Effects of preparation and structure of cerium-zirconium mixed oxides on diesel soot catalytic combustion

Applied Catalysis A: General 413–414 (2012) 292–300

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Effects of preparation and structure of cerium-zirconium mixed oxides on diesel soot catalytic combustion Claudinei F. Oliveira, Fillipe A.C. Garcia, Daniel R. Araújo, Julio L. Macedo, Sílvia C.L. Dias, José A. Dias ∗ Laboratório de Catálise, Instituto de Química, Universidade de Brasília, Brasília, DF 70910-970, Brazil

a r t i c l e

i n f o

Article history: Received 8 July 2011 Received in revised form 11 November 2011 Accepted 16 November 2011 Available online 25 November 2011 Keywords: Cerium-zirconium mixed oxide Sol–gel method Diesel soot combustion Zirconia Ceria

a b s t r a c t Mixed oxides of Cex Zr1−x O2 (0.1 ≤ x ≤ 0.9) were prepared by sol–gel method, in aqueous ammonia solution with CeCl3 ·7H2 O and ZrOCl2 ·8H2 O as precursors, and employed in diesel soot combustion. The catalysts were characterized by XRF/EDX, nitrogen adsorption, TGA/DTG, powder XRD, FTIR/DRIFTS and Raman. In addition, the acidity was evaluated by adsorption and desorption of pyridine. XRD indicated the formation of solid solutions that progressively distorted from cubic into tetragonal lattices. Raman studies confirmed that the Ce–O bonding was stronger in the mixed oxide series because of the cell contraction, as a result of the zirconium insertion. Ce0.8 Zr0.2 O2 was the most acidic and active material, shifting the combustion temperature (Tm ) from 622 to 547 ◦ C (loose contact) or 404 ◦ C (tight contact). The calculated activation energy for the catalytic combustion of this optimized oxide attested that the combustion temperature was lower under all conditions, compared to the thermal process. The catalyst was utilized five times without any appreciable loss of activity and maintained its structural properties. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Research aimed to solve the ambient pollution caused by emission of diesel soot particulates is currently one of the most difficult tasks. One of the challenges is to regenerate the catalyst in the diesel filter to retain particulate (DPF), as this device undergoes broad reaction conditions for regeneration. Another drawback is the slow reaction process due to the loose contact between the particulates and the catalyst. The low mobility and relative high diameter (dp ) between 10 and 100 nm inhibit its penetration in microporous catalysts [1,2]. Moreover, there is a wide variation in the temperature of exhausted gases (200–600 ◦ C), depending on the motor design and power output [3]. Therefore, an efficient catalyst must work at low temperatures and be thermally stable for regeneration purposes. In the mid 1970s, the technology of three-way catalysts (TWC) symbolized a major breakthrough in the development of devices for automotive pollution control [4,5]. The devices consisted of CeO2 as the main component to promote oxygen storage and release capacity (OSC), which is closely related to the efficiency of TWC system [4,5]. Another breakthrough was the advent of new pro-

∗ Corresponding author at: Universidade de Brasília, Campus Darcy Ribeiro, Asa Norte, Instituto de Química, Laboratório de Catálise (A1-62/21), caixa postal 4478, Brasília, DF 70904-970, Brazil. Tel.: +55 61 3107 3846; fax: +55 61 3368 6901. E-mail address: [email protected] (J.A. Dias). URL: http://www.unb.br/iq/labcatalise (J.A. Dias). 0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2011.11.020

moters based on Cex Zr1−x O2 mixed oxides. It was discovered that these solid solutions improved the thermal stability and OSC of the TWC [4]. One of the key properties of these promoters is the fast redox cycle of Ce3+ and Ce4+ and the high mobility of O2− ions in the lattice of ceria compounds [6]. Thus, studies on oxides are very relevant for understanding their fundamental characteristics and suitability in the automotive industry. Preparation of mixed oxides of different proportions of Ce:Zr results in materials with tunable properties compared to CeO2 or ZrO2 . They show high thermal stability, oxygen storage, fluidity, and mobility in a modified lattice [7–11]. These enhanced properties can be related to the structure of the mixed oxide. The substitution of Ce by Zr is considered a modification of the CeO2 lattice upon insertion of zirconium, because of the smaller size of Zr4+ (84 pm) versus Ce4+ (97 pm). Increasing the amount of Zr, the CeO2 lattice (cubic fluorite structure, Fm3m space group) distorts to form tetragonal P42 /nmc space group (t , t and t-phase) [12–14]. Accordingly, no well-defined frontiers between these phases exist, as the distortion continuously changes with sample composition. In addition, this distortion is sensitive to the particle size of the mixed oxide [5]. It is known that different preparation methods of Cex Zr1−x O2 lead to materials with variable structure and physico-chemical properties [15,16]. For instance, nanotubes of Cex Zr1−x O2 were prepared using membrane templates of polycarbonate leading to a mixture of phases (cubic and tetragonal), when x = 0.5, 0.7 and 0.9 [15]. In another work, the mixed oxides were synthesized by sol–gel method, and different phases were observed, such as cubic (x = 0.16)

