Int. J. Appl. Ceram. Technol., 8 [2] 467–481 (2011) DOI:10.1111/j.1744-7402.2009.02452.x
Ceramic Product Development and Commercialization
Preparation, Structural Characterization, and Thermal Ammonolysis of Two Novel Dimeric Transition Silylamide Complexes Fei Cheng, Christopher N. Hope, Stephen J. Archibald, John S. Bradley, Stephen Clark, M. Grazia Francesconi, Stephen M. Kelly,* and Nigel A. Young Department of Chemistry, The University of Hull, Hull HU6 7RX, United Kingdom
Fre´de´ric Lefebvre Laboratoire de Chimie Organome´tallique de Surface, 69616 Villeurbanne Cedex, France
The preparation and molecular structures of two novel dimeric transition metal silylamide complexes fLi0.5Zr[NHSi(NMe2)3]1.5[NSi(NMe2)3]0.5[m-NSi(NMe2)3]g2 and fLi0.5Hf[NHSi(NMe2)3]1.5[NSi(NMe2)3]0.5[m-NSi(NMe2)3]g2 with a tetrahedral coordination environment are reported. Thermal ammonolysis of fLi0.5Zr[NHSi(NMe2)3]1.5[NSi(NMe2)3]0.5 [m-NSi(NMe2)3]g2 in an autoclave yields a mesoporous partially lithiated silicon zirconium imide powder Si3Zr(N)(NH)x(NH2)y (NMe2)z with a surface area of 440 m2/g. A microporous partially lithiated silicon hafnium imide powder Si3Hf(N)(NH)x(NH2)y (NMe2)z with a surface area of 232 m2/g was obtained via a similar ammonolysis process of fLi0.5Hf[NHSi (NMe2)3]1.5[NSi(NMe2)3]0.5[m-NSi(NMe2)3]g2. Both of these silicon zirconium and hafnium imide powders have a disordered octahedral coordination environment. Pyrolysis of these zirconium and hafnium silicon imide powders leads to the formation of mixtures of porous zirconium or hafnium lithium silicon nitride ceramics with a regular octahedral coordination environment. They contain some residual lithium and exhibit a much reduced surface area due to an almost total collapse of the pores during the pyrolysis process.
Introduction This work was financially supported by Engineering and Physical Sciences Research Council (EPSRC). *
[email protected] r 2009 The American Ceramic Society
Microporous and mesoporous solids with a high effective surface area and a narrow pore-size distribution have attracted considerable attention as size-selective cata-
468
International Journal of Applied Ceramic Technology—Cheng, et al.
lysts, catalyst supports, and filters.1,2 The synthetic flexibility and low reaction temperatures of sol–gel techniques render this technique an attractive approach for preparing micro/mesoporous oxide materials as powders, membranes, and filters with high purity and homogeneity.3 A number of highly porous, nonoxide ceramics prepared by sol–gel processes have also been reported, that is boron nitride,4–8 metal (carbo)nitrides,9–14 silicon carbonitride,15–21 and boron carbonitride.22,23 Polymeric boron-, titano,- and tantalo-silazanes have been prepared by a sol– gel process by controlled coammonolysis of elemental alkylamides.24 The synthesis of B/C/N xerogels by the reaction of B-trichloroborazine with bis(trimethylsilyl)carbodiimide has also been reported.25 The authors have developed an efficient nonaqueous sol–gel process to prepare a high-surface-area porous silicon diimide gel by the acid-catalyzed ammonolysis of tris(dimethylamino)silylamine H2NSi(Me2N)3 (TDSA) in a manner similar to the synthesis of silica from orthosilicate esters.26 Pyrolysis of silicon diimide gel under a flowing atmosphere of ammonia at 10001C yields a mesoporous silicon nitride with a high surface area and a narrow pore size distribution.27 We have also reported the preparation of a microporous silicon imido nitride with a degree of pore-size control by pyrolysis of gels prepared via a hot ammonolysis of TDSA in a concentrated solution of n-alkyl amines.2 The advantageous properties, for example a very high melting point, good thermal conductivity and stability, good hydrothermal and mechanical stability, high resistance to corrosion and chemical inertness of porous silicon nitride and its multinary or metal-doped derivatives render them attractive as porous ceramics in high-temperature applications to inhibit sintering and coking due to hot spots. The basic nature of mesoporous silicon nitride and its derivatives induces a distinctly different physical property profile than that of the corresponding acidic silicates and may induce higher selectivity and activity as heterogeneous catalysts or catalyst supports in reactions where basicity promotes the reaction.28–31 Silicon nitride is a useful support for catalytically active phases, for example the dehydrogenation of propane using Pt/Si3N4,30 the selective hydrogenation of butadiene to butene using Pd/Si3N4,31 the oxidation of methane using Pd/Si3N4 and Pt/Si3N432–34 and in heterogeneous catalysis, for example Heck and Suzuki aryl–aryl coupling reactions (Cheng et al.,35 F. Cheng, unpublished results). Nitrogen-based porous catalysts may be superior to commercially available metal oxide catalysts for some gas-phase and liquid-phase reac-
Vol. 8, No. 2, 2011
tions. Porous nonoxide materials may also be of advantage in a wide range of applications, for example as selective gas filters for solid-state semiconductor gas sensors.36–38 Composite mesoporous ceramics consisting of two or more binary nitrides may exhibit properties superior to those of single-component nitrides.39 The most attractive approach to ternary mesoporous ceramic nitrides uses single-source molecular precursors possessing heterogeneous linkages [M–N–M0 ] due to the presence of an atomiclevel homogeneity of the ceramic components.40 The reaction intermediate TDSA and its lithium salt are versatile reaction intermediates due to reactive SiNH2 or SiNHLi groups for the preparation of a wide range of multinary silylamides, which contain, in addition to peripheral Si(NMe2)3 groups, Si–N–M backbones (M 5 Al, B, Ti, etc.). The exclusively MNn environment of the ceramogenic metal centers renders them attractive as single-source precursors to multinary silicon-based nitride ceramics. The presence of peripheral Si(NMe2)3 groups leads to multinary imidosilicate gels with high atomic-level homogeneity as precursors to micro- and mesoporous homogenous ternary nitride-based materials.41–48 In this paper, we report attempts to prepare two novel dimeric transition silylamide complexes, fZr[NHSi(NMe2)3]2 [m-NSi(NMe2)3]g2 and fHf[NHSi(NMe2)3]2[mNSi(NMe2)3]g2, their structural characterization using single-crystal X-ray crystallography, and their thermal ammonolysis to give a mesoporous silicon zirconium imide powder Si3Zr(N)(NH)x(NH2)y(NMe2)z and a microporous silicon hafnium imide powder. The structure of these porous, ternary, silicon- and nitrogen-based derivatives incorporating hafnium and zirconium is elucidated. Unexpectedly, it was found that these hafnium and zirconium silylamide complexes and imides also incorporate lithium in their structures. Structural data are provided primarily for the zirconium derivatives due to the very strong similarity in structure observed for each of the zirconium and hafnium materials. The partially lithiated silicon metal imide powders Si3M(N)(NH)x(NH2)y(NMe2)z (M 5 Zr and Hf) were then pyrolyzed to form crystalline ceramic materials incorporating silicon, nitrogen, lithium, and zirconium or hafnium, respectively, with an almost nonporous structure. Experimental Procedure General Comments All procedures were performed under an anhydrous nitrogen atmosphere using standard Schlenk techniques
www.ceramics.org/ACT
Preparation, Structural Characterization, and Thermal Ammonolysis of Silylamide Complexes
or in a nitrogen-filled glove box. The solvents, pentane and tetrahydrofuran (THF), were freshly distilled from sodium/benzophenone before use. Zirconium chloride (ZrCl4) and hafnium chloride (HfCl4) were obtained from Aldrich (St. Louis, MO). Liquified ammonia was from Energas (Hull, U.K.) Lithium tris(dimethylamino)silylamide LiHNSi(Me2N)3 was prepared according to a previously reported procedure.44,45 Fourier transform infrared spectra were recorded on a Nicolet (Madison, WI) Magna-500 FTIR spectrometer.49 29Si and 13 C cross-polarization (CP) NMR spectra with magicangle spinning (MAS) were obtained with a Bruker (Karlsruhe, Germany) DSX-300 spectrometer operating at frequencies of 59.6 and 75.5 MHz, respectively, with tetramethylsilane (TMS) as a reference. Nitrogen adsorption isotherms were obtained at 77 K using a Micromeretics (Norcross, GA) Tristar 3000 instrument and surface area was determined from BET analysis.
