC. Crystallized β-SiC mixed with partly a-SiC (ceramic yield: 29.1%) was formed in the present work (PS1–SiC-1200). Oxygen (continuously distributed within the sample) and a small amount of excess free carbon (graphite and amorphous modification) were also available. The latter appears to be located as top layer on the particles, as HR-SEM and EDS analysis demonstrated. The pyrolysis program might be optimized in order to achieve maximum ceramic yields and best results in terms of crystal growth and low excess of free carbon via the pyrolysis of PS1-II.

Acknowledgements

The authors thank Wacker Chemie AG for financial support. K. D. gratefully acknowledges Katia Rodewald for recording HR-SEM images and for her patience analyzing my samples with EDS. Carina Dressel (TUM, Chair of Inorganic Chemistry with Focus on Novel Materials) is acknowledged for her help performing the XRPD measurements. Ulrike Ammari, Petra Ankenbauer and Bircan Dilki of the Chair of Inorganic Chemistry (TUM) are thanked for performing the elemental analyses.

References

1. M. A. Brook, Silicon in Organic, Organometallic and Polymer Chemistry, Wiley, New York, 2000.
2. R. West, J. Organomet. Chem., 1986, 300, 327–346.
3. R. D. Miller and J. Michl, Chem. Rev., 1989, 89, 1359–1410.
4. R. West, in The Chemistry of Organic Silicone Compounds, ed. S. Patai and Z. Rappoport, Wiley, Chichester, UK, 1989, vol. 2, pp. 1207–1240.
5. K. Matyjaszewski, M. Cypryk, H. Frey, J. Hrkach, H. K. Kim, M. Moeller, K. Ruehl and M. White, J. Macromol. Sci., Chem., 1991, 28, 1151–1176.
6. F. S. Kipping and J. E. Sands, J. Chem. Soc., Trans., 1921, 119, 830–847.
7. F. S. Kipping, J. Chem. Soc., Trans., 1924, 125, 2291–2297.
8. W. G. Boberski and A. L. Allred, J. Organomet. Chem., 1974, 71, C27–C28.
9. W. G. Boberski and A. L. Allred, J. Organomet. Chem., 1975, 88, 65–72.
10. J. P. Wesson and T. C. Williams, J. Polym. Sci., Polym. Chem. Ed., 1979, 17, 2833–2843.
11. R. E. Trujillo, J. Organomet. Chem., 1980, 198, C27–C28.
12. J. P. Wesson and T. C. Williams, J. Polym. Sci., Polym. Chem. Ed., 1980, 18, 959–965.
13. P. Trefonas, P. I. Djurovich, X.-H. Zhang, R. West, R. D. Miller and D. Hofer, J. Polym. Sci., Polym. Lett. Ed., 1983, 21, 819–822.
14. R. D. Miller, D. Hofer, J. Rabolt and G. N. Fickes, J. Am. Chem. Soc., 1985, 107, 2172–2174.
15. J. F. Rabolt, D. Hofer, R. D. Miller and G. N. Fickes, Macromolecules, 1986, 19, 611–616.
16. J. M. Zeigler, Polym. Prepr., 1986, 27, 109–110.
17. K. S. Mazdiyasni, R. West and L. D. David, J. Am. Ceram. Soc., 1978, 61, 504–508.
18. R. West, L. D. David, P. I. Djurovich, K. L. Stearley, K. S. V. Srinivasan and H. Yu, J. Am. Chem. Soc., 1981, 103, 7352–7354.
19. A. R. Wolff, I. Nozue, J. Maxka and R. West, J. Polym. Sci., Part A: Polym. Chem., 1988, 26, 701–712.
