Structural Evolution of Polymer-Derived Amorphous SiBCN Ceramics

Nov 10, 2011 - NanoScience Technology Center, Department of Chemistry and. ‡. Advanced Materials Processing and Analysis Center and Department of Me...
0 downloads 9 Views 934KB Size
ARTICLE pubs.acs.org/JPCC

Structural Evolution of Polymer-Derived Amorphous SiBCN Ceramics at High Temperature Sourangsu Sarkar,† Zhehong Gan,§ Linan An,*,‡ and Lei Zhai*,† †

NanoScience Technology Center, Department of Chemistry and ‡Advanced Materials Processing and Analysis Center and Department of Mechanical, Materials and Aerospace Engineering, University of Central Florida, Orlando, Florida 32826, United States § Center of Interdisciplinary Magnetic Resonance, National High Magnetic Field Laboratory, Tallahassee, Florida 32310, United States

bS Supporting Information ABSTRACT: Polymer-derived amorphous SiBCN ceramics are synthesized through a simple dehydrocoupling and hydroboration reaction of an oligosilazane containing amine and vinyl groups and BH3 3 Me2S, followed by pyrolysis. Two types of ceramics, denoted as Si2B1 and Si4B1, are produced from preceramic polymers with Si/B ratios of 2/1 and 4/1, respectively. The structural evolution of these ceramics with respect to the pyrolysis temperature and boron concentration is investigated using solid-state NMR, Raman, and EPR spectroscopy. Solid-state NMR suggests the presence of three major components in the ceramics: (i) hexagonal boron nitride (h-BN), (ii) turbostratic boron nitride (t-BN), and (iii) BN2 C groups. Increasing pyrolysis temperature leads to the transformation of BN2C groups into BN3 and “free” carbon. A thermodynamic model is proposed to explain such transformation. Raman spectroscopy measurements reveal that the concentration of the “free” carbon cluster decreases with increasing pyrolysis temperature, and Si4B1 contains more “free” carbon cluster than Si2B1. EPR studies reveal that the carbon (C)-dangling bond content also decreases with increasing pyrolysis temperature. It appears that the complete decomposition of the metastable BN2C groups to the BN3 groups and the “free” carbon affects the crystallization of SiBCN, which leads to Si4B1 ceramics crystallized at 1500 °C, whereas Si2B1 ceramics crystallized at 1600 °C.

’ INTRODUCTION Polymer-derived ceramics (PDCs) are a class of multifunctional ceramics synthesized by the thermal decomposition of polymeric precursors.13 The direct polymer-to-ceramic processing route of PDCs makes the materials very suitable for the fabrication of complex shaped components and devices with hybrid processing/ shaping technologies.49 Such process has led to significant technology breakthrough in ceramics researches including the development of ceramic fibers, coatings, composites, and MEMS (microelectromechanical systems). PDCs exhibited many advantageous properties over conventional polycrystalline ceramics. For example, PDCs exhibit excellent thermal stability,10 high-temperature mechanical properties,11 and oxidation/corrosion resistance.1215 PDCs also possess tunable semiconducting behavior and significant piezoresistivity up to ultrahigh temperatures,1619 allowing the fabrication of high-temperature microsensors. Among various PDCs, siliconboron carbonitride (SiBCN) ceramics have received special attention due to their exceptional high-temperature stability,10,2023 creep resistance,24 and electrical properties,25,26 granting them potential applications in coatings,27,28 microelectronics,28 membranes,29 and high-temperature stable ceramic fibers for composites.30,31 Understanding the structures and structural evolution of SiBCN ceramics at high r 2011 American Chemical Society

temperatures is essential to comprehend the structure/property relationship that holds the key to realize these applications. Previous studies on the structures and structural evolution of polymer-derived SiBCN suggest that the polymer-to-ceramic conversion is accompanied by the formation of several amorphous intermediates.32,33 The continuous transformation from preceramic polymers to amorphous ceramics takes place above 600 °C. At 1400 °C, amorphous SiBCN contains three major structural components: (i) amorphous (graphite-like) carbon clusters in nanoscale, called “free” carbon, (ii) boron nitride (BN) domains in the form of hexagonal BN, and (iii) an amorphous SiCN matrix.32,33 Further annealing of the SiBCN at higher temperatures leads to the formation of silicon carbide (SiC) nanocrystallites within the amorphous microstructure.34,35 It has also been suggested that boron atoms are embedded in turbostratic BCN layers within amorphous SiBCN ceramics, decreasing the grain boundary mobility and suppressing the crystallization.36 The turbostratic stacked BCN layers act as encapsulation layers that retard the decomposition of silicon nitride (Si3N4) to form Received: April 8, 2011 Revised: November 9, 2011 Published: November 10, 2011 24993

dx.doi.org/10.1021/jp203287h | J. Phys. Chem. C 2011, 115, 24993–25000

The Journal of Physical Chemistry C SiC nanocrystallites. Such function is attributed to the reduced activity of carbon in integrated BCN layers compared with “free” carbon37 and the development of covalent BN bonds that generates a rigid SiBCN network and causes the depression in crystallization of the materials.38 Thereby, the increase in boron content leads to the increased resistance to the crystallization of SiBCN ceramics.38 A number of theoretical models regarding Gibbs free energy of the amorphous SiCN domains and the completely crystallized states have been proposed to explain the crystallization behavior of SiBCN ceramics and the impact of the addition of boron on the crystallization behavior. Although unable to enunciate the crystallization of SiBCN ceramics completely, these thermodynamic models suggest that the addition of boron increases the driving energy required for the crystallization process.39,40 The models of metastable phase equilibria including the Gibbs energies of the amorphous SiCN phase and the nanocrystalline phases41 can predict the crystallization behavior in good agreement with the experimental results.42 However, the evolution of BN and possible “free” carbon formation during the amorphous to crystalline transformation of SiBCN ceramics has not been investigated. “Free” carbon formed during the pyrolysis of PDCs is an important structural feature that significantly influences the thermal stability of PDCs.43,44 Studying the generation and evolution of “free” carbon will shed light on the amorphous intermediates that closely relate to ceramic properties. In this Article, we report the systematic study on the structural evolution of polymer-derived amorphous SiBCN ceramics upon the pyrolysis at different temperatures using ceramic precursors with different Si/B ratios. Two ceramic precursors with Si/B ratios of 4/1 and 2/1 were synthesized according to the previously reported synthetic procedure.45 These ceramic precursors have cross-linked structures of covalently bonded borazine and linear polysilazane chains.45 SiBCN ceramics were obtained through the pyrolysis of these precursors, and their structures were examined using various approaches, including solid-state nuclear magnetic resonance (NMR), Raman spectroscopy and electron paramagnetic resonance (EPR). The investigation provides a new perspective on the amorphous microstructures of SiBCN ceramics. Whereas previous studies10,37 of the thermal stability of SiBCN ceramics suggested that the formation of SiC from the reaction of carbon with Si3N4 induced the crystallization at high temperature, our studies suggest that BN2C reacts with SiCN matrix to generate SiC and “free carbon”, which can further react with Si3N4 to form SiC, causing the transformation of the ceramics from amorphous to crystalline. In addition, the thermodynamic data derived from the structural evolution indicate that the conversion of BN2C to BN3 completes at lower temperature for ceramics with lower boron concentration. These results provide important information about the factors that contribute to the thermal stability of SiBCN ceramics.

’ EXPERIMENTAL METHODS Synthesis of SiBCN Preceramic Polymer Precursors. In a typical procedure, 3.0 g polysilazane (VL20, Kion, Charlotte, NC) (Figure S1 of the Supporting Information) was dissolved in 20 mL of toluene (Alfa Aesar, Ward Hill, MA) in a 100 mL twonecked round-bottomed flask equipped with rubber septa, a magnetic stirrer, and a reflux condenser. The flask was purged with N2 gas for 20 min. Borane dimethyl sulfide (BH3 3 Me2S, Aldrich, St. Louis, MO) was then added dropwise into the flask at room temperature through

ARTICLE

Scheme 1. Possible Chemical Structure of the Polymer Precursors

syringes. The mixture was stirred for 3 h at room temperature, then slowly heated to reflux and heated at reflux for 3 h. The solvent and Me2S was removed under vacuum. A white odorless solid was obtained and dried for overnight under vacuum. 1.0 (10.2 mmol) and 2.0 mL (20.4 mmol) of BH3 3 Me2S were used to produce precursors with Si/B ratios of 4/1 and 2/1, respectively. It is expected that borane reacts with the vinyl groups (hydroboration reaction forms BC bonds) and the amine groups (dehydrocoupling reaction forms BN bonds.) and lead to a cross-linked polymer. (Scheme 1; see the Results and Discussion). Pyrolysis. The ceramic samples were prepared by the pyrolysis of the obtained polymer powders in a quartz tube under a steady flow of argon (Ar) in a tube furnace. The following heating cycle was used: (i) an initial 3 °C/min ramp up to 900 °C, (ii) a 1 °C/min ramp to the desired pyrolysis temperatures, (iii) a 4 h dwelling time at the desired pyrolysis temperatures, and (iv) cooling of samples to room temperature at a cooling rate of 5 °C/min. Characterization. Solid polyborosilazane precursors were characterized by FTIR spectrometer (Perkin Elmer, Waltham, MA). The thermogravimetric analysis (TGA) was performed on a Q5000 thermogravimetric analyzer (Texas Instruments, Dallas, TX), at a heating rate of 3 °C/min to 1000 °C in an inert atmosphere of N2. The crystallization behavior was characterized by X-ray diffraction (XRD) analysis (XRD Rigaku, Tokyo, Japan) using a monochromatic Cu Kα radiation with a wavelength of λ/2 = 154.06 pm. The solid-state NMR experiments were carried out at 14.1 T (600 and 192.6 MHz Larmor frequencies for 1H and B11, respectively) using a Bruker Avance 600 spectrometer at the National High Magnetic Field Laboratory (NHMFL, Tallahassee, FL). A 4 mm magic-angle spinning (MAS) probe with a sample rotation frequency of 12.5 kHz was used for the measurement. 11B NMR spinecho was recorded at 96.2 kHz at a pulse length of 1.30 and 2.60 μs with a recycle delay of 15 s and a sample rotation frequency of 12.5 kHz. The spectra were calibrated relative to an aqueous solution of H3BO3 (δ 19.6) as an external standard and are given relative to BF3 3 OEt2 (δ 0). High-field, high-frequency EPR spectra were recorded at room temperatures on a home-built spectrometer at the EMR facility of the NHMFL. The instrument is a transmission-type device in which microwaves are propagated in cylindrical lightpipes. The microwaves are generated by a phase-locked Virginia Diodes source generating frequency of 13 ( 1 GHz and producing its harmonics of which the 2nd, 4th, 6th, 8th, 16th, 24th, and 32nd are available. A superconducting magnet (Oxford Instruments) capable of reaching a field of 17 T was employed. To characterize 24994

dx.doi.org/10.1021/jp203287h |J. Phys. Chem. C 2011, 115, 24993–25000

The Journal of Physical Chemistry C

ARTICLE

Figure 2. Thermogravimetric analysis (TGA) of Ceraset VL20 and SiBCN preceramic polymers (Si/B of ratios 4/1 and 2/1) showing that ceramic yield increases with increased B concentration. Figure 1. FTIR spectra of ceraset VL20 and SiBCN preceramic polymers.

the samples for Raman spectroscopic measurements, we ground the ceramic materials into fine powders using high-energy ball milling (8000M-115, Spex Certiprep Group, Metuchen, NJ) for 30 min. The obtained powders were then pressed into disks of 12.5 mm in diameter and 3 mm in thickness. The Raman spectra were obtained for the disks using Renishaw inVia Raman microscopy (Renishaw, Gloucestershire, U.K.). The excitation source used was the 532 nm line of a silicon-solid laser, and the size of the focused laser beam was ∼10 μm.

’ RESULTS AND DISCUSSION The preceramic precursors used in this study were synthesized using a commercially available cyclosilazane and BH3 3 Me2S. Previous studies demonstrated that the reaction of cyclic oligosilazane, (CH3SiHNH)n, with BH3 3 Me2S results in the evolution of hydrogen (H2) on amine groups, called dehydrocoupling reaction, and the formation of the cross-linked products that contain borazine rings with boron atoms attaching to three nitrogen atoms.45 These precursors produced SiBCN ceramics through a pyrolysis in an inert atmosphere. Similar dehydrocoupling strategy was utilized in our studies to synthesize the SiBCN preceramic polymers. Borane reacts with the vinyl groups (hydroboration reaction forms BC bonds) and the amine groups (dehydrocoupling reaction forms BN bonds) of VL20, generating colorless, glass-like powders after a complete removal of the solvent. The resultant preceramic polymers were analyzed by Fourier transform infrared (FTIR) spectroscopy (Figure 1). In VL20 FTIR spectrum, the bands related to Si-NHSi include NH stretching and deformation vibration at 3370 and 1171 cm1. A stretching vibration for SiN stretching vibration appears at 859 cm1. The bands ascribed to vinyl groups (CH2dCHSi) are CdC stretching vibration at 1582 cm1 and CH stretching vibrations at 3034 cm1. A strong absorption band assigned to SiH vibration appears at 2104 cm1. The characteristic band of Si-CH3 groups locates at 1250 cm1, coupled to CH stretching and methyl (CH3) group deformation vibrations at 2950 and 1398 cm1, respectively. The FTIR spectra of the SiBCN polymer precursors indicate the formation and disruption of various covalent bonds. The spectra indicate that NH stretching (3370 cm1) and deformation vibration (1171 cm1) decrease with a consequent increase in a broad absorption band in the range of 13301450 cm1 (assigned to BN vibration) overlapping with CH3 deformation vibration. This observation is in accordance with the dehydrocoupling reaction between amine (NH) groups of VL20 and BH functional groups of borane. The absorption band

Figure 3. X-ray diffraction (XRD) of (a) XRD of SiCN ceramics pyrolyzed at1425 °C, (b) Si4B1 ceramics pyrolyzed at 1500 °C, and (c) Si2B1 ceramics pyrolyzed at 1500 and 1600 °C.

at 1582 cm1 assigned to vinyl group decreases, suggesting that the hydroboration reaction between BH and CH2dCH proceeds simultaneously with the dehydrocoupling reaction. The absorption band at 2415 cm1 clearly confirms the presence of BH vibrations. These BH groups can act as active functional sites for further cross-linking reactions.29 The boron concentration of two samples was investigated by FTIR spectroscopy, where the peak area of BH absorption of each polymer sample was divided by the peak area of SiC of the same sample. The BH/SiC absorption peak area ratio of Si2B sample is about 1.6 to 2.2 times that of Si4B sample, suggesting that boron concentration in Si2B samples is about twice the boron concentration in Si4B sample. The conversion of preceramic polymer precursors into ceramics was investigated by TGA (Figure 2). Cross-linked 24995

dx.doi.org/10.1021/jp203287h |J. Phys. Chem. C 2011, 115, 24993–25000

The Journal of Physical Chemistry C

Figure 4. 11B magic-angle spinning (MAS) NMR spectra recorded for Si2B1 ceramics pyrolyzed at the temperature ranging from 1000 to 1600 °C.

