Applied Catalysis A: General 233 (2002) 21–33 Oxidative dehydrogenation of propane over Cr2 O3/Al2 O3 and Cr2 O3 catalysts: effects of loading, precursor and surface area Maymol Cherian a , Musti Someswara Rao a , Wei-Tin Yang b , Jih-Mirn Jehng b , Andrew M. Hirt c , Goutam Deo a,∗ a b Department of Chemical Engineering, Indian Institute of Technology, Kanpur 208 016, India Department of Chemical Engineering, National Chung Hsing University, Taichung 402, Taiwan c Material Research Laboratory Inc., 290 North Bridge Street, Struthers, OH 44471, USA Received 12 October 2001; received in revised form 26 January 2002; accepted 5 February 2002 Abstract Several alumina supported chromium oxide catalysts were prepared by varying the chromium oxide loading, precursors and surface areas of the support. The prepared catalysts were characterized using BET, XRD, XPS and UV–VIS spectroscopic techniques. The monolayer limit was observed to be ∼9 mol Cr/m2 . Below monolayer limits, surface chromium oxide species were present irrespective of precursors and surface area of the support. The activity of the prepared samples was tested for ODH of propane. It was observed that the supported chromium oxide samples were active for the ODH reaction and that propene was the major product. The activity and selectivity increased with loading up to monolayer limits and decreased for higher loadings. Bulk Cr2 O3 was also studied for the ODH reaction and found to behave differently than the Cr2 O3 /Al2 O3 catalysts. The efficiency of each surface chromium oxide species on alumina to carry out the ODH of propane, the turn over frequency (TOF), was relatively independent of chromium oxide loading for the monolayer catalysts. However, it appears that a constant fraction of the surface chromium oxide species is active for the ODH of propane. This constant fraction of the surface chromium oxide phase is given by the polymeric chromium oxide species. The inactive fraction is given by the monomeric chromium oxide species that is not reducible. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Oxidative dehydrogenation; Chromia; Alumina; Propane; Propene; Surface active sites; Catalyst; Precursor; Surface area; Loading 1. Introduction Oxidative dehydrogenation (ODH) or non-ODH of hydrocarbons into their corresponding alkenes is an industrially important process due to the increased demand of alkenes [1]. Alkenes, especially ethene and propene, are used as the feedstock for the petrochemical industry. Catalytic dehydrogenation is accepted ∗ Corresponding author. Fax: +91-512-590104. E-mail address: [email protected] (G. Deo). as a method for the production of alkenes more selectively even at higher conversion. Many transition and noble metals and metal oxides have been successfully tried for several dehydrogenation reactions. Widely used dehydrogenation catalysts include supported or unsupported Pt, Pd, Sn, V–Mg or Mo based oxides, Co, Bi or Sb based oxides [2,3]. However, the deactivation of these catalysts with time-on-stream poses a problem. In addition, obtaining high conversion with high selectivity over many of these catalysts was not always possible. Furthermore, thermodynamic 0926-860X/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 8 6 0 X ( 0 2 ) 0 0 1 3 2 - 1 22 M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 constraints limited the universal application of the dehydrogenation reaction. ODH provides an attractive alternative [4]. Therefore, development of a catalyst that kinetically controls the ODH reaction is a challenge in recent times. Chromia-based catalysts have been studied for various petrochemical and environmental applications [5,6]. Cr2 O3 supported on Al2 O3 is employed as the commercial catalyst for dehydrogenation of alkanes. Various studies were performed to gain an insight about the activity of Cr2 O3 /Al2 O3 for dehydrogenation reactions [7–11]. Bulk Cr2 O3 was studied for the oxidation of hydrocarbons and it was observed that Cr2 O3 is efficient for complete combustion. However, when chromia is supported on another metal oxide (e.g. Al2 O3 , TiO2 , SiO2 , ZrO2 , etc.), the structure and reactivity properties are altered and an improvement in the activity/selectivity of these chromia-based catalysts is observed [12–15]. Various characterization techniques have been applied to study the supported chromium oxide catalysts: these include XRD, Raman, UV–VIS, TPR, XPS and ESR studies [16–20]. Results of these characterization techniques show that the supported chromia catalysts possess a two-dimensional surface species below monolayer loadings that are the active sites in several reactions. The support-surface interaction stabilizes the Cr species into different oxidation states, such as Cr3+ , Cr5+ and Cr6+ . The relative stability of different chromium oxidation states depends upon the extent of support-surface interaction, which is decided by the acid–base properties of the support and the interaction energy between the surface-support phases [21]. Several recent studies focus on the ODH of alkanes over alumina supported chromia catalysts. Flick and Huff [22] studied the ODH of ethane over Pt-Cr2 O3 /Al2 O3 catalysts and observed an increase in C2 H4 selectivity and C2 H6 conversion compared to the results over Pt/Al2 O3 catalysts. Grzybowska et al. [23] and Al-Zahrani et al. [24] studied the influence of loading and preparation methods on the structure-reactivity properties of Cr2 O3 /Al2 O3 catalysts for ODH of isobutane. These studies reported that the catalytic properties of supported chromium oxide are due to the surface chromium oxide species formed as a result of chromium-support interactions. To the best of our knowledge only four studies have reported the ODH of propane over chromium supported on alumina catalysts [25–28]. These studies are limited to analysis of a single Cