Facile Synthesis of Pure Monoclinic and Tetragonal Zirconia

Pure monoclinic (m) and tetragonal (t) zirconia nanoparticles were readily synthesized ... zirconium salts (e.g., hydrated zirconyl nitrate) and urea ...
0 downloads 0 Views 1MB Size
Download PDF
8358

Langmuir 2008, 24, 8358-8366

Facile Synthesis of Pure Monoclinic and Tetragonal Zirconia Nanoparticles and Their Phase Effects on the Behavior of Supported Molybdena Catalysts for Methanol-Selective Oxidation Weizhen Li, Hua Huang, Hongjia Li, Wei Zhang, and Haichao Liu* Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Stable and Unstable Species, College of Chemistry and Molecular Engineering, Green Chemistry Center, Peking UniVersity, Beijing 100871, China and Shanghai Key Laboratory of Green Chemistry and Green Processes, Department of Chemistry, East China Normal UniVersity, Shanghai 200062, China ReceiVed February 2, 2008. ReVised Manuscript ReceiVed April 28, 2008 Pure monoclinic (m) and tetragonal (t) zirconia nanoparticles were readily synthesized from the reaction of inorganic zirconium salts (e.g., hydrated zirconyl nitrate) and urea in water and methanol, respectively, via a facile solvothermal method. The role of the solvents was crucial in the formation of the pure ZrO2 phases, whereas their purity was essentially insensitive to other variables, including reaction temperature, reactant concentration, pH, and zirconium salts. Water as the solvent led to the transformation of hydrous ZrO2 precipitates initially formed with tetragonal structures to thermodynamically more stable m-ZrO2 via the dissolution-precipitation process, whereas methanol favored the removal of water molecules from the precursors via their reaction with urea, consequently maintaining the tetragonal structures. The obtained tetragonal samples were found to possess superior hydrothermal stability compared to those reported previously, which provides the possibility for systematically studying the effects of ZrO2 phases on many catalytic reactions involving water as a reactant or product. Using these pure m- and t-ZrO2 phases as supports, dispersed MoOx catalysts were synthesized at MoOx surface densities of ∼5.0 Mo/nm2, which is close to one monolayer of coverage. Characterization by X-ray diffraction and Raman spectroscopy confirmed that the pure ZrO2 phases remained unchanged in the presence of the MoOx domains and the MoOx domains existed preferentially as 2D polymolybdate structures. The catalysts were subsequently examined for selective methanol oxidation as a test reaction. m-ZrO2 support led to 2-fold greater oxidation rates than for t-ZrO2 support, reflecting the higher intrinsic reactivity of the MoOx domains on m-ZrO2. This is consistent with their higher reducibility probed by temperatureprogrammed reduction with H2 (H2 TPR). These observed effects of the ZrO2 phases provide the basis for designing catalysts with tunable redox properties and reactivity.

1. Introduction Zirconia(ZrO2), because of its surface characteristics of possessing both acid-basic and redox functions,1 has been widely used as a catalyst and catalyst support in numerous reactions ranging from alcohol dehydration, CO/CO2 hydrogenation, and alkane isomerization to the selective oxidation of alcohols and alkanes.1–7 The crystal phase of ZrO2 (monoclinic and tetragonal) strongly influences the catalyst activities and selectivities.3–5,7 Li et al. found that monoclinic ZrO2 (m-ZrO2) favors the synthesis of isobutene from CO hydrogenation, whereas ethylene and propylene are the main products on tetragonal ZrO2 (t-ZrO2).5 Bell and co-workers reported that Cu/ZrO2 catalysts with m-ZrO2, compared to those with t-ZrO2, are nearly an order of magnitude more active for methanol synthesis from CO and H2 and also exhibit higher methanol selectivity.3 For acid-catalyzed n-alkane isomerization, sulfated t-ZrO2 was found to be 2-5 times more active than sulfated m-ZrO2.4 These marked differences in catalytic performance clearly necessitate the synthesis of pure m- and t-ZrO2 phases, for which * To whom correspondence should be addressed. E-mail: [email protected]. Tel/Fax: 86-10-6275 4031. (1) (a) Tanabe, K. Mater. Chem. Phys. 1985, 13, 347–364. (b) Yamaguchi, T. Catal. Today 1994, 20, 199–217. (2) Asakura, K.; Aoki, M.; Iwasawa, Y. Catal. Lett. 1988, 1, 395–403. (3) Rhodes, M. D.; Bell, A. T. J. Catal. 2005, 233, 198–209. (4) Stichert, W.; Schuth, F.; Kuba, S.; Knozinger, H. J. Catal. 2001, 198, 277–285. (5) Li, W.; Yin, Y. Q.; Gao, R. X.; Hou, R. L.J. Mol. Catal. (China) 1999, 13, 186–192. (6) Li, W.; Liu, H.; Iglesia, E. J. Phys. Chem. B 2006, 110, 23337–23342. (7) He, D.; Ding, Y.; Luo, H.; Li, C. J. Mol. Catal. A 2004, 208, 267–271.