VL20 shows four distinct stages of mass losses: (i) 0.9% between 25 and 300 °C, (ii) 5.4% between 300 and 500 °C, (iii) 10% between 500 and 800 °C, and (iv) 1.3% between 800 and 1000 °C. Previous results suggested that these losses are due to the evolution of methane (CH4) and H2 along with the evolution of low-molecular-weight oligomers.46 In contrast, Si2B1 ceramic precursor shows a distinctively four-step mass-loss process: (i) an initial mass loss of 2% due to the evolution of solvent and the oligomers, (ii) subsequent 3% mass loss between 300 to 500 °C probably due to the release of molecules containing two and four carbon units similar to the report by Hauser and coworkers,29 (iii) a major 7% mass loss step in the 500800 °C range by the release of CH4 and H2 as the decomposition products, and (iv) a final 1% mass loss due to the evolution of H2, leaving ∼87% of the original sample mass. Additionally, Si2B1 preceramic polymer gives a higher ceramic yield than Si4B1 preceramic polymer precursor (∼85%), mainly due to the less mass loss in the early stage of decomposition, suggesting that boron helped the cross-link of VL20. For investigation of the crystallization behavior of the SiBCN ceramics, polymer precursors were pyrolyzed under Ar and were characterized by XRD (Figure 3). As a comparison, VL20 was also pyrolyzed under similar conditions. It was found that VL20 forms SiCN-based amorphous ceramics at ∼1000 °C and crystallizes at 1425 °C (Figure 3a). In contrast, XRD spectrum reveals that Si4B1 ceramics become crystalline at 1500 °C (Figure 3b). Sharp peaks of β-SiC were observed at 2θ = 35, 50, and 72°, corresponding to the reflections from (111), (220), and (311) β-SiC.47 Si2B1 ceramics remain amorphous up to 1600 °C,

ARTICLE

with no noticeable crystalline features (Figure 3c). When pyrolyzed at 1650 °C, Si2B1 ceramics start to crystallize to form β-SiC crystallites. The observed improved resistance to crystallization by boron is consistent with previous studies.37,38 As previously mentioned, the pyrolysis of PDCs leads to different amorphous intermediates at different pyrolysis temperatures. Understanding the structural evolution holds a key to enlightening the relationship between the PDC structure and the thermal stability. The structural changes of the SiBCN ceramics pyrolyzed at the temperature ranging from 1000 to 1600 °C were examined by solid-state NMR spectroscopy. In this Article, 11B was chosen as the nucleus of interest because the coordination around boron provides important information about chemical reactions during the pyrolysis. Figure 4 shows the 11B NMR spectra of the Si2B1 ceramics obtained at different pyrolysis temperatures. It is seen that these spectra of the Si2B1 ceramic samples, except the sample pyrolyzed at 1600 °C, can be fitted with three Lorentzian peaks at δ 19.6, 24.2, and 30.4. The peaks at 19.6 and 24.2 ppm agree well with the reported chemical shifts of two BN3 forms, turbostratic boron nitride (t-BN) and hexagonal boron nitride (h-BN), respectively.48 BN3 groups are generated from the groups obtained through the dehydrocoupling reaction in preceramic polymers during the pyrolysis. The peak at 30.4 ppm can be attributed to the BN2C groups, resulting from the pyrolysis of the groups obtained through both dehydrocoupling reaction and hydroboration reaction in preceramic polymers. The spectra reveal the structural evolution from BN2C to BN3. As shown in Figure 4, the peak area ratio between BN3 and BN2C increases with increased pyrolysis temperature, suggesting that BC bonds in BN2C groups are not stable at higher temperatures and decompose into BN and “free” carbon until all BN2C groups are converted to BN3 groups and “free” carbon, as shown by the NMR spectrum of SiBCN ceramic obtained at 1600 °C. Similar results were obtained for Si4B1 ceramic samples except that BN2C disappears at a lower temperature of 1500 °C. It is interesting to note that the BN2C disappearance temperatures are coincident with those of crystallization temperatures. We propose the following reaction forming BN3, SiC, and the “free” carbon (C) to account for the conversion of BN2C to BN3 demonstrated in 11B MAS NMR spectra BN2 C þ SiCN S BN3 þ SiC þ C The equilibrium constant, K, for this reaction can be expressed as K ¼

½BN3 ½SiC½C ½BN2 C½SiCN

ð1Þ

where [BN3], [SiC], [C], [BN2C], and [SiCN] are the respective concentrations. On the basis of an assumption that [SiCN] is equal to unity and [BN3] = [SiC] = [C], eq 1 can be rearranged as K ¼

½BN3 3 ½BN2 C

ð2Þ

From van’t Hoff equation on the temperature dependence of the equilibrium constant ln K ¼ 

