佐藤 智司

The liquid-phase diffusion coefficient of nickel nitrate in wet silica gels with different Si content as well as in mesoporous silica gel was measured by the flow-cell method, where absorbance of a gel plate placed in a flow cell was monitored as a function of time. For diffusion in the mesoporous silica gel with pore diameter of 12 run, the tortuosity was as low as 1.2, suggesting that the liquid-phase diffusion of nickel nitrate is not restricted in the large mesopores. In contrast, the diffusion coefficient in wet silica gels monotonously decreased with increasing Si content in the gels, and the tortuosity in the wet gel, 1.7-3.5, was larger than that in the mesoporous silica gel. The wet gel is considered to consist of crosslinked polymeric network of silica. The tortuosity in the wet gel would vary depending on the mesh size in the network and on the number of free side chain.

A porous titania film was prepared on a glass substrate by a spin-coating technique from the precursor solution containing titanium tetraisopropoxide and stearic acid. Stearic acid formed a 2 : 2 complex with titanium in the dried film, and the complex melted before decomposition. Before the coating of the precursor solution, methylsilane precoating on the hydrophilic glass surface was effective in preparing a uniform film because of hydrophobicity of the complex melt. The resultant film was transparent and composed of anatase after heating at 500 degreesC. Micrometer-scale rough surface with aggregates of secondary particles composed of primary crystallites was observed in the film, and mesopores seem to be located at the inter-crystallites. It is probable that decomposition of stearic acid and aggregation of anatase crystallites occur simultaneously to provide the porous structure during heating.

An amorphous silica with distinct bimodal pore structure containing both macropores and mesopores was prepared from an acidified solution of water glass in the presence of poly(acrylic acid) by inducing phase separation together with gelation. Both of pore sizes of the macropores and the mesopores can be controlled in the ranges from 1 to 10 mum and from 4 to 30 mm, respectively.

A Ni/SiO2 catalyst was prepared by homogeneous precipitation of nickel hydroxide in a sol-gel-derived wet silica gel. The preparation process consists of two successive steps: gelation of silica in the presence of nickel nitrate and urea at 50 degreesC, followed by aging at higher temperature, typically at 80 degreesC, to decompose the urea. The decomposition of urea increases the pH of the solution in the wet gel, leading to the concurrence of structural rearrangement of silica gel and deposition of nickel species. As a result, the structure of the silica changes from a ramified polymeric network into particle aggregates that entrap the nickel cations in the particles. The resulting Ni/SiO2 contains large mesopores that have high thermal stability up to 1000 degreesC and highly dispersed Ni metal particles with typical crystallite size of 4 nm even at high Ni content at 20 wt%.

The dehydrogenation of cyclohexene over carbon deposited on alumina, C/Al2O3, which was prepared by contacting hydrocarbons such as cyclohexane, cyclohexene and cyclohexanone with alumina surface at greater than or equal to 823 K, was investigated under non-oxidative conditions. The deposited carbon was identified as condensed aromatic hydrocarbon with graphite-like structure in the results of XRD and C-13 NMR measurements. It was found that the dehydrogenation activity was generated on alumina by carbon deposition: the dehydrogenation of cyclohexene proceeded over C/Al2O3 at 773 K, while the isomerization to methylcyclopentenes proceeded over pure alumina without carbon. The selectivity to dehydrogenated products such as benzene and 1,3-cyclohexadiene was maximized at a specific carbon content. The disproportionation of cyclohexene into benzene and cyclohexane proceeded over C/Al2O3 above the specific carbon content. Unpaired electrons observed in the EPR measurement are related to the ratio of dehydrogenation and disproportionation, whereas the dehydrogenation activity is not related with the unpaired electron content.

A supported type of silica-alumina catalyst prepared by depositing alumina on silica was treated under steaming conditions. Physical properties of the silica-alumina were characterized by the N-2 adsorption-desorption isotherm, small-angle X-ray scattering and solid-state NMR measurements. The particle size and the mean pore size increased slightly with increasing steam pressure up to 2.0 MPa. The specific surface area of the silica-alumina steamed at 2.0 MPa was half of that before steaming. The number of four-coordinate Al ions increased with increasing steam pressure and was saturated at a steam pressure of 0.5 MPa. Temperature-programmed desorption of chemisorbed 2,6-dimethylpyridine elucidated that the number of Bronsted acid sites is maximized at 0.5 MPa.

