3962
J. Phys. Chem. A 2010, 114, 3962–3968
Preparation of Multiwalled Carbon Nanotube-Supported Nickel Catalysts Using Incipient Wetness Method† Pooya Azadi,‡ Ramin Farnood,*,‡ and Emanuel Meier§ Department of Chemical Engineering and Applied Chemistry, UniVersity of Toronto, Toronto, Canada, Department of Chemistry, ETH, Zurich, Switzerland ReceiVed: July 31, 2009; ReVised Manuscript ReceiVed: August 29, 2009
In this paper, a systematic study on preparation of multiwalled carbon nanotube (MWCNT)-supported nickel catalyst is pursued. Functional groups are introduced on the surface of MWCNTs using nitric acid, sulfuric acid, and partial oxidation in air. Nickel oxide nanoparticles are formed on the surface of functionalized multiwalled carbon nanotubes by incipient wetness impregnation of nickel nitrate, followed by calcination in air. The effects of acid type and concentration, acid treatment time, partial oxidation, nickel loading, precursor solvent, and calcination temperature on the size of the nickel nanoparticles and homogeneity of the composite material are evaluated. Characteristics of the Ni/MWCNT catalysts were examined using BET, scanning transmission electron microscopy, X-ray diffraction, thermogravimetric analysis in air and nitrogen, temperatureprogrammed reduction, X-ray photoelectron spectroscopy, acid-base titration, and ζ-potential analyzer. Results of this work are useful for formulating CNT-supported nickel catalysts for a wide range of different applications, such as reforming of hydrocarbons, catalytic hydrothermal gasification of biomass, and energy storage. Introduction Since their discovery in 1991, many applications have been suggested for carbon nanotubes (CNTs). Carbon nanotubes offer excellent properties as a catalyst support, such as proper pore sizes, moderate to high specific area, great thermal stability and stability in acidic or basic environments. Due to the novel properties of these cylindrical carbon molecules, they can act as a catalyst support. However, due to lack of oxygen functional groups on their outer surface and their hydrophobicity, formation of bonds between a metal precursor and a CNT is not an easy task as compared to the metal oxide-supported catalysts, such as alumina. Once synthesized, CNTs contain some impurities. Hence, some pretreatment steps are usually implemented in preparation of CNT-supported catalysts (such as mild acid treatment) for removal of amorphous carbon and the metal nanoparticles that are used to catalyze the CNT synthesis. Then a controlled number of oxygen functional groups are added to the outer surface by either partial oxidation in air or by means of a liquid oxidizing agent, such as hydrogen peroxide, nitric acid, etc.1 A few techniques2 have been developed for decoration of carbon nanotubes. Among them are electrochemical methods,3–5 wet impregnation,6 and incipient wetness.7 Each method leads to a certain degree of control over nanoparticle size and metal dispersion. The functionalized CNTs readily interact with the metal precursors, which either can be reduced directly to metal in a reducing atmosphere or it can be calcined to create metal oxides, followed by a reduction in hydrogen flow. Regarding the oxidation of CNT, it is found that acid concentration and treatment time play a significant role in functionalization of the nanotubes. In addition, it has been reported that the carboxylic groups are the dominant groups added on the CNT during the
oxidative treatment by nitric acid. Furthermore, acid treatment may increase the specific area of the CNT, particularly for multiwalled carbon nanotubes by oxidizing and dissolving the outer walls of CNTs, as well as breaking the CNT particles. Many researchers have investigated the activity of CNTsupported catalysts for variety of applications, including synthesis of more carbon nanotubes,1 reforming of hydrocarbons,7 hydrogen storage,8 and hydrogenation,9 but there is a lack of systematic study on different aspects of particle formation on the surface of carbon nanotubes. Experimental methods Multiwalled carbon nanotubes with average length and diameter of 7 µm and 120 nm were obtained from SigmaAldrich, Oakville, Canada. An oxidative pretreatment in boiling nitric or sulfuric acid is conducted to introduce oxygenated functional groups onto the surface of the nanotubes. The following conditions were used throughout catalyst preparation unless mentioned otherwise. The nitric acid concentration and treatment time were 10 M and 5 h, respectively. In all experiments, 0.2 g of MWCNT was dispersed in 50 mL of acid, and the mixture was boiled using a hot plate and reflux system. The functionalized CNTs were washed with water, centrifuged twice, and dried at 110 °C overnight. Following this step, the MWCNTs were impregnated with a nickel precursor, followed by calcination at 350 °C for 3 h. The nickel loadings were controlled by changing the concentration of the nickel nitrate. A Quantachrome catalyst characterization unit TABLE 1: Elemental Composition of Nickel Decorated MWCNT Obtained by XPS Nitric Acid Concentration preparation step
carbon
oxygen
nickel
nitrogen
no treatment acid treated impregnated calcined
98.5 89 75 82
1.5 11 19 11.6
0 0 2.8 6.4
0 0 3.2 0
†
Part of the special issue “Green Chemistry in Energy Production Symposium”. * Corresponding author. E-mail:
[email protected]. ‡ University of Toronto. § ETH.