ΔG ΔH o ΔSo ¼  þ RT RT R

ð3Þ

where ΔH° and ΔS° are the standard enthalpy and entropy of the reaction, respectively, R is the universal gas constant, and T is the reaction temperature in Kelvin. Combining eqs 2 and 3, 24996

dx.doi.org/10.1021/jp203287h |J. Phys. Chem. C 2011, 115, 24993–25000

The Journal of Physical Chemistry C

ARTICLE

conversion of BN2C to BN3 should be more favored for Si4B1 than Si2B1, which is consistent with our results that BN2C disappeared at lower temperature for Si4B1 than Si2B1. Another effect of the BN2C to BN3 conversion reaction is the generation of temperature-dependent “free” carbon, the concentration of which is equal to that of BN3. These “free” carbons subsequently react with Si3N4 to form crystalline SiC and N2 according to the following reaction.37,49 Si3 N4 þ 3C f 3SiC þ 2N2 v

Figure 5. Plot of ln([BN3]3/[BN2C]) versus (1000/T) for SiBCN ceramic samples with Si/B ratios (a) 2/1 and (b) 4/1.

we obtain ln

½BN3 3 ΔH o ΔSo ¼  þ ½BN2 C RT R

ð4Þ

Equation 4 suggests that the ln[BN3]3/[BN2C] should exhibit a linear relationship with 1/T. Figure 5 plots ln([BN3]3/[BN2C]) versus 1/T for both Si2B1 and Si4B1. A well-defined straight line with a negative slope is obtained for both the ceramics, as predicted by eq 4. Such a linear relationship supports the proposed reaction pathway. The negative slope suggests that the reaction is endothermic. Endothermic reactions are favored by the increase in temperature, and hence the conversion of BN2C to BN3 proceeds to completion with the increasing of the temperature. Slope and intercept of the straight lines can be used to calculate ΔH° and ΔS° for the reaction ð5Þ ΔH o ¼  slopeR ΔSo ¼ interceptR

ð6Þ

The values of ΔH° and ΔS° for Si2B1 ceramics were found to be 38.06 kcal/mol and 30.08 cal/mol, respectively, whereas the ΔH° and ΔS° values for Si4B1 ceramics were calculated to be 45.84 kcal/mol and 35.90 cal/mol, respectively. From thermodynamics, the Gibbs free energy (ΔG°) of the reaction can be written as ΔGo ¼ ΔH o  ΔSo T

ð7Þ

Substituting the values of ΔH° and ΔS° in eq 7, the Gibbs free energy of the reaction for the two ceramics can be expressed as ΔGo2=1 ¼ 38060  30:08T and ΔGo4=1 ¼ 45840  35:90T

ð8Þ

Equation 8 reveals that at the pyrolysis temperature higher than 1000 °C, ΔG°4/1 is more negative than ΔG°2/1, suggesting a

Temperature-dependent “free” carbon formation is believed to be one of the causes of higher thermal stability of the PDCs with higher boron doping level. Therefore, it would be interesting to investigate the dependence of “free” carbon concentration on annealing temperature and boron doping level. In this Article, Raman spectroscopy, a standard nondestructive and one of the most sensitive tools for the characterization of different forms of carbon,5052 was used to study the structural change of carbon in SiBCN ceramics. Figure 6a,b shows the Raman spectra recorded from the Si2B1 and Si4B1 ceramic samples pyrolyzed at different temperatures. The Raman spectra of these two SiBCN ceramic samples have similar structural features. Two Raman bands were observed at ∼1350 and 1600 cm1, which are assigned to the D and G band, respectively. These two bands are the most striking features of disordered amorphous carbon. Also, the presence of these two peaks suggests the formation of significant fraction of carbon atoms bonded as ring-like configurations.53 The G band has E2g symmetry and is generated by the in-plane bond stretching of sp2 hybridized carbon atoms.54,55 This band is active at all sites and not necessarily limited to those arranged in six-fold symmetry. The D-band has A1g symmetry and is assigned to the breathing mode of aromatic rings.54,55 This band is forbidden in perfect graphitic structures and becomes actively only in the presence of local disorder. Therefore, the carbon atoms in disordered amorphous carbon are covalently bonded to form hexagonal ring-like configurations similar to graphite-like structures with defect sites resulting in the appearance of D-band in the Raman spectra. Further characterization of the obtained ceramics was carried out using EPR, which is a useful nondestructive tool to monitor the dangling bond. Figure 7a shows the room-temperature EPR spectra of Si2B1 ceramics pyrolyzed in the temperature ranging from 1000 to 1600 °C. The spectra show only one intense line, predominantly Lorentzian in shape, regardless of the pyrolysis temperature with a g value equal to 2.0032 ( 0.0001. This g factor suggests that the observed EPR line is due to the carbon dangling bonds (unpaired electrons) presented in the defects in the bulk SiBCN ceramic matrices.57 The difference of this g value from other g values (2.018, 2.02, and 2.012) that are corresponsive to ordered carbon networks such as planar graphite phase,56 turbostratic carbon,33 and multiwalled carbon nanotubes58 indicates that the synthesized SiBCN ceramics do not contain any ordered networks. In addition, the absence of hyperfine lines (boron or nitrogen dangling bonds) and hyperfine satellites (silicon dangling bonds) indicates that no dangling bonds are formed on boron, nitrogen, or silicon. Figure 7b plots the change in line width, that is, peak-to-peak width, with the pyrolysis temperature and indicates that the line width decreases initially up to 1350 °C and then increases with increasing temperature.52 This phenomenon is probably caused by the unresolved proton hyperfine couplings resulting from the loss of hydrogen.59 Similar 24997

dx.doi.org/10.1021/jp203287h |J. Phys. Chem. C 2011, 115, 24993–25000

The Journal of Physical Chemistry C

ARTICLE

Figure 6. Raman spectra of SiBCN ceramic samples pyrolyzed in the temperature range from 1100 to 1350 °C with Si/B-ratios of (a) 2/1 and (b) 4/1. Signal intensity decreases with increasing temperature and the intensity increases with the increase in Si/B-ratio from 2/1 to 4/1.