A porous titania film was prepared on a glass substrate by a spin-coating technique from the precursor solution containing titanium tetraisopropoxide and stearic acid. Stearic acid formed a 2:2 complex with titanium in the dried film, and the complex melted before decomposition. Before the coating of the precursor solution, methylsilane precoating on the hydrophilic glass surface was effective in preparing a uniform film because of hydrophobicity of the complex melt. The resultant film was transparent and composed of anatase after heating at 500℃. Micrometer-scale rough surface with aggregates of secondary particles composed of primary crystallites was observed in the film, and mesopores seem to be located at the inter-crystallites. It is probable that decomposition of stearic acid and aggregation of anatase crystallites occur simultaneously to provide the porous structure during heating.

Mesoporous titania with regulated pore size as well as high specific surface area was prepared from titanium alkoxide and various carboxylic acids with different alkyl-chain length [CH3(CH2)(n)COOH : n = 0-20]. Although the pore size and pore volume of titania changed slightly at n < 10, they increased with increasing the alkyl-chain length of carboxylic acid at n greater than or equal to 10. Each carboxylic acid forms a complex with titanium alkoxide in an organic-inorganic precursor, while the precursor prepared by using carboxylic acid with n greater than or equal to 10 consists of a lamella-type mesophase. The interplanar distance of the lamella structure increases with increasing the alkyl-chain length of carboxylic acid. During calcination, the lamella structure collapses together with the combustion of organic moieties and titania particles crystallized to anatase aggregate to provide pores. Although carboxylic acid does not act as a true template, the formation of the lamella-type mesophase in the precursor plays an important role in controlling the pore size of titania.

Structural formation process of Ni/SiO2 and Cu/SiO2 catalysts prepared by solution exchange of wet silica gel was investigated. Microstructures of Cu/SiO2 and Ni/SiO2 were quite different from each other. In the case of Cu/SiO2, Cu particles with diameter of ca. 3-5 nm dispersed homogeneously at less Cu content, and the particle size of Cu as well as pore size of silica gel support increased with increasing Cu content. In the Ni/SiO2, the Ni particles with diameter of ca. 6-10 nm gathered densely to form aggregates in silica matrix resulting in sea-island structure, whereas the size of Ni particle slightly increased with increasing Ni content. The difference in the structure of the metal-silica composites is probably caused by the difference in interaction between silica gel network and metal ions during drying and heating processes.

The pore formation process of silica gel from tetraethoxysilane in the presence of citric acid (CA) was investigated. The silica gel prepared by pyrolysis of a composite of CA and silica had high specific surface area, ca. 1000 m(2) g(-1) and its pore volume increased linearly with increasing CA content. In addition, the average pore size successively increased from the micropore range to the mesopore range of ca. 10 nm with the increase in the CA content. The silica and the CA-silica composite were amorphous without long-range ordering. In the composite, amorphous CA and polymeric silica gel are mixed on a nanometer scale without the aid of specific chemical bonding, except for weak hydrogen bonding. Because of the nature of the composite, the silica gel network becomes rigid through the additional formation of Si-O-Si bonds during heating, irrespective of the existence of CA. Therefore, the bulky structure is retained without shrinkage after elimination of CA, and silica gel with high specific surface area and pore volume results. Thus, CA is considered to provide mesopores as a template.

A Cu metal surface was evaluated by a novel technique combining temperature-programmed reduction (TPR) measurement with N2O oxidation. The technique consists of three steps: the usual TPR measurement, the oxidation of the Cu surface by N2O, and the subsequent TPR measurement. The surface Cu oxidized by N2O was determined as a ratio of the peak area of the second TPR profile to that of the first one. It was found that bulk oxidation gradually proceeds after surface oxidation even at 30 degreesC. After the surface oxidation, the Cu2O produced by N2O oxidation varied with N2O exposure time (t) and had a linear correlation with ta at temperatures below 100 degreesC. The linear correlation in the parabolic plot proves that bulk oxidation proceeds through the diffusion process, and the Y-intercept corresponds to the surface oxidation. Both the dispersion and the Cu metal surface area of the sample were calculated from the intercept in the parabolic plot for the Cu2O produced by N2O oxidation. In addition, we found that a very large Cu metal surface area, as high as 32 m(2) g(cat)(-1), was created on a Cu-MgO catalyst through a citrate process using a molten mixture of copper nitrate, magnesium nitrate, and citric acid. (C) 2000 Academic Press.