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

and a minor tetragonal phase (0.50 ≤ x ≤ 0.75) in a predominant cubic structure [16]. This work deals with preparation of a wide range of ceriumzirconium mixed oxides (Cex Zr1−x O2 , 0 ≤ x ≤ 1.0) synthesized by the sol–gel method. The materials were characterized by elemental and textural analysis, XRD, Raman, and DRIFTS, as well as by pyridine adsorption/desorption to evaluate the acidity. All catalysts were tested in the combustion of diesel soot under loose and tight contact, using Printex® U as model particulate. The most active catalyst was studied further to obtain the kinetic oxidation parameters using model free kinetics method, thermal stability data and to determine its reutilization in this process. 2. Experimental 2.1. Preparation of mixed oxides of Cex Zr1−x O2 Cex Zr1−x O2 (x = 0; 0.1; 0.2; 0.3; 0.4; 0.5; 0.6; 0.7; 0.8; 0.9 and 1.0) materials were synthesized by the sol–gel method from 0.3 mol L−1 aqueous solutions of ZrOCl2 ·8H2 O and CeCl3 ·7H2 O (both reagents obtained from Sigma and used without further purification). The solutions, with their respective calculated molar ratios, were added at a rate of 2 mL min−1 to aqueous ammoniacal solution (6.0 mol L−1 ) under magnetic stirring in an ultrasonic bath at room temperature (25 ◦ C). The ratio of precursor volumes to mineralizer agent was 1:1. The formed gel was completely dried under magnetic stirring at 80 ◦ C in a fume hood. The remaining solid was washed with demineralized water to a neutral pH level. Then, the solid was filtered, dried at 100 ◦ C for 2 h in a vacuum oven, ground to finer particles and finally calcined at 650 ◦ C for 4 h to remove all residues of chloride and ammonium ions. Additional tests aiming the stability study of the materials were carried out with calcinations at 850 and 1000 ◦ C for 4 h as well as in the presence of different amounts of water. 2.2. Gas phase adsorption of pyridine Gas phase pyridine (Py) adsorption was conducted in a customized apparatus [17]. Platinum crucibles loaded with about 40 mg of the samples were placed in shallow porcelain crucibles and inserted into a glass reactor adapted to a tubular furnace (Model F21135, Thermolyne). The materials were dried under analytical N2 (99.999% at 100 mL min−1 ) at 300 ◦ C for 1 h, cooled to 100 ◦ C, and then gaseous pyridine diluted in N2 was passed through the samples for another hour. Finally, the system was held at 150 ◦ C under N2 for 2 h to ensure all physically adsorbed pyridine was removed. After this, the materials were analyzed by TG/DTG and DRIFTS.

293

1 s step−1 ) with Cu K␣ radiation of 1.5418 A˚ (40 kV and 30 mA) and a graphite monochromator. Raman spectra were obtained in the samples packed into sample cups in a Perkin Elmer Raman Station 400F equipped with a high sensitivity open electrode CCD detector. The laser excitation (Nd–YAG) wavelength and power were 785 nm and 50 mW, respectively. FT-infrared spectra were obtained with a Thermo Scientific spectrometer model Nicolet 6700 FTIR with a DTGS detector using 256 scans and a spectral resolution of 4 cm−1 . In the transmittance mode, each sample was pressed in dried 1-wt% KBr (Merck) pellets. DRIFTS measurements were collected in the Smart Diffuse Reflectance accessory using the powder with no KBr. The reflectance was acquired against an alignment mirror (256 scans with a 4 cm−1 resolution) and then converted to Kubelka–Munk spectra after baseline corrections. Specific surface area, volume of pores and porosity of samples were measured based on adsorption and desorption isotherms of nitrogen obtained at −196 ◦ C in a Micromeritics model ASAP 2020C instrument. The mixed oxides were previously calcined at 650 ◦ C for 4 h prior to measurements. The degasification was performed under vacuum at 200 ◦ C for 22 h before adsorption measurements. The BET specific surface area was calculated from the adsorption branches in the relative pressure range of 0.06–0.20. The average pore size and the pore volume were calculated from the desorption branches using the Barrett–Joyner–Halenda (BJH) method. Thermal analysis was conducted in a simultaneous TG–DSC model SDT 2960 from TA Instruments, with scan rate of 10 ◦ C min−1 , from room temperature up to 800 ◦ C under nitrogen (99.999%) or synthetic air flow of 100 mL min−1 . Synthetic air was acquired from White Martins Praxair Inc. at 99.999% (O2 + N2 ) with 20 ± 0.5% O2 . The number of acid sites was determined by quantitative analysis of TG/DTG curves of the materials after pyridine adsorption. The method, developed in our laboratory, involves the mass loss analysis of the materials before and after pyridine adsorption, taking into consideration the hydration of each sample and has been described elsewhere [18,19]. The activation energy of the catalytic and non-catalytic soot oxidation was obtained by TG experiments, which were performed at 2, 5, 10, 15 and 20 ◦ C min−1 . The degree of conversion was recorded at 20, 40, 50, 60 and 80%. To calculate the activation energies and the pre-exponential factor of the reactions, the model-free kinetics was used. The equipment was calibrated for each heating rate under the same flow, to obtain accurate data.

2.4. Catalytic combustion of diesel soot 2.3. Analytical techniques for characterization of the mixed oxides The catalysts were characterized by elemental analysis using XRF/EDX, powder X-ray diffraction (XRD), infrared spectroscopy (FTIR/DRIFTS), Raman, textural properties using low temperature nitrogen adsorption–desorption and thermal analysis (TG/DTG/DTA). The compositions of the samples were evaluated by X-ray fluorescence in a Shimadzu EDX 720 spectrometer using a rhodium X-ray tube. The XRF spectra were collected under vacuum (<45 Pa) using 2-channels where the X-ray source was set at 50 and 15 kV for Ti–U and Na–Sc ranges, respectively. The analyses were performed on powders placed inside the sample holder using a polypropylene film. The Ce and Zr content were determined by the Quali-Quant method of fundamental standards. Powder XRD was obtained in a Bruker D8 FOCUS diffractometer between 5◦ < 2 < 70◦ (0.02◦ step−1 and integration time of

The catalytic activities of the mixed oxides were evaluated in the oxidation reaction of standard soot produced in a fuel burner (Printex® U by Evonik). This model soot was analyzed and had a composition in agreement with the literature [20]. The Printex® U was mixed with the catalyst at a mass ratio of 1:20, respectively. Considering that the contact between the catalyst and the soot is crucial in the reaction [21], it was tested at tight and loose contacts. In the tight mode, the mixing process was carried out in an agate mortar and pestle for 5 min to promote firm contact between the components. The loose contact was conducted in the agate mortar with a spatula for 5 min to completely mix the components. For each reaction, about 15 mg of the mixture was placed in the platinum crucible and placed into the TG–DSC equipment. The mixture was subjected to a temperature ramp at a rate of 10 ◦ C min−1 from room temperature (∼26 ◦ C) to 700 ◦ C under synthetic air flow (100 mL min−1 ). All experiments used ␣-Al2 O3 as the reference.