a variety of synthesis approaches have been accordingly employed to date. m-ZrO2 is frequently prepared via the hydrolysis of zirconyl chloride under reflux or hydrothermal conditions.4,8 Jung and Bell obtained m-ZrO2 under reflux at pH 1.5 for 240 h,8 and Stichert et al. obtained m-ZrO2 by hydrothermal treatment at pH ∼5 for about 1500 h.4 Unlike m-ZrO2, t-ZrO2 is metastable at room temperature and tends to transform to m-ZrO2 upon thermal treatment or contact with water vapor,9,10 which imposes a difficulty in synthesizing pure t-ZrO2 in the absence of any stabilizing components such as Y3+ and Mg2+.11 The synthesis of undoped t-ZrO2 has been achieved generally by using zirconium alkoxides as precursors via sol-gel, solvothermal, and spraypyrolysis methods.11–14 The use of inorganic zirconium salts (e.g., zirconyl chloride and zirconium nitrate) was also explored to prepare pure t-ZrO2.8,15 For example, Jung and Bell obtained t-ZrO2 by refluxing an aqueous solution of zirconyl chloride at (8) Jung, K. T.; Bell, A. T. J. Mol. Catal. A 2000, 163, 27–42. (9) Xie, S.; Iglesia, E.; Bell, A. T. Chem. Mater. 2000, 12, 2442–2447. (10) Murase, Y.; Kato, E. J. Am. Ceram. Soc. 1983, 66, 196–200. (11) Djurado, E.; Meunier, E. J. Solid. State. Chem. 1998, 141, 191–198. (12) (a) Shukla, S.; Seal, S.; Vanfleet, R. J. Sol-Gel Sci. Technol. 2003, 27, 119–136. (b) Shukla, S.; Seal, S.; Vij, R.; Bandyopadhyay, S. J. Nanopart. Res. 2002, 4, 553–559. (c) Shukla, S.; Seal, S. Int. Mater. ReV. 2005, 50, 45–64. (d) Shukla, S.; Seal, S. J. Phys. Chem. B 2004, 108, 3395–3399. (e) Shukla, S.; Seal, S.; Vij, R.; Bandyopadhyay, S.; Rahman, Z. Nano Lett. 2002, 2, 989–993. (f) Shukla, S.; Seal, S. ReV. AdV. Mater. Sci. 2003, 4, 123–126. (13) Joo, J.; Yu, T.; Kim, Y. W.; Park, H. M.; Wu, F.; Zhang, J. Z.; Hyeon, T. J. Am. Chem. Soc. 2003, 125, 6553–6557. (14) Sun, Q.; Zhang, Y.; Deng, J.; Chen, S.; Wu, D. Appl. Catal., A 1997, 152, L165-L171. (15) (a) Chuah, G. K.; Jaenicke, S.; Cheong, S. A.; Chan, K. S. Appl. Catal., A 1996, 145, 267–284. (b) Chuah, G. K.; Jaenicke, S.; Pong, B. K. J. Catal. 1998, 175, 80–92.

10.1021/la800370r CCC: $40.75  2008 American Chemical Society Published on Web 06/27/2008

Synthesis of Pure Zirconia Nanoparticles

pH 10 for 240 h.8 However, most of these synthesis routes to pure m- and t-ZrO2 are complicated and require tedious procedures, close control of reaction parameters such as temperature, pressure, and pH, or the use of relatively expensive and toxic zirconium alkoxides. In this context, simple, easy synthesis routes are highly desired, especially for the practical applications of ZrO2 nanoparticles of pure phases. In this work, we report the facile and reproducible synthesis of pure m- and t-ZrO2 phases via a solvothermal method. We choose inexpensive and nontoxic inorganic zirconium salts (e.g., zirconyl nitrate) and obtain pure m- and t-ZrO2 phases with average diameters of ∼5.9 and 5.3 nm and narrow particle size distributions under mild hydrothermal and methanothermal conditions, respectively. These pure ZrO2 phases are then used to prepare ZrO2-supported MoOx catalysts with 2D polymolybdate structures, aimed at examining intrinsic effects of the ZrO2 crystal phase on the catalytic oxidation of methanol, which are effects that have not yet been reported in the literature.

2. Experimental Section 2.1. Sample Synthesis. ZrO2 samples were synthesized at 120-200 °C under autogenous pressure for 2-20 h in a Teflonlined stainless-steel autoclave (100 mL) containing solutions (80 mL) of urea (CO(NH2)2; >99.9%, Beijing Chemicals) and zirconyl nitrate (ZrO(NO3)2 · 2H2O; >45.0% ZrO2, Beijing Chemicals). Deionized water and methanol (>99.9%, Beijing Chemicals) were used as solvents for synthesizing m-ZrO2 and t-ZrO2, respectively. The concentrations of Zr4+ in the solutions varied in the range of 0.2-0.6 M, and the urea/ Zr4+ molar ratios varied in the range of 0-10. The resulting precipitates were washed thoroughly with water and methanol and treated at 110 °C overnight in ambient air and then at 400 °C for 4 h in dry air and nitrogen for m-ZrO2 and t-ZrO2, respectively. ZrO2-supported MoOx (MoOx/ZrO2) catalysts were prepared by incipient wetness impregnation of as-synthesized ZrO2 (after drying in ambient air at 110 °C overnight) with aqueous solutions of ammonium heptamolybdate ((NH4)6Mo7O24 · 4H2O; >99.0%, Beijing Chemicals), followed by treatment at 110 °C overnight and then at 600 °C for 6 h in air. The resulting catalysts were denoted as MoOx/ m-ZrO2 and MoOx/t-ZrO2, depending on the crystal phases of the ZrO2 supports, m-ZrO2 and t-ZrO2. The nominal surface density of MoOx (Mo/nm2) for the catalysts was estimated from their measured Mo contents (7.5-8.5 wt % Mo) and BET surface areas (88-94 m2/g). 2.2. Sample Characterization. BET surface areas of the samples were measured by nitrogen physisorption at 78.3 K on an ASAP 2010 analyzer (Micromeritics). Prior to the measurements, the samples were treated at 200 °C for 4 h under dynamic vacuum conditions. X-ray diffraction (XRD) measurements were obtained using a Rigaku D/Max-2000 diffractometer with a slit width of 0.5° using Cu KR radiation (λ ) 1.5406 Å), operated at 30 kV and 100 mA. The 2θ angles were scanned from 10 to 80° at a rate of 2° min-1. The average crystallite sizes (D) were estimated by the Scherrer equation,18 D ) 0.90λ/β cos θ, where θ is the diffraction angle and β is the fwhm. The volume fractions of monoclinic (Vm) and tetragonal (Vt) phases were calculated in the way reported previously,3,8,16 on the basis of the integrated intensities of (111) and (-111) peaks of m-ZrO2 and the (111) peak of t-ZrO2, (16) Toraya, H.; Yoshimura, M.; Somiya, S. J. Am. Ceram. Soc. 1984, 67, C119-C121. (17) Fernandez, L. E.; Sanchez, E. V.; Panizza, M.; Carnasciali, M. M.; Busca, G. J. Mater. Chem. 2001, 11, 1891–1897.

