208
Langmuir 2004, 20, 208-213
Study of Support Effects on the Reduction of Ni2+ Ions in Aqueous Hydrazine Abdel-Ghani Boudjahem, Serge Monteverdi, Michel Mercy, and Mohammed M. Bettahar* Catalyse He´ te´ roge` ne, UMR CNRS 7565, Faculte´ des Sciences, Universite´ Henri Poincare´ , Nancy-I, BP 239, 54506 Vandœuvre Ce´ dex, France Received June 24, 2003. In Final Form: October 23, 2003 We have studied the effect of silica of quartz-type on the reducibility of nickel acetate in aqueous hydrazine (80 °C, pH ) 10-12) and metal particle formation. The obtained materials were characterized by X-ray diffraction, transmission electron microscopy, and thermodesorption experiments. With nickel acetate alone, the reduction was partial (45%) and a metal film at the liquid-gas interface or a powdered metal precipitate with an average particle size of 120 nm was obtained. In the presence of silica as the surfactant, the reduction of nickel acetate was total and the nickel phase deposited as a film on the support with an average particle size of 25 nm. Supported nickel acetate was also totally reduced. Crystallites of a mean particle size of about 3 nm were obtained. Decreasing the nickel content or increasing the hydrazine/nickel ratio decreased the metal particle size. Whiskers were formed for low nickel loadings. Hydrogen thermal treatment of the reduced phase showed that the organic acetate fragment, belonging to the precursor salt, still remained strongly attached to the nickel phase. The amount of the retained organic matrix depended on the metal particle size. Surface defects are suggested as active sites, which enhanced nickel ion reduction in the presence of silica as the surfactant or support. Metal-support interactions and the nucleation/ growth rate were the main factors determining the size and morphology of the supported metal particles formed. The organic matrix covered the reduced nickel phase.
1. Introduction The performances of metal-supported catalysts strongly depend on the method of preparation, which determines the size and shape of the metal particles formed.1-7 The challenge in heterogeneous catalysis is the control of the morphology of the final material, notably for impregnated catalysts.1-7 Conventional supported metal catalysts are prepared by in situ reduction of a metal salt or oxide. An alternative method used to obtain supported catalysts with well-defined metal particles is the preparation via metal colloids. In recent years, a great deal of attention has been paid to metal nanoparticle research because of their unusual properties as compared to those of the bulk metal. They exhibit unusual electronic, optical, magnetic, and chemical properties as a result of their extremely small size and large surface-to-volume ratio.8-12 These properties are determined much more by the surface atoms than by the lattice atoms. They have potential applications as advanced materials8-14 and also have been utilized in * Author to whom correspondence should be addressed. (1) Tautster, S. J.; Fung, S. C. J. Catal. 1978, 5, 29. (2) Bartholomew, C. H.; Pannell, R. B.; Fowler, R. W. J. Catal. 1983, 79, 34-46. (3) Richardson, J. T.; Lei, M.; Turk, B.; Forster, K.; Twigg, M. V. Appl. Catal. 1994, 110, 217-237. (4) Molina, R.; Poncelet, G. J. Catal. 1998, 173, 257-267. (5) Miyazaki, A.; Balin, I.; Aika, K.; Nakano, V. J. Catal. 2001, 204, 364-371. (6) Reinen, D.; Selwood, P. W. J. Catal. 1963, 2, 109-120. (7) Houalla, M.; Lemaıˆtre, J.; Delmon, B. J. Chem. Soc., Faraday Trans. 1982, 78, 1389-1400. (8) Halperin, W. P. Rev. Mod. Phys. 1998, 58, 533-. (9) Schmid, G. Chem. Rev. 1992, 92, 1709-1727. (10) Lewis, L. N. Chem. Rev. 1993, 93, 2693-2730. (11) Volotkin, Y.; Sinzig, J.; De Jong, L. J.; Schmid, G.; Vargaftik, M. N.; Moisseev, I. I. Nature 1996, 384, 621-623. (12) Colvin, V. N.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370, 354-357. (13) Whetten, R. L. Acc. Chem. Res. 1999, 32, 397-406.