294

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

Table 1 Elemental and physico-chemical properties of Cex Zr1−x O2 (0 ≤ x ≤ 1) materials. Mixed oxides Nominal

Actual

ZrO2 Ce0.1 Zr0.9 O2 Ce0.2 Zr0.8 O2 Ce0.3 Zr0.7 O2 Ce0.4 Zr0.6 O2 Ce0.5 Zr0.5 O2 Ce0.6 Zr0.4 O2 Ce0.7 Zr0.3 O2 Ce0.8 Zr0.2 O2 Ce0.9 Zr0.1 O2 CeO2

ZrO2 Ce0.09 Zr0.91 O2 Ce0.17 Zr0.83 O2 Ce0.28 Zr0.72 O2 Ce0.33 Zr0.67 O2 Ce0.45 Zr0.55 O2 Ce0.55 Zr0.45 O2 Ce0.67 Zr0.33 O2 Ce0.78 Zr0.22 O2 Ce0.91 Zr0.09 O2 CeO2

a b c d e f

SBET (m2 g−1 )

Smicro a (m2 g−1 )

Vp b (cm3 g−1 )

Ps c (nm)

DXRD d (nm)

DBET e (nm)

APy f (mmol g−1 )

19.4 22.2 41.1 42.3 40.7 34.6 39.9 37.4 41.2 38.6 30.3

3.1 2.2 2.0 3.2 3.9 2.1 2.8 3.1 3.1 2.8 5.4

0.16 0.08 0.12 0.12 0.12 0.11 0.12 0.11 0.12 0.12 0.14

19.3 7.9 7.9 7.9 7.8 8.3 7.9 8.2 7.8 7.9 13.7

19.6 11.0 9.3 8.4 8.0 5.8 5.1 8.8 11.8 12.7 19.6

50.0 42.7 22.7 24.0 23.2 27.1 23.9 24.0 21.5 21.7 28.8

n.a. 0.034 0.031 0.028 0.025 0.036 0.028 0.039 0.043 0.041 n.a.

Surface area of micropores calculated by t-plot method. BJH desorption cumulative volume of pores. Average pore size diameter (4 V/A) calculated by BJH desorption method. Average domain of particle size calculated by the Scherrer equation at 2 ∼ = 29◦ . Average particle diameter (DBET ) = 6000/( × SBET ) Total adsorbed pyridine obtained by TG/DTG analysis.

3. Results and discussion 3.1. Elemental and physico-chemical properties of the materials As the stoichiometries of the mixed oxides were close, for incremental changes in the oxide composition, the elemental analysis of each prepared material was of fundamental importance. Table 1 (columns 1 and 2) shows the nominal and the real values of Ce and Zr obtained by XRF-EDX. The actual values obtained were very close to the nominal values. For simplicity, the nominal formula will be referred to in this work. The approximated values matched, mainly due to the previous determination of the hydration degree of each precursor and the rapid weighting of the salts during preparation. Residues of chloride and ammonium were not detected either by XRF or CHN analyses, respectively. Significant information about the textural properties of the mixed oxides was obtained by the analyses of nitrogen adsorption–desorption isotherms (Fig. S1). According to IUPAC classification, the BET isotherms for all the synthesized materials were type IV with capillary condensation at relative low pressures (P/P0 ) and pore filling with the liquid adsorbate molecules at higher pressures. This behavior was characteristic of mesoporous materials. These materials also showed different types of hystereses depending on the stoichiometry. The main type was H3, which is related to split-shaped pores formed by non-rigid plate aggregates. Nonetheless, the hysteresis of the type H1 was observed for CeO2 and ZrO2 , whereas type H2 predominated for Ce0.1 Zr0.9 O2 and Ce0.7 Zr0.3 O2 . Type H1 is related to tubular pores with circular sections or polygons, while H2 hysteresis is indicative of the presence of inkpot-shaped pores having narrow necks and large cavities, but irregular in terms of shape and size distribution [22]. Other textural properties of Cex Zr1−x O2 (0 ≤ x ≤ 1.0) are listed in Table 1 (columns 3–8). The specific surface areas for these oxides ranged from 19.4 to 42.3 m2 g−1 , pore volumes from 0.08 to 0.16 cm3 g−1 , and average pore size diameters from 7.8 to 19.3 nm. Considering only the mixed oxides, the variation in these parameters was much lower: the average surface area was about 37.6 m2 g−1 , the pore volumes about 0.11 cm3 g−1 , and the pore sizes about 8 nm. These values indicated the relative structural similarities of these materials. Notably, the contribution of microporous areas was very low (average 2.8 m2 g−1 ) for all materials, which is less than 7.5% of total surface area. Although the variation of these parameters was not completely regular, they demonstrated the major mesoporous structure of the synthesized oxides.

The domain of crystallite sizes (DXRD ) were on average below 8.9 nm for all mixed oxides. Synthesized ceria and zirconia showed sizes of about 19.6 nm. In addition, the average particle sizes (DBET ) exhibited values about three times greater than the ones obtained by the Scherrer’s equation. These differences can be attributed to the formation of small weakly aggregated nanocrystallites (average around 23 nm). The sizes, calculated by XRD, decreased as the amount of cerium increased up to Ce0.6 Zr0.4 O2 forming crystallites with sizes as low as 5.1 nm. Nonetheless, the calculated (DBET ) aggregates had about the same size (∼23 nm), with the exception of ZrO2 and Ce0.1 Zr0.9 O2 that were around 46 nm. The results are consistent with the nanostructured character of the prepared materials and well in agreement with others in the literature. The acidity of these materials was obtained by adsorption of pyridine (Table 1, column 9). Compared with other solid acids studied in our laboratory, e.g., Ce/USY, CuO/Nb2 O5 /SiO2 -Al2 O3 and H3 PW12 O40 /ZrO2 [17,18,23], these mixed oxides had a much lower adsorption of this base. Generally, the amount of acid sites was similar among the series, but a slightly higher value was obtained by Ce0.8 Zr0.2 O2 with 0.043 mmol g−1 . This acidity was about 42% higher than the lowest value in the series (0.025 mmol g−1 for Ce0.4 Zr0.6 O2 ). The nature of the acid sites was followed by DRIFTS under ambient conditions (Fig. 1). The strong absorption bands at 1442 cm−1 were assigned to pyridine interacting through hydrogen bonds [24] with Ce and/or Zr on the mixed oxide surfaces (Fig. S2). No band related to Brønsted or Lewis (absorption above 1580 cm−1 )

Fig. 1. DRIFTS spectra of Cex Zr1−x O2 (0.1 ≤ x ≤ 0.9) after gas phase pyridine adsorption.