Langmuir, Vol. 24, No. 15, 2008 8359

Vm ) x)

1.311x , V ) 1 - Vm 1 + 0.311x t Im(111) + Im(111)

Im(111) + Im(111) + It(111)

Transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) images were taken on a Philips Tecnai F30 FEGTEM operated at 300 kV. The samples were prepared by uniformly dispersing in ethanol and then placing onto carbon-coated copper grids. The average size of ZrO2 particles and their size distributions were obtained by averaging not less than 200 particles randomly distributed in TEM images. Ultraviolet (UV) Raman spectra were obtained at ambient temperature on a Jobin-Yvon T64000 spectrometer equipped with a He-Cd laser at a wavelength of 325 nm. Visible Raman spectra were measured at ambient temperature on a Bruker Senterra confocal microscope spectrometer equipped with an Nd:YAG laser at a wavelength of 532 nm and a CCD camera. The resolution was 3-5 cm-1. Raman shifts for all samples were measured in the range of 100-1500 cm-1 in ambient air. H2 temperature-programmed reduction (H2 TPR) profiles were obtained on a flow unit (TP5000, Tianjin Xianquan) in a 5% H2/N2 flow (50 cm3 min-1; Beijing Huayuan, certified mixture). MoOx/ ZrO2 samples containing ca. 4 mg of Mo were placed in a quartz cell, and the temperature was ramped linearly from 30 to 800 at 10 °C min-1. The H2 concentration in the effluent was measured by online mass spectrometry (Hiden HPR 20), and the response was calibrated by the reduction of CuO powder (Beijing Chemicals, >99.0%) in H2 (5% H2/N2). 2.3. Catalytic Methanol Oxidation Reactions. Methanol oxidation reactions were carried out in a packed-bed quartz mircoreactor (6 mm i.d.). MoOx/ZrO2 samples (80-100 mesh, 0.2 g) were diluted with quartz (∼2 g) to prevent temperature gradients and then treated in a 10% O2/N2 (Beijing Huayuan, 99.999%) flow (30 cm3 min-1) at 400 °C for 1 h before introducing the reactant mixture consisting of 3.5 kPa of CH3OH (Beijing Chemicals, 99.99%) and 30 kPa of O2 (Beijing Huayuan, 99.999%) with balance N2 (86.5 kPa; Beijing Huayuan, 99.999%). Reactants and products were analyzed by online gas chromatography (Shimadzu 2010 GC) using two packed columns (Carbosieve B and Porapak N) and thermal conductivity detectors (TCD). Steady-state reaction data were collected after 2 h on stream. Selectivities are reported on a carbon basis, and rates are reported as molar CH3OH conversion rates per mole of Mo per h. Blank experiments were conducted on pure ZrO2 supports, and no CH3OH conversions were detected under any condition in this study.

3. Results and Discussion 3.1. Synthesis of Pure ZrO2 Phases. Figure 1a shows a representative XRD pattern for the ZrO2 sample hydrothermally synthesized at 160 °C and calcined at 400 °C in flowing air or nitrogen. Four diffraction peaks in the 2θ range of 20-40° were detected at 24.2, 28.2, 31.5, and 34.4 with a shoulder at 35.3°; these are attributed to m-ZrO2 structures (JCPDS card no. 371484). When methanol substituted for water as the solvent, only the tetragonal phase was formed under identical conditions, as shown in Figure 1b, where only two characteristic t-ZrO2 peaks were present at 30.3 and 34.8° (JCPDS card no. 17-0923) in the 2θ range of 20-40°. The sizes of the m-ZrO2 and t-ZrO2 crystallites, estimated from X-ray line broadening, were 5.9 and 5.3 nm, and their BET surface areas were 137 and 106 m2/g, respectively. UV Raman spectra (Figure 2) confirmed the synthesis of pure ZrO2 phases. In the ZrO2 sample synthesized in water, Raman bands were detected at 759 (w), 637 (s), 618 (s), 558, 537 (w), 504, 477 (s), 381, 343, 335, 305, 221 (w), 187 (s), and 177 (s)

8360 Langmuir, Vol. 24, No. 15, 2008

Li et al.

Figure 1. X-ray diffraction (XRD) patterns for ZrO2 synthesized in water (a) and methanol (b) at 160 °C for 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10), and for ZrO2 synthesized in methanol at 160 °C for 20 h and treated at 400 °C for 4 h, followed by hydrothermal treatment at 160 °C for 1 h (c) and by treatment with water vapor at 350 °C for 4 h (d).

Figure 3. (a) TEM (scale bar 10 nm) and (b) HRTEM (scale bar 5 nm) images for ZrO2 synthesized in water at 160 °C for 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10); the inset represents its size distribution histogram.