heterogeneous catalysis.14-18 The chemical route of preparation of such materials is of specific interest because it allows better structural control on the microscopic level.19-22 The chemical methods have generally involved the reduction of the relevant metal salt in the presence of stabilizers such as linear polymers,23 ligands,23,24 and surfactants,25 notably tetraalkylammonium salts26 or heterogeneous supports,27,28 which prevent the nanoparticles from agglomerating. In a previous work, we have reported results on the chemisorptive and catalytic properties of nickel metal nanoparticles supported on silica of a low surface area, prepared by reduction of nickel acetate by hydrazine in aqueous media.29 In the present paper, we report more details on the preparation method used. We have studied the influence of the support and preparation conditions (14) Gates, B. C. Chem. Rev. 1995, 95, 511-522. (15) Paulus, U. A.; Endrushchat, U.; Feldmeyer, G. J.; Schmidt, T. J.; Bonnemann, H.; Behm, R. J. J. Catal. 2000, 195, 383-393. (16) Sales, E. A.; Benhamida, B.; Caizergues, V.; Lagier, J. P.; Fievet, F.; Bozon-Verduraz, F. Appl. Catal. 1998, 172, 273-283. (17) Franquin, D.; Monteverdi, S.; Molina, S.; Bettahar, M. M.; Fort, Y. J. Mater. Sci. 1999, 34, 4481-4488. (18) Lefondeur, S.; Monteverdi, S.; Molina, S.; Bettahar, M. M.; Fort, Y. J. Mater. Sci. 2001, 36, 2633-2638. (19) Fendler, J. H.; Meldrum, F. C. Adv. Mater. 1995, 7, 607-632. (20) Fievet, F.; Fievet-Vincent, F.; Lagier, J. P.; Dumont, B.; Filgarz, M. J. Mater. Chem. 1993, 3, 627-632. (21) Rieke, R. D. Acc. Chem. Res. 1997, 10, 377. (22) Ozin, G. A. Adv. Mater. 1992, 4, 612-649. (23) Poulin, J. C.; Kagan, H. B.; Vargavtik, M. N.; Stolarov, I. P.; Moisseev, I. I. J. Mol. Catal. 1995, 95, 109-113. (24) Amiens, C.; de Caro, D.; Chaudret, B.; Bradley, J. S.; Mazel, R.; Roucau, C. J. Am. Chem. Soc. 1993, 115, 11638-11639. (25) Esumi, K.; Matzuhita, K.; Torigoe, K. Langmuir 1995, 11, 32853287. (26) Reetz, M. T.; Helbig, W. J. Am. Chem. Soc. 1994, 116, 74017402. (27) Tauster, S. J. Acc. Chem. Res. 1987, 20, 389. (28) Brayner, R.; Viau, G.; da Cruz, G. M.; Fie´vet-Vincent, F.; Fie´vet, F.; Bozon-Verduraz, F. Catal. Today 2000, 57, 187-192. (29) Boudjahem, A.; Monteverdi, S.; Mercy, M.; Ghanbaja, G.; Bettahar, M. M. Catal. Lett. 2002, 84, 115-122.
10.1021/la035120+ CCC: $27.50 © 2004 American Chemical Society Published on Web 12/09/2003
Reduction of Ni2+ Ions
Langmuir, Vol. 20, No. 1, 2004 209
Table 1. Amount of the Surface Organic Fragments of the Fresh Nickel Particles, after H2 Treatment at the Programmed Temperature from 25 to 600 °C catalyst
Nisurfa (µmol‚g-1 Ni)
CH4b (µmol‚g-1 Ni)
0.86%Ni/SiO2 1.10%Ni/SiO2 1.40%Ni/SiO2 2.73%Ni/SiO2 4.30%Ni/SiO2 1.10%Ni+SiO2 unsupported Ni
3012 1666 1116
9037 3825 6380
504
3345 4840 176
a
7
COb (µmol‚g-1 Ni)
CH3COOc (µmol‚g-1 Ni)
CH3COO/Nisurf
4518 1912 3190
1.5 1.2 2.9
1728 2420 88
3.4
112
b
13.0
c
Surface nickel atom numbers from XRD calculation. Amount of CH4 and CO evolved. Equivalent amount calculated from eq 2.