20. R. G. Jones and S. J. Holder, Polym. Int., 2006, 55, 711–718.
21. C. Krempner, Polymers, 2012, 4, 408–447.
22. A. Feigl, A. Bockholt, J. Weis and B. Rieger, in Advances in Polymer Science, Silicon Polymers, ed. A. M. Muzafarov, Springer, Berlin, Heidelberg, 2011, vol. 235, pp. 1–31.
23. R. D. Miller, Angew. Chem., Int. Ed., 1989, 28, 1733–1740.
24. A. R. Wolff and R. West, Appl. Organomet. Chem., 1987, 1, 7–14.
25. R. G. Kepler, J. M. Zeigler, L. A. Harrah and S. R. Kurtz, Phys. Rev. B: Condens. Matter Mater. Phys., 1987, 35, 2818–2822.
26. M. Stolka, H. J. Yuh, K. McGrane and D. M. Pai, J. Polym. Sci., Part A: Polym. Chem., 1987, 25, 823–827.
27. S. Yajima, J. Hayashi and M. Omori, Chem. Lett., 1975, 4, 931–934.
28. S. Yajima, Y. Hasegawa, J. Hayashi and M. Iimura, J. Mater. Sci., 1978, 13, 2569–2576.
29. Y. Hasegawa, M. Iimura and S. Yajima, J. Mater. Sci., 1980, 15, 720–728.
30. M. Birot, J.-P. Pillot and J. Dunoguès, Chem. Rev., 1995, 95, 1443–1477.
31. R. M. Laine and F. Babonneau, Chem. Mater., 1993, 5, 260–279.
32. K. Shiina and M. Kumada, J. Org. Chem., 1958, 23, 139.
33. I. M. T. Davidson, J. Organomet. Chem., 1988, 341, 255–265.
34. S. Yajima, K. Okamura, J. Hayashi and M. Omori, J. Am. Ceram. Soc., 1976, 59, 324–327.
35. A. F. Holleman, E. Wiberg and N. Wiberg, Lehrbuch der Anorganischen Chemie, de Gruyter, Berlin, New York, 2007, pp. 984–985.
36. J. Bullot and M. P. Schmidt, Phys. Status Solidi B, 1987, 143, 345–418.
37. W. Uhlig, J. Organomet. Chem., 1996, 516, 147–154.
38. C. A. Burkhard, J. Am. Chem. Soc., 1949, 71, 963–964.
39. P. A. Bianconi and T. W. Weidman, J. Am. Chem. Soc., 1988, 110, 2342–2344.
40. P. A. Bianconi, F. C. Schilling and T. W. Weidman, Macromolecules, 1989, 22, 1697–1704.
41. H. K. Kim and K. Matyjaszewski, Polym. Prepr., 1989, 30, 131–132.
42. R. West, R. Menescal, T. Asuke and C.-H. Yuan, Polym. Prepr., 1990, 31, 248–249.
43. W. Uhlig, Z. Naturforsch., B: J. Chem. Sci., 1995, 50, 1674–1678.
44. W. L. Wilson and T. W. Weidman, J. Phys. Chem., 1991, 95, 4568–4572.
45. J. Maxka, J. Chrusciel, M. Sasaki and K. Matyjaszewski, Macromol. Symp., 1994, 77, 79–92.
46. P. P. C. Sartoratto, C. U. Davanzo, I. Valéria and P. Yoshida, Eur. Polym. J., 1997, 33, 81–88.
47. A. Watanabe, H. Miike, Y. Tsutsumi and M. Matsuda, Macromolecules, 1993, 26, 2111–2116.
48. R. Richter, G. Roewer, U. Böhme, K. Busch, F. Babonneau, H. P. Martin and E. Müller, Appl. Organomet. Chem., 1997, 11, 71–106.
49. W. Uhlig, Z. Anorg. Allg. Chem., 1996, 51, 703–710.
50. K. Deller and B. Rieger, J. Organomet. Chem., 2015, 794, 188–196.
51. W. L. F. Armarego and C. L. L. Chai, Purification of Laboratory Chemicals, Butterworth-Heinemann, Oxford, U.K., 2003.
52. P. Czubarow, T. Sugimoto and D. Seyferth, Macromolecules, 1998, 31, 229–238.
53. R. G. Jones, U. Budnik, S. J. Holder and W. K. C. Wong, Macromolecules, 1996, 29, 8036–8046.