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

295

Fig. 3. Raman spectra of Cex Zr1−x O2 (0 ≤ x ≤ 1) calcined at 650◦ for 4 h. Fig. 2. Powder XRD of Cex Zr1−x O2 (0 ≤ x ≤ 1) calcined at 650◦ for 4 h.

was observed, except at 1640 cm−1 , which was attributed to O–H bending of moisture on the samples. However, these interactions have also been considered as Lewis type [25]. It should be noted that under a controlled atmospheric experiment, the dehydration of the surface oxide led to unsaturated coordinated Ce and Zr, which might be exposed to the direct interaction of the electron pair of pyridine forming a Lewis type site [25]. 3.2. Structural aspects of the mixed oxides The diffractograms of the mixed oxides are shown in Fig. 2. The synthesized ZrO2 was monoclinic (PDF-2: 01-0708739) corresponding to the planes (−1 1 1) and (1 1 1) at 2 = 28.2◦ and 31.5◦ , respectively, as well as tetragonal (PDF-2: 00-0501089) with the main peak at 2 = 30.4◦ (1 1 1); whereas CeO2 showed typical reflections of fluorite cubic structure (PDF-2: 0431002) with peaks at 2 = 28.5◦ , 33.1◦ , 47.6◦ , 56.5◦ and 59.2◦ correspondent to the (h k l) planes: (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2), respectively [26]. On the other hand, the mixed oxides had a unique pattern that progressively distorted to a tetragonal phase as the content of cerium was ≤0.4 with the main peak at 2 = 30.2◦ [27]. The X-ray diffraction peaks confirmed the preparation of solid solutions for the mixed oxides. The formation of solid solutions was evidenced by the absence of phase segregation with the incorporation of Zr4+ ions in the structural framework of ceria. The substitution of Ce+4 (0.097 nm) by a smaller ion such as Zr4+ (0.084 nm) resulted in a contraction of the cubic cell parameter, decreasing from 0.544 nm to 0.520 nm. No cell expansion was observed due to the possible reduction of Ce4+ to Ce3+ , as a result of the bigger ionic radium (0.107 nm) of the later. It was observed for Ce0.5 Zr0.5 O2 large peaks between 2 = 28.6◦ up to 29.3◦ and 2 = 33.1◦ up to 33.8◦ . This result has been interpreted as a single phase of zirconium atoms doped into ceria [28], although the existence of another phase based on cerium doped in tetragonal zirconia cannot be ruled out [29,30]. In addition, as the ratio Zr:Ce was close to 1, a decrease of the peak intensities and increased half width peaks was observed. Thus, the mixed oxides formed smaller crystals than the parent oxides, as already calculated. Consequently, nanoparticles of these mixed oxides showed a lower degree of aggregation with similar domains [31]. Another important characterization of the mixed oxides was based on the Raman spectra. This method can identify the symmetric vibrational stretching bands of O–Ce–O of cubic ceria (fluorite type). The symmetry of the optical absorption modes of fluorite is reported to be  optico = F1u (IR) + F2g (Raman) [32,33]. The assignment of F2g mode is observed at 465–470 cm−1 . This band is directly

related to the oxygen positions in the lattice and are sensitive to crystal symmetry [34,35]. The Raman spectra (Fig. 3) of the mixed oxides showed two main trends: Cex Zr1−x O2 oxides rich in cerium (0.6 ≤ x ≤ 1.0) exhibited a single band in the 465–470 cm−1 range, while the oxides rich in zirconium (0 ≤ x < 0.4) did not show this characteristic band. For ZrO2 , synthesized by sol–gel (650 ◦ C), the main bands appeared at 337, 349, 382 and 474 cm−1 , which are characteristic of zirconia in the monoclinic phase [36]. A weak band at about 265 cm−1 also confirmed the presence of tetragonal phase for ZrO2 . The observed Raman bands were found for CeO2 (465 cm−1 ), Ce0.9 Zr0.1 O2 (466 cm−1 ), Ce0.8 Zr0.2 O2 (470 cm−1 ), Ce0.7 Zr0.3 O2 (469 cm−1 ), Ce0.6 Zr0.4 O2 (468 cm−1 ), Ce0.5 Zr0.5 O2 (466 cm−1 ) and Ce0.4 Zr0.6 O2 (466 cm−1 ). The incorporation of Zr4+ into the cerium lattice led to a Raman shift for higher wavenumbers, compared to ceria. These shifts are a result of cell parameter contraction by substitution of Ce4+ for smaller cation Zr4+ . The Zr–O bonding is shorter and stronger than Ce–O. Hence, this partial substitution produced a growing distortion in the cubic lattice of the ceria to tetragonal structure, as observed by XRD. This lattice distortion caused a decrease in the intensity of the F2g Raman band characteristic of cubic ceria reaching a non-detectable signal below Ce0.4 Zr0.6 O2 . The Raman data were in agreement with XRD observations and confirmed the incorporation of the zirconium ions without phase segregation. It was also noted that among the series of mixed oxides, Ce0.8 Zr0.2 O2 underwent the highest Raman shift (465–470 cm−1 ) that might be related to a maximum contraction degree of the cubic lattices in this material, which leads to slight structural distortions. These distortions are important factors that help explain the high catalytic activities of these materials in diesel soot oxidation, as they increase oxygen flow and mobility in the lattice (see next section). 3.3. Catalytic activity of Cex Zr1−x O2 oxides The activity experiments were conducted in a model combustion reaction of particulate material (Printex® U) from diesel engines. The combustion process was simulated by studying the profile of the derivative thermogravimetric (DTG) curves obtained under oxidative conditions. The comparison was made using the temperature of the maximum derivative mass loss (Tm ), which was used for evaluation of catalyst performance [2]. The DTG curve of the Printex® U displayed distinct peaks at maximum temperatures of about 532, 566, 600, and 622 ◦ C. These events are associated with the initial oxidation step of organic compounds adsorbed on the particulates (532 ◦ C), their complete oxidation (566 and 600 ◦ C) and oxidation of particulates (622 ◦ C), which was considered the temperature of complete oxidation of the particulate [2].

296

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

Fig. 4. DTG curves for the mixtures of 1:20 Printex® U:Cex Zr1−x O2 (0 ≤ x ≤ 1) under air at 110 mL min−1 and rate of 10 ◦ C min−1 prepared at tight contact. Maximum temperatures (Tm ) are listed for each catalyst.