(temperature, nickel acetate or hydrazine concentration) on the size and morphology of the nickel nanoparticles formed. The materials obtained were characterized by X-ray diffraction (XRD) and transmission electron microscopy (TEM). Their thermal stability was also examined by temperature-programmed desorption. 2. Experimental Section 2.1. Catalyst Preparation. Doubly distilled water was used as the solvent. Nickel acetate tetrahydrate (g99.0%, Fluka) and aqueous hydrazine 24-26% (g99.0%, Fluka) were used as received. The silica support (Chempur, 99.99%) of 15 m2‚g-1 and grains of 325 mesh was pretreated in air at 500 °C for 12 h after a heating rate of 10 °C‚min-1 and stored under argon. The silica support (3.0-4.0 g) was impregnated with 25 mL of an aqueous solution of nickel acetate. The nickel acetate concentration in the solution was calculated to obtain the nominal composition in the range 1.0-5.0 wt % Ni. The suspension was stirred for 16 h at room temperature. The solvent was then evaporated at 80 °C under vacuum and the obtained solid dried 16 h at 100 °C. The reduction of the supported nickel acetate was performed under an argon atmosphere in a three-necked reaction flask of 500 mL plunged in an oil bath for heating. The reaction flask was fitted with a reflux condenser, a thermocouple for measurement of the reaction temperature, and a gas buret for the measurement of the volume of nitrogen evolved (see the following). The supported nickel acetate precursor was introduced in the reaction flask filled with doubly distilled water. The obtained suspension was stirred for 20 min at room temperature, and then the hydrazine solution was added. The reaction mixture was slowly heated from room temperature to 80 °C. The suspension became blue then translucid as the reaction temperature was raised and changed to the dark black color of colloidal nickel. The degree of reduction was determined from the nitrogen volume according to the following reaction:30
2Ni2+ + N2H4 + 4OH- f 2Ni0 + N2 + 4H2O
(1)
Preliminary experiments showed that no hydrazine decomposed in the conditions used. The pH of the solution was 10-12 and remained almost constant during the reduction process. The black suspension was maintained at 80 °C for 30 min then cooled to room temperature. It was filtered and washed several times with water until a neutral pH was obtained. The filtrate was dried at 60 °C under a vacuum. The resulting solid was stored under argon. The samples are denoted x%Ni/SiO2 (for their nickel contents see Table 1). A sample of supported nanoparticles was prepared by the simultaneous introduction of nickel acetate and hydrazine at room temperature in a suspension of silica and then reducing at 80 °C with the same procedure as that just described. The final material is denoted 1.10%Ni+SiO2. For unsupported nickel acetate, the same reduction procedure was also used. 2.2. Equipment and Procedures. The nickel composition and specific area of the catalysts were determined on a Varian AA1275 atomic absorption spectrophotometer and Carlo Erba Sorptomatic 1900 equipment, respectively. The TEM images were (30) Chen, D.- H.; Hsieh, C.-H. J. Mater. Chem. 2002, 12, 2412.
recorded with a Phillips CM20 STEM. The spectra were obtained by placing a drop of the nanoparticle suspension on the carboncoated copper grid. XRD patterns I(θ) were recorded on a classical θ/2θ diffractometer using Cu KR or Mo KR radiation. Thermal treatment experiments were carried out on a pulse chromatography quartz microreactor equipped with the catharometric detector of a microchromatograph (AT M200, HewlettPackard) fitted with molecular sieve columns and MTI software. The gases utilized were purchased from Air Liquide. Oxygen traces were eliminated from argon (99.995%) and hydrogen (99.995%) by using a manganese oxytrap (Engelhardt). The catalyst sample was flowed with pure H2 (50 mL‚min-1) at a heating rate of 7.5 °C‚min-1 from room temperature to 600 °C.
