Modulating the Selectivity for CO and Butane Oxidation over

Jul 1, 2008 - ... Johnson Matthey Technology Centre, Sonning Common, Reading, U.K., and Department of Chemical Engineering, University of Cambridge, ...
0 downloads 0 Views 966KB Size
Download PDF
10968

J. Phys. Chem. C 2008, 112, 10968–10975

Modulating the Selectivity for CO and Butane Oxidation over Heterogeneous Catalysis through Amorphous Catalyst Coatings R. N. Devi,† F. C. Meunier,*,†,‡ T. Le Goaziou,† C. Hardacre,† P. J. Collier,*,§ S. E. Golunski,§ L. F. Gladden,| and M. D. Mantle| School of Chemistry & Chemical Engineering, Queen’s UniVersity Belfast, BT9 5AG, U.K., Laboratoire de Catalyse et Spectrochimie, CNRS-ENSICAEN, UniVersity of Caen, 14050 Caen, France, Johnson Matthey Technology Centre, Sonning Common, Reading, U.K., and Department of Chemical Engineering, UniVersity of Cambridge, Cambridge CB2 3RA, U.K. ReceiVed: February 8, 2008; ReVised Manuscript ReceiVed: May 14, 2008

The preparation of porous films directly deposited onto the surface of catalyst particles is attracting increasing attention. We report here for the first time a method that can be carried out at ambient pressure for the preparation of porous films deposited over 3 mm diameter catalyst particles of silica-supported Pt-Fe. Characterization of the sample prepared at ambient pressure (i.e., open air, OA) and its main structural differences as compared with a Na-A (LTA) coated catalyst made using an autoclave-based method are presented. The OA-coated material predominantly exhibited an amorphous film over the catalyst surface with between 4 and 13% of crystallinity as compared with fully crystallized LTA zeolite crystals. This coated sample was highly selective for CO oxidation in the presence of butane with no butane oxidation observed up to 350 °C. This indicates, for the first time, that the presence of a crystalline membrane is not necessary for the difference in light off temperature between CO and butane to be achieved and that amorphous films may also produce this effect. An examination of the space velocity dependence and adsorption of Na+ on the catalysts indicates that the variation in CO and butane oxidation activity is not caused by site blocking predominantly, although the Pt activity was lowered by contact with this alkali. 1. Introduction Metallic particles catalyze a large number of important reactions such as the hydrogenation, dehydrogenation and combustion of hydrocarbons. Traditional metal-based catalysts do not exhibit shape or size selective reactivity, which would allow discriminating reagents in a mixture. Shape selective metal-containing composites were first reported by Weisz et al.,1,2 who synthesized platinum particles encapsulated in Na-A zeolite (LTA-type structure) used for the selective hydrogenation of n-butene in n-butene/isobutene mixtures. The preparation of such encapsulated materials is highly complex and requires several synthetic steps. Another strategy that has been used to prepare shape selective Pt-based composites consists in the deposition of a polycrystalline zeolitic layer on monocrystalline surfaces.3 A composite material containing ZSM-5/Pt supported on TiO2(001) prepared according to this technique showed a preferential hydrogenation of linear alkenes over that of the branched counterparts. The preparation of noble metals encapsulated in hollow zeolitic spheres was also recently reported and these catalysts were shown to selectively oxidize small alcohols and to protect the encapsulated metals from large poison molecules.4 The preparation of porous films directly deposited onto the surface of catalyst particles (up to several mm large) is attracting * Corresponding authors. E-mail: [email protected] (F.C.M.), [email protected] (P.C.). † School of Chemistry & Chemical Engineering, Queen’s University Belfast. ‡ Laboratoire de Catalyse et Spectrochimie, CNRS-ENSICAEN, University of Caen. § Johnson Matthey Technology Centre. | Department of Chemical Engineering, University of Cambridge.

increasing attention.5–8 This type of composite materials displayed unusual activity and selectivity patterns in a number of reactions of academic and industrial interest including amine5 and maleic anhydride6 syntheses, selective hydrogenation reactions,7 and Fischer-Tropsch synthesis.8 In most cases, zeolitebased films deposited on macroscopic beads or supports were prepared by in situ synthesis of the zeolite. However, postsynthesis deposition techniques have also been developed.9,10 Foley et al.5 demonstrated that it was possible to obtain shape selectivity effect by creating a layer of amorphous carbon-based molecular sieve on top of silica-alumina particles. Kunkeler et al. also proposed a method to prepare amorphous silica layers over zeolites to neutralize surface acidity.11 A zeolite (BEAtype) was coated by a 20-30 nm thick porous layer prepared from tetraethyl orthosilane (TEOS), but no work dealing with potentially shape-selective catalysis was reported on this material. Collier et al. showed that it was possible to prepare Pt/SiO2 catalysts coated with Na-A zeolite that are active for the selective oxidation of CO in the presence of butane.6 The membrane was shown to be a thin monocrystalline layer of zeolite covering the platinum particles on the surface. The close contact between the zeolite seeds and the catalyst surface was achieved through a charge reversal step using a polyelectrolyte cationic polymer, which allowed the electrostatic repulsion between the negatively charged oxide surfaces surface present at the typical basic pH used for the hydrothermal synthesis to be overcome.12 The methods reported to date used to prepare shape selective Pt-based catalysts are complex, requiring the use of autoclaves to carry out the hydrothermal synthesis and repetition of the coating step in order to achieve a satisfactory coating of the