The DTG curves for tight contact are shown in Fig. 4. Synthesized ZrO2 and CeO2 showed Tm at 600 and 474 ◦ C, respectively. In the presence of all mixed oxides (Cex Zr1−x O2 ), Tm for Printex® U had a lower value than ZrO2 (Tm = 600 ◦ C), as well as for x > 0.5, the mixed oxides showed a better performance in the oxidation of soot than CeO2 (Tm = 474 ◦ C). Besides, the profile of soot oxidation was significantly enhanced with onset temperatures starting around 350 and offset temperatures lower than 520 ◦ C (i.e., T about 170 ◦ C). It was note worthy that, in general, as the mole fraction of Ce increased the Tm decreased with the best performance achieved with Ce0.8 Zr0.2 O2 . Fig. 5 shows the profile of the DTG curves for the oxidation of particulates under loose contact conditions. As observed for different materials in the literature, the soot oxidation was less effective (higher Tm ) than the tight contact. For instance, the Tm of Ce0.8 Zr0.2 O2 under loose contact condition is 143 ◦ C higher than that under tight contact. The difference on Tm under tight and loose contact clearly demonstrates the effect of the catalyst contact area on the kinetics of the combustion reaction, as already observed for other catalysts in the literature [2]. Furthermore, it was noted that the Tm in the Cex Zr1−x O2 series was similar (variation was only 63 ◦ C between the best and the worst catalyst using loose contact); whereas for tight contact, this difference was 196 ◦ C. Nonetheless, the observed trend in the temperatures was regular, i.e., the Tm decreased as the molar fraction of cerium approximated to one (maximum at Ce0.8 Zr0.2 O2 ). These observations are supported by corresponding reports in the

Fig. 5. DTG curves for the mixtures of 1:20 Printex® U:Cex Zr1−x O2 (0 ≤ x ≤ 1) under air at 110 mL min−1 and rate of 10 ◦ C min−1 prepared at loose contact. Maximum temperatures (Tm ) are listed for each catalyst.

literature [20,37,38], which have attributed the rate of oxidation to active oxygen being faster than gas-phase oxygen used in the flow. In addition, in this work, the best catalyst did not show either the highest surface area or the lowest crystallite size, even though the average values for these parameters were fairly similar in all materials. In recent studies involving these mixed oxides, it was concluded that the soot oxidation was independent of the particle size within the size range of 29–93 nm [39]. Thus, these parameters by themselves did not explain the best activity for Ce0.8 Zr0.2 O2 catalyst. The oxidation activity of Cex Zr1−x O2 mixed oxides is known to depend on the presence of active oxygen in these catalysts [11,26]. The active oxygen is considered to be highly reactive species of oxygen, such as radicals, superoxide radicals, etc. [5]. In order to show this dependence, an additional experiment was performed. The most active catalyst (Ce0.8 Zr0.2 O2 ) was reacted with Printex® U (tight contact) under nitrogen atmosphere (Fig. S3). The result showed that Tm under nitrogen was 418 ◦ C and under air was 404 ◦ C, which was consistent with the apparent negligible role of external oxygen in the oxidation process. Actually, this achievement demonstrated the preferential utilization of the lattice oxygen in the oxidation of soot. The main reason for this higher activity of the catalyst compared to the soot by itself may be attributed to the higher amount of active oxygen in their composition. This active oxygen present in the catalyst is able to oxidize the organic residues adsorbed on the soot and the particulate itself under more favorable temperature conditions. The amount of active oxygen is estimated from the OSC for the Ce0.8 Zr0.2 O2 catalyst at about 390 ␮mol O2 g−1 [40]. Accordingly, if the access to the active oxygen of the catalyst framework is limited, the reaction relies on the oxygen present in the air flow to oxidize the particulate, which is not effective because it requires higher activation energies. Actually, the role of external oxygen is to replace lattice oxygen creating new active oxygen species [20,41] for further oxidation of the particulate. Other interesting property observed for the most active catalyst was the amount of hydrogen bonding or Lewis sites. The amount of Lewis sites was important in the activity of the best catalyst. Particularly, high cerium molar fraction concentration was the parameter that brought the most active materials to this reaction. The kinetic redox cycle involving Ce4+ /Ce3+ is proportional to the amount of available cerium sites. This is consistent with the proposed Mars–Van Krevelen mechanism [20,40] described usually as follow. Soot (represented here as carbon, C) is oxidized to CO using lattice oxygen and leaving vacancies that reduce cerium(IV) to cerium(III). In the presence of external oxygen, which is adsorbed on the surface of the catalyst, cerium(III) is reoxidized to cerium(IV). Then, CO is oxidized to CO2 using oxygen of the lattice forming cerium(III). The cycle is completed by the oxidation of cerium(III) with reposition of lattice oxygen by external oxygen forming cerium(IV). These main steps involve adsorption and desorption of different species (e.g., CO, CO2 ) on the surface that might have diverse activation energies. The proposed mechanism scheme (Fig. 6) is consistent with the acidity results, since the most acidic material has the higher amount of available cerium responsible for the enhancement of the redox cycle. Another important parameter related to the kinetic of soot oxidation is the oxygen mobility. The oxygen mobility will facilitate the mechanism of soot oxidation via formation and regeneration of the Ce4+ /Ce3+ redox pair. The enhanced contact of the particulate with the catalyst smoothes the progress of the reaction with low participation of external oxygen, even though it is necessary to re-oxidize the metal site and complete the catalytic cycle [31]. The mobility is directly related to structural features of Cex Zr1−x O2 materials. The presence of Zr4+ makes the lattice distorted and defective. The

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

297

method [48]. This approach has been described in ASTM standard method E1641 and takes into consideration the results of multiple experiments performed by thermogravimetric analysis under variable heating rates. In general, kinetic analysis of solid state reactions of decomposition is based on a single kinetic equation [49]: d˛ = k(T )f (˛) dt

(1)

where ˛ is the extension of conversion, t is the time, k is the rate constant, T is the absolute temperature and f(˛) is the reaction model. Usually, the relation of temperature with the rate constant is brought in by substituting k(T) with the Arrhenius equation (k = A exp(−Ea /RT)), leading to: d˛ = A exp dt Fig. 6. Scheme for redox process involving Ce4+ /Ce3+ under O2 atmosphere for mixed oxide catalysts of Cex Zr1−x O2 .