469
Ammonolysis
The ammonothermal reaction was carried out at 1601C in a stainless-steel Parr autoclave fitted with a pressure gauge. The autoclave was charged with either the zirconium precursor 2 (M 5 Zr) (3.2 g) or the hafnium precursor 2 (M 5 Hf) (2.9 g) in a glove box and then ammonia (6 mL) was condensed into the autoclave at 781C. The autoclave was sealed and heated to 1601C in an oil bath reaching a pressure of 1.5 MPa. After 24 h, the autoclave was cooled to room temperature, flushed with nitrogen for 0.5 h, and then evacuated under reduced pressure for 1 h to remove ammonia and the resultant dimethylamine. The partially lithiated yellow silicon zirconium imide solid Si3M(N)(NH)x(NH2)y(NMe2)z (3; M 5 Zr) (yield 1.4 g) and the white silicon hafnium imide solid Si3M(N)(NH)x(NH2)y(NMe2)z (3; M 5 Hf) (yield 1.65 g) were collected in a glove box. Pyrolysis
Synthesis of Lithiumbis(Dimethylamino)Silylamino-lbis[(Dimethylamino)Silylimino]-Zirconium fLi0.5M[NHSi(NMe2)3]1.5[NSi(NMe2)3]0.5[lNSi(NMe2)3]g 2 (2; M 5 Zr)
A solution of ZrCl4 (0.51 g, 2.27 mmol) in pentane (20 mL) was added to a suspension of LiHNSi(Me2N)3 (1.54 g, 8.69 mmol) in 40 mL of pentane. After 15 h of stirring at room temperature, the resulting precipitate of lithium chloride was filtered off and the filtrate was evaporated down to about half its initial volume and then cooled to 801C overnight to yield white crystals, which were filtered off under suction and dried under vaccuum (yield: 47%). IR (KBr, cm1): 3255 (w) (n(NH)); 2974 (s), 2875(s), 2833(s), 2783(s), (n(CH3)); 1187(s); and 986(s).
Synthesis of Lithiumbis(Dimethylamino)Silylamino-lbis[(Dimethylamino)Silylimino]-Hafnium fLi0.5M[NHSi(NMe2)3]1.5[NSi(NMe2)3]0.5 [l-NSi(NMe2)3]g 2 (2; M 5 Hf)
This was prepared as white crystals in a manner similar to 2 (M 5 Zr). Yield: 42%; IR (KBr, cm1): 3254 (w) (n(NH)); 2971 (s), 2872(s), 2831(s), 2780(s), (n(CH3)); 1180(s); and 982(s).
The silicon zirconium and hafnium imides 3 (M 5 Zr and Hf) were pyrolyzed in a tube furnace. About 0.5 g of 3 (M 5 Zr and Hf) was placed in an Al2O3 boat, which was then introduced into a quartz tube in a glove box. The ammonolyzed product was heated to 2001C at a ramp rate of 51C/min, held at 2001C for 2 h, and then heated at 10001C for 2 h under an NH3 flow. The brown Si–Zr–N ceramic solid 4 (M 5 Zr) and light-yellow Si–Hf–N solid 4 (M 5 Hf) were obtained by pyrolyzing the 3 (M 5 Zr and Hf) precursors, respectively, under an atmosphere of ammonia. X-Ray Crystallography
The diffraction data sets were collected on a Sto¨e IPDS-II imaging plate diffractometer using graphite monochromated MoKa radiation (l 5 0.71073 A˚). The crystals were kept at 150 K during data collection using the Oxford (Buckinghamshire, U.K.) Cryosystems 700 series cryostream cooler. Numerical absorption corrections were applied using X-RED. The structures were solved using direct methods (SHELXS 97) and refined by full-matrix least squares against F2 (SHELXL 97).50,51 All non-H atoms were refined anisotropically. H atoms were placed in idealized positions and refined using a riding model with C–H 5 0.97 A˚ and Uiso(H) 5 1.2 times Ueq of the carrier atom. The WinGX package was used for the refinement and production of data tables and ORTEP-3 was used for struc-
470
International Journal of Applied Ceramic Technology—Cheng, et al.
ture visualization.52,53 All ORTEP representations show ellipsoids at the 50% probability level. Crystallographic data for the structural analysis have been deposited with the Cambridge Crystallographic Centre, CCDC nos. 685268 for compound 2 (M 5 Zr) and 685267 for compound 2 (M 5 Hf). Copies of this information may be obtained free of charge from e-mail:
[email protected] or http:// www.ccdc.cam.ac.uk.
Vol. 8, No. 2, 2011
(a)
(b)
X-Ray Absorption Spectroscopy
The XANES and EXAFS spectra were obtained at the Daresbury Labortaory SRS operating at 2 GeV and ca. 100–200 mA. The Zr K-edge data were collected at ca. 80 K on station 9.2 using a Si(220) monochromator and detuning for harmonic rejection. The Hf L3-edge data were collected on station 7.1, also at ca. 80 K, with a Si(111) monochoromator and harmonic rejection both by a mirror and detuning of the second crystal. The data were calibrated using the first maximum in the first derivative of metal foils (Zr K, 17998 eV; Hf L3, 9561 eV).54 Background subtraction used PAXAS with a quadratic for the preedge and either high-order polynomials or splines for the postedge.55 Curved wave fitting was performed using Excurv98 and for multiple scattering pathways the cluster formalism was used.56
Results and Discussion Synthesis of Precursor Molecules
Ideally, the reaction of four equivalents of lithium tris(dimethylamino)silylamine LiHNSi(Me2N)344 with zirconium tetrachloride ZrCl4 or hafnium tetrachloride HfCl4 in pentane should yield M[NHSi(NMe2)3]4 (1; M 5 Zr and Hf) (Eq. 1). Both the M[NHSi(NMe2)3]4 (1; M 5 Zr and Hf) intermediates decompose readily and should form fM[NHSi(NMe2)3]2[m-NSi(NMe2)3]g2, (2; M 5 Zr and Hf) by releasing H2NSi(NMe2)3 (Eq. 2) according to the behavior observed for the corresponding intermediate Ti[NHSi(NMe2)3]4 incorporating titanium instead of zirconium or hafnium.44 The formation of white crystals and subsequent loss of H2NSi(NMe2)3 was confirmed by 1H NMR and FTIR analyses. Unfortunately, the chemical composition of the crystalline compounds 2 (M 5 Zr and Hf) could not be characterized by solution NMR due to their complete insolubility in
3600
3000
2400
1800
1200
600
Wavenumber (cm–1) Fig. 1. IR spectra of (a) complex 2 (M 5 Zr) and (b) complex 2 (M 5 Hf).