Figure 2. UV Raman spectra for ZrO2 synthesized in water (a) and methanol (b) at 160 °C for 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10), and for ZrO2 synthesized in methanol at 160 °C for 20 h and treated at 400 °C for 4 h, followed by hydrothermal treatment at 160 °C for 1 h (c) and by treatment with water vapor at 350 °C for 4 h (d).

cm-1 (Figure 2a), which are all characteristic of m-ZrO2. The sample synthesized in methanol exhibited Raman features at 639 (s), 604 (sh), 462 (s), 313 (s), 271 (s), and 149 (s) cm-1 (Figure 2b), which all arise from the Raman-active modes for the tetragonal phase of ZrO2.17 These XRD and Raman results are consistent with the following TEM results shown in Figures 3 and 4. To examine the hydrothermal stability of the ZrO2 sample synthesized in methanol, it was further treated under hydrothermal conditions (at 160 °C for 1 h) or with water vapor (at 350 °C for 4 h). The corresponding XRD and UV Raman results are shown in Figures 1c,d and 2c,d, respectively, which will be discussed below. (18) Tani, E.; Yoshimura, M.; Somiya, S.; Shigeyuki, J. Am. Ceram. Soc. 1983, 66, 11–14.

The TEM images (Figures 3a and 4a) show that the particles for the two ZrO2 samples are mostly primary particles with an average diameter of ∼6.0 ( 0.25 nm and a narrow particle size distribution. HRTEM images show clear lattice fringes with lattice spacings of 0.317 and 0.284 nm (Figure 3b) and 0.297 nm (Figure 4b), respectively, for the ZrO2 samples synthesized in water and methanol, corresponding to (1j11) and (111) planes of m-ZrO2 (JCPDS card no. 37-1484) and to (111) planes of t-ZrO2 (JCPDS card no. 17-0923). Taken together, these results demonstrate that the obtained crystalline ZrO2 nanoparticles exist in pure monoclinic and tetragonal phases. Parallel studies showed that these pure ZrO2 phases can be synthesized via the current solvothermal method under a wide range of reaction conditions. Figure 5a shows the XRD patterns for the ZrO2 crystal phases synthesized hydrothermally at varying temperatures in the range of 120-200 °C for 20 h. At 120 °C, m-ZrO2 was formed with a fraction of the t-ZrO2 phase (ca. 24%). Such mixed phases converted to the pure m-ZrO2 phase as the temperature increased to 140 °C, indicating that higher temperatures favor the formation of the m-ZrO2 phase in water. It was noted that varying the temperature from 140 to 200 °C slightly changed the average crystallite size and BET surface area of the m-ZrO2 phase from 5.3 to 9.1 nm (calculated from the XRD results) and from 140 to 130 m2/g, respectively (Table 1.) Under the methanothermal conditions, only the pure t-ZrO2 phase was formed over the whole temperature range of 120-200

Synthesis of Pure Zirconia Nanoparticles

Figure 4. (a) TEM (scale bar 10 nm) and (b) HRTEM (scale bar 5 nm) images for ZrO2 synthesized in methanol at 160 °C for 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10); the inset represents its size distribution histogram.

°C, as shown in Figure 5b; the average crystallite sizes of the t-ZrO2 phases, however, decreased from 5.3 to 3.9 nm with increasing the reaction temperature whereas their BET surface areas were slightly varied in the range of 90-107 m2/g (Table 1). These observed temperature effects are different from those under hydrothermal conditions, which are likely related to the greater steric hindrance of methanol relative to that of water, and the consequent effects on the aggregation of the ZrO2 precipitates or their intermediates involved in the ZrO2 growth mechanism, as discussed below. Similar results were observed with changing Zr4+ and urea concentrations and consequent reaction pH values. The 3-fold variation of Zr4+ concentration in the range of 0.2 -0.6 M (at a given urea to Zr4+ molar ratio of 10/1 and a constant pH of ∼10), as shown in Figure 6, led to the exclusive formation of pure m- and t-ZrO2 phases in water and methanol, respectively. But it was found that higher Zr4+ concentrations favor the formation of the pure ZrO2 phases with smaller crystallite sizes and greater surface areas. Upon increasing the Zr4+ concentration from 0.2 to 0.6 M, the size decreased from 11.2 to 6.6 nm and 11.0 to 3.9 nm, and the surface area increased from 112 to 130 m2/g and from 82 to 139 m2/g for m-ZrO2 and t-ZrO2, respectively. Such effects on the crystallite size reflects the classical nucleation and growth mechanism, from which it is known that higher Zr4+ concentrations favor faster nucleation relative to their growth, thus decreasing the size of the final particles. These effects may

Langmuir, Vol. 24, No. 15, 2008 8361

Figure 5. X-ray diffraction (XRD) patterns for ZrO2 synthesized in water (a) and methanol (b) at different temperatures in the range of 120-200 °C for 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10). Table 1. Average Crystallite Sizes and BET Surface Areas of the ZrO2 Samples Synthesized at Different Temperature for 20 h and Then Treated at 400 °C for 4 h m-ZrO2a temperature (°C) 140 160 180 200

t-ZrO2b

average size (nm)c

surface area (m2/g)

average size (nm)c

surface area (m2/g)

5.3 5.9 6.5 9.1

140 137 138 130

5.3 5.3 4.9 3.9

97 106 94 90

a Synthesized in water and treated in dry air or nitrogen (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10). b Synthesized in methanol and treated in dry nitrogen (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10). c Calculated from X-ray diffraction patterns for these samples by the Sherrer equation.

also reflect the fact that higher Zr4+ concentrations facilitate the formation of nuclei with uniform sizes from the reaction between Zr4+ and OH- ions (generated from urea hydrolysis), which accordingly minimizes the effects of the Ostwald ripening process that would otherwise lead to larger crystallites. Figure 7a shows the results upon variation of urea to Zr4+ ratios (at a given Zr4+ concentration of 0.4 M) in the wide range of 0-10 at 200 °C under hydrothermal conditions, where pure m-ZrO2 phases were exclusively detected even in the absence of urea (i.e., for a ratio of 0). The pH of the reaction solution was measured after the hydrothermal reaction, and it was changed