coordination number for Ce4+ decreased from 8 (CeO2 ) to lower number, which facilitates the migration of oxygen positions into the bulk structure and hence the release of this oxygen to oxidize soot. In summary, the results of these catalytic tests lead to general principles that apply to the oxidation of particulates in the presence of the Ce–Zr mixed oxide catalysts: (i) oxygen storage capacity (CeO2  CeO2−x + x/2O2 ); (ii) oxygen mobility; (iii) higher concentration of accessible Lewis sites; (iv) lower activation energy of adsorbed species. Generally, one may associated a lower activation energy process with a higher rate reaction. As a result, a lower Tm would be obtained. To confirm this correlation, a kinetic study was developed to calculate the activation energy for the most active mixed oxide catalyst (vide next section). Before the discussion about the kinetics parameters, a short comparison with other catalysts for soot oxidation in the literature can be provided. It should be noted that there are many differences in the cited publications related to experimental conditions limiting a straight evaluation, such as: (a) preparation method; (b) type of precursor used in the syntheses; (c) calcination temperature; (d) actual stoichiometry of Ce:Zr oxide; (e) system used for the oxidation reaction (e.g., TGA system; fixed bed flow reactor; TPO apparatus); (f) ratio soot:catalyst; (g) source and flow of gas used (e.g., air; diluted oxygen; O2 /NOx ). Therefore, much care should be taken to compare results of different laboratories [42]. There have been tested different classes of catalysts used for diesel soot oxidation, for example: pure oxides [2,43], mixed oxides [44,45] and supported mixed oxides [2,10,46]. So far, the most promising catalysts have involved mixed oxides usually in the presence of some promoter. The Cex Zr1−x O2 systems have been studied by a number of researchers [25,26,31,40]. There is a consensus that these catalysts enhance the activity on diesel soot oxidation compared to CeO2 . However, the exact stoichiometric formula of the mixed oxide has not been definitely pointed. Also, the calculated activation energy is not the same, even though the general trend is in good agreement [44,47]. The main parameter to measure activity has been Tm , which strongly depends on the experimental set up conditions as already pointed. One of the important accomplishments in this work is that the whole series of Ce–Zr mixed oxides were tested under the same experimental conditions, so that the most promising catalyst (Ce0.8 Zr0.2 O2 ) was indicated unequivocally. 3.4. Kinetics of the catalytic combustion for Ce0.8 Zr0.2 O2 In this work, the kinetic parameters for the catalytic and noncatalytic soot oxidation were calculated using the Flynn and Wall

 −E  a

RT

f (˛)

(2)

where A is the pre-exponential factor, Ea is the activation energy and R is the universal gas constant. Considering the TG experiments were conducted under nonisothermal conditions, i.e., at linear heating rates (ˇ = dT/dt), Eq. (2) takes the following form: d˛ = dT

A ˇ

f (˛) exp

 −E  a

RT

(3)

In this study the kinetic data were analyzed based on the mass loss produced during the thermal process. The temperature shifted higher as the heating rate increased, as a kinetic phenomenon well known in the literature. Kinetic parameters, such as activation energy and pre-exponential factor are important for comparison of different catalysts as well as for the design of catalytic soot converters [40,43]. The DTG curves for soot oxidation are shown in Fig. 7(A–C) with tight and loose contact with Ce0.8 Zr0.2 O2 , as well as without catalyst, respectively. For each conversion extent and based upon the linearization of Eq. (3), a plot of ln [RA/Ea␣ g(˛)] versus 1/T was obtained. Accordingly, the activation energy was obtained from the angular coefficient. Representative data for 50% conversion are shown in Fig S4 for soot, for its tight and loose contact with the catalyst. The actual activation energy was considered as an average of the values obtained at each conversion (20, 40, 50, 60 and 80%). It could be noted that the values decreased from soot to loose and tight contact catalytic system, respectively (Ea = 181.7 ± 21.1; 149.4 ± 3.4 and 142.5 ± 8.1 kJ mol−1 ). Also, the pre-exponential factors were 9.7 × 109 ; 9.2 × 107 and 7.8 × 109 s−1 in the same order, respectively. These data were obtained using the model-free kinetics for nonisothermal experiments. The great advantage of this method is that the activation energy can be determined as a function of the extent of conversion without making any assumptions about the reaction model [50]. Making use of the same temperature regions, data from isothermal and nonisothermal kinetics may be comparable [50]. For example, for the soot used in this work (Printex® U) the activation energy is in reasonable agreement with others reported as about 160 kJ mol−1 [40,43,47,51]. The main difference observed is related to the kinetic model used for calculation as well as the particular type of soot tested. Another source of the higher calculated activation energy was due to the oxidation of soot involves more than one step in the interval of temperature studied. This is one of the major reasons for disagreement between isothermal and nonisothermal data [50]. On the other hand, the activation energies for the Ce0.8 Zr0.2 O2 catalyst are in excellent agreement with others in the literature (e.g., for loose contact, Ea = 149.5 kJ mol−1 [40]; Ea = 138 kJ mol−1 [47]). This demonstrates that the model-free kinetics was able to predict a good estimation of the activation energy for the catalytic reaction. The model-free kinetics does not demand a model of reaction to calculate the activation energy. However, it has been attributed an

298

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

Fig. 8. DTG curves of Printex® U combustion with Ce0.8 Zr0.2 O2 (tight contact) calcined at different temperatures.

Besides, the transference of oxygen from the surface to the bulk of the solid will establish how well the catalyst works. 3.5. Thermal stability of Ce0.8 Zr0.2 O2 oxide

Fig. 7. DTG curves of Printex® U combustion with Ce0.8 Zr0.2 O2 obtained at variable linear heating rates (ˇ) under: (A) tight contact; (B) loose contact and (C) without catalyst.

order reaction between 0.3 and 1.0 for the soot combustion, with and without catalyst, using different approaches for the calculation of global kinetics parameters [40,43,47,51]. It is important to note that the choice of model may lead to some apparent contradictions about the influence of a catalyst in this oxidation process (e.g., effect of the catalyst on lowering the activation energy of soot oxidation. This is most related to the calculation of the kinetics parameters by dynamic methods such as TGA, which are considered controversial in some aspects [52]. However, it is worth to mention that the approximate results are in good agreement with the proposed mechanisms in the literature. The calculated activation energy is related to the process of interaction of oxygen with soot. The most prominent source is the lattice oxygen at the contact point with the catalyst, so that the regeneration of the vacancies produced during soot oxidation should be effective to accelerate the reaction.