10.1021/jp801158d CCC: $40.75  2008 American Chemical Society Published on Web 07/01/2008

Selectivity for CO and Butane Oxidation

J. Phys. Chem. C, Vol. 112, No. 29, 2008 10969

sample surface in some cases.7 We report here for the first time a method that can be carried out at ambient pressure for the preparation of porous films deposited over a pelletized Pt-based oxidation catalyst. Characterization of this open-air (OA) sample and its main structural differences as compared to a material (AU) made using the autoclave-based method as reported in ref 6 will be presented. The oxidation of butane could be selectively prevented by the coating, while CO oxidation was essentially identical over the parent and coated catalysts. 2. Experimental Section 2.1. Material Synthesis and Characterization. The catalyst active phase was made up of 5% Pt + 0.5% Fe deposited onto a low surface area silica (Norton), consisting of 3 mm diameter spheres, with a specific surface area of 0.81 m2 g-1 (BET method) and mesopore volume 1.5 × 10-4 cm3 g-1. (BJH method). The catalysts were prepared by the incipient wetness method using Pt(NO3)2 and Fe(NO3)3 · 9H2O (Johnson Matthey, Pt assay; 15.8%) precursors. The impregnated pellets were dried at 105 °C and calcined at 500 °C for 2 h. The dispersion of the low surface area catalyst was evaluated at around 0.6%. For the synthesis of the coating over the low surface area catalyst, 5 g of catalyst was added to a diluted ammonia solution (0.1 M) of the polyelectrolyte 2,2-dimethylaminoethyl acrylate (Ciba). After 20 min, the catalyst spheres were removed from this solution and added to the zeolite gel prepared from 6.07 g of sodium aluminate (Sigma), 15.51 g of sodium silicate pentahydrate (Sigma) in 104 cm3 of distilled water.6 The gel composition was 3.6Na2O:Al2O3:2.6SiO2:223H2O. The mixture was heated either in an autoclave at 100 °C for 18 h (denoted “AU” sample) or in open-air (denoted “OA” sample) for 5 h at 105 °C in a glass beaker covered with a watch glass (the water level was maintained by the regular addition of boiling water). The pellets were subsequently removed from the solution and separated from the zeolite powder formed during the reaction by hand, washed, dried, and calcined at 500 °C. 2.2. Material Characterization. The samples were characterized by N2 adsorption isotherms at 77 K (BET method for the surface area and BJH method for the mesoporosity 13), chemisorption, XRD, SEM, FIB-TEM, and 27Al MAS NMR. The BET specific surface areas of the supports and catalysts were measured using a Micromeritics ASAP 2010 system from Micromeritics Instrument Corporation. X-ray diffraction studies were carried out using a Philips Panalytical X’Pert Pro diffractometer set in Bragg-Brentano geometry equipped with a Cu KR X-ray source (1.5405 Å) at 40 kV and 40 mA. The 2θ range scanned was between 5 and 65 °, with a step size of 0.0167° and a dwell time of 106.045 s for each step. The rotation speed of the sample was set at 15 rpm. All samples were crushed prior to analysis. The SEM used was a Jeol JSM-6500F field emission scanning electron microscope. The TEM observations were carried out by an FEI Tecnai F20 field emission high-resolution transmission electron microscope. The TEM operating conditions were as follows: voltage, 200 kV; C2 aperture, 30 or 50 m. For a fine powder or crushed sample, the sample was dispersed and then observed directly by TEM. When the sample was too large to be observed by transmission, the sample was prepared by ion beam milling and wedge polishing with a gallium focused ion beam (FIB). A standard lift-off technique was used to transfer the milled sections on to holey carbon grid for further HRTEM experiments. 27Al MAS NMR studies were performed on a Bruker Biospin AV 400 MHz spectrometer fitted with a standard 4 mm magic

Figure 1. CO (0) and butane (×) conversions measured over the uncoated 5% Pt + 0.5% Fe/SiO2 catalyst.