Preparation of high-surface-area SiO2-ZrO2 by deposition of silicate species on zirconia under hydrothermal conditions in an aqueous ammonia solution was investigated. A fresh precipitate of ZrO(OH)(2) was heated in a pressure vessel with several pieces of quartz glass tube at 100 degreesC. In the measurement of ICP-AES and solid-state Si-29 NMR, it was elucidated that a silica component dissolved from the glass in the basic solution was deposited on the ZrO(OH)(2) precipitate and that the silica content increased with increasing hydrothermal period. The resulting SiO2-ZrO2 had a specific surface area higher than 240 m(2) g(-1) even after calcination at 500 degreesC. In contrast to the pure ZrO2 particles readily aggregating, the silicate species deposited on the surface of ZrO2 particles prevented the agglomeration of the primary particles during calcination, resulting in high-surface-area SiO2-ZrO2. (C) 2000 Academic Press.

Liquid-phase hydrogenation of cyclohexanone, acetone, 2-butanone, 3-pentanone, and 4-heptanone to the corresponding secondary alcohols was investigated over various porous Ni catalysts at 0 degrees C under a hydrogen pressure of 1.1 MPa. Pore size distributions as well as Ni surface area of Ni-MgO catalysts, which were prepared from a melt of the corresponding nitrates and citric acid, with high Ni contents of 60-80 wt% were controlled by the calcination temperature of the precursors. For the hydrogenation of acetone, reaction rate constants were directly proportional to the Ni surface areas of the catalysts, and Raney nickel which had the largest Ni surface area showed the highest reaction rate. For the hydrogenation of other reactants larger than acetone in molecular size, however, rate constants do not have a simple linear correlation with Ni surface area. Ni-MgO catalysts with large mesopores exhibited reaction rates higher than those of Raney nickel catalysts with the largest Ni surface areas. Assuming that diffusion of both reactants and products is restricted in small pores such as in Raney nickel, we tried to evaluate an effective pore size for the liquid-phase mass transfer in porous materials by a novel approach analyzing reaction rate data coupled with pore size distribution and hydrogen chemisorption data. Cumulative Ni surface areas were calculated by multiplying the Ni surface area by a fraction of cumulative surface area located in pores larger than a specific size to the total surface area, and relationship between the cumulative Ni surface areas and the reaction rate constants were examined. It was found that the rate constants for the hydrogenation of 2-butanone, cyclohexanone, 3-pentanone, and 4-heptanone were proportional to cumulative Ni surface areas in pores larger than critical sizes of 2.0, 2.3, 3.2, and 3.7 nn in radius, respectively. It has been consequently elucidated that the mass transfer of the reactants is restricted in pores smaller than a critical size that depends on the size of the reactants. (C) 2000 Academic Press.

A microporous silica gel was prepared by a sol-gel process in the presence of ethylene glycol (EG) and its oligomers (oligoEG), and the pore formation process was investigated by using NMR and TG-DTA. The micropore volume changed with the polymerization degree of oligoEG, and showed a minimum at the trimer of EG. An attractive interaction between silanol and organic additives plays a key role in micropore formation: the interaction inhibits the bond formation of Si-O-Si during drying and heating, and heterogeneous domains which convert into micropore are formed after the combustion of additives. There are two kinds of interaction: one is eater (Si-OR) bond formation of a silanol with an end hydroxy group of additives; the other is hydrogen bonding between a silanol and an ether oxygen of additives. The former interaction is significant for EG. With increasing length of oligoEG, however, the former interaction rapidly disappears, while the latter becomes stronger. Because of a minimum in the combination of these interactions, the minimal volume of the micropore is observed.

Silica-supported nickel catalysts with both macropores and mesopores were prepared in an alkoxide-derived sol-gel process in the presence of poly(ethylene oxide) (PEO) with an average molecular weight of 100000. In this process, the interconnected macroporous morphology is formed when transitional structures of spinodal decomposition are frozen by the sol-gel transition of silica. The addition of nickel into a silica-PEO system has negligible effect on the morphology formation suggesting that phase separation in the nickel-containing system proceeds by repulsive interaction between the solvent and PEO adsorbed on the silica, as observed in the system without nickel. In gel formation, it was found that the Ni was distributed selectively in mesopores in the silica gel skeleton as fine particles rather than aggregated in macropores. It is considered that PEO interacts with both silica and nickel cations. The interaction between PEO and Ni makes nickel distribute in the silica phase and keeps Ni from aggregation during drying, resulting in a high dispersion of Ni.

A novel technique for the measurement of the diffusion coefficient of a solute in a transparent wet-gel monolith was developed by analyzing the decay behavior of the absorbance of the gel plate in a flow cell. The diffusion coefficient of a Ni(NO3)(2) in a wet silica gel is estimated by fitting the absorbance-time data with an exponential function. The efficiency of the technique was ensured by testing the effects of several measurement conditions, such as the flow rate of the external solution and the thickness of the gel plate. The diffusion coefficient of Ni(NO3)(2) in wet silica gel is 6.0x10(-10) s(-1) at 20 degrees C, which is ca. 60% of that in bulk solution. The data suggest that the silica gel network allows relatively fast transport of the liquid phase.