organic solvents. MCl4 þ 4LiHNSiðNMe2 Þ3 ! M½NHSiðNMe2 Þ3 4 þ 4LiCl
ð1Þ
M½NHSiðNMe2 Þ3 4 ! fM½NHSiðNMe2 Þ3 2 ½NSiðNMe2 Þ3 g2 þ 2H2 NSiðNMe2 Þ3 M ¼ Zr; Hf
ð2Þ The FTIR spectra of 2 (M 5 Zr and Hf) are very similar (Fig. 1). The weak absorption at 3276 cm1 can be assigned to n(NH) of the MNHSi groups. The strong absorption at 929 cm1 is attributable to the presence of n(Si–N) of SiNMe2 groups.49 The results of elemental analysis were unsatisfactory because these crystalline products are very air sensitive. The molecular structures of the compounds 2 (M 5 Zr and Hf) were determined by single-crystal X-ray diffraction (XRD) analysis (Table I). Views of the molecules 2 (M 5 Zr and Hf) are presented in Figs 2 and 3, and selected bond lengths and angles are presented in Table II. The compounds are isostructural dimers with two bridging and two terminal amides to yield a distorted tetrahedral geometry around each metal center. The analogous titanium compound has been characterized previously.44 The asymmetric unit for each compound comprises half of the molecule with a crystallographic C2 axis present in the dimer. There is little variation in the structural parameters between the two compounds. The metal-to-metal distances across the dimer are 3.23(8) and 3.20(5) A˚ for
www.ceramics.org/ACT
Preparation, Structural Characterization, and Thermal Ammonolysis of Silylamide Complexes
Table I. Crystal Data for the Single-Crystal X-Ray Structures 2 (M 5 Zr)
2 (M 5 Hf)
Formula C36H108N24LiSi6Zr2 C36H108N4LiSi6Hf2 Mr 1235.38 1409.92 Crystal system Monoclinic Monoclinic P 21/c Space group P 21/c A (A˚) 13.3059(15) 13.2582(17) b (A˚) 13.0146(8) 12.9918(8) c (A˚) 20.427(2) 20.389(3) b (1) 111.333(8) 111.222(10) 3295.0(5) 3273.8(7) Volume (A˚3) Z 2 2 1.245 1.430 Density (calc.) (mg/m3) 0.471 3.324 mMoKa (mm1) T (K) 150 150 y range (1) 1.90–27.50 2.66–34.83 Measured 13246 38761 reflections 7565 [0.0916] 14040 [0.0773] Unique reflections (Rint) Completeness 99.9% (27.50) 98.7% (34.83) (y1) 0.768 0.975 Goodness of fit on F2 0.0556, 0.1251 0.0328, 0.0555 R1, wR2 (I42s(I)) 0.0925, 0.1344 0.1417, 0.0696 R1, wR2 (all data) Extinction 0.026(3) 0.00071(5) coefficient 1.012 and 1.152 2.015 and 3.802 Largest diff. peak and hole (e/A˚3)
the zirconium and hafnium structures, respectively, with slightly shorter bond distances overall for the hafnium complex, as would be expected for the small reduction in ionic radius. Amide-bridged zirconium dimers have been structurally characterized previously, showing similar M–M distances (in the range 3.04–3.12 A˚).57,58 There are very few equivalent amide-bridged hafnium dimers reported that are appropriate for comparison, but they also have M–M distances that fall within this
471
range.59,60 There is asymmetry in the bond lengths around the four-membered M2N2 ring for both structures, which was previously observed for the titanium complex and attributed to dppp interactions.44 The distances to the amide ligands are consistent with previously published data on zirconium and hafnium structures with terminal amide ligands, which have M–N bond lengths in the range 1.98–2.04 A˚.61,62 When carrying out the crystallographic refinement, residual electron density was observed in close proximity to the bridging amide donor (N2) for both structures, strongly suggesting that it was not anomalous and should be assigned. The best model was achieved by including a half-occupancy lithium ion. The compounds were analyzed by ICP-MS, which showed that lithium had been incorporated into the structure of compounds 2 during the synthesis, providing evidence for its presence in the crystalline form. The distances between the lithium and metal center or amide/amine N donors are consistent with known related structures. In the compound 2 structures, the metal–Li distances are Hf–Li (2.70(2) A˚) and Zr–Li (2.749(17) A˚) and comparable to structures of lithium– amine complexes in the presence of transition metals (M– Li 2.70–2.96 A˚, M 5 Mn, Zr, W).63–65 The lithium interactions with the closest amide and amine donors are at distances of ca. 2.04 and 2.30 A˚, respectively, which correlate well with typical bond lengths.63–65 Both the compound 2 structures show disorder at the amine group, which coordinates to the lithium ion (0.5 occupancy for each), indicating that the alternate amine position is only occupied when the lithium ion is present. This suggests that the structures of 2 (M 5 Zr and Hf) should be formulated as fLi0.5M[NHSi(NMe2)3]1.5[NSi(NMe2)3]0.5 [m-NSi(NMe2)3]g2 and that compounds 1 (M 5 Zr and Hf) also contain lithium. Ammonolysis
The ammonolysis of the complexes 2 (M 5 Zr and Hf) was carried out in an autoclave at 160 1C based on an ideal reaction expressed as in the following equation fLi0:5 ½NHSiðNMe2 Þ3 1:5 ½NSiðNMe2 Þ3 0:5 ½NSiðNMe2 Þ3 g2 þ 18NH3 ! 2Msi3 ðNÞðNHÞ2 ðNH2 Þ9 þ 17:5HNMe2 þ 0:5LiNMe2 ðM ¼ Zr; HfÞ
ð3Þ Ammonolysis of the zirconium intermediate 2 (M 5 Zr) yielded a light yellow, partially lithiated silicon imide
472
International Journal of Applied Ceramic Technology—Cheng, et al.
Vol. 8, No. 2, 2011
Fig. 2. ORTEP plot of the crystal structure of 2 (M 5 Zr). Hydrogen atoms are omitted for clarity. The disorder and half-occupancy lithium ion are not shown.
Fig. 3. ORTEP plot of the crystal structure 2 (M 5 Hf). Hydrogen atoms are omitted for clarity. The disorder and half-occupancy lithium ion are not shown.
www.ceramics.org/ACT
Preparation, Structural Characterization, and Thermal Ammonolysis of Silylamide Complexes
Table II.
473
Selected Bond Lengths [A˚] and Angles [1] for 2 (M 5 Hf and Zr)
Hf(1)–N(1) Hf(1)–N(2)#1 Hf(1)–N(3) Hf(1)–N(2)
1.971(4) 2.058(3) 2.083(3) 2.146(3)
N(1)–Hf(1)–N(2)#1 Zr(1)–N(1) Zr(1)–N(2)#1 Zr(1)–N(3) Zr(1)–N(2)
103.77(15) 1.994(4) 2.075(3) 2.098(3) 2.187(3)
N(1)–Zr(1)–N(2)#1
103.70(16)
N(1)–Hf(1)-N(3) N(2)#1–Hf(1)–N(3) N(1)–Hf(1)–N(2) N(2)#1–Hf(1)–N(2) N(3)–Hf(1)–N(2) Hf(1)#1–N(2)–Hf(1) N(1)–Zr(1)–N(3) N(2)#1–Zr(1)–N(3) N(1)–Zr(1)–N(2) N(2)#1–Zr(1)–N(2) N(3)–Zr(1)–N(2) Zr(1)#1–N(2)–Zr(1)
115.00(14) 101.82(14) 121.65(14) 81.04(15) 120.75(12) 99.01(14) 115.43(15) 101.90(14) 121.16(14) 81.30(15) 120.69(13) 98.70(15)
Symmetry transformations used to generate equivalent atoms: #1x, y11, z for 2 (M 5 Hf) and #1 x, y, z for 2 (M 5 Zr).
powder Si3M(N)(NH)x(NH2)y(NMe2)z (3; M 5 Zr). The ammonolysis of the corresponding hafnium intermediate 2 (M 5 Hf) yielded partially lithiated Si3M(N)(NH)x(NH2)y(NMe2)z (3; M 5 Hf) as a white solid. Compared with the IR spectra (Fig. 1) of the precursor compounds 2 (M 5 Zr and Hf), the FTIR spectra (Fig. 4) of both of the ammonolyzed products 3 (M 5 Zr and Hf) exhibit a lower intensity of the n(CH) bands from 2795 to 2993 cm1. This indicates a substantial, but incomplete, elimination of dimethylamino groups. Longer ammonolysis times could reasonably be expected to lead to a more complete reaction. The presence of the broad n(N–H) bands centered at about
3322 cm1 and the low-intensity bands at about 1591 cm1, due to the presence of NH2, suggests that NH groups are present in different environments, such as Si–NH–Si, Si–NH–M (M 5 Zr, Hf), and SiNH2. The peak at about 1190 cm1 is due to the presence of d(N–C) and d(N–H).3 The broad band from 900 to 1005 cm1 can be assigned to n(Si–N). The 13C CPMAS NMR spectra show a single resonance at 37.7 ppm for the ammonolyzed hafnium intermediate 3 (M 5 Hf) and two resonance peaks at 37.8 and 39.3 ppm for the ammonolyzed zirconium intermediate 3 (M 5 Zr), indicating the presence of Si(NMe2)x (Fig. 5). This result is consistent with the IR spectra of these compounds. Both of the 29Si CP-MAS NMR spectra of
(a)
(a)
(b)
(b)
4000
3000
2000
1500
1000
500
Wavenumber (cm–1) Fig. 4. IR spectra of (a) the ammonolyzed zirconium product 3 (M 5 Zr) and (b) the ammonolysized hafnium product 3 (M 5 Hf).
8
6
40
20
0
–
–40
ppm Fig. 5. 13C CP MAS NMR of ammonolyzed products (a) 3 (M 5 Zr) and (b) 3 (M 5 Hf).