Figure 2. Conversion of butane over the uncoated catalyst (×), the sample coated by the AU method with (b) or without (O) contacting the plain catalyst with the polyelectrolyte.

angle spinning (MAS) probe operating at a resonance frequency of 104.28 MHz for 27Al. Al(NO3)3 · 9H2O was used as 0 ppm standard and the samples were spun at the magic angle at a speed of 13 kHz. A simple pulse-acquire sequence was used to detect the 27Al NMR signal with the following parameters: 1.25 µs excitation pulse of 45 deg; a recycle delay of 1.0 s; 64 averages. 2.3. Catalyst Testing. The catalytic tests were performed in a fixed bed reactor. The composition of the gas mixture was n-butane/CO/O2/He ) 1/0.5/18/80.5 vol %. The total flow was 100 cm3 min-1. The mass of the catalyst used was 400 mg, unless otherwise stated. The sample was used as small pellets (2.5 mm < diameter < 3 mm). The reactor temperature was controlled by a Eurotherm temperature controller. CO, CO2, O2, and n-butane were analyzed by online gas chromatography (Perkin-Elmer Clarus 500) fitted with a Chromorsorb 102 and a molecular sieve 5 Å columns. Product detection was carried out using both a TCD and FID detectors. The butane and CO conversions were measured on reaction isotherms, the temperature being typically increased by 25 or 50 °C steps. Five GC analyses were carried out at each temperature and the last four runs were essentially identical (the first run was different because of the time needed to get the reactor temperature constant). No deactivation could be observed during the isotherms. 3. Results 3.1. Importance of the Utilization of the Polyelectrolyte Modifier. The importance of contacting the plain catalyst with the polyelectrolyte before carrying out the coating step to obtain a shape selective catalyst was assessed by measuring the activity for the oxidation of CO in the presence of butane over various samples (Figures 1 and 2). The CO and butane conversions measured over the uncoated Pt-Fe/SiO2 catalyst are shown in Figure 1 as a reference. For this catalyst, the CO light-off occurred at around 125 °C and CO was fully combusted at 150 °C. The plot representing the conversion of butane was essentially parallel to that of CO and was shifted to higher temperatures by ca. 50 °C. In contrast, the butane light-off temperature was shifted to higher temperatures (i.e., butane oxidation was inhibited) by

10970 J. Phys. Chem. C, Vol. 112, No. 29, 2008

Devi et al.

Figure 3. Conversion of CO (full symbols) and butane (open symbols) over coated catalysts prepared by the hydrothermal method AU (triangles) and the open air method OA (circles).

more than 200 °C with respect to the CO light-off curve over the AU-coated sample (Figure 2). It should be noted that the coating procedure did not significantly change the CO oxidation profile with full CO conversion at 200 °C; these are not shown in Figure 2 for the sake of clarity. The inhibition of butane oxidation was significantly reduced when the sample prepared by the hydrothermal method was not contacted first with the polyelectrolyte (Figure 2). These data show that the use of the polyelectrolyte is instrumental in obtaining a shape selective AU-coated catalyst. The polyelectrolyte is thought to act as a charge reversal agent making the surface of the catalyst positively charged under the coating synthesis conditions, facilitating the anchoring of the negatively charges zeolite crystal seeds.6,12 3.2. Open-Air versus Hydrothermal Method. All the samples prepared in this section were obtained by contacting the plain catalyst with the polyelectrolyte. Thereafter, a coating method similar to the AU method except that the coating step took place under ambient pressure (OA, open-air) was performed. The reactivity patterns obtained over the Pt-Fe/SiO2 coated by the autoclave-based AU method and coated by the OA method were almost identical (Figure 3). In both cases, CO conversion was complete at 200 °C, while no butane conversion was found up to 350 °C, showing that the utilization of an autoclave was not a requirement to obtain a selective catalyst. The increase of butane conversion with temperature was less sharp in the case of the AU and OA coated samples (Figure 3) as compared with that observed in the case of the plain sample (Figure 1). This fact is consistent with a mechanism whereby the butane conversion is limited by mass transport associated with its larger mass and kinetic diameter (430 pm) compared with CO (kinetic diameter ) 376 pm) for both coated samples. This latter point is in agreement with the observation that the AU material consisted of a thin layer of Na-A (LTA-type) zeolite over the surface of the catalyst particles.6 3.3. Exchange of Counterions in the AU and OA samples. In order to examine whether decreasing the pore size would result in further increases in the temperature difference between CO and butane oxidation, the Na+ in the OA and AU-prepared sample were exchanged with larger monovalent cations, K+ and Rb+ using the corresponding chloride salts. The coated beads were heated at 95 °C with 25 mL of salt solution (1M) for 3 h after which the beads were separated and immersed in 25 mL of fresh salt solution and heated at 95 °C for another 3 h. This was repeated and kept in contact with fresh salt solutions for 18 and 24 h. In each case the pellets were washed until the washings were free of Cl- ions as determined using aqueous silver nitrate. The level of cation exchange was not determined. Figure 4 compares the catalytic activity of the samples. The CO oxidation curve was essentially identical for all the samples (i.e., full CO combustion was achieved at 200 °C) and, therefore, only that of the parent sample is reported for the sake of clarity.