Figure 8. Plot of EPR integrated intensity per milligram (spin concentration) with pyrolysis temperature and with Si/B-ratios 2/1 and 4/1. Spin concentration decreases with increasing thermolysis temperature and increasing Si/B-ratio.

Figure 7. (a) Experimental EPR spectra of the SiBCN ceramics with Si/ B-ratio 2/1 pyrolyzed in the temperature range from 1000 to 1600 °C. (b) Line width variation of the same sample with pyrolysis temperature.

EPR signal features were also observed with Si4B1 ceramic samples. (See Figure S4 in the Supporting Information.) EPR integrated intensities were calculated from the EPR spectra as the product of the square of peak-to-peak width and peak-to-peak height. In the absence of peak fitting and the spin concentration of the reference sample materials, the integrated intensity values are an approximate and can only be used for estimating the relative differences in the paramagnetic concentration. It is seen that the integrated intensities per milligram of

the samples decrease with increasing pyrolysis temperature (Figure 8). This result seems controversial with the results of NMR and Raman that suggest that more carbon dissolves in SiCN matrix, which should lead to more C-dangling bonds. Such controversy could be caused by two possible factors: (i) the decrease in the concentration of carbon-dangling bond by the recombination of the paramagnetic centers or (ii) the consumption of the carbon paramagnetic centers by the electron-deficient boron atom forming metastable BC bond. The latter seems to be more likely because Figure 8 shows that the Si4B1 with less boron content has integrated intensity several folds of magnitude higher than Si2B1. It is observed that the integrated spin intensity becomes independent of the Si/B ratio with the pyrolysis temperature at T > 1300 °C. This could possibly mean that all boron atoms are involved in the bond formation through either BN or BC at a certain temperature, leading to the saturation of carbon dangling bonds.

’ CONCLUSIONS In summary, polymer-derived SiBCN ceramics with Si/B ratios of 2/1 and 4/1 were synthesized through dehydrocoupling and hydroboration reaction of VL20 with BH3 3 Me2S, followed by pyrolysis. The structural evolution of the pyrolyzed ceramics was examined by XRD, solid-state NMR, Raman, and EPR 24998

dx.doi.org/10.1021/jp203287h |J. Phys. Chem. C 2011, 115, 24993–25000

The Journal of Physical Chemistry C spectroscopy. Three major structural components, such as, h-BN, t-BN, and BN2C groups, were found in the amorphous matrix. Pyrolysis of the ceramic samples at higher temperatures shows the evolution of boron nitride (BN3) and “free” carbon with the disappearance of BN2C groups. A thermodynamic model involving equilibrium constants and Gibbs free energies is proposed to describe the transitions from BN2C to BN3 with a possible evolution of “free” carbon. Gibbs free energy calculations suggest that conversion of BN2C to BN3 completes at lower temperature in Si4B1 ceramic samples compared with Si2B1 ceramics. Qualitative Raman spectra measurements confirmed the presence of high concentration of “free” carbon in Si4B1 ceramics. Finally, the EPR spectra suggest the presence of carbon centered radical as the source for paramagnetic center and a possible formation of BC bonds, which further dissociates into stable BN bonds. The study on the relationship between the structure and the thermal stability of SiBCN ceramics and thermodynamic modeling of the structural evolution and characterization of the boron-rich phase could provide a useful guideline to design amorphous SiBCN ceramics.

’ ASSOCIATED CONTENT

bS

Supporting Information. Chemical structure of Ceraset VL 20, 11B MAS NMR spectra recorded for SiBCN ceramics (Si/B = 4/1) and experimental EPR spectra of the SiBCN ceramics with Si/B-ratio 4/1. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected] (L.Z.); [email protected] (L.A.).