International Journal of Applied Ceramic Technology—Cheng, et al.
474
(a)
(b)
60
20
–20
–60
–100
–140
–180
ppm Fig. 6. 29Si CP MAS NMR of ammonolyzed product (a) 3 (M 5 Zr) and (b) 3 (M 5 Hf).
the ammonolyzed products of the ceramic precursors 3 (M 5 Zr and Hf) (Fig. 6) exhibit a single broad resonance at about 41 ppm, indicating the formation of SiN4 in different environments.44 The Zr K-edge XANES spectrum of 3 (M 5 Zr) is shown in Fig. 7a and this can be used to identify the local coordination of the Zr as the XANES spectra are sensitive to coordination environments. For example, tetrahedral and other noncentrosymmetric Zr compounds have characteristic preedge features in their Zr K-edge XANES spectra, which are much weaker and usually absent in centrosymmetric geometries, such as octahedral or dodecahedral.66 In addition, the intensity and structure of the white line can also be used to help
Normalised Intensity
1.5
1.0 (d) 0.5
(c) (b)
0.0
(a)
17960 17980 18000 18020 18040 18060 18080 18100
Photon Energy /eV Fig. 7. Zr.
Zr K-edge XANES of M 5 Zr (a) 3, (b) 4, (c) ZrN, (d)
Vol. 8, No. 2, 2011
identify the Zr environment as it has been shown that for octahedral coordination, for example in BaZrO3, a splitting of the white line is observed, while in sevenfold coordination, there is no splitting and, although the white line can be split in eightfold coordination, its intensity is greater than that for octahedral coordination.67–70 Therefore, the lack of a preedge feature in the Zr K-edge XANES spectrum of 3 (Fig. 7a) means that the initial tetrahedral coordination in 2 has increased and the intensity and splitting of the white line are characteristic of Zr in a distorted octahedral coordination environment. Our spectra of 3 are very similar to those observed for SiO2–ZrO2 nanosols in pillared clays71 and dimeric Zr(OnBu)4,70 both of which have a distorted octahedral environment. The Hf L3-edge XANES of the analogously treated material were also consistent with an octahedral environment as they were lacking a characteristic bump just after the white line indicative of seven or higher coordination.72 The Zr K-edge EXAFS (Table III) confirms that the coordination environment around the Zr is at least octahedral and the best fit to the first peak in the FT is to a Zr–N of 2.19(3) A˚ with a coordination number between 6 and 7. The Debye–Waller factor for this shell is quite large for both 6 and 7 coordination numbers and while this may indicate a lower coordination number, we believe that it is much more likely to be due to the disorder within the Zr–N distances indicated by the XANES spectra. While it was possible to obtain more reasonable Debye–Waller factors by introducing a split Zr–N shell, this did not result in a statistically significant improvement in the fit. Therefore, Fig. 8a shows the first shell fit to Zr–N6 with a Zr–N bond length of 2.19(3) A˚. The second peak in the FT is best fit by a Zr– Zr interaction of 3.24(4) A˚ with an occupation number between 3 and 4, and the fit shown in Fig. 8a shows the fit using 3. Although a Zr–Si interaction of about 2.94 A˚ could be included, detailed analysis showed that its inclusion was not statistically significant and it was therefore not included in the final fit shown in Fig. 8a. The Zr–N and Zr y Zr distances are close to those expected for edge rather than face or vertex-sharing octahedra. The analysis of the Hf L3-edge EXAFS data revealed a local environment very similar to that for Zr. The first shell refined to a Hf–N distance of 2.17(3) A˚ with a coordination number of about 6. The second shell was best fit by 3 Hf y Hf interactions at 3.27(4) A˚. As in the Zr case, a Si shell could be included in the refinement but not with any great certainty.
Zr–N6 Zr–Zr12 Zr–Zr6 Zr–Zr24 Zr–Zr12 Ef y Rz
2.289(3) 3.244(3) 4.579 (2xr1) 5.622(7) 6.488 (2r2) 4.3(6) 29.1
0.013(3) 0.0095(3) 0.0109(7) 0.012(1) 0.0127(9)
˚2 z
2r (A )
2
Zr–Zr6 Zr–Zr6 Zr–Zr6 Zr–Zr2 Zr–Zr12 Zr–Zr6 Zr–Zr12 Zr–Zr6
Zr
Zr–N6 Zr–Zr9 Zr–Zr6 Zr–Zr18 Zr–Zr9
4
z
Estimated systematic errors in XAFS bond lengths are 71.5% for well-defined coordination shells. 2s2 is the Debye–Waller factor. y Ef is aRsingle refined parameter to reflect differences in the theoretical and experimental Fermi levels. R z R ¼ jwT wE jk3 dk= jwE jk3 dk 100%. J Christensen.75 Lichter.82 XRD, X-ray diffraction.
w
Standard deviation in parentheses.
2.2925 3.2421 4.5850 5.6155 6.4842
ZrN
EXAFS r (A˚) w
4.5(6) 29.4%
Efy Rz
XRD J r (A˚)
0.0217(13) 0.0128(5)
2.187(7) 3.238(4)
Zr–N6 Zr–Zr3
2s (A )
˚2 z
r (A˚ ) 2
3.232 3.179 4.534 5.148 5.568 5.598 6.078 6.464
XRD r (A˚ )
EXAFS r (A˚) w
3.248(4) 3.151(4) 4.527(8) 5.095(33) 5.604(9) 5.499(19) 6.107(8) 6.496 (2 r1) Efy Rz
EXAFS r (A˚) w
Ef y Rz
5.655(9) 6.470 (2 r2)
2.214(8) 3.235(3)
Refined EXAFS Parameters for 3, 4,(M 5 Zr), ZrN, and Zr
3
w
EXAFS
Table III.
0.0030(6) 0.0038(6) 0.0126(14) 0.0123(69) 0.0060(24) 0.00637(47) 0.0105(15) 0.0160(19) 0.8(6) 19.6%
2r2 (A˚2 ) z
4.4(6) 26.3%
0.017(2) 0.020(2)
0.014(2) 0.0109(3)
2s 2 (A˚2 ) z
www.ceramics.org/ACT Preparation, Structural Characterization, and Thermal Ammonolysis of Silylamide Complexes 475
International Journal of Applied Ceramic Technology—Cheng, et al.
476
6
Experiment Theory
25 20
2
15
|FT|
k3χ(k)
Experiment Theory
(a)
4
0 –2
10 5
–4 –6
0 3
4
5
6
7
8
9 10 11 12 13 14 15
0
1
2
3
4
Experiment Theory
14 12 10 8 6 4 2 0 –2 –4 –6 –8 –10 –12 –14
(b)
3
4
5
6
7
8
6
7
80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0
8
9
10
9 10 11 12 13 14 15
Experiment Theory
0
1
2
3
4
k / Å–1
5
6
7
8
9
10
r /Å Experiment Theory
30
5
r /Å
|FT|
k3χ(k)
k / Å–1
20
120
Experiment Theory
100
(c)
10
80
|FT|
k3χ(k)
Vol. 8, No. 2, 2011
0
60
–10
40
–20
20
–30
0 3
4
5
6
7
8
9 10 11 12 13 14 15
0
1
2
3
4
k / Å–1 Experiment Theory
20
5
6
7
8
9
10
r /Å Experiment Theory
120
(d)
100 80
|FT|
k3χ(k)
10 0
60
–10
40
–20
20 0 3
4
5
6
7
8
k/
Fig. 8.
9 10 11 12 13 14 15
0
1
2
3
4
Å–1
Zr K-edge EXAFS and FTs at ca. 80 K of M 5 Zr (a) 3, (b) 4, (c) ZrN, and (d) Zr foil.