Figure 4. (a) Conversion of CO (0) and butane conversions measured over the AU-coated sample prepared with NaOH (×) and subsequently ion-exchanged with K+ (O) and Rb+ (9). (b) Conversion of CO (0) and butane conversions measured over the OA-coated sample prepared with NaOH (×) and subsequently ion-exchanged with K+ (O) and Rb+ (9).

As expected in the case of the AU sample (i.e., Na-LTAcoated Pt-Fe/SiO2), the substitution of Na+ with cations of increasing size led to catalysts with decreasing activity for butane oxidation (at least up to 480 °C), that is in terms of activity: Na-AU > K-AU > Rb-AU. This was most likely due to narrowing of the coating pore size, hindering butane diffusion (Figure 4a). Although changing the cation also had an effect on the butane oxidation activity for the OA samples (Figure 4b) no clear tendency was found with cation size, Rb-OA > Na-OA > K-OA. The fact that no correlation between activity and expected pore size was observed in the case of the OA materials suggests that the porosity of these samples was different from that of the LTA crystal-based coated AU catalysts, possibly due to a less crystalline structure. Because of the presence of carbon dioxide in the reaction effluent and the low temperature used, it is likely that the basic sites of the alumino-silicate coating were covered with CO2 and carbonate species were present.14 Therefore, the effective pore size of the samples was probably related to the presence of these (bulky) carbonate species. 3.4. XRD Characterization of the AU and OA Samples. The crystallinity of the coating could not be studied by XRD, due to the low weight proportion of coating present on the catalyst (ca. 2 wt %, Vide infra) and the high signal intensity associated with the crystalline silica support. When the AU and OA coated samples were crushed and analyzed by XRD, only peaks corresponding to quartz and cristobalite were observed. The XRD of the powdered material formed in the solution during the coating of the pellets coated revealed the presence of well-crystallized LTA crystals as determined by a comparison with simulated XRD data 15) for both OA- and AU- prepared catalysts. 3.5. SEM Characterization of the AU and OA Samples. Scanning electron micrographs (SEM) of the uncoated catalyst and the coated materials were examined to assess the structural differences between the materials. The SEM analysis was focused onto the internal surface of the macroporous catalyst pellet for two reasons. First, the external surface of each sample

Selectivity for CO and Butane Oxidation

Figure 5. (a) SEM picture of the cross section of a bead of the plain Pt-Fe/SiO2 catalyst. (b) SEM picture of the internal surface of the pore circled in white in the left picture.

was contaminated by loosely bound crystals formed in solution along the beads. Second, the internal pores are the most difficult to coat because it requires a deep penetration of the gel, and therefore, the quality of the coating in the internal pores is the best indicator of a successful coating method. A cross section of a pellet of the uncoated Pt-Fe/SiO2 catalyst is shown in Figure 5 and the pore analyzed by SEM is circled in white (Figure 5a). A magnified picture of the internal pore surface (Figure 5b) revealed a smooth surface populated with submicron needles embedded into the surface. A SEM picture of the typical internal macropores of the AU samples is shown in Figure 6a. The internal surface of the AU sample showed a covering of small crystals, which are thought to be Na-A zeolite (Figure 6a). This observation is consistent with the TEM analysis reported elsewhere,6 which indicated that the surface of the AU sample was covered with a Na-A layer. On the contrary, the OA catalyst showed no evidence of the presence of any zeolite crystals (Figure 6b). The surface of the OA sample internal pores was also smooth but showed a different texture compared with the uncoated catalyst (Figure 5b). The surface of the OA sample seemed to reveal the presence of a continuous submicrometer-thick coating tight to the surface, with some minor gaps possibly being present (Figure 6b).

J. Phys. Chem. C, Vol. 112, No. 29, 2008 10971

Figure 6. SEM pictures of an internal pore of the coated sample prepared by (a) the AU method and (b) the OA method.