3. Results and Discussion 2+ Ion Reduction in the Hydrazine Aqueous
3.1. Ni Solution. In the absence of a support and at room temperature, the pale green nickel acetate solution became blue, the color of the complex formed between the nickel salt and hydrazine: [Ni(N2H4)3]2+.31 This complex was formed by the substitution of water ligands by hydrazine ligands. When the temperature was increased, the reaction medium progressively changed to a dark black colloidal solution and gaseous N2 evolved. The degree of reduction depended on the reaction conditions. Thus, the reduction was total at 80 °C, pH ) 10-12, [N2H4] ) 0.83 M, and [Ni(OAc)2] ) 0.06 M. During the reaction, the nickel particles formed a film at the liquid-gas interface or precipitated as a powdered solid. The XRD spectra (Figure 1a) showed a face-centered cubic (fcc) pattern for the obtained Ni0 particles with a mean size of about 120 nm. TEM images (Figure 2a) showed mainly spherical particles with a mean size ranging from 60 to 180 nm. However, a close inspection of the micrographs showed that particles with different geometrical shapes were also found. High-resolution TEM studies are necessary to obtain more information on the morphology of the observed agglomerates. 3.2. Effect of the Presence of a Suspension of SiO2 on Ni2+ Ion Reduction in the Hydrazine Aqueous Solution. The reduction of Ni2+ ions was carried out in the presence of a suspension of silica of a low surface area (15 m2‚g-1) and of a quartz-type structure, as shown by XRD. The obtained mixture (denoted 1.10%Ni+SiO2) changed in color as the reaction temperature was increased; the pale green color of nickel acetate changed to blue, then translucid, and dark black. The black colloidal particles precipitated on the silica support. The XRD study confirmed the formation of fcc metallic nickel (Figure 1b). The morphology of the reduced nickel phase was not too much changed as compared to that obtained in the absence of silica, as shown by the TEM micrographs (Figure 2b). However, in the presence of silica, on one hand, the reduction was deeper (Figure 3); in the same reaction conditions (temperature, pH, nickel or hydrazine amount), (31) Li, Y. D.; Li, L. Q.; Liao, H. W.; Wang, H. R. J. Mater. Chem. 1999, 9, 2675-2677.
210
Langmuir, Vol. 20, No. 1, 2004
Figure 1. XRD spectra of reduced nickel nanoparticles: (a) unsupported Ni; (b) 1.10%Ni+SiO2 ; and (c) 1.10%Ni/SiO2.
the reduction was total in the presence of silica and only 45% in its absence. On the other hand, the metal mean particle size strongly decreased to around 20 nm according to the XRD study. These results strongly indicated that the nickel phase was interacting with the silica support during the Ni2+ ion reduction and metal particle formation processes. A possible explanation of the increase of the degree of reduction may reside in the structure of the used silica, a highly crystallized material. It most probably contains a certain number of defects that would have played the role of active centers for Ni2+ ion reduction. In other words,
Boudjahem et al.
electrons are transferred from the reducing agent to the metal ion in the vicinity of the surface defect. A similar conclusion was reported in the case of the electroless reduction of palladium by aqueous hydrazine in the presence of laser or thermally prereduced ceria;32 for given operating conditions, a metallic palladium deposit of nanosize occurred only in the presence of ceria. Palladium ion reduction was ascribed to the presence of the defects induced or increased by the reduction pretreatment of ceria. A similar explanation of the activating role of silica may also reside in the presence of metal impurities incorporated in the support or defects created in situ by metal-impurity removal by aqueous hydrazine. These impurities or created defects would also have played the role of active centers for Ni2+ ion reduction, inducing a deeper reduction of the precursor and a lower metal particle size. It is worth noting that, in the same operating conditions, the reduction of nickel acetate by aqueous hydrazine was prevented in the presence of silica of a high BET surface area (255 m2‚g-1): no reduction was observed, even at 100 °C. A similar observation was reported for PdCl2 reduction by aqueous hydrazine in the presence of silica of 380 m2‚g-1 of BET surface area and was attributed to the lack of defects for the catalytic reduction of Pd2+ ions.32 This probably also holds for our Ni2+ ions in the presence of a high-surface-area silica support. The surface acidity of this silica may also have prevented nickel ion reduction. The silica support also decreased the particle size growth, most probably as a result of the relative nucleation and growth rates. This point is discussed in the following. 