8362 Langmuir, Vol. 24, No. 15, 2008

Figure 6. X-ray diffraction (XRD) patterns for ZrO2 synthesized in water (a) and methanol (b) at different Zr4+ concentrations in the range of 0.2-0.6 M at 200 °C for 20 h, followed by treatment at 400 °C for 4 h (urea/Zr4+ molar ratio ) 10).

from acidic (pH 1-2) to neutral (pH 7-8) and finally to basic (pH ∼10) as the urea to Zr4+ ratio increased from 0 to 1/1 and then up to 5/1 and 10/1 as a result of the formation of OH- ions from the hydrolysis of urea. Following this increase in the urea to Zr4+ ratio and pH, the size of the m-ZrO2 crystallites increased from 3.5 nm to a maximum value of 13.2 nm (at the ratio of 1/1), and then decreased to 9.1 nm. Such size dependences are likely relevant to the higher solubility of the ZrO2 precipitates or their intermediates in acidic or basic solutions because of the amphoteric nature of ZrO2. In the case of methanol as the solvent, pure t-ZrO2 phases were formed when the urea to Zr4+ ratios were greater than 2, as shown in Figure 7b. Ratios smaller than 2 led to the formation of a small fraction of m-ZrO2 impurity (decreasing from 19% for a ratio of 0 to only ∼4% for a ratio of 2/1), apparently due to the effects of the crystal water present in the ZrO(NO3)2 · 2H2O precursor. Water has higher solubility than methanol for the ZrO2 precipitates and their intermediates, which are responsible for the effective operation of the dissolution and precipitation mechanism. As a result, more stable m-ZrO2 is formed in Figure 7a under the hydrothermal conditions irrespective of the urea concentration. However, under the methanothermal conditions, water is provided only by the crystal water from the precursor. When urea is not added or at lower urea to Zr4+ ratios, the crystal water is available for dissolving the ZrO2 precipitates or their

Li et al.

Figure 7. X-ray diffraction (XRD) patterns for ZrO2 synthesized in water (a) and methanol (b) at different urea/Zr4+ molar ratios in the range of 0-10 at 200 °C for 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+).

intermediates to some extent, which results in the formation of small amount of m-ZrO2 along with the major t-ZrO2. With the increasing addition of urea, the crystal water is consumed in the hydrolysis of urea (i.e., (NH2)2CO + H2O f 2NH3 + CO2); consequently, only the metastable t-ZrO2 is formed as a result of the ineffectiveness of the dissolution and precipitation mechanism in operating under these conditions. Furthermore, it was found that such effects of trace water were much weaker at lower temperatures under the methanothermal conditions; for example, the m-ZrO2 impurity (∼9%) was detected only in the absence of urea, and at urea to Zr4+ ratios greater than 1/1, the pure t-ZrO2 phase was obtained at 140 °C (Supporting Information, Figure S1), in agreement with the effects of temperature and solvent on the ZrO2 formation discussed above. It was noted that the pure m-ZrO2 and t-ZrO2 phases can also be obtained using other zirconium salts, such as ZrOCl2 · 8H2O and Zr(NO3)4 · 5H2O, under similar reaction conditions, as shown in Figure 8. Taken together, these results clearly show the lower sensitivity of the current solvothermal method (in terms of the phase purity) to the reaction parameters, compared to that of previously reported methods that generally require careful control of the reaction temperature, pH, pressure, and zirconium

Synthesis of Pure Zirconia Nanoparticles

Langmuir, Vol. 24, No. 15, 2008 8363

Figure 8. X-ray diffraction (XRD) patterns for ZrO2 synthesized using Zr(NO3)4 · 5H2O in water (a) and methanol (c) and ZrOCl2 · 8H2O in water (b) and methanol (d) at 200 °C for 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10).

precursors.8,12,14,15,18,19 This advantage of our current method provides its potential for large-scale production of pure ZrO2 phases. Metastable t-ZrO2 tends to transform to m-ZrO2 upon treatment with water vapor.9,10 For instance, Xie et al. reported that 80% of the t-ZrO2, prepared from refluxing aqueous solutions of zirconyl chloride and ammonia, transformed to m-ZrO2 at room temperature after contact with water.9 In sharp contrast, our t-ZrO2 samples remained intact both in the bulk and on the surface after hydrothermal treatment (tested in autoclave) at 160 °C or even after thermal treatment with flowing water vapor at 350 °C, as evidenced from the XRD patterns and UV Raman spectra shown in Figures 1c,d and 2c,d, respectively. Such hydrothermal stability of the obtained t-ZrO2 offers the possibility to study the rigorous correlations between the ZrO2 phases and their catalytic performance in many reactions involving water as a reactant or product, as exemplified by the following methanol-selective oxidation. 3.2. Plausible Mechanism of Solvent Effects on ZrO2 Phases. Clearly, the effects of solvent on the formation of the pure ZrO2 phases are crucial. The same results obtained with ZrO(NO3)2 · 2H2O, ZrOCl2 · 8H2O, and Zr(NO3)4 · 5H2O led us to envisage that the ZrO2 synthesis from these different zirconium salts may follow a similar mechanism. Previous studies showed that ZrOCl2 · 8H2O exists mainly in the form of cyclic [Zr(OH)2 · 4H2O]48+ tetramers in solutions at room temperature, of which the four Zr4+ ions are arranged in a square and linked by four bridging OH groups per Zr4+ ion.20,21 Upon heating or the addition of bases, these tetramers condense via olation and oxolation to form corresponding polymers (e.g., [ZrOx(OH)4 - 2x · mH2O]n) and ultimately hydrous ZrO2 precipitates (ZrOx(OH)4 - 2x) depending on their supersaturation concentrations under the control of the general dissolution-precipitation equilibrium.21 Our results show that the ZrO2 phases evolved with time from predominantly t-ZrO2 to pure m-ZrO2 during the hydrothermal synthesis, and such evolution was facilitated at higher reaction (19) Stefanic, G.; Music, S.; Grzeta, B.; Popovic, S.; Sekulic, A. J. Phys. Chem. Solids 1998, 59, 879–885. (20) Clearfield, A.; Vaughan, P. A. Acta Crystallogr. 1956, 9, 555–558. (21) (a) Johnson, J. S.; Kraus, K. A. J. Am. Chem. Soc. 1956, 78, 3937–3943. (b) Chen, F.; Huang, L.; Zhong, Z.; Gan, G. J.; Kwan, S. M.; Kooli, F. Mater. Chem. Phys. 2006, 97, 162–166.