’ ACKNOWLEDGMENT The financial support from National Science Foundation (DMR0706526) is gratefully acknowledged. The author acknowledges the use of the NMR and EPR facilities at the National High Magnetic Field Laboratory (NHMFL), which is supported by the NSF through the Cooperative Agreement no. DMR-0654118, the State of Florida and the DOE. We thank Dr. A. Ozarowski at the NHMFL for performing the high-field EPR measurements. ’ REFERENCES (1) Riedel, R.; Passing, G.; Schonfelder, H.; Brook, R. J. Nature 1992, 355, 714–717. (2) Kroke, E.; Li, Y. L.; Konetchny, C.; Lecomte, E.; Fasel, D.; Riedel, R. Mater. Sci. Eng., R. 2000, 26, 97–199. (3) Colombo, P.; Mera, G.; Riedel, R.; Soraru, G. D. J. Am. Ceram. Soc. 2010, 93, 1805–1837. (4) Liew, L.; Zhang, W.; An, L.; Shah, S.; Lou, R.; Liu, Y.; Cross, T.; Anseth, K.; Bright, V.; Raj, R. Am. Ceram. Soc. Bull. 2001, 80, 25–30. (5) Liu, Y.; Liew, L.; Lou, R.; An, L.; Bright, V. M.; Dunn, M. L.; Daily, J. W.; Raj, R. Sens. Actuators, A 2002, 95, 143–151. (6) Liew, L.; Liu, Y.; Luo, R.; Cross, T.; An, L.; V., M.; Dunn, M. L.; Daily, J. W.; Raj, R. Sens. Actuators, A 2002, 95, 120–134. (7) Liew, L.; Zhang, W.; Bright, V. M.; An, L.; Dunn, M. L.; Raj, R. Sens. Actuators, A 2001, 89, 64–70. (8) An, L.; Xu, W.; Rajagopalan, S.; Wang, C.; Wang, H.; Kapat, J.; Chow, L.; Fan, Y.; Zhang, L.; Jiang, D.; Guo, B.; Liang, J.; Vaidyanathan, R. Adv. Mater. 2004, 16, 2036–2040.

ARTICLE

(9) Yajima, S.; Hasegawa, Y.; Okamura, K.; Matsuzawa, T. Nature 1978, 273, 525–527. (10) Reidel, R.; Kienzle, A.; Dressler, W.; Ruwisch, L.; Bill, J.; Aldinger, F. Nature 1996, 382, 796–798. (11) An, L.; Riedel, R.; Konetachny, C.; Kleebe, H. J.; Raj, R. J. Am. Ceram. Soc. 1998, 81, 1349–1352. (12) An, L.; Wang, Y.; Bharadwaj, L.; Fan, Y.; Zhang, L.; Jiang, D.; Sohn, Y.; Desai, V.; Kapat, J.; Chow, L. Adv. Eng. Mater. 2004, 6, 337–340. (13) Bharadwaj, L.; Fan, Y.; Zhang, L.; Jiang, D.; An, L. J. Am. Ceram. Soc. 2004, 87, 483–486. (14) Wang, Y.; Fan, Y.; Zhang, L.; Zhang, W.; An, L. Scr. Mater. 2006, 55, 295–297. (15) Wang, Y.; Fei, W.; Fan, Y.; Zhang, L.; Zhang, W.; An, L. J. Mater. Res. 2006, 21, 1625–1628. (16) Wang, Y.; Zhang, L.; Xu, W.; Jiang, T.; Fan, Y.; Jiang, D.; An, L. J. Am. Ceram. Soc. 2008, 91, 3971–3975. (17) Wang, Y.; Jiang, T.; Zhang, L.; An, L. J. Am. Ceram. Soc. 2009, 92, 1603–1606. (18) Ryu, H. Y.; Wang, Q.; Raj, R. J. Am. Ceram. Soc. 2010, 93, 1668–1676. (19) Zhang, L.; Wang, Y.; Wei, Y.; Xu, W.; Fang, D.; Zhai, L.; Lin, K.; An, L. J. Am. Ceram. Soc. 2008, 91, 1346–1349. (20) Wideman, T.; Cortez, E.; Remsen, E. E.; Zank, G. A.; Carrol, P. J.; Sneddon, L. G. Chem. Mater. 1997, 9, 2218–2230. (21) Srivastava, D.; Duesler, E. N.; Paine, R. T. Eur. J. Inorg. Chem. 1998, 1998, 855–859. (22) Aldinger, F.; Weinmann, M.; Bill, J. Pure Appl. Chem. 1998, 70, 439–448. (23) Wang, Z. -C.; Aldinger, F.; Riedel, R. J. Am. Ceram. Soc. 2001, 84, 2179–2183. (24) Riedel, R.; Ruswisch, L. M.; An, L.; Raj, R. J. Am. Ceram. Soc. 1998, 81, 3341–3344. (25) Hermann, A. M.; Wang, Y.-T.; Ramakrishnan, P. A.; Balzar, D.; An, L.; Haluschka, C.; Riedel, R. J. Am. Ceram. Soc. 2001, 84, 2260–2264. (26) Ramakrishnan, P. A.; Wang, Y. -T.; Balzar, D.; An, L.; Haluschka, C.; Riedel, R.; Hermann, A. M. Appl. Phys. Lett. 2001, 78, 3076–3078. (27) Rooke, M. A.; Sherwood, P. M. A. Chem. Mater. 1997, 9, 285–296. (28) Hegemann, D.; Riedel, R.; Oehr, C. Chem. Vap. Deposition 1999, 5, 61–65. (29) Hauser, R.; Borchard, S. -N.; Riedel, R.; Ikuhara, Y. H.; Iwamoto, Y. J. Ceram. Soc. Jpn. 2006, 114, 524–528. (30) Lu, L.; Song, Y. C.; Feng, C. X. J. Mater. Sci. Lett. 1998, 17, 599–602. (31) Bernard, S.; Weinmann, M.; Gerstel, P.; Miele, P.; Aldinger, F. J. Mater. Chem. 2005, 15, 289–299. (32) Schuhmacher, J.; Berger, F.; Weinmann, M.; Bill, J.; Aldinger, F.; Muller, K. Appl. Organomet. Chem. 2001, 15, 809–819. (33) Berger, F.; Muller, A.; Aldinger, F.; Muller, K. Z. Anorg. Allg. Chem. 2005, 631, 355–363. (34) M€uller, A.; Zern, A.; Gerstel, P.; Bill, J.; Aldinger, F. J. Eur. Ceram. Soc. 2002, 22, 1631–1643. (35) M€uller, A.; Gerstel, P.; Weinmann, M.; Bill, J.; Aldinger, F. J. Eur. Ceram. Soc. 2000, 20, 2655–2659. (36) Jalowiecki, A.; Bill, J.; Aldinger, F.; Mayer, J. Composites, Part A 1996, 27, 717–721. (37) Weimann, M.; Schumacher, J.; Kummer, H.; prinz, S.; Peng, J.; Seifert, H. J.; Christ, M.; Muller, K.; Bill, J.; Aldinger, F. Chem. Mater. 2000, 12, 623–632. (38) Yamamoto, K.; Tsuganezawa, N.; Makimura, S.; Sawa, D.; Nakhigashi, S.; Kojima, K.; Hasegawa, Y. J. Mater. Res. 2008, 23, 1642–1646. (39) Golczewski, J. A.; Aldinger, F. J. Non-Cryst. Solids 2004, 347, 204–210. (40) Tavakoli, A. H.; Gerstel, P.; Golczewski, J. A.; Bill, J. J. NonCryst. Solids 2009, 355, 2381–2389. (41) Golczewski, J. A. J. Ceram. Soc. Jpn. 2006, 114, 950–957. 24999