3.6. TEM Characterization of the OA Sample. An in-depth characterization of the structure of the coating obtained in the case of the OA sample was carried out by analyzing a thin section of the sample perpendicular to the internal pore surface. The methodology used during the preparation of the slice of the sample is depicted in Figure 7. The section was first protected by a Pt strap and then milled with a gallium ion beam. The thin slice obtained was then cut and lifted off the surface with a rod before being deposited on a copper grid for further TEM analysis. TEM pictures of the OA slice at two different magnifications are reported in Figure 8. The darker band on the top of the images was associated with the Pt protective strap. The marblelike region at the bottom of the images was typical of a crystalline phase and represented the crystalline SiO2 support. The region in between these two zones was found to have little structure and was typical of an amorphous phase (note: the large number of dark spots observed on the picture with the higher magnification were due to the sputtering of some Pt from the protective strap during FIB milling). The structureless region is due to the coating formed over the OA sample. These pictures show that the coating closely adheres to the surface of the Pt-Fe/SiO2 catalyst and its thickness was lower than 500 nm (Figure 8).

10972 J. Phys. Chem. C, Vol. 112, No. 29, 2008

Devi et al.

Figure 7. Schematic representation of the preparation by focused ion beam milling of a slice of the OA sample to be analyzed by TEM.

Figure 8. TEM pictures of a FIB-milled cross-section of the surface of an internal macropore of the OA sample (two different magnifications are shown). The Pt strap was used to protect the section of interest during the milling of the surface with Ga ions.

Figure 10. TEM microdiffraction patterns associated with the flakelike particle collected during the handling of the OA sample: (a) pattern observed initially. (b) no crystalline diffraction pattern observed after 40 s.

Figure 9. TEM pictures of materials detached from the OA sample during sample handling (two different magnifications are shown). The circled area on the right-hand-side picture highlights the negative crystal.

Pellets of the coated catalyst were crushed with a mortar and pestle and the fragments were also analyzed by TEM. Some fragments displaying a flake-like shape were observed (Figure 9, left), which appeared to have a Si/Al ratio of about 1 as determined by energy-dispersive X-ray analysis, similar to that expected for Na-A zeolite. Since these fragments contained aluminum, it is obvious that those were formed during the coating step. The TEM pictures also show the presence of negative crystals (see circled area, Figure 9, right), as defined by Valtchev et al. during a study of the template-free synthesis of Na-A.16 These negative crystals are Na-rich liquid regions occluded in the gel and the gel crystallization begins at the corresponding liquid-solid interface. A microdiffraction pattern collected over these solids was consistent with a LTA-type structure (Figure 10a). However, the sample quickly became damaged under the electron beam and no crystalline pattern

could be observed after 40 s exposure (Figure 10b). The similarity of the structure of these flake-shaped particles with those observed by Valtchev et al.16 indicates that these residues are poorly crystallized LTA-type solid still in the first stage of the crystallization process, which takes place via gel reorganization.17 3.7. 27Al-MAS NMR Characterization of the OA Sample. In order to evaluate the crystallinity of the coating 27Al-MAS NMR was performed by examining the shape of the 27Al NMR peak which is related to the aluminum environment in the sample. Al atoms in a well-crystallized zeolitic material lead to a sharper peak than those observed in a poorly crystallized sample. The 27Al NMR of the OA-sample presented a broadband (Figure 11a) centered around 55 pm and is attributed to 27Al atoms in a tetrahedral Al(SiO4)4 environment,16 The broadness of the peak (approximately 2180 Hz) confirms the low crystallization of the OA-coating. To be able to quantify the crystallinity of the sample, Na-LTA powder samples at different stages of crystallization were prepared by varying the temperature and reaction time of the synthesis. The extent of crystallization was determined for each sample by normalizing the intensity of the main XRD peak at 2θ values of 30.2 ° and 24.1 ° to that of the well-crystallized

Selectivity for CO and Butane Oxidation

J. Phys. Chem. C, Vol. 112, No. 29, 2008 10973 TABLE 1: Elemental Analysis of the Plain and OA-Coated Catalyst composition (wt %) material

Pt

Fe

Al

Na

plain Pt/Fe LSA OA-coated sample (OA-method)

3.8 3.5

0.37 0.43

0.13 0.37

0.01 0.6

TABLE 2: BET Surface Areas of the LTA Sample, Uncoated, and OA- and AU-Coated Catalysts

Figure 11. (a) 27Al MAS NMR signal (normalized) of the OA sample. (b) 27Al MAS NMR signals of of materials hydrothermally treated for 1, 2, 3 and 4 h at 100 °C (bottom to top lines).