Figure 9. X-ray diffraction (XRD) patterns for ZrO2 synthesized in water (a) and methanol (b) at 140 °C for 2-20 h, followed by treatment at 400 °C for 4 h. (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10).

temperatures. As shown in Figure 9a, two weak diffraction peaks were mainly observed at around 30.3 and ∼35° after the hydrothermal synthesis at 140 °C for 2 h, showing predominant formation of the t-ZrO2 phase with low crystallinity. About 78% t-ZrO2 phase was formed after 4 h, which transformed completely to m-ZrO2 after 6 h. Such a transformation was expedited at higher temperatures and was completed after 4 h at 200 °C (Figure S2). However, XRD patterns and UV Raman spectra show that the pure t-ZrO2 phase was always obtained during the methanothermal synthesis (Figures 9b, 10, and S3), irrespective of the reaction time and temperature employed in this study. These results indicate that the hydrous ZrO2 precipitates are formed preferentially with tetragonal structures in the early stage of the synthesis process. Under the hydrothermal conditions, the hydrous ZrO2 precipitates underwent the dissolution and precipitation steps with time, especially in the presence of OH- ions (generated from the hydrolysis of urea), which could induce the thermodynamically favorable restructuring and transformation of metastable t-ZrO2 to more stable m-ZrO2. In contrast, during the methanothermal synthesis, urea hydrolysis readily occurred via its reaction with the bridging OH groups and coordinated H2O molecules of the Zr tetramers, leading to the formation of wellordered tetramers and ZrO2 precipitates presumably with more rigid oxo bridges (Zr-O-Zr) than hydroxyl bridges (Zr-OHZr) that are otherwise prevalent in the presence of water. Meanwhile, compared to H2O, the steric hindrance of methanol

8364 Langmuir, Vol. 24, No. 15, 2008

Figure 10. UV Raman spectra for ZrO2 synthesized in methanol at 140 °C for 2, 6, and 20 h, followed by treatment at 400 °C for 4 h (0.4 M Zr4+, urea/Zr4+ molar ratio ) 10).

and its weaker ability to dissolve the ZrO2 precipitates or their intermediates, as mentioned above, could prevent the occurrence of the dissolution-precipitation process and thus the structural change in the ZrO2 precipitates. Taken together, these factors are supposed to facilitate the formation and stability of the ZrO2 precipitates with tetragonal structures in methanol. These propositions on the solvent effects and also the role of urea are confirmed by the experiments that were carried out by the solvothermal treatment of preprepared hydrous ZrO2 (ZrOx(OH)4 - 2x) precipitates in water and methanol, respectively. The hydrous ZrO2 precipitates were obtained by the hydrolysis of an aqueous solution of ZrO(NO3)2 · 2H2O at ∼pH 10 using NH4OH at room temperature, as discussed above, followed by thoroughly washing with deionized water and then naturally drying at 20 °C. These preprepared ZrO2 precipitates contained a mixture of m-ZrO2 (∼21%) and t-ZrO2 (∼79%) with average particle sizes of ∼21.9 nm, as shown in Figure 11a. Methanothermal treatment of the ZrO2 precipitates at 160 °C led to no obvious change in their phase compositions and particle sizes in the absence of urea (Figure 11b) but to the detection of only t-ZrO2 phases (with similar average particle sizes of ∼19.0 nm) in the presence of urea (urea/contained water molar ratio ) 5; Figure 11c and also the UV Raman spectrum shown in Figure S4a). Such a marked difference is consistent with the above discussion concerning the role of urea in removing water and the trace water effects on t-ZrO2 formation under the methanothermal conditions. In contrast, hydrothermal treatment at 160 °C led to the significant transformation of t-ZrO2 to m-ZrO2; pure m-ZrO2 and ∼81% m-ZrO2 (with ∼19% t-ZrO2) were obtained, respectively, in the presence and absence of urea with average particle sizes of ∼7.6 and ∼11.4 nm (Figure 11e,d and also Figure S4b). Such a decrease in the particle sizes of the ZrO2 phases upon hydrothermal treatment apparently reflects the occurrence of the dissolution-precipitation process in water, especially accelerated by OH- ions generated from the hydrolysis of urea. A study is in progress to further clarify the genesis of the solvent effects on the ZrO2 phases and improve their thermal and/or hydrothermal stability. 3.3. Effects of ZrO2 Crystal Phases on the Selective Oxidation of Methanol. Next, we examine the effects of the ZrO2 crystal phases on the selective oxidation of methanol, a frequently used probe reaction that is sensitive to the catalyst

Li et al.