dx.doi.org/10.1021/jp203287h |J. Phys. Chem. C 2011, 115, 24993–25000

The Journal of Physical Chemistry C

ARTICLE

(42) Tavakoli, A. H.; Gerstel, P.; Golczewski, J. A.; Bill, J. J. Am. Ceram. Soc. 2010, 93, 1470–1478. (43) Raj, R.; Riedel, R.; Soraru, G. D. J. Am. Ceram. Soc. 2001, 84, 2158–2159. (44) Varga, T.; Navrotsky, A.; Moats, J. L.; Morcos, R. M.; Poli, F.; Muller, K.; Saha, A.; Raj, R. J. Am. Ceram. Soc. 2007, 90, 3213–3219. (45) Seyferth, D.; Plenio, H. J. Am. Ceram. Soc. 1990, 73, 2131–2133. (46) Li, Y. -L.; Kroke, E.; Riedel, R.; Fasel, C.; Gervais, C.; Babonneau, F. Appl. Organomet. Chem. 2001, 15, 820–832. (47) Wang, Z. -C.; Kamphowe, T. W.; Katz, S.; Peng, J.; Seifert, H. J.; Bill, J.; Aldinger, F. J. Mater. Sci. Lett. 2000, 19, 1701–1704. (48) Marchetti, P. S.; Kwon, D.; Schmidt, W. R.; Interrante, L. V.; Maciel, G. E. Chem. Mater. 1991, 3, 482–486. (49) Bill, J.; Aldinger, F. Adv. Mater. 1995, 7, 775–787. (50) Wang, Y.; Alsmeyer, D. C.; McCreery, R. L. Chem. Mater. 1990, 2, 557–563. (51) Tunistra, F.; Koenig, J. L. J. Chem. Phys. 1970, 53, 1126–1130. (52) Trassl, S.; Motz, G.; Rossler, E.; Ziegler, G. J. Non-Cryst. Solids 2001, 293295, 261–262. (53) Janakiraman, N.; Aldinger, F. J. Eur. Ceram. Soc. 2009, 29, 163–173. (54) Gregori, G.; Kleebe, H. J.; Brrequel, H.; Enzo, S.; Ziegler, G. J. Non-Cryst. Solids 2005, 351, 1393–1402. (55) Ferrari, A. C.; Robertson, J. Phys. Rev. B: Condens. Matter 2000, 61, 14095–14107. (56) Jiang, T.; Wang, Y.; Wang, Y.; Orlovskaya, N.; An, L. J. Am. Ceram. Soc. 2009, 92, 2455–2458. (57) Andrenenko, S. I.; Stiharu, I.; Misra, S. K. J. Appl. Phys. 2006, 99, 113907. (58) Bardeleben, H. J.; Cantin, J. L.; Zeinert, A.; Racine, B.; Zellama, K.; Hai, P. N. Appl. Phys. Lett. 2001, 78, 2843–2845. (59) Jeschke, G.; Kroschel, M.; Jansen, M. J. Non-Cryst. Solids 1999, 260, 216–227.

25000

dx.doi.org/10.1021/jp203287h |J. Phys. Chem. C 2011, 115, 24993–25000