Figure 12. (a) XRD of materials hydrothermally treated for 1, 2, 3, and 4 h at 100 °C (bottom to top lines). (b) Relation between the fwhm of the 27Al MAS NMR signal and the OA sample crystallinity determined by XRD.

samples obtained at long synthesis time (Figure 12a). The crystallization of Na-A started at 100 °C after 1 h and was complete after 4 h. The same samples were then also analyzed by 27Al MAS NMR (Figure 11b) and the peak full-widths at half-maximum (fwhm) measured. This value was related to the percentage of crystallinity determined from the XRD patterns. The fwhm of the 27Al MASNMR peak of the catalyst coated by the OA method was then determined from the calibration curve reported in Figure 12b. According to this method, the crystallinity of the OA sample was around 4%. 3.8. BET Specific Surface Area of the OA Sample. The chemical analysis by ICP-AES of the plain Pt-Fe/SiO2 and the OA-coated sample revealed that no significant loss of platinum occurred as a result of the coating step (Table 1). The increase

samples

surface area (BET) (m2 g-1)

LTA powder sample uncoated catalyst OA sample

370 ( 10 0.74 ( 0.1 1.3 ( 0.2

in the proportion of Al was due to the formation of the Alcontaining coating on the catalyst pellets. BET surface area measurements were carried out on the uncoated catalyst and the OA-sample to evidence any potential increase of the surface area in relation to the formation the coating. These data (Table 2) indicate that a small but significant increase in the surface area of the coated sample was observed as compared with that of the uncoated catalyst. The data reported in Table 1 and 2 can be used to determine what the increase of surface area would be if the coating were a perfect LTA film. It was assumed for the calculation that the coating had the same composition as that of pure LTA, i.e. typical formula Na12Al12Si12O48 (MW ) 1704 g mol-1), the corresponding weight fraction of Al being about 19 wt %. For 1 g of OA-coated catalyst, the added Al content was 2.4 mg. Therefore, ca. 12.6 mg ()2.4/0.19) of coating would be present on the coated catalyst. 12.6 mg of LTA zeolite would then exhibit a surface area of 4.7 m2. However, the increase of surface area measured over the OA-prepared sample was only 0.6 () 1.3 -0.74) m2 g-1. Assuming that the added surface area is proportional to the crystallinity of the coating, the OA-coating crystallinity would be ca. 13%. (This assumption arbitrarily supposes that any aluminosilicate other than purely crystallized LTA has a negligible surface area). It should be reminded that the determination of BET surface area derived from N2 adsorption in zeolites is flawed because of potential capillary condensation. Therefore the numbers representing the BET surface area should be taken as a semiquantitative evaluation of the surface area and porosity of the sample, which is useful here for our rough evaluation of the film crystallinity. It should be stressed that the determination of the sample crystallinity described in the previous paragraph is marred by a large imprecision, because of the possible errors on the measure of the elemental composition and the fact that other Al- or Nacontaining phases could have been formed. If the calculation were based on the Na or Pt proportions in the fresh and coated samples, a degree of crystallinity of around 4% and 2% would be found, respectively. Note that this is closer to the value obtained from the NMR analysis. In any case, the main conclusion remains the same, which is that the degree of crystallinity is low. 3.9. Effect of Synthesis Gel Contact with the Uncoated Catalyst. In order to examine the chemical effect of the components of the gel used to prepare the LTA zeolite, the uncoated catalyst was contacted with the synthesis mixture (following treatment with the polyelectrolyte) under mild conditions. The treatments used were of 1 h at 50 and 100 °C and 40 h at room temperature, all under atmospheric pressure. No crystalline LTA was observed alongside the beads under these conditions. The conversion of CO measured over these

10974 J. Phys. Chem. C, Vol. 112, No. 29, 2008

Devi et al.

Figure 14. Conversion of CO (O, b) and butane (9, 0) measured over 4 mg of uncoated catalyst (open symbol) and 400 mg of the OAcoated sample (solid symbol).

Figure 13. (a) CO onversions measured over the uncoated catalyst contacted with the synthesis gel at (b) 20 °C for 40 h, (2) 50 °C for 1 h, and (9) 100 °C for 1 h and (]) NaOH treatment. (b) Butane onversions measured over (0) the plain catalyst, (×) the OA-coated sample and the uncoated catalyst contacted with the synthesis gel at (b) 20 °C for 40 h, (2) 50 °C for 1 h, and (9) 100 °C for 1 h and (O) NaOH treatment.