Figure 11. X-ray diffraction (XRD) patterns for ZrO2 synthesized by the hydrolysis of ZrO(NO3)2 · 2H2O using NH3 · H2O at room temperature, followed by natural drying at 20 °C and then treatment at 400 °C for 4 h (a), by methanothermal treatment of the naturally dried ZrO2 precipitates (ZrOx(OH)4 - 2x) (prepared for part a) in the absence (b) and presence (c) of urea, and by hydrothermal treatment of the naturally dried ZrO2 precipitates (ZrOx(OH)4 - 2x) (prepared in part a) in the absence (d) and presence (e) of urea, respectively, at 160 °C for 20 h, followed by treatment at 400 °C for 4 h.

surface structures, and their redox and acid-base properties.22 To our knowledge, no such studies have been reported in the literature although the effects of ZrO2 crystal phases on nonoxidation reactions (e.g., CO hydrogenation and alkane isomerization) have been extensively studied,2–5,7 as mentioned above. Table 2 shows methanol oxidation rates (normalized per Mo atom) and selectivities at 210-250 °C on MoOx domains dispersed on m-ZrO2 and t-ZrO2 (MoOx/m-ZrO2 and MoOx/t-ZrO2), respectively. The methanol oxidation rates (per Mo atom) were calculated by not considering dimethyl ether (DME) as a product (i.e., on a DME-free basis) because of the fact that its formation is merely from the bimolecular dehydration of methanol on strong acid sites22 and is thus irrelevant to the catalyst oxidation activities. The two catalysts contained similar MoOx surface densities of ∼5.0 Mo/nm,2 which is close to the theoretical monolayer coverage of 2D molybdate structures.23 Dimethyl ether (DME), formaldehyde (HCHO), methylformate (MF), dimethoxymethane (DMM), and CO2 were detected as products. The oxidation rates and selectivities in Table 2 were obtained and compared at similar methanol conversions (∼20%), which were achieved by varying the reactant space velocities. HCHO was the major product, and its selectivities increased slightly upon increasing the reaction temperature from 210 to 250 °C, reaching about 90%, whereas the DMM selectivities monotonically decreased to zero on both catalysts. MF (ca. 5-8%) and CO2 (ca. 1-2% or 1-3%) selectivities were low and slightly varied within the temperature range. DME was detectable only on MoOx/t-ZrO2 with selectivities as low as 2%, indicating the weaker acidity of the two catalysts. The oxidation rates increased from 92.4 to 330.3 mol/mol Mo h and from 39.7 to 132.3 mol/ mol Mo h on MoOx/m-ZrO2 and MoOx/t-ZrO2, respectively, as the reaction temperatures increased from 210 to 250 °C. At the given temperatures, the rates were always about 2 times greater for MoOx/m-ZrO2 than for MoOx/t-ZrO2 (Table 2). By referring to the structures of the two catalysts, such an observed difference (22) Tatibouet, J. M. Appl. Catal., A 1997, 148, 213–252.

Synthesis of Pure Zirconia Nanoparticles

Langmuir, Vol. 24, No. 15, 2008 8365

Table 2. Methanol Oxidation Rates and Selectivities on MoOx/m-ZrO2 and MoOx/t-ZrO2a MoOx/m-ZrO2

MoOx/t-ZrO2

selectivity (% carbon) reaction temperature (°C) 210 220 230 240 250

CH3OH oxidation rateb (mol/mol Mo-h) 92.4 140.2 179.6 233.2 330.3

DME 0 0 0 0 0

selectivity (% carbon)

HCHO

DMM

MF

CO2

CH3OH oxidation rateb (mol/mol Mo-h)

89.0 91.7 90.8 90.5 90.0

3.1 0 0 0 0

5.4 5.8 6.6 6.7 7.0

2.5 2.5 2.5 2.8 3.0

39.7 58.8 90.1 114.2 132.3

DME 1.9 1.8 1.9 2.0 1.9

HCHO

DMM

MF

CO2

83.1 85.8 86.0 86.9 89.9

6.6 5.5 3.6 1.5 0

6.6 5.9 7.6 8.6 7.2

1.8 1.0 0.8 1.0 1.1

a ∼5.0 Mo/nm2, 3.5 kPa of CH3OH, 30 kPa of O2, balance N2, ∼20% methanol conversions. b Rate for CH3OH conversion not including DME (i.e., on a DME-free basis).

Figure 12. X-ray diffraction (XRD) patterns (a) and Raman spectra (b) for MoOx domains supported on m-ZrO2 or t-ZrO2 and, for comparison, bulk MoO3.

of m-ZrO2 (759 (w), 637 (s), 618 (s), 558, 537 (w), 504, 477 (s), 343, 381, 335, 305, 221 (w), 187 (s), and 177 (s) cm-1) and t-ZrO2 (645 (s), 604, 455, 314 (s), 281 (s), and 149 (s) cm-1),17 clearly showing no detectable transformation of the pure ZrO2 phases. In the range of 1000-800 cm-1, the two bands at 970 and 863 cm-1 on MoOx/m-ZrO2 and at 981 and 875 cm-1 on MoOx/t-ZrO2 correspond to ModO and Mo-O-Mo stretching vibrations in 2D polymolybdates, respectively.23,24 MoOx/t-ZrO2 showed an additional band at 817 cm-1 that is characteristic of crystallite MoO3, which suggests the presence of crystalline MoO3 structure. The fraction of Mo present as crystalline MoO3 for this sample was estimated by Raman spectroscopy using a method previously reported,25 which was only ∼6.7%. Thus, we conclude that MoOx/t-ZrO2 contains predominately 2D polymolybdates, as also evidenced by the absence of any MoO3 diffractions (Figure 12a). Such low content of crystalline MoO3 and its relatively intense Raman peak (at 819 cm-1 relative to 875 cm-1) reflects the consensus that Raman scattering cross sections are 10-103 greater for crystallite MoO3 than for dispersed MoOx species.23,26,27 These 2D MoOx structures mean that most Mo sites on the two catalysts are accessible to reactants; therefore, the methanol oxidation rates per Mo in Table 2 represent turnover rates and reflect the intrinsic surface reactivity of the MoOx domains. This led us to conclude that the ∼2-fold greater oxidation rates of MoOx/m-ZrO2 relative to those of MoOx/t-ZrO2 at similar methanol conversions (∼20%) (Table 2) arise from the higher intrinsic reactivities of the MoOx domains deposited on m-ZrO2 than on t-ZrO2. This is consistent with the more reducible nature of the MoOx domains on m-ZrO2 than on t-ZrO2, which was probed by temperature-programmed reduction using H2 (H2 TPR). As shown in Figure 13, the peak temperature for the MoOx reduction, corresponding to the reduction of Mo6+ to Mo4+,23 on m-ZrO2 was 36 °C below than that on t-ZrO2 (428 vs 464 °C). Such correlation between the reducibility and methanol oxidation rates is consistent with the Mars-van Krevelen redox mechanism, which involves the rate-determining activation of C-H bonds in CH3OH using lattice oxygen atoms within the MoOx domains.22 Similar reactivity-reducibility correlation was observed for selective oxidation reactions of methanol,6,28,29 ethanol,30 and