3.3. Reduction of Ni2+ Ions Supported on SiO2 in the Hydrazine Aqueous Solution. Support Effect. The reducibility of nickel acetate supported on silica in aqueous hydrazine was investigated with various nickel loadings. The XRD study of the metal precursors shows almost flat XRD spectra, indicating a high dispersion of the nickel ions on the support. The initial aqueous suspension became translucid as the reaction temperature was raised and changed to the dark black color of colloidal reduced nickel. Typical results are given for the supported 1.10%Ni/SiO2 catalyst in Figure 3. The reduction was total within less than 30 min at 80 °C, as it was for the 1.10%Ni+SiO2 system (Figure 3). Thus, the reducibility order of the nickel ions was as follows: free nickel acetate , nickel acetate + SiO2 ≈ SiO2-supported nickel acetate. The supported reduced nickel obtained was also of fcc structure as the XRD study showed (Figure 1c). The XRD crystallite size was about 3 nm, a domain much smaller than that given by the 1.10%Ni+SiO2 system (25 nm) or free nickel acetate (120 nm). The particle size order was as follows: SiO2-supported nickel acetate < nickel acetate + SiO2 , free nickel acetate. The difference in the particle size may have arisen from the difference in the number of nuclei formed and the growth rate. For the free nickel acetate system, nucleation and growth took place in solution where large particles formed. In contrast, for the supported system, nucleation and growth probably mainly took place on the support where smaller metal particles formed. Indeed, in the precursor, the nickel ions are widely spread over and strongly attached to the support; also, the growth processes of the metal particles formed are expected to occur through surface diffusion rather than solution diffusion: specific (32) Bensalem, A.; Shafeev, G.; Bozon-Verduraz, F. Catal. Lett. 1993, 18, 165-171.
Reduction of Ni2+ Ions
Langmuir, Vol. 20, No. 1, 2004 211
Figure 2. TEM micrographs of reduced nickel nanoparticles: (a) unsupported Ni; (b) 1.10%Ni+SiO2 ; (c) 1.10%Ni/SiO2; and (d) after H2/300 °C treatment of the nanoparticles shown in part c.
Figure 3. Degree of reduction with time on the stream of nickel acetate. [N2H4] ) 0.5 M; Ni2+ ion amount ) 1.2 mmol; T ) 80 °C.
nickel interactions with support active sites slowed the migration of the Ni0 atoms for almost all the nuclei formed. In other words, the final size would be determined by the growth of the primary Ni0 particles formed. For the 1.10%Ni+SiO2 system, where nickel acetate was only in contact with a suspension of the silica support, mediumsize particles were produced: the metal-support interac-
tions were probably looser than those for the supported nickel acetate. In the absence of silica, no specific interaction inhibited the metal particle growth in the solution; also, much larger sizes were observed. Strikingly, the effect of the support was also to change the morphology of metallic nickel in the case of the 1.10%Ni/SiO2 catalyst: it drastically changed from mainly spherical particles (Figure 2a,b) to whiskers (Figure 2c), as shown by the TEM technique. To the best of our knowledge, this is the first example of supported nickel nanowhiskers obtained by chemical reduction. Nickel-Loading Effect. The effect of nickel loading on the reduction of supported nickel was investigated. The nickel percentage was varied in the range of 1-5% at a constant hydrazine concentration (0.5 M). The results obtained are reported in Figures 3 and 4. The reduction was total for all catalysts within 20-30 min. The apparent reduction rate appreciably increased with increasing nickel loading. This is reported in Figure 3, where it can be seen that the 1.40%Ni/SiO2 precursor is more rapidly reduced than the 1.10%Ni/SiO2 precursor. In contrast, the particle size increased with the nickel percentage (Figure 4): the higher the metal loading, the lower the number of particles. More strikingly, the particle morphology changed with the nickel loading: whiskers were formed for the supported metallic phase of low nickel content (0.86-1.40%) only (Figure 2c).
212
Langmuir, Vol. 20, No. 1, 2004
Figure 4. Effect of the nickel content on the particle size in the reduction of the supported catalysts. [N2H4] ) 0.5 M; T ) 80 °C.