The stability is an important issue in the development of a catalyst for industrial applications. In this sense, the most active material (Ce0.8 Zr0.2 O2 ) was tested under different calcinations temperature: 650, 850 and 1000 ◦ C for 4 h under tight contact with soot. The results are shown in Fig. 8. It could be seen that Tm shifted to higher values (404, 431 and 462 ◦ C, respectively) as the calcination of the catalysts increased. This is probably related to sintering of the catalysts, since the same trend have been observed for similar systems in the literature [26,37]. Nonetheless, it should be mentioned that no apparent structural change was observed for Ce0.8 Zr0.2 O2 before and after soot oxidation by XRD and FTIR spectra. However, calculation by the Scherrer’s equation of the average domain of particle size showed that Ce0.8 Zr0.2 O2 increased from about 12 nm to 30 nm when it was calcined from 650 to 1000 ◦ C, which is in agreement with the sintering explanation. Another important aspect of any catalyst is the resistance to deactivation in the presence of water. Ce0.8 Zr0.2 O2 catalyst was mixed with Printex® U under inert and regular atmosphere with moisture. The amount of water determined by TG was about 0.8 wt% under inert atmosphere, since no pre-thermal treatment was performed either in catalyst or soot. On the other hand, under regular atmosphere, an amount of 1.2 wt% and 2.5 wt% was obtained under different time of exposition to air. The results were practically the same for the content of water of 0.8, 1.2 and 2.5 wt% (Tm = 405, 404 and 406 ◦ C, respectively). These achievements were consistent with other, which pointed that for these Ce–Zr mixed oxides the addition of water up to 7 wt% had no influence in their activity [40]. 3.6. Reutilization studies Reutilization is one of the most advantages for a heterogeneous catalyst. To verify this important property, further studies were conducted with the best catalyst (Ce0.8 Zr0.2 O2 ) for soot combustion. The monitoring of Tm for five experiments were obtained and shown in Fig. 9. It could be seen that successive runs were very similar and the average Tm for this catalyst was about 403 ◦ C. The activity was maintained as well as the general external aspects (e.g., color, grain size). The structure was also checked by XRD (Fig. 10). The diffractograms confirmed the same initial structure for Ce0.8 Zr0.2 O2 . Besides, the size of the crystallites (DXRD ) was kept practically the same (11.8 nm and 12.0 nm for the fresh and fourtime reutilized catalyst, respectively). DRIFTS results also displayed the same spectrum obtained for the initial material (figure not

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

299

catalyst treated at 650 ◦ C was the most active and stable. It also showed water tolerance for up to 2.5 wt% content in the mixtures with Printex® U. In addition, this catalyst demonstrated to be active and reutilizable (five times). This catalyst could be recovered without any thermal or chemical treatment and still kept its basic structural features. Acknowledgments We acknowledge CNPq and CAPES for scholarships to research and graduate students and the financial support provided by DPP/IQ/UnB, FINATEC, CAPES, FAPDF, MCT/CNPq, FINEP/CTInfra, and Petrobras. Also, we want to thank Mr. Jove Lobo Vila Verde and Mr. Thiago Fatobene from PerkinElmer-Brazil for Raman spectra. Fig. 9. DTG curves of Printex® U combustion with Ce0.8 Zr0.2 O2 for five utilizations.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.apcata.2011.11.020. References [1] [2] [3] [4] [5] [6] [7] [8] [9]

Fig. 10. XRD of fresh and after second and fourth reutilization of Ce0.8 Zr0.2 O2 in the combustion of Printex® U.

shown). Thus, this catalyst showed high efficiency for soot combustion without any calcination treatment of the samples between the runs and showed the same basic structural parameters. It should be mentioned that during the reutilization of the catalysts (second to fourth time) the material was actually thermally treated at 700 ◦ C, since the TG experimental run end up at that temperature. In addition, the average water content was 1.4 wt% for each run, which reiterated the water resistance property of this catalyst. Accordingly, the results confirmed the reason for utilization of these Ce–Zr mixed oxides materials in TWC devices with good performance.

[10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

4. Conclusions Mixed oxides based on Cex Zr1−x O2 , prepared by the sol–gel method, produced solid solutions active to diesel soot combustion. It was confirmed that the insertion of Zr4+ into the ceria framework formed progressively distorted cubic lattices. Higher the Zr4+ content resulted in higher lattice distortion, consequently changing the crystalline symmetry from cubic to tetragonal. The mixed oxides formed catalysts with only hydrogen bonding sites, which under dehydration conditions led to Lewis sites. In the studied series, Ce0.8 Zr0.2 O2 had the highest number of Lewis sites, which are important for the kinetic redox cycle involving Ce4+ /Ce3+ . Raman studies demonstrated that the Ce–O bonding was stronger in this mixed oxide composition, because of the unit cell contraction caused by zirconium insertion. Accordingly, Ce0.8 Zr0.2 O2 was the most active in the range of explored compositions (0 ≤ x ≤ 1.0) for the combustion of diesel soot. Calculation of the activation energy for the catalytic combustion process by the model-free kinetics showed good agreement with others in the literature. Tested at different calcination temperatures (650 to 1000 ◦ C), the Ce0.8 Zr0.2 O2

[21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34]