10.1021/jp907403b 2010 American Chemical Society Published on Web 10/12/2009
Preparation of MWCNT-Supported Nickel Catalysts
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3963
Figure 1. TGA of fresh and FMWCNT in air.
average size of nickel oxide nanoparticles was calculated using the Scherrer equation and corrected to account for the instrument line-broadening obtained from standard LaB6.10 The metal dispersion is defined as the percentage of metal atoms exposed on the surface, and it can be calculated based on crystallite size. To determine the number of carboxyl groups on the MWCNTs, 20 mg of functionalized carbon nanotubes was dispersed in 50 mL of 0.001 M NaOH solution by an ultrasonic bath and then titrated with 0.001 M HCL solution. Results and Discussions
Figure 2. ζ-Potential of FMWCNT vs nitric acid concentration.
In this section, physical properties of MWCNTs are given, and different aspects of the functionalization of carbon nanotubes and their impact on surface and structural properties are discussed. Following this part, the effect of different parameters on the impregnation and calcination of functionalized multiwalled carbon nanotubes (FMWCNTs) will be presented. Physical Properties of MWCNT. According to the manufacturer, the average length, diameter and density of the multiwalled carbon nanotubes are 7 µm, 120 nm, and 2.1 g/mL, respectively. On the basis of our nitrogen adsorption tests, the specific area of these multiwalled carbon nanotubes is equal to 14.3 m2/g. The total pore volume and average pore diameter are equal to 0.1 mL/g and 27.7 nm, respectively. Functionalization of MWCNTs
Figure 3. Concentration of carboxyl groups vs treatment time.
TABLE 2: Weight Percent of the Removed Functional Groups Obtained from TGA in Nitrogen sample
acid
no treatment acid treated acid treated acid treated acid treated
nitric nitric nitric sulfuric
concn [M] treatment time [h] % removed 10 10 16 10
5 24 5 5
0.15 1.23 1.43 13.2 2.01
was utilized to study the pore size distribution of the carbon nanotubes as well as reduction of nickel oxide nanoparticles in flowing hydrogen at a heating rate of 10 °C/min. Scanning transmission electron microscopy (STEM) was performed on a Hitachi HD-2000 STEM microscope. Thermogravimetric analysis (TGA) scans were obtained using a TGA Q500 apparatus from TA Instruments. X-ray diffraction (XRD) patterns obtained from a Philips XRD system at a scanning rate of 0.015°/s. The
To be able to uniformly attach metal nanoparticles on the carbon nanotubes, the surface of MWCNTs should be chemically modified by an oxidizing agent. It is known that oxidative treatment not only introduces oxygen-containing functional groups, such as carboxyl and hydroxyl, onto the outer surface of carbon nanotubes, but also, it increases the specific surface area of the multiwalled carbon nanotubes. The BET surface area of the MWCNTs used in this study changed from 14.3 to 19.5 m2/g after 5 h of treatment in 10 M nitric acid. The extent of this surface treatment can be determined by different analytical methods, such as UV spectroscopy, X-ray photoelectron spectroscopy (XPS), and acid-base titration. Table 1 summarizes the elemental composition of the catalysts at different preparation steps obtained from XPS. As it is expected, the amount of oxygen increases after the acid treatment. Then, addition of nickel nitrate hexahydrate further increases the oxygen content of the surface. Finally, nickel nitrate hexahydrate is decomposed to nickel oxide, and as a result, the surface concentration of the oxygen decreases. Figure 1 depicts TGA of fresh and FMWCNT in air. Expectedly, due to an increase in defective sites and addition of oxygen-containing functional groups, the cracking temper-
3964
J. Phys. Chem. A, Vol. 114, No. 11, 2010
Azadi et al.
Figure 4. TEM micrographs of 20% Ni/FMWCNT, 5 h oxidative treatment with 4 M (left) and 16 M (right) nitric acid.
raised to 800 °C at a heating rate of 10 °C/min and kept at 800 °C for 1 h. Table 2 illustrates the weight percent of removed functional groups by thermal annealing of functionalized carbon nanotubes. It should also be noted that almost no weight loss is observed below 700 °C; therefore, the weight loss corresponds to removal of functional groups rather than the adsorbed species. From this table, increasing acid concentration has a more dramatic effect on the surface concentration of the functional groups. In addition, it is realized that at the same concentration, sulfuric acid leads to a higher concentration of functional groups than nitric acid does. Figure 5. Nickel crystallite size vs nitric acid concentration for 20% Ni/FMWCNT.