5
r /Å
6
7
8
9
10
Preparation, Structural Characterization, and Thermal Ammonolysis of Silylamide Complexes
Therefore, the Zr and Hf XANES and EXAFS data confirm that on ammonolysis, the initial Zr/Hf tetrahedral coordination in 2 is converted to a higher coordination of at least 6 and that while the M y M distances remain very similar, the occupation numbers increase from 1 to around 3. These data therefore indicate that the Zr is octahedrally coordinated and that these octahedra are edge sharing. On heating the ammonolyzed samples to 6001C, there was very little difference in the Zr K-edge and Hf L3-edge XANES or EXAFS, except for slight increases in the Debye– Waller factors. All these spectral features indicate that ammonolysis of the precursors fLi0.5M[NHSi(NMe2)3]1.5[NSi (NMe2)3]0.5[m-NSi(NMe2)3]g2 (2; M 5 Zr and Hf) leads to the formation of the analogous partially lithiated zirconium or hafnium silicon imides Si3M(N)(NH)x (NH2)y(NMe2)z (3; M 5 Zr and Hf), containing residual dimethylamino groups. These imide powders have threedimensional networks mainly connected by Si–N(H)–Si backbones formed during the ammonolysis and condensation reactions (Eqs. 4–6). The presence of residual dimethylamino groups shows that the transamination reactions were not yet complete under these reaction conditions. However, compared with the silicon titanium imide powder prepared from fTi[NHSi(NMe2)3]2[m-NSi (NMe2)3]g2 by a similar method,44 the 3 (M 5 Zr and Hf) have many more NMe2 groups according to the FTIR analysis (Fig. 4). This suggests that under similar reaction conditions, the ammonolysis and condensation reactions of precursor compounds 2 (M 5 Zr and Hf) progressed slower than those of the analogous fTi[NHSi(NMe2)3]2 [m-NSi(NMe2)3]g2. Si NMe2 þ NH3 ! Si NH2 þ HNMe2 ð4Þ Si NH2 þ Me2 N Si ! Si NH Si þHNMe2 2 Si NH2 ! Si NH Si þNH3
ð5Þ ð6Þ
Physisorption analysis shows that the ammonolyzed zirconium product 3 (M 5 Zr) exhibits a type IV nitrogen adsorption isotherm (Fig. 9), typical of mesoporous materials, with a large surface area of 440 m2/g. The pore size varies from 27 to 50 A˚, with an average pore diameter of 34 A˚ (Fig. 9). On the other hand, the ammonolyzed hafnium product 3 (M 5 Hf) displays a type I nitrogen adsorption isotherm typical of micro-
350
Adsorption
477
Desorption
300 Volume Adsorbed cm3/g STP
www.ceramics.org/ACT
250
200
150 1.2 0.8
100
0.4 0
50
10
30
50
70
90
Pore Diameter (A)
0 0
0.2
0.4 0.6 0.8 Relative Pressure (P/P0)
1
Fig. 9. Nitrogen adsorption isotherm and pore-size distribution for 3 (M 5 Zr).
porous materials (Fig. 10). The surface area of 232 m2/g of the 3 (M 5 Hf) is just above half of that of the corresponding ammonolyzed zirconium product 3 (M 5 Zr). The pore size and pore size distribution of the 3 (M 5 Hf) are similar to that of the corresponding zirconium derivative 3 (M 5 Zr). Pyrolysis
The ceramic yields of the gels 3 (M 5 Zr and Hf) pyrolyzed at 10001C are very similar (62–65%); see Fig. 11 for the thermogravimetric analysis under argon flow 3 (M 5 Zr). ICP analysis of both pyrolyzed products 4 (M 5 Zr and Hf) confirms the presence of silicon and zirconium or silicon and hafnium in the respective ceramics 4 (M 5 Zr and Hf). The IR spectra of the products 4 (M 5 Zr and Hf) formed by pyrolysis at 10001C indicate that dimethylamino groups have not been removed completely as shown by the presence of the n(CH) bands from 2800 to 3000 cm1. This is illustrated in the IR spectrum (Fig. 12) of the pyrolyzed zirconium-containing product 4 (M 5 Zr). The broad n(N–H) bands centered at 3420 cm1 and d(N–H) band at 1200 cm1 suggest that a huge number of NH
International Journal of Applied Ceramic Technology—Cheng, et al.
478
Volume Adsorbed cm3/g STP
100
Adsorption
Desorption
80
60
40
20
0 0
Fig. 10.
Vol. 8, No. 2, 2011
0.2 0.4 0.6 0.8 Relative Pressure (P/P0)
1
Fig. 12. Infra-red spectrum of the pyrolyzed product 4 (M 5 Hf) formed from the pyrolysis of the silicon hafnium imide 3 (M 5 Hf) at 10001C.
Nitrogen adsorption isotherm for 3 (M 5 Hf).
groups are present in the product. The broad band centered at 960 cm1 can be ascribed to n(Si–N). The imide powders start to lose material at temperatures as low as 501C and the pyrolysis is substantially complete at 4001C; see Fig. 11. This thermal instability precludes the use of these imide powders for some practical applications, for example many catalysis applications. There is a substantial change in structure after heating the ammonolyzed product 3 (M 5 Zr) at 10001C to yield the corresponding partially lithiated Si–M–N ceramic 4 (M 5 Zr) as a nitridic phase. There is no significant change in edge position in the Zr K-edge XANES spectra of 3 and 4 (Figs. 7a and b), indicating
Fig. 11. Thermogravimetric analysis of the pyrolysis of the silicon hafnium imide powder 3 (M 5 Hf) at 1000 1C to form the pyrolyzed product 4 (M 5 Hf).
that the Zr has not been reduced when heated to 10001C under NH3. In contrast, the edge for ZrN (Fig. 7c) is about 2 eV lower in energy than for 4 and the white line structure is different. There is a further larger shift for elemental Zr, which also has a different edge structure (Fig. 7d). The improved resolution of the splitting of the white line in the Zr K-edge XANES spectrum of 4 (Fig. 7b) indicates that the Zr environment is closer to being regular octahedral in 4 than in 3,70 and is similar to that observed for BaZrO3,69,73 which has a regular octahedral Zr environment.74 The Hf L3-edge XANES was not sufficiently sensitive to identify any structural modifications between 3 and 4. Therefore, the XANES spectra indicate that the Zr oxidation state remains the same after pyrolysis and that the local Zr environment has become more regular, but that 4 does not contain either ‘‘bulk’’ ZrN or Zr. At first glance, there appears to be quite a strong similarity between the Zr K-edge EXAFS of 4 (Fig. 8b) and that of ZrN (Fig. 8c) at the same temperature with a first coordination shell of Zr–N6 and the remainder resembling an fcc/ccp lattice of Zr.50,51 However, the XANES indicates that while the immediate Zr environment is octahedral, substantial differences are to be expected in the Zr–Zr interactions. The EXAFS confirms this as the first shell in the FT corresponds to Zr–N at 2.21(3) A˚ with a coordination number of around 6. The second peak in the FT is best fit to a Zr y Zr interaction of 3.23(4) A˚ with an occupation number of 9, compared with 3.24(4) in ZrN with an occupation number of 12. A peak at 4.58(6) A˚ in ZrN is absent
www.ceramics.org/ACT
Preparation, Structural Characterization, and Thermal Ammonolysis of Silylamide Complexes