in the rates can be correlated to the effects of the ZrO2 crystal phases. The catalyst structures were characterized by XRD and visible Raman spectroscopy. As shown in Figure 12a, only diffraction peaks of m-ZrO2 and t-ZrO2 were observable, demonstrating that the pure crystal phases of the ZrO2 supports remained intact in the presence of MoOx domains, and these MoOx domains were highly dispersed on the ZrO2 surfaces. These observations were confirmed by Raman spectra (Figure 12b). The Raman bands appearing in the range of 800-100 cm-1 are characteristic

(23) Liu, H.; Cheung, P.; Iglesia, E. J. Catal. 2003, 217, 222–232. (24) Xie, S.; Chen, K.; Bell, A. T.; Iglesia, E. J. Phys. Chem. B 2000, 104, 10059–10068. (25) Baltrus, J. P.; Makovsky, L. E.; Stencel, J. M.; Hercules, D. M. Anal. Chem. 1985, 57, 2500–2503. (26) Radhakrishnan, R.; Reed, C.; Oyama, S. T.; Seman, M.; Kondo, J. N.; Domen, K.; Ohminami, Y.; Asakura, K. J. Phys. Chem. B 2001, 105, 8519–8530. (27) Williams, C. C.; Ekerdt, J. G.; Jehng, J. M.; Hardcastle, F. D.; Wachs, I. E. J. Phys. Chem. 1991, 95, 8791–8797. (28) Deo, G.; Wachs, I. E. J. Catal. 1994, 146, 323–334. (29) Liu, H.; Iglesia, E. J. Phys. Chem. B 2005, 109, 2155–2163. (30) Zhang, W.; Desikan, A.; Oyama, S. T. J. Phys. Chem. 1995, 99, 14468– 14476.

8366 Langmuir, Vol. 24, No. 15, 2008

Figure 13. H2 temperature-programmed reduction profiles (H2 TPR) for MoOx domains supported on m-ZrO2 and t-ZrO2.

dimethylether23,31 on MoOx, VOx, and RuOx-based materials, also following Mars-van Krevelen redox cycles. Taken together, these results clearly show the significant effects of the ZrO2 crystal phases on the redox properties of the supported MoOx catalysts. Such effects were also found preliminarily for other metal oxides, such as VOx and RuOx, which are being examined in the methanol selective oxidation.

4. Conclusions Pure t-ZrO2 and m-ZrO2 nanocystallites have been conveniently synthesized via a solvothermal method using cheap inorganic zirconium salts and urea in methanol and water, respectively. Although the solvents are crucial in dictating the relative amounts of ZrO2 phases formed, other reaction parameters, including reaction temperature, reactant (urea and zirconium salts) concentration, pH, and hydrous zirconium salts, can be varied over wide ranges. During the solvothermal synthesis, hydrous ZrO2 (31) Liu, H.; Cheung, P.; Iglesia, E. J. Phys. Chem. B 2003, 217, 222–232.

Li et al.

precipitates are formed preferentially with tetragonal structures; they are then transformed to thermodynamically more stable monoclinicstructuresinwatermostlikelyviathedissolution-precipitation process as a result of their higher solubility in water than in methanol, whereas methanol as the solvent favors the removal of water from the ZrO2 precipitates via urea hydrolysis and retention of their tetragonal structures. The obtained t-ZrO2 phases are superior to the samples reported in the literature in terms of their hydrothermal stability, and this made it possible to systematically examine the effects of ZrO2 crystal phases on the selective oxidation reactions generally involving water as an inevitable product. Two-dimensional MoOx domains dispersed on pure m-ZrO2 relative to t-ZrO2 exhibit higher reducibility and intrinsic reactivity for the selective oxidation of methanol, showing that variation in the ZrO2 crystal phases can significantly tune the redox properties of the ZrO2-supported MoOx catalysts. This simple solvothermal approach may be potentially extended to facile syntheses of pure crystal phases of other transition-metal oxides, and the observed effects of the ZrO2 crystal phases on the redox properties of catalysts may provide a basis for the design of catalysts for the selective oxidation reactions of alkanes, alkenes, and alcohols. Acknowledgment. This work was supported by the National Natural Science Foundation of China (grant nos. 20573004, 20673005, and 20733009) and the National Basic Research Project of China (no. 2006CB806100). This work was also supported in part by the Program for New Century Excellent Talents in University (NECT-05-0010), State Education Ministry of China. We thank Professors Can Li and Zhaochi Feng (Dalian Institute of Chemical Physics, CAS) for their great help in obtaining the UV Raman spectra. Supporting Information Available: XRD patterns for ZrO2 synthesized in methanol at urea/Zr4+ molar ratios of 0-10 and 140 °C in water and 200 °C in methanol (urea/Zr4+ ) 10) for 2-20 h. UV Raman spectra for ZrO2 synthesized by the hydrolysis of ZrO(NO3)2 · H2O using NH3 · H2O at room temperature, followed by treatment. This material is available free of charge via the Internet at http: //pubs.acs.org. LA800370R