Pore size and pore volume of an amorphous silica with high specific surface area, ca, 1000 m(2)g(-1), can be controlled in meso scale by using citric acid as a template in sol-gel reaction of tetraethoxysilane.

Vapor-phase alkylation of phenol with methanol was investigated over CeO2-MgO catalysts prepared utilizing a molten mixture of the corresponding nitrates and citric acid. The CeO2-MgO had attractive catalytic performance without decay of activities at the temperature range between 450 and 550 degrees C, and it had excellent selectivities to the sum of o-cresol and 2,6-xylenol higher than 98%. The CeO2-MgO catalysts were found to be mixtures of MgO and an interstitial solid solution of MgxCe1-x/2O2 asa result of XRD measurement. It is confirmed that citric acid used in the preparation heightens the dispersion of the solid solution in the MgO matrix. The pure CeO2, which also exhibited efficient ortho-selectivity, had only weak basic sites in the TPD experiment of adsorbed CO2, while the pure MgO with strong basicity showed very low reaction rate in the methylation. The solid solution of MgCe1-x/2O2 in the CeO2MgO catalyst probably provides active centers for the methylation of phenol. In the results of methanol decomposition, methanol was converted into CO, CO2, and CH4 over the CeO2-MgO catalysts, without producing dimethyl ether. The reaction mechanism of the ortho-methylation over the CeO2-MgO catalyst is speculated: the ortho position of phenol adsorbed perpendicularly on weak basic sites on the MgxCe1-x/2O2 solid solution is selectively alkylated by methanol which is possibly activated in the form of formyl or hydroxy methyl group rather than methyl cation. (C) 1998 Academic Press.

An application of Ni/MgO catalyst which was prepared by using citric acid to the hydrogenation of carbon dioxide was investigated. The Ni/MgO catalysts with Ni content higher than 30 wt% were found to exhibit high catalytic activities for the hydrogenation of carbon dioxide due to the large Ni metal surface area. The usage of citric acid realizes the formation of NiO-MgO solid solution at a low temperature of 773 K. Aggregation of Ni metal particle is probably depressed by MgO highly dispersed in the Ni particle during the reduction of the NiO-MgO solid solution. A structural model of the Ni catalyst is proposed.

CuO-Al2O3 systems with different Cu/Al ratios were prepared by the amorphous citrate process, and their catalytic activities for the dehydrocoupling of methanol to methyl formate were examined, In the CuO-Al2O3 systems, CuAl2O4 samples calcined at temperatures below 1000 degrees C were partially reduced during the reaction at 310 degrees C and decreased the selectivity to methyl formate (MF) with increasing Cu(0) content. Cu(0) species catalyzed the dehydrocoupling reaction without decomposition of methyl formate at temperatures below 210 degrees C, whereas Cu(0) decomposed methyl formate into methanol and CO above 210 degrees C. On the other hand, CuAl2O4 calcined at 1100 degrees C showed the highest activity of the formation of methyl formate at a reaction temperature of 310 degrees C. Since the CuAl2O4 sample was not reduced even at 310 degrees C, Cu(II) species in the CuAl2O4 were found to be effective sites for the dehydrocoupling at temperatures above 250 degrees C. In addition, Cu(I) species in CuAlO2 was also found to be active with keeping high MF selectivity up to 290 degrees C. (C) 1997 Academic Press.

Various acetal formation reactions were examined by a Keggin-type heteropolyacid. The reaction kinetics of the acetal formation was also investigated through comparison of heteropolyacids with other acid catalysts in the reaction between cyclohexanone and 1,2-butandiol to produce cyclohexanone ethylethylene acetal. Tungstosilicic acid was found to have a superior catalytic activity compared to other acids. For several reactions between a carbonyl compound and a dialcohol, the reaction mixture was spontaneously separated into two liquid phases, a product phase and a catalyst phase, with high yields of acetal. The recovered catalyst phase was repeatedly applicable to the reaction for the formation of cyclohexanone ethylethylene acetal. The phase separation is concluded to be caused by exclusion of the product acetal from the catalyst phase. Although the catalytic activity was strongly depressed by water produced during the reaction, tungstosilicic acid was found to have a high resistance to water, compared to other acid catalysts.