Impregnation of Functionalized MWCNT
ature of nanotubes decreases upon subjecting them to an oxidative treatment. This result is consistent with the previous literature. It is also found that treatment in boiling nitric acid can decrease the ζ-potential of the carbon nanotubes to more negative values. This effect is more pronounced for treatment in higher acid concentrations. Figure 2 shows the ζ-potential after 5 h treating MWCNTs at different acid concentrations. Increasing the negative surface charge of the FMWCNTs improves their dispersion in water and enhances the impregnation and adsorption of precursor on their surfaces. Acid Treatment Time. The concentration of carboxyl groups on the surface of FMWCNTs is measured by acid-base titration. Figure 3 shows the relationship between acid treatment time and the concentration of the carboxyl groups. On the basis of this figure, the concentration of carboxyl groups significantly increases at prolonged treatment times. To better evaluate the effectiveness of different oxidative treatments, a thermogravimetric analysis in an inert environment is performed to thermally remove the functional groups from the surface of FMWCNTs. For all samples, the temperature is
Effect of Acid Concentration. It has been shown that the number of oxygen-containing functional groups has a positive relationship to the concentration of the utilized acid. Figure 4 demonstrates two TEM micrographs of carbon nanotubes decorated with nickel on their outer surface. On the basis of this image, it is apparent that dispersion of nickel nanoparticles is considerably better when MWCNT is functionalized in a more concentrated oxidizing agent. Figure 5 shows the effect of surface modification using different acid concentrations on the average nickel crystallite size as determined using XRD. In all cases, nickel loading is 20% by weight, and treatment time is set to 5 h. Treatment Time. Titration of functionalized MWCNT indicated that prolonged treatment times result in a higher concentration of functional groups. Therefore, it is anticipated that nickel dispersion increases with the extent of treatment. Figure 6 exhibits nickel nanoparticles on 20% Ni/MWCNT catalysts with and without surface functionalization. Using the Scherrer equation, nickel crystallite size were calculated, and the results are plotted in Figure 7. The average crystallite size decreases from 30 to about 15 nm upon surface treatment of nanotubes for 24 h in 10 M nitric acid.
Figure 6. TEM micrographs of 20% Ni/MWCNT, with no treatment (left) and with 24 h treatment (right) in 10 M nitric acid.
Preparation of MWCNT-Supported Nickel Catalysts
Figure 7. Nickel crystallite size vs treatment time for 20% Ni/ FMWCNT.
J. Phys. Chem. A, Vol. 114, No. 11, 2010 3965
Figure 9. Nickel crystallite size vs treatment type for 20% Ni/ FMWCNT.
Figure 8. Effect of oxidative treatment on the cracking temperature of 20% Ni/MWCNT with and without acid treatment.
To better assess the effect of functionalization, oxidation of 20% Ni/MWCNT with and without surface modification is studied by TGA (Figure 8). It is obsereved that 20% Ni/ FMWCNT has a single cracking temperature at about 550 °C, whereas the weight derivative curve of the composite made of 20% Ni/FMWCNT can be deconvoluted into two separate peaks whose centers are at 570 and 740 °C. This further confirms that without functionalization, a significant number of carbon nanotubes have no nickel attached to their surface while other ones are decorated with nickel nanoparticles. This observation illustrates the nonuniformity of nonchemically modified catalysts, as opposed to modified ones. Treatment Method. Four different oxidizing agents (HNO3, H2SO4, a mixture of HNO3 and HCl, and partial oxidation in air) are considered for addition of functional groups onto MWCNTs. Partial oxidation is performed in air at 600 °C until 20% of the weight is burned. As can be seen in Figure 9, partial oxidation in air cannot improve the dispersion of nanoparticles. For other treatment methods, under the same conditions (concentration and time), a sulfuric acid-treated sample has the smallest crystallite size as compared to the other two. This finding suggests that at the same acid molar concentration, sulfuric acid treatment can lead to a more uniform Ni nanoparticle dispersion. Nickel Loading. Ni/FMWCNT catalysts with five different nickel loadings are prepared using incipient wetness method. The XRD pattern of these samples is given in Figure 10. The first peak at 26° corresponds to CNTs; the other two peaks at 37° and 43° are NiO(111) and NiO(200), respectively. However, there is a very small MWCNT peak at around 43° that coincides with NiO(200).
Figure 10. XRD patterns of decorated carbon nanotubes at different nickel loading in wt %.
Figure 11. Nickel crystallite size and dispersion vs nickel loading in wt %. Treatment time: 5 h.
The nickel oxide crystallite size and its corresponding dispersion for FMWCNTs are presented in Figure 11. It is noted that nickel nanoparticles become larger at higher loadings, and as a result, better dispersions can be achieved only at low nickel percentages. Figure 12 shows TEM of decorated carbon nanotubes with 5%, 20%, and 50% nickel. As was indicated by XRD, the nickel crystallites become larger at higher nickel loadings, resulting in a poor dispersion. Furthermore, it is noticed that at high nickel
3966
J. Phys. Chem. A, Vol. 114, No. 11, 2010
Azadi et al.
Figure 12. TEM micrographs of Ni/FMWCNT at different metal loadings: (a) 5%, (b) 20%, (c) 50%, and (d) porous nickel crystals formed on 50% Ni/FMWCNT.
Figure 14. ζ-Potential of Ni/FMWCNT vs nickel loading in wt %.
Figure 13. Derivate of weight loss vs temperature for different nickel loadings.
loadings, nickel crystallites are somewhat porous with a pore diameter of about a few nanometers. This phenomenon has been observed for nickel nanobelts, and it has been suggested that it could be an artifact of the interaction effect of the TEM electron beam with nickel crystalls.11,12 Thermogravimetric analysis of the samples with different nickel loadings (Figure 13) indicated that at lower nickel loadings (