samples was identical to that obtained over the uncoated and OA-coated catalysts (see Figures 1 and 3). However, the plots corresponding to butane conversion were shifted to higher temperatures compared with the uncoated catalyst, albeit to a smaller degree with respect to the change in activity found for the OA-coated sample (Figure 13). This observation reveals that the simple contact with the gel, even if the gel is not heated, resulted in a significant inhibition of the oxidation of the alkane, with no apparent effect on CO oxidation. Figure 13 also illustrates the effect of adsorption of Na+ by contacting the catalyst with NaOH. This treatment inhibited both the CO and butane light off temperature as well as limiting the maximum butane conversion which are not seen for any of the coated materials. This results indicates that the change between the uncoated and coated materials or contacted with the synthesis gel, cannot be simply associated with an alkali metal ion site blocking effect. In conclusion, the data reported in Figure 13 suggest that simple gel contact may be enough to produce a sufficiently robust, amorphous aluminosilicate membrane at the catalyst surface which can modify the catalyst and lead to the enhanced discrimination between the CO and butane oxidation observed. 3.10. Effect of the Space Velocity on the Oxidation of CO and Butane. The space velocity of the reaction was changed in order to examine whether the selectivity observed in the case of the coated material could have simply arisen from a total blocking of a fraction of the Pt sites during the synthesis. This effect was simulated by using a reduced number of Pt sites in the form of a smaller mass of uncoated catalyst. To facilitate the precise weighing of the mass of sample used, the plain catalyst beads were crushed and homogeneously dispersed in an inert SiO2 powder in a known proportion. The CO/butane oxidation profiles reported in Figure 14 were obtained using a total mass of 4 mg of uncoated catalyst corresponding to a reduction in the number of Pt sites of 99% as compared to the data reported in Figure 1 (in which 400 mg of the same catalyst was used). The catalytic data obtained over the AO-coated sample (using the usual mass of 400 mg) are also reported for

the sake of comparison. A significant difference in the light off temperature for the CO and butane oxidation was observed when the number of Pt sites was reduced (Figure 14). The butane conversion decreased, as expected, but the CO conversion was also strongly affected, contrary to the case of the coated sample. The measured butane conversion was still higher over the 4 mg of plain catalyst as compared to the 400 mg of OA sample, while the CO conversion was already much lower in the case of the former material. Therefore, in agreement with the Na+ adsorption studies, the effect of the coating cannot be explained by a decrease in the surface area of Pt sites. 4. Discussion The material prepared here by the novel open-air (OA) method presents a high selectivity for the combustion of CO in the presence of butane. In contrast to the case of a catalyst prepared by an autoclave-based method, for which the selectivity was based primarily on size-exclusion due to the formation of a crystalline Na-LTA layer directly on the Pt,6 the size selectivity can also be achieved by the formation of an amorphous coating on the catalyst surface. This amorphous layer may be achieved simply by contacting the gel and the catalyst with mild aging with the amorphous coating functioning in a manner similar to a zeolite coating (Figures 6b and 8) and will increase the length of the pathway over which the molecules will have to travel before reaching the Pt active sites. Since the diffusivity of butane is lower than that of CO, the latter will be preferentially oxidized. It is important to recognize that (i) the mere contact with the synthesis gel is sufficient to modify the activity of the catalyst with respect to CO and butane oxidation and (ii) the oxidation of butane appeared more inhibited than that of CO (Figure 13). A chemical effect due to a poisoning of the noble metal with Na cannot be excluded completely as contact of the catalyst with NaOH also leads to an increase in the butane oxidation lightoff temperature. The effect of small concentration of alkali metals on the oxidative properties of Pt have been shown previously.18,19 This effect must be relatively small compared with the molecular discrimination or not being reproduced using NaOH versus synthesis gel as the CO light is unaffected by the gel treatment and the butane lightoff reaches 100% conversion whereas with NaOH the butane lightoff peaks at ∼40%. It is possible that the speciation of the sodium is important which cannot be simulated in a simple system easily. It is yet difficult to quantify the relative contribution of the increase of the diffusional pathway and the modification of the Pt leading to the inhibition of butane oxidation. Importantly, although gel contact at low temperatures or short exposure time was able to achieve some discrimination, this effect was maximized on treating the catalyst under OA (and AU) conditions. This may