At high Ni contents ranging from 70 to 80 wt.%, Ni-MgO catalysts prepared by using a melt of the corresponding nitrates and citric acid were found to have high nickel surface area, and exhibited catalytic activities for the hydrogenolysis of cyclohexane at 350 degrees C.

Both surface borate structures and acidities of alumina-boria catalysts prepared by supporting B2O3 (boria) on an alumina support were investigated by using both B-11 magic angle spinning (MAS) NMR measurements and temperature-programmed desorption (TPD) of adsorbed pyridines, together with the catalytic properties of these alumina-boria materials for l-butene isomerization. The concentrations of tetrahedral oxygen-coordinated borate (BO4) species as well as number of Bronsted acid sites were found to increase with increasing boria content. The catalytic activity for l-butene isomerization was found to be caused by the strong Bronsted acid sites generated from BO4 species on the surface of alumina.

The structures of silica-grafted aluminum chloride catalysts were investigated by solid-state NMR measurements, together with their catalytic properties in the alkylation of benzene with cyclohexene. Catalyst samples prepared by contacting aluminum trichloride vapor with a silica gel surface at temperatures between 140 and 250 degrees C were found to be active for the alkylation, while other samples prepared at higher temperatures were inactive. The grafted aluminum chloride species were classified into two structural types, corresponding to monomeric and dimeric aluminum chloride. The combination of Al-27 and Si-29 NMR experiments elucidates important structural features of these aluminum chloride/silica systems. The active catalysts were found to have a dimeric aluminum chloride structure, with chemical shifts of 85 and 42 ppm in the Al-27 magic-angle spinning NMR spectra. Those peaks correspond to aluminum species coordinated to chloride and/or oxygen with coordination numbers of 4 and 5, respectively. It is speculated that the catalytically active species is the product of reaction of the aluminum trichloride dimer with one hydroxyl on the silica surface.

The acid-catalysed cyclotrimerization of aliphatic aldehydes has been examined through comparison of heteropoly acids with other acid catalysts. A Keggin-type heteropoly acid such as phosphomolybdic acid catalyses the cyclotrimerization of aldehydes, such as ethanal, propanal, butanal, 2-methylpropanal, 2,2-dimethylpropanal, hexanal, octanal, and decanal, to produce the respective 2,4,6-trialkyl-1,3,5-trioxanes in high yields. Catalyst turnover number of the heteropoly acid is more than 10 000 for propanal cyclotrimerization. In addition to the high catalytic activities, the reaction mixture spontaneously separates into two phases, a product phase and a catalyst phase, at high conversions of aldehyde. For propanal cyclotrimerization, the reaction mixture separates into two liquid phases, and the recovered catalyst phase may be repeatedly applied to the reaction without additional care in isolation of the catalyst. The phase separation phenomenon has been concluded to be caused by the insolubility of the heteropoly acid coordinated with propanal in the product 2,4,6-triethyl-1,3,5-trioxane.

Various Cu-Ni-SiO2 catalysts were prepared by alkoxide method. These catalysts showed higher activities for hydrogenation of benzene or dehydrogenation of cyclohexane than those by impregnation method. The surface properties of catalysts were characterized by means of various instruments such as XPS, SEM, EPMA, XRD, and TPR.

A supported type of silica-alumina catalyst prepared by depositing silica on an alumina (CVD SiO2/Al2O3) was characterized by solid-state magic-angle spinning (MAS) NMR measurement. The P-31 MAS NMR of trimethylphosphine chemisorbed on the catalyst elucidated a change in acid properties (Bronsted or Lewis type) with silica loading. Lewis acid sites on the alumina surface were converted into Bronsted acid sites by the deposition of silica. At lower silica loadings, both Bronsted and Lewis acid sites existed. Above 12 wt.% silica loading, Lewis acid sites disappeared and only Bronsted acid sites were observed. The catalytic activities of CVD SiO2/Al2O3 for cumene cracking and 1-butene isomerization varied with silica loading, and the catalytic behaviors were correlated with the amount of Bronsted acid sites generated by silica deposition. The Si-29 MAS NMR of silica deposited on the alumina surface revealed the microstructures of silica. The silica which was highly dispersed, forming a silica bilayer, exhibited high catalytic activities and strong Bronsted acidity. Further silica layering decreased both the catalytic activities and the acidity.

A temperature-programmed desorption of adsorbed dimethylpyridine clarified the characteristic changes in acid type, acid strength, and acid amount of silicaalumina prepared by depositing silica on an alumina support. Lewis acid sites of the alumina changed into Bronsted acid sites together with a change in the acid strength with increasing silica deposition.