In the nucleation step, the reduction increased with the number of nickel atoms available in the reaction media; also, the higher the nickel loading the higher the reduction rate (Figure 3). In the growth step, the support played a more important role. Indeed, we have seen previously that the final size would be determined by the growth of the primary Ni0 particles formed in the presence of the support. In such a case, the increase in the particle size with increasing loadings can also be correlated to the intervention of metal-support interactions. At low nickel loadings, a high proportion of nickel atoms are in close contact with the support; also, the specific nickel-support interactions, invoked previously, are expected to be important. These interactions tend to slow the growth processes, and small metal particles are formed. In contrast, at higher loadings, the interactions with the support are decreased, notably for those nickel atoms not directly or indirectly attached to the support; consequently, these atoms tend to agglomerate and larger metal particles are formed. These trends are those observed in heterogeneous catalysis for the reduction of supported Ni2+ ions by molecular hydrogen in gas-phase media.2-4,33 It was reported that the unsupported nickel particle size increased with the concentration of the nickel chloride precursor reduced by aqueous hydrazine in the presence of polyvinyl pyrrolidone.34 It was ascribed to the increase in the growth rate of the crystallites formed.34 The same observation was reported for unsupported nickel boride nanoparticles synthesized in a water-in-oil microemulsion system.35 These results are in good accordance with the apparent effect of nickel loading on the particle size for supported nickel catalysts (Figure 4). Hydrazine Concentration Effect. In the present case, at the constant hydrazine concentration of 0.5 M, the hydrazine/nickel ratio in fact changed as the nickel content changed from 1 to 5%. Also, the variation in the particle size would be rather correlated to the variation in the hydrazine/nickel ratio. This effect was studied by varying the hydrazine concentration at a constant nickel content using the 1.10%Ni/SiO2 catalyst. Figure 5 shows that the average diameters of nickel nanoparticles decreased with the increase of the hydrazine (33) Sholten, J. J.; Pijpers, A. P.; Husting, A. M. L. Catal. Rev. Sci. Eng. 1985, 27, 151-206. (34) Chou, K.-S.; Huang, K.-C. J. Nanoparticle Res. 2001, 3, 127132. (35) Chen, D.-H.; Wu, S.-H. Chem. Mater. 2000, 12, 1354-1360. (36) Nagy, B. J. Colloids Surf. 1989, 35, 201-213.
Boudjahem et al.
Figure 5. Effect of hydrazine concentration on the particle size in the reduction of 1.10%Ni/SiO2. T ) 80 °C.
Figure 6. Treatment of 4.30%Ni/SiO2 catalyst under H2 flow at the programmed temperature.
concentration and seemed to approach a constant value when the hydrazine concentration was above 0.6 M. This is in good accordance with the results of hydrazine reduction of nickel chloride in a microemulsion system.36 The authors suggest that the phenomenon could be explained by the influence of the reduction rate on the nucleation. This could also hold for our supported nickel catalysts. In such a case, at a low hydrazine concentration, the reduction rate was slow and, consequently, a small amount of nuclei formed, whereas the nuclei formed rapidly grew, leading to larger particles. With the increase of the hydrazine concentration, the enhanced reduction favored the generation of many nuclei and the formation of smaller nickel particles. For a large enough hydrazine concentration (N2H4/Ni2+ > 10), the reduction rate was much faster than the nuclei formation, so the number of nuclei tended to a constant value. Therefore, the size of the resultant nickel nanoparticles was not further reduced and remained at a constant value. 3.4. Thermal Treatment of the Nickel-Supported Nanoparticles. Decomposition of the Organic Matrix of the Fresh Nickel Particles. The thermal study of the nickel nanoparticles evidenced the presence of an organic matrix and gave some structural information on the fresh samples. Indeed, for all catalysts, the treatment under H2 at the programmed temperature led to the formation of methane
Reduction of Ni2+ Ions
Langmuir, Vol. 20, No. 1, 2004 213
and, to some extent, carbon monoxide (Figure 6). Molecular hydrogen was also formed when the fresh particles were flowed under argon instead of hydrogen (not shown). Methane and carbon monoxide molecules resulted from the hydrogenation/hydrogenolysis of the acetate fragment of the catalyst precursor. For methane, the overall process would be as follows:
CH3COO + 9/2H2 f 2CH4 + 2H2O
(2)