Selectivity for CO and Butane Oxidation be due to the OA (and AU) treatments forming an essentially gap-free film. It must be noted that the possible poisoning effect of the gel constituent on the oxidation activity of catalysts is typically not discussed in the case of materials treated in similar conditions,7 while our work clearly shown that it does contribute to discriminating the reactants, albeit to a small extent. The structureless morphology of the coating observed by SEM (Figure 6b) was due to the fact that it was formed directly from the adsorption of the aluminosilicate gel onto the catalyst surface. The thermal treatment induced a partial crystallization by bond rearrangement, similarly to the early stage of the lowtemperature crystallization of Na-LTA as described by Valtchev et al.17 This resulted in the initial film morphology being kept while some porosity was created. The use of longer crystallization time would result in a more crystalline coating, due to the solution-mediated crystallization process,17 as reported in the case of the AU sample prepared by Collier et al. over longer synthesis periods.6 The use of longer crystallization time also favored the formation of large cuboidal crystals due to Ostwald ripening, as observed for the AU sample (Figure 6a). It can therefore be proposed that the origin of the intimate contact between the LTA crystals and the catalyst surface reported in the case of the AU sample reported elsewhere6 derives directly from the fact that the crystallization starts from the rearrangement of a gel that perfectly wetted the catalyst surface. To this respect, the use of the polyelectrolyte was shown to be paramount, as a sample prepared without it led to a poor inhibition of butane oxidation (Figure 2). Conclusions This work shows that it is possible to prepare a homogeneous porous coating about 500 nm thick deep into the bulk of 3 mm diameter catalyst beads by a simple ambient pressure-method not requiring the use of autoclaves. A mostly amorphous film (between 4 and 13% of crystallinity as compared to fully crystallized LTA zeolite crystals) was formed over the catalyst surface. The coated sample obtained by this method led to a catalyst selective for CO oxidation in the presence of butane. No butane oxidation was observed up to 350 °C. The discrimi-

J. Phys. Chem. C, Vol. 112, No. 29, 2008 10975 nation between CO and butane oxidation was not only due to pore size exclusion or diffusional constraints, but there was also some effect from the chemical and/or structural modification of Pt induced by the contact with the synthesis gel. Acknowledgment. This work was partly supported by the CARMAC project (EPSRC). We thank Dr. Dogan Ozkaya (JM) for his efforts with electron microscopy. References and Notes (1) Weisz, P. B.; Frilette, V. J. J. Phys. Chem. 1960, 64, 382. (2) Weisz, P. B.; Frilette, V. J.; Maatman, R. W.; Mower, E. B. J. Catal. 1962, 1, 307. (3) VanderPuil, N.; Creyghton, E. J.; Rodenburg, E. C.; Sie, T. S.; vanBekkum, H.; Jansen, J. C. J. Chem. Soc., Faraday Trans 1996, 92, 4609. (4) Ren, N.; Yang, Y.-H.; Shen, J.; Zhang, Y.-H.; Xu, H.-L.; Gao, Z.; Tang, Y. J. Catal. 2007, 251, 182. (5) Foley, H. C.; Lafyatis, D. S.; Mariwala, R. K.; Sonnichsen, G. D.; Brake, L. D. Chem. Eng. Sci. 1994, 49, 4771. (6) Collier, P.; Golunski, S.; Malde, C.; Breen, J.; Burch, R. J. Am. Chem. Soc. 2003, 125, 12414. (7) Zhong, Y.; Chen, L.; Luo, M.; Xie, Y.; Zhu, W. Chem. Commun. 2006, 27, 2911. (8) He, J.; Yoneyama, Y.; Xu, B.; Nishiyama, N.; Tsubaki, N. Langmuir 2005, 21, 1966. (9) Ke, C.; Yang, W. L.; Ni, Z.; Wang, Y. J.; Tang, Y.; Gua, Y.; Gao, Z. Chem. Commun. 2001, 783. (10) Berenguer-Murcia, A.; Morallo´n, E.; Cazorla-Amoro´s, D.; Linares´ . Micro. Meso. Mater. 2005, 78, 159. Solano, A (11) Kunkeler, P. J.; Moeskops, D.; van Bekkum, H. Microp. Mater 1997, 11, 313. (12) Goodman, K. E.; Hayes, J. W.; Malde, C. N.; Petch, M. I. European Patent 0 878 233, 1998. (13) Groen, J. C.; Pe´rez-Ramı´rez, J. Appl. Catal. A: Gen. 2004, 268, 121. (14) Diaz, E.; Munoz, E.; Vega, A.; Ordonez, S. Ind. Eng. Chem. Res. 2008, 47, 412. (15) Gramlich, V.; Meier, W. M. Z. Kristallogr. 1971, 133, 134. (16) Valtchev, V. P.; Bozhilov, K. N. J. Am. Chem. Soc. 2005, 127, 16171. (17) Smaihi, M.; Barida, O.; Valtchev, V. Eur. J. Inorg. Chem. 2003, 4370. (18) Yentekakis, I. V.; Moggridge, G.; Vayenas, C. G.; Lambert, R. M. J. Catal. 1994, 146, 292. (19) Vernoux, P.; Leinekugel-Le-Cocq, A.-Y.; Gaillard, F. J. Catal. 2003, 219, 247.

JP801158D