S. Braum, L.G. Appel, M. Schmal, Quim. Nova 27 (2003) 472–482. V.S. Braga, F.A.C. Garcia, J.A. Dias, S.C.L. Dias, J. Catal. 247 (2007) 68–77. M.S. Gross, M.A. Ulla, C.A. Querini, Appl. Catal. A 360 (2009) 81–88. M. Shelef, R.W. McCabe, Catal. Today 62 (2000) 35–50. R. Di Monte, J. Kaˇspar, Top. Catal. 28 (2004) 47–57. V.B. Mortola, J.A.C. Ruiz, L.V. Mattos, F.B. Noronha, C.E. Hori, Catal. Today 133135 (2008) 906–912. A.I. Kozlov, Do H. Kim, A. Yezerets, P. Andersen, H.H. Kung, M.C. Kung, J. Catal. 209 (2002) 417–426. A.B. Hungría, N.D. Browning, R.P. Erni, M. Fernández-García, J.C. Conesa, J.A. Pérez-Omil, A. Martínez-Arias, J. Catal. 235 (2005) 251–261. Y. Nagai, T. Yamamotob, T. Tanaka, S. Yoshida, T. Nonaka, T. Okamoto, A. Suda, M. Sugiura, Catal. Today 74 (2002) 225–234. C. Wang, W.B. Fan, Z.T. Liu, J. Lu, Z.W. Liu, Z.F. Qin, J.G. Wang, J. Mol. Catal. A: Chem. 329 (2010) 64–70. R. Di Monte, J. Kaˇspar, Catal. Today 100 (2005) 27–35. R.D. Shannon, Acta Crystallogr. A 32 (1976) 751–767. M. Yashima, H. Arashi, M. Kakihana, M. Yoshimura, J. Am. Ceram. Soc. 77 (1994) 1067–1071. M. Yashima, M. Kakihana, M. Yoshimura, Solid State Ionics 86–88 (1996) 1131–1149. ˜ M.G. Zimicz, D.G. Lamas, J.G. Sacanell, A.G. Leyva, R.T. R.O. Fuentes, L.M. Acuna, Baker, Chem. Mater. 20 (2008) 7356–7363. S. Rossignol, Y. Madier, D. Duprez, Catal. Today 50 (1999) 261–270. G.F. Ghesti, J.L. Macedo, V.C.I. Parente, J.A. Dias, S.C.L. Dias, Microporous Mesoporous Mater. 100 (2007) 27–34. F.A.C. Garcia, V.S. Braga, J.C.M. Silva, J.A. Dias, S.C.L. Dias, J.L.B. Davo, Catal. Lett. 119 (2007) 101–107. J.S. Santos, J.A. Dias, S.C.L. Dias, F.A.C. Garcia, J.L. Macedo, F.S.G. Sousa, L.S. Almeida, App. Catal. A 394 (2011) 138–148. A. Bueno-López, K. Krishna, M. Makkee, J.A. Moulijn, J. Catal. 230 (2005) 237–248. L. Hensgen, K. Stöwe, Catal. Today 159 (2011) 100–107. J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids: Principle, Methodology and Applications, Academic Press, San Diego, 1999. C.F. Oliveira, L.M. Dezaneti, F.A.C. Garcia, J.L. de Macedo, J.A. Dias, S.C.L. Dias, K.S.P. Alvim, Appl. Catal. A 372 (2010) 153–161. E.P. Parry, J. Catal. 2 (1963) 371–379. B. Rivas, R. López-Fonseca, M.A. Gutiérrez-Ortiz, J.I. Gutiérrez-Ortiz, App. Catal. B 104 (2011) 373–381. I. Atribak, B. Azambre, A. Bueno-López, A. García-García, Appl. Catal. B 92 (2009) 126–137. F.A. Silva, D.S. Martinez, J.A.C. Ruiz, L.V. Mattos, C.E. Hori, F.B. Noronha, Appl. Catal. A 335 (2008) 145–152. J.A.C. Ruiz, F.B. Passos, J.M.C. Bueno, E.F. Souza-Aguiar, L.V. Mattos, F.B. Noronha, Appl. Catal. A 334 (2008) 259–267. C.E. Hori, H. Permana, K.Y. Ng Simon, A. Brenner, K. More, K.M. Rahmoeller, D. Belton, Appl. Catal. B 16 (1998) 105–117. N. Guillén-Hurtado, I. Atribak, A. Bueno-Lopez, A. García-García, J. Mol. Catal. A: Chem. 323 (2010) 52–58. F.C. Gennari, A.C. Ramos, A. Condó, T. Montini, S. Bengió, A. Cortesi, J.J.A. Gamboa, P. Fornasiero, Appl. Catal. A 398 (2011) 123–133. G.R. Rao, T. Rajkumar, J. Colloid Interface Sci. 324 (2008) 134–141. L.N. Ikryannikova, A.A. Aksenov, G.L. Markayan, G.P. Muravieva, B.G. Kostyuk, A.N. Kharlanov, E.V. Linina, Appl. Catal. A 210 (2001) 225–235. M. Fernandez-García, A. Martínez-Arias, A. Iglesias-Juez, C. Belver, A.B. Hungría, J.C. Conesa, J. Soria, J. Catal. 194 (2000) 385–392.

300

C.F. Oliveira et al. / Applied Catalysis A: General 413–414 (2012) 292–300

[35] I. Atribak, A. Bueno-López, A. García-García, J. Catal. 259 (2008) 123–132. [36] J. Miciukiewicz, T.M. Knözinger, Appl. Catal. A 122 (1995) 151–159. [37] E. Aneggi, M. Boaro, C. de Leitenburg, G. Dolcetti, A. Trovarelli, Catal. Today 112 (2006) 94–98. [38] A. Bueno-López, K. Krishna, M. Makkee, J.A. Moulijn, Catal. Lett. 99 (2005) 203–205. [39] I. Atribak, A. Bueno-López, A. García-García, Catal. Commun. 9 (2008) 250–255. [40] L. Zhu, J. Yu, X. Wang, J. Hazard. Mater. 140 (2007) 205–210. [41] G. Preda, G. Pacchioni, Catal. Today 177 (2011) 31–38. [42] B.A.A.L. van Setten, J.M. Schouten, M. Makkee, J.A. Moulijn, Appl. Catal. B 28 (2000) 253–257. [43] B. Dernaika, D. Uner, Appl. Catal. B 40 (2003) 219–229. [44] D. Fino, N. Russo, E. Cauda, G. Saracco, V. Specchia, Catal. Today 114 (2006) 31–39.

[45] E. Aneggi, C. de Leitenburg, G. Dolcetti, A. Trovarelli, Catal. Today 114 (2006) 40–47. [46] F.A.C. Garcia, J.C.M. Silva, J.L. de Macedo, J.A. Dias, S.C.L. Dias, G.N.R. Filho, Microporous Mesoporous Mater. 113 (2008) 562–574. [47] B. Azambre, S. Collura, P. Darcy, J.M. Trichard, P. Da Costa, A. García-García, A. Bueno-López, Fuel Process. Technol. 92 (2011) 363–371. [48] J.H. Flynn, L.A. Wall, Polym. Lett. 4 (1996) 323–328. [49] N.A. Soto, W.R. Machado, D.L. López, Qui. Nova 33 (2010) 1500–1505. [50] S. Vyazovkin, C.A. Wight, Thermochim. Acta 340-341 (1999) 53–68. [51] F.S. Toniolo, E. Barbosa-Coutinho, M. Schwaab, I.C.L. Leocadio, R.S. Aderne, M. Schmal, J.C. Pinto, Appl. Catal. A 342 (2008) 87–92. [52] M.E. Brown, Introduction to Thermal Analysis – Techniques and Applications, Chapman and Hall, New York, 1988.