These results are in good agreement with previous work, which showed that the fresh supported nickel particles are inactive toward hydrogen adsorption and in benzene hydrogenation.29 They became active only after a hydrogen thermal treatment. This confirms that the supported particles are most probably covered by the organic matrix. In addition, the high temperature of decomposition of the organic fragment (Figure 6) indicated that the latter was strongly attached to the reduced nickel phase. The amount of methane formed after hydrogen thermal treatment (Table 1) was much lower for the unsupported (176 µmol‚g-1 Ni) than for the supported (3345-9037 µmol‚g-1 Ni) particles. This was accounted for by the metal coalescence extent during the reduction process; once formed, the large unsupported metal particles adsorbed low amounts of the organic matrix, whereas the much smaller supported metal particles adsorbed much greater amounts. For the supported particles, the amount of methane evolved increased as the nickel content decreased: it was 9037µmol‚g-1 Ni for 0.86%Ni/SiO2 against 3345µmol‚g-1 Ni for 4.30%Ni/SiO2. These results were also accounted for by the correlation between the particle adsorption capabilty and size: the lower the particle size, the greater the amount of the organic fragment adsorbed. From XRD particle size and eq 2 stoechiometry, we calculated the CH3COO/Nisurf ratio, that is, the number of acetate entities attached per surface nickel atom (Table 1). It can be seen from Table 1 that the nonsupported particles were embedded in about 13.0 monolayers of acetate species, whereas the supported particles adsorbed approximately 1.0-3.5 monolayers only. It appears that the smaller the particle size, the lower the number of organic monolayers adsorbed on their surface. It can be concluded that the support more efficiently stabilized the nickel particles than the organic matrix did. Effect of the Thermal Treament on the Metal Particle Size and Morphology. After hydrogen treatment at the constant temperature of 300 °C for 2 h, the supported particles exhibited the characteristic band of metallic nickel. Both XRD spectra and TEM micrographs showed no change in the particle size for low nickel-loading catalysts (0.86-1.40%) after the thermal treatment. In contrast, for the higher nickel loadings of 2.73 or 4.30% the particle size increased: from 11.8 or 13.8 nm to 16.6 or 24.7 nm, respectively. In addition, the TEM experiments indicated that the metal phase morphology changed after this treatment for catalysts of low nickel content: spherical
nanoparticles homogeneously dispersed (Figure 2d) on the support were formed in place of whiskers (Figure 2c). The thermal treatment effect on the metal particle size could also be correlated to metal-support interactions. The metal nickel particles coalesced through surface diffusion on the support. Such a mechanism probably holds for those Ni0 particles not in strong interaction with the support. For a high nickel content, part of the nickel phase is not directly or less directly interacting with the support and will probably migrate more easily, accounting for a bigger average particle size during the heat treatment. In contrast, for a lower nickel content, the metal particles were rather stabilized by the support and less prone to coalescence. Conclusions The preparation of nickel nanoparticles was carried out by the reduction of nickel acetate with aqueous hydrazine at a temperature of 80 °C and a pH of 10-12. The effect of silica of quartz-type on the reducibilty of nickel acetate and metal particle formation was studied. In the absence of silica, the reduction was partial (45%) and the metal particles obtained were spherical in shape. They formed a film at the liquid-gas-phase interface or precipitated as a powdered solid with an average particle size of 120 nm. In the presence of a suspension of silica, the reduction was total and a nickel film was deposited on the support with an average particle size of 25 nm only. When nickel acetate was previously impregnated on silica, a further decrease of the mean particle size to about 3 nm was observed. On the other hand, the supported particle size increased with increasing the nickel content and, for a given loading, with a decreasing hydrazine/nickel ratio. In addition, strikingly, the nickel phase morphology changed to whiskers at low loadings. Hydrogen thermal treatment of the nickel particles led to the formation of gaseous methane arising from the decomposition of the organic acetate residue, through hydrogenation/hydrogenolysis reactions. This treatment also changed the nickel phase shape from whiskers to spherical particles. The effect of the support on the reduction of nickel acetate was ascribed to surface defects of the highly crystallized silica, which played the role of reduction sites: electrons are transferred from the reducing agent to the metal ion in the vicinity of the surface defect. The particle growth is believed to occur through surface diffusion. The variation of the particle size with the nickel content was ascribed to metal-support interactions, which were stronger at low loadings than at high loadings; consequently, small and large particles were obtained, respectively. As to the variation of the particle size with the hydrazine/nickel ratio, it could be explained by the influence of the reduction rate on the nucleation and nuclei growth. Finally, it appeared that the organic matrix covered the fresh metal particles formed. LA035120+