in the FT of 4 and as this is due to a linear Zr–N–Zr scattering pathway, its intensity would be enhanced by multiple scattering. Although Zr y Si and Zr y Zr interactions can be added to the refinement, they are not statistically significant. The two remaining features in the FT due to Zr y Zr at 5.66(6) and 6.47(7) A˚ fit much better to an fcc lattice of Zr (e.g., in ZrN, see Fig. 8c) rather than an hcp lattice (as in Zr, see Fig. 8d), but with reduced occupation numbers of 18 and 9. The presence of features at these distances indicates that the particles containing Zr are reasonably large. In the case of Hf, the first shell refined to Hf–N6 with a bond length of 2.21(3) A˚. The first Hf y Hf shell at 3.21(4) A˚ refined to an occupation number of 6, compared with 9 in the Zr case, but as in the Zr data, there is no feature in the FT around 4.5 A˚ expected for linear Hf–N–Hf scattering. As for Zr, the more distant features in the FT fit best to an fcc/ccp lattice of Hf with distances of 5.60(6) and 6.41(7) A˚, but with lower coordination numbers. In addition to cubic ZrN and HfN with a NaCl structure of interpenetrating ccp/fcc arrays with six-coordinate Zr75,76 or six-coordinate Hf,77 there are a number of other nitride phases, such as orthorhombic Zr3N4 with six- and seven-coordinate Zr78,79 or cubic high-pressure phases of Zr3N4 and Hf3N4 with eightfold zirconium or hafnium coordination.80 There are also nitrogen-deficient hafnium nitride phases such as the trigonal Hf3N2 and Hf4N3, both with the nitrogens located in octahedral interstitial holes.81 Although 4 (M 5 Hf) was sufficiently crystalline to obtain XRD data, it has not been possible to identify the phases present, although ICP analysis confirms the presence of lithium and zirconium, or hafnium in the pyrolyzed products. Therefore, the Zr K-edge and Hf L3-edge XANES and EXAFS of 4 data indicate that around each metal, there is a fairly regular octahedral coordination of N, with second (and higher) shells of Zr or Hf with greater occupation numbers than in 3, indicating some clustering of metal during the pyrolysis stage. The metal–metal distances in all of the samples are very similar and it is the number of such interactions that increases on ammonolysis and pyrolysis. Both the EXAFS and the XANES show that 4 does not contain ZrN or Zr, but they indicate the presence of a structure that is deficient in Zr compared with ZrN. The porous structure of both of the ammonolysized products 3 (M 5 Zr and Hf) is substantially changed after being heated at 10001C for 2 h under an atmo-
479
sphere of ammonia to form Si–Zr–N and Si–Hf–N ceramic products 4 (M 5 Zr and Hf), respectively. The surface areas of the pyrolyzed products 4 (M 5 Zr and Hf) were both below 5 m2/g. These results are very different from those observed during the pyrolysis of the analogous silicon titanium imide Si3Ti(N)(NH)x(NH2)y(NMe2)z prepared from fTi[NHSi(NMe2)3]2[mNSi(NMe2)3]g2. A mesoporous Si–Ti–N composite ceramic material with a relatively large effective surface area 163 m2/g was obtained after pyrolysis of the silicon titanium imide gel under an NH3 flow at 10001C.44 The significant reduction in the active surface area after being heated to a high temperature to produce the pyrolyzed products 4 indicates that the cross-linked structures of the ammonolysized products 3 (M 5 Zr and Hf) are not strong enough to withstand the pyrolysis conditions. This is consistent with a relatively high concentration of residual dimethylamino groups in the ammonolyzed precursors 3 (M 5 Zr and Hf) and their pyrolyzed products 4 (M 5 Zr and Hf) and, perhaps, to the presence of lithium. A lower relative concentration of amino groups would lead to a lower degree of crosslinking during the pyrolysis procedure and contribute to collapse of the pores. It is also possible that the structural weakness of the Zr–N–Si bonds and Hf–N–Si bonds in the zirconium and hafnium silicon imides 3 (M 5 Zr and Hf) compared with the Ti–N–Si bonds of the corresponding titanium silicon imide Si3Ti(N)(NH)x(NH2)y(NMe2)z leads to shrinkage and then collapse of the pores during the pyrolysis procedure. Preliminary PXRD experiments reveal the formation of a mixture of crystalline Si–M–N products.
Conclusions Two novel dimeric transition silylamide complexes, that is lithium-bis(dimethylamino)silylamino-m-bis[(dimethylamino)silylimino]-zirconium fLi0.5M[NHSi(NMe2)3] 1.5[NSi(NMe2)3]0.5[m-NSi(NMe2)3]g2 (2; M 5 Zr) and lithium-bis(dimethylamino)silylamino-m-bis[(dimethylamino)silylimino]-hafnium fLi0.5M[NHSi(NMe2)3]1.5[NSi (NMe2)3]0.5[m-NSi(NMe2)3]g2 (2; M 5 Hf), could be prepared by a reaction of lithium tris(dimethylamino)silylamide LiHNSi(NMe2)3 with the corresponding metal chlorides. The inclusion of the lithium in these silylamine complexes is unexpected, because it does not occur in the synthesis of the corresponding titanium silylamide. Partially lithiated silicon zirconium imide Si3M(N)(NH)x(N-
480
International Journal of Applied Ceramic Technology—Cheng, et al.
H2)y(NMe2)z (3; M 5 Zr) was obtained as a high surfacearea mesoporous powder by ammonolysis of the zirconium intermediate 2 (M 5 Zr). Similar ammonothermal treatment of the corresponding hafnium intermediate 2 (M 5 Hf) yielded partially lithiated silicon hafnium imide Si3M(N)(NH)x(NH2)y(NMe2)z (3; M 5 Hf) as a microporous powder with a lower surface area. The dimethylamino groups present in 2 are not completely removed in 3, suggesting that the reaction is much slower than that used to prepare the equivalent silicon titanium imide. The K-edge spectra of 3 suggest that the tetrahedral coordination environment of 2 has increased to a disordered octahedral coordination environment for 3. The large surface area of both of the zirconium and hafnium imide intermediates 3 (M 5 Zr and Hf) is reduced significantly during pyrolysis at 10001C under an NH3 flow to form the corresponding partially lithiated zirconium and hafnium silicon nitride ceramics 4 with a regular octahedral coordination environment. Acknowledgments CCLRC and STFC are thanked for access to synchrotron and computing facilities at Daresbury Laboratory, and Dr. Fred Mosselmans is thanked for his assistance with the setting up of stations 7.1 and 9.2. References J. M. Thomas and W. J. Thomas, Heterogeneous Catalysis, VCH, Weinheim, 1997. 2. D. Farrusseng, et al., ‘‘Pore-Size Engineering of Silicon Imido Nitride for Catalytic Applications,’’ Angew. Chem. Int. Ed. Eng., 40 4204–4207 (2001). 3. C. J. Brinker, and G. W. Scherer, (eds). Sol, Academic Press, London, 1990. 4. D. A. Lindquist, et al., ‘‘Formation and Microstructure of Boron Nitride Aerogels,’’ J. Am. Ceram. Soc., 73 757–760 (1990). 5. C. K. Narula, R. Schaeffer, A. Datye, and R. T. Paine, ‘‘Synthesis of Boron Nitride Ceramics from 2,4,6-Triaminoborazine,’’ Inorg. Chem., 28 4053– 4055 (1989). 6. R. T. Paine, C. K. Narula, R. Schaeffer, and A. K. Datye, ‘‘Formation of Boron Nitride Coatings on Metal Oxides,’’ Chem. Mater., 1 486–489 (1989). 7. C. K. Narula, et al., ‘‘Models and Polyborazine Precursors for Boron Nitride Ceramics,’’ Chem. Mater., 2 377–384 (1990). 8. Th. T. Borek, X. Qiu, L. M. Rayfuse, A. K. Datye, R. T. Paine, and L. F. Allard, ‘‘Boron Nitride Coatings on Oxide Substrates: Role of Surface Modifications,’’ J. Am. Ceram. Soc., 74 2587–2591 (1991). 9. B. W. Pfennig, A. B. Bocarsly, and R. K. Prud’homme, ‘‘Synthesis of a Novel Hydrogel Based on a Coordinate Covalent Polymer Network,’’ J. Am. Chem. Soc., 115 2661–2662 (1993). 10. M. Heibel, G. Kumar, C. Wyse, P. Bukovec, and A. B. Bocarsly, ‘‘Use of SolGel Chemistry for the Preparation of Cyanogels as Ceramic and Alloy Precursors,’’ Chem. Mater., 8 1504–1511 (1996). 11. S. L. Sharp, A. B. Bocarsly, and G. W. Scherer, ‘‘Bulk Properties of a Cyanogel Network: Toward an Understanding of the Elastic, Mechanical, and Physical Processes Associated with SolGel Processing of Cyanide-Bridged Gel Systems,’’ Chem. Mater., 10 825–832 (1998).
1.