54. L. A. Vermeulen, K. Smith and J. Wang, Electrochim. Acta, 1999, 45, 1007–1014.
55. K. Matyjaszewski, J. Chrusciel, J. Maxka and M. Sasaki, J. Inorg. Organomet. Polym., 1995, 5, 261–279.
56. K. Matyjaszewski and H. Kim, Polym. Bull., 1989, 22, 253–259.
57. R. G. Jones, R. E. Benfield, R. H. Cragg, A. C. Swain and S. J. Webb, Macromolecules, 1993, 26, 4878–4887.
58. E. A. Williams, in The Chemistry of Organic Silicone Compounds, ed. S. Patai and Z. Rappoport, Wiley, Chichester, UK, 1989, vol. 1, pp. 511–554.
59. D. A. Stanislawski and R. West, J. Organomet. Chem., 1981, 204, 295–305.
60. Y. Nakadaira, T. Kobayashi, T. Otsuka and H. Sakurai, J. Am. Chem. Soc., 1979, 101, 486–487.
61. T. Y. Gu and W. P. Weber, J. Organomet. Chem., 1980, 195, 29–38.
62. G. Engelhardt, R. Radeglia, H. Kelling and R. Stendel, J. Organomet. Chem., 1981, 212, 51–58.
63. U. Herzog, N. Schulze, K. Trommer and G. Roewer, J. Organomet. Chem., 1997, 547, 133–139.
64. A. L. Smith, Spectrochim. Acta, 1960, 16, 87–105.
65. T. R. Crompton, in The Chemistry of Organic Silicone Compounds, ed. S. Patai and Z. Rappoport, Wiley, Chichester, UK, 1989, vol. 1, pp. 393–444.
66. U. G. Stolberg and H. P. Fritz, Z. Anorg. Allg. Chem., 1964, 330, 1–7.
67. T. F. Tenisheva, A. N. Lazarev and R. I. Uspenskaya, J. Mol. Struct., 1977, 37, 173–186.
68. P. Trefonas, R. West, R. D. Miller and D. Hofer, J. Polym. Sci., Polym. Lett. Ed., 1983, 21, 823–829.
69. K. Furukawa, M. Fujino and N. Matsumoto, Macromolecules, 1990, 23, 3423–3426.
70. A. Watanabe, J. Organomet. Chem., 2003, 685, 122–133.
71. Y.-T. Shieh and S. P. Sawan, J. Appl. Polym. Sci., 1995, 58, 2013–2024.
72. D. J. Carlsson, J. D. Cooney, S. Gauthier and D. J. Worsfold, J. Am. Ceram. Soc., 1990, 73, 237–241.
73. G. Fritz, W. Kemmerling, G. Sonntag, H. J. Becher, E. A. V. Ebsworth and J. Grobe, Z. Anorg. Allg. Chem., 1963, 321, 10–20.
74. J.-M. Bind, Mater. Res. Bull., 1978, 13, 91–96.
75. P. Trucano and R. Chen, Nature, 1975, 258, 136–137.
76. S. Yajima, J. Hayashi, M. Omori and K. Okamura, Nature, 1976, 261, 683–685.
77. G. D. Soraru, F. Babonneau and J. D. Mackenzie, J. Mater. Sci., 1990, 25, 3886–3893.
78. P. Colombo, T. E. Paulson and C. G. Pantano, J. Am. Ceram. Soc., 1997, 80, 2333–2340.
79. Y. Hasegawa and K. Okamura, J. Mater. Sci., 1983, 18, 3633–3648.
80. S. Yajima, K. Okamura, T. Matsuzawa, Y. Hasegawa and T. Shishido, Nature, 1979, 279, 706–707.
81. S. Yajima, Am. Ceram. Soc. Bull., 1983, 62, 893–903.

---

Footnote

† Electronic supplementary information (ESI) available: Analytical data of the polysilanes PS1-II, PS2-I and PS2-II (GPC, NMR, IR, UV/Vis, PL, TGA, DSC), of the polycarbosilanes PS1–PCS-330 and PS1–PCS-365 (¹H NMR) and of the pyrolytic residue PS1–SiC-1200 (IR). See DOI: 10.1039/c5ra19266h

---