Vol. 8, No. 2, 2011
12. R. S. Deshpande, et al., ‘‘Morphology and Gas Adsorption Properties of PalladiumCobalt-Based Cyanogels,’’ Chem. Mater., 15 4239–4246 (2003). 13. M. Vondrova, et al., ‘‘Supported Superparamagnetic Pd/Co Alloy Nanoparticles Prepared from a Silica/Cyanogel Co-Gel,’’ Chem. Mater., 17 6216– 6218 (2005). 14. M. Vondrova, P. W. Majsztrik, S. Gould, and A. B. Bocarsly, ‘‘Solid-State Chemistry in a Microwave Oven: Preparation of Pd/Co Alloy from Cyanogel Coordination Polymers,’’ Chem. Mater., 17 4755–4757 (2005). 15. A. O. Gabriel and R. Riedel, ‘‘Preparation of Non-Oxidic Silicon Ceramics by an Anhydrous Sol–Gel Process,’’ Angew. Chem. Int. Ed. Eng., 36 384–386 (1997). 16. A. O. Gabriel, R. Riedel, S. Stork, and W. F. Maier, ‘‘Synthesis and Thermally Induced Ceramization of a Poly(methlysilsesquicarbodi-imide) Gel,’’ Appl. Organomet. Chem., 11 833–841 (1997). 17. R. Riedel, et al., ‘‘Inorganic Solid-State Chemistry with Main Group Element Carbodiimides,’’ Chem. Mater., 10 2964–2979 (1998). 18. A. O. Gabriel, et al., ‘‘Thermal Decomposition of Poly(methylsilsesquicarbodiimide) to Amorphous SiCN Ceramics,’’ Chem. Mater., 11 412–420 (1999). 19. D. S. Kim, E. Kroke, R. Riedel, A. O. Gabriel, and S. C. Shim, ‘‘An Anhydrous Sol–Gel System Derived from Methyldichlorosilane,’’ Appl. Organomet. Chem., 13 495–499 (1999). 20. S. Nahar-Borchert, E. Kroke, R. Riedel, B. Boury, and R. J. P. Corriu, ‘‘Synthesis and Characterization of Alkylene-Bridged Silsesquicarbodiimide Hybrid Xerogels,’’ J. Organomet. Chem., 686 127–133 (2003). 21. K. W. Volger, R. Hauser, E. Kroke, R. Riedel, Y. H. Ikuhara, and Y. Iwamoto, ‘‘Synthesis and Characterization of Novel Non-Oxide Sol–Gel Derived Mesoporous Amorphous Si–C–N Membranes,’’ J. Ceram. Soc. Jpn., 114 567–570 (2006). 22. E. Kroke, K. W. Vo¨lger, A. Klonczynski, and R. Riedel, ‘‘A Sol–Gel Route to B4C,’’ Angew. Chem. Int. Ed. Eng., 40 1698–1700 (2001). 23. K. W. Vo¨lger, et al., ‘‘B/C/N Materials and B4C Synthesized by a Non-Oxide SolGel Process,’’ Chem. Mater., 15 755–764 (2003). 24. J. Lo¨ffelholz, J. Engering, and M. Jansen, ‘‘Sol–Gel Process in the AmmonoSystem—a Novel Access to Silicon Based Nitrides,’’ Z. Anorg. Allg. Chem., 626 963–968 (2000). 25. O. Gabriel, R. Riedel, S. Storck, and W. F. Maier, ‘‘Synthesis and Thermally Induced Ceramization of a Poly(methlysilsesquicarbodi-imide) Gel,’’ Appl. Organomet. Chem., 11 1–10 (1997). 26. R. Rovai, C. W. Lehmann, and J. S. Bradley, ‘‘Non-Oxide Sol–Gel Chemistry: Preparation from Tris(dialkylamino)silazanes of a CarbonFree, Porous Silicon Diimide Gel,’’ Angew. Chem. Int. Ed., 38 2036–2038 (1999). 27. F. Cheng, S. Clark, S. M. Kelly, J. S. Bradley, and F. Lefebvre, ‘‘Preparation of Mesoporous Silicon Nitride Gels Via a Non-Aqueous Sol–Gel Route,’’ J. Am. Ceram. Soc., 87 1413–1414 (2004). 28. Y. Xia and R. Mokaya, ‘‘Ordered Mesoporous MCM-41 Silicon Oxynitride Solid Base Materials with High Nitrogen Content: Synthesis, Characterisation and Catalytic Evaluation,’’ J. Mater. Chem., 14 2507–2515 (2004). 29. Y. Xia and R. Mokaya, ‘‘Highly Ordered Mesoporous Silicon Oxynitride Materials as Base Catalysts,’’ Angew. Chem. Int. Ed. Eng., 42 2639–2644 (2003). 30. D. Hullmann, G. Wendt, U. Singliar, and G. Ziegenbalg, ‘‘Propane Dehydrogenation over Supported Platinum Silicon Nitride Catalysts,’’ Appl. Catal. A: Gen., 225 261–270 (2002). 31. G. Garcia Cervantes, F. J. Cadete Santos Aires, and J. C. Bertolini, ‘‘Compared Properties of Pd on Thermo-Conductor Supports (SiC, Si3N4) and Pd on Oxide Supports (Al2O3, SiO2) for the 1,3-Butadiene Hydrogenation Reaction,’’ J. Catal., 214 26–32 (2003). 32. C. Me´thivier, J. Massardier, and J. C. Bertolini, ‘‘Pd/Si3N4 Catalysts: Preparation, Characterization and Catalytic Activity for the Methane Oxidation,’’ Appl. Catal. A: Gen., 182 337–344 (1999). 33. F. Monnet, Y. Schuurman, F. J. Cadete Santos Aires, J. C. Bertolini, and C. Mirodatos, ‘‘New Si3N4 Supported Platinum Catalysts for the Partial Oxidation of Methane at High Temperature,’’ Catal. Today, 64 51–58 (2001). 34. I. Kurzina, F. J. Cadete Santos Aires, G. Bergeret, and J. C. Bertolini, ‘‘Total Oxidation of Methane over Pd Catalysts Supported on Silicon Nitride Influence of Support Nature,’’ Chem. Eng. J., 107 45–53 (2005).
www.ceramics.org/ACT
Preparation, Structural Characterization, and Thermal Ammonolysis of Silylamide Complexes
35. F. Cheng, et al., ‘‘A New Class of Transition-Metal-Doped Mesoporous Silicon Nitride Catalysts Via a Non-Aqueous Sol–Gel Route,’’ Chem. Commun. 5662–5563 (2005). 36. M. Baumbach, et al, ‘‘New Type of Filter for Micro-Machined Silicon Sensors: Novel Mesoporous Filters for Selective Gas Absorption,’’ Proceedings IEEE Sensors, Vienna, 2004. 37. A. Maisonat, et al, ‘‘New Generation of Micro-Machined Silicon Sensors: Nano-Structured Pd- and Pt-Doped Tin Dioxide Sensitive Layers for the Detection of Hazardous Gases,’’ Technical Digest Eurosensors, XVIII, Rome, 2004. 38. F. Cheng, S. M. Kelly, S. Clark, J. S. Bradley, M. Baumbach, and A. Schu¨tze, ‘‘Preparation and Characterisation of a Supported Si3N4 Membrane Via a Non-Aqueous Sol–Gel Process,’’ J. Membr. Sci., 280 530–535 (2006). 39. K. K. Chawla, Ceramic Matrix Composites, Chapman & Hall, London, 1993. 40. R. T. Paine, J. F. Janik, and M. Fan, ‘‘Some Recent Developments in Precursor Routes to Ceramic Nanocomposites,’’ Polyhedron, 13 1225–1232 (1994). 41. F. Cheng, B. Toury, F. Lefebvre, and J. S. Bradley, ‘‘Preparation of a Mesoporous Silicon Boron Nitride Via a Non-Aqueous Sol–Gel Route,’’ Chem. Commun. 657 242–243 (2003). 42. F. Cheng, B. Toury, S. J. Archibald, and J. S. Bradley, ‘‘Synthesis and Structure of 2,4,6-Tris[tris(dimethylamino)silylamino]Borazine: f[(CH3)2N]3 SiNHg3B3N3H3,’’ J. Organomet. Chem., 657 71–74 (2002). 43. F. Cheng, et al., ‘‘Preparation of Mesoporous Silicon Boron Imide Gels from Single Source Precursors Via a Non-Aqueous Sol–Gel Route,’’ Chem. Mater., 15 4651–4657 (2003). 44. F. Cheng, et al., ‘‘Ammonothermal Synthesis of a Mesoporous Si–Ti–N Composite Material from a Single Source Precursor,’’ Chem. Mater., 17 5594–5602 (2005). 45. J. S. Bradley, et al., ‘‘Synthesis and Structures of [(Me2N)3SiNHLi]4, ((Me2N)3SiNH)3Al and Li(THF)1 (C4H8O)Al[NHSi(NMe2)3]3, 2 [((Me2N)3SiNH)4Al],’’ J. Chem. Soc. Dalton Trans. 1846–1851 (2003). 46. F. Cheng, S. M. Kelly, F. Lefebvre, S. Clark, R. Supplit, and J. S. Bradley, ‘‘Preparation of a Mesoporous Silicon Aluminum Imide Gel Via a NonAqueous Sol–Gel Route,’’ J. Mater. Chem., 15 772–777 (2005). 47. F. Cheng, S. Clark, S. J. Archibald, S. M. Kelly, and J. S. Bradley, ‘‘Synthesis and Structure of Tetrakis(isopropylimino)bis-[tris(dimethylamino)silylamino]Alane,’’ J. Organometal. Chem., 690 1205–1208 (2005). 48. F. Cheng, S. M. Kelly, S. Clark, J. S. Bradley, and F. Lefebvre, ‘‘Catalytic Ammonolytic Sol–Gel Preparation of a Mesoporous Silicon Aluminium Nitride from a Single-Source Precursor,’’ J. Organomet. Chem., 692 3816–3822 (2007). 49. K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, John Wiley & Sons Inc., New York, 1997. 50. G. M. Sheldrick, SHELXS-97, ‘‘Program for Structure Solution’’, University of Go¨ttingen, Germany, 1997. 51. G. M. Sheldrick, SHELXL-97: ‘‘Program for the Refinement of Crystal Structures’’, University of Go¨ttingen, Germany, 1997. 52. L. J. Farrugia, ‘‘ORTEP-3 for Windows—a Version of ORTEP-III with a Graphical User Interface (GUI),’’ J. Appl. Cryst., 30 565–566 (1997). 53. L. J. Farrugia, ‘‘WinGX Suite for Single Crystal Small Molecule Crystallography,’’ J. Appl. Cryst., 32 837–838 (1999). 54. A. C. Thompson, et al. Center for X-Ray Optics and Advanced Light Source XRay Data Booklet, Lawrence Berkeley National Laboratory, University of California, Berkeley, CA, 2001. 55. N. Binsted, PAXAS, Program for the Analysis of X-Ray Absorption Spectra, University of Southampton, U.K., 1988. 56. N. Binsted, EXCURV98, CCLRC Daresbury Laboratory Computer Program, CCLRC, Daresbury Laboratory, U.K., 1998. 57. R. Bettenhausen, W. Milius, and W. Schnick, ‘‘Bis(trimethylsilyl)-Ammonium Salts Obtained by Reaction of Hexamethyldisilazane with TiCl4, ZrCl4, or SnCl4,’’ Chem. Eur. J., 3 1337–1341 (1997). 58. J. Hagadorn and J. Arnold, ‘‘Zirconium Chemistry Involving Benzamidinate Ligands. Reduction of [PhC(NSiMe3)2]2ZrCl2 to Form an Imido-Iminoacyl Compound by Carbon–Nitrogen Bond Cleavage,’’ Organometallics, 13 4670–4672 (1994). 59. I. Castillo and T. D. Tilley, ‘‘Hafnium Imido Complexes Containing Silyl Ligands,’’ J. Organomet. Chem., 643 431–440 (2002).
481
60. Y. Bai, H. W. Roesky, M. Noltemeyer, and M. Witt, ‘‘Synthese und Strukturen von (Monoorganyl)amiden und -imiden des Zirkoniums und Hafniums,’’ Chem. Ber., 125 825–830 (1992). 61. C. P. Schaller, C. C. Cummins, and P. T. Wolczanski, ‘‘Hydrocarbon Activation via Reversible 1,2-RH-Elimination from (tBu3SiNH)3ZrR) Synthetic, Structural, and Mechanistic Investigations,’’ J. Am. Chem. Soc., 118 591–611 (1996). 62. C. Airoldi, D. C. Bradley, H. Chudzynska, M. B. K. Hursthouse, K. M. A. Malik, and P. R. Raithby, ‘‘Preparation, Properties, and Crystal Structures of the Monochlorotris[bis(trimethylsilyl)amido]-Compounds of Titanium, Zirconium, and Hafnium,’’ J. Chem. Soc. Dalton Trans. 110 2010–2015 (1980). 63. R. J. Morris and G. S. Girolami, ‘‘Permethylmanganates. Synthesis and Characterization of Divalent [MnMe42-], Trivalent [MnMe52-], and Tetravalent [MnMe62-],’’ J. Am. Chem. Soc., 110 6245–6246 (1988). 64. P. M. Morse and G. S. Girolami, ‘‘Are d0 ML6 complexes always octahedral? The X-Ray Structure of Trigonal-Prismatic [Li(tmed)]2[ZrMe6],’’ J. Am. Chem. Soc., 111 4114–4116 (1989). 65. Q. D. Shelby and G. S. Girolami, ‘‘Synthesis, Dynamic Behavior, and X-Ray Crystal Structure of [Li(tmed)]3[WH(C2H4)4],’’ Organometallics, 18 2297– 2299 (1999). 66. S. Imamura, T. Yamashita, K. Utani, H. Kanai, and K. Hamada, ‘‘The Structure of Zirconium Oxide in Silica Matrix and its Epoxidation Activity,’’ React. Kinet. Catal. Lett., 72 11–20 (2001). 67. P. Li, I. W. Chen, and J. E. Penner-Hahn, ‘‘X-Ray Absorption Studies of Zirconia Polymorphs. I. Characteristic Local Structures,’’ Phys. Rev. B, 48 10063–10073 (1993). 68. F. Farges, G. E. Brown, and D. Velde, ‘‘Structural Environment of Zr inTwo Inosilicates from Cameroon: Mineralogical and Geochimical Implications,’’ Am. Mineral., 79 838–847 (1994). 69. G. Mountjoy, et al., ‘‘Changes in the Zr Environment in ZirconiaSilica Xerogels with Composition and Heat Treatment as Revealed by ZrK-Edge XANES and EXAFS,’’ Phys. Chem. Chem. Phys., 2 2455–2460 (2000). 70. M. Bauer, C. Gastl, C. Koppl, G. Kickelbick, and H. Bertagnolli, ‘‘EXAFS Spectroscopy of the Alkoxide Precursor Zr(OnBu)4 and its Modification in Solution,’’ Monatsh. Chem., 137 567–581 (2006). 71. J. H. Choy, J. B. Yoon, H. Jung, and J. H. Park, ‘‘Structural Evolution of SiO2–ZrO2 Nano-Sol Intercalated Clays Upon Pillaring Reaction,’’ J. Mater. Chem., 13 557–562 (2003). 72. M. Bauer, S. Muller, G. Kickelbick, and H. Bertagnolli, ‘‘The Structures of the Precursor Hf(OnBu)4 and its Modification in Solution: EXAFS-Investigation in Combination with XANES- and IR-Spectroscopy,’’ New J. Chem., 31 1950–1959 (2007). 73. G. Mountjoy, et al., ‘‘Structural Characterization of Mixed (TiO2)(x)(ZrO2)(y)(SiO2)(1xy) Sol–Gels (0.05ox, yo0.15) by a Combination of X-Ray and Spectroscopy Techniques,’’ J. Phys. Chem. B, 107 7557–7566 (2003). 74. I. Ahmed, et al., ‘‘Proton Conductivity and Low Temperature Structure of InDoped BaZrO3,’’ Solid State Ionics, 177 2357–2362 (2006). 75. A. N. Christensen, ‘‘A Neutron Diffraction Investigation on Single Crystals of Titanium Carbide, Titanium Nitride and Zirconium Nitride,’’ Acta Chem. Scand. A, 29 563–568 (1975). 76. T. W. Baker, ‘‘The Coefficient of Thermal Expansion of Zirconium Nitride,’’ Acta Cryst., 11 300–301 (1958). 77. A. N. Christensen, ‘‘A Neutron Diffraction Investigation on Single Crystals of Titanium Carbide, Titanium Nitride, and Zirconium Nitride,’’ Acta Chem. Scand., 44 851–852 (1990). 78. M. Lerch, E. Fueglein, and J. Wrba, ‘‘Synthesis, Crystal Structure, and High Temperature Behavior of Zr3N4,’’ Z. Anorg. Allg. Chem., 622 367–372 (1996). 79. T. Bredow and M. Lerch, ‘‘Anion Distribution in Zr2ON2,’’ Z. Anorg. Allg. Chem., 630 2262–2266 (2004). 80. A. Zerr, G. Miehe, and R. Riedel, ‘‘Synthesis of Cubic Zirconium and Hafnium Nitride Having Th3P4 Structure,’’ Nat. Mater., 2 185–189 (2003). 81. E. Rudy, ‘‘The Crystal Structures of Hf3N2 and Hf4N3,’’ Metall. Trans., 1 1249–1252 (1970). 82. B. D. Lichter, ‘‘Precision Lattice Parameter Determination of Zr–O Solid Solutions,’’ Trans. Met. Soc. AIME, 218 1015–1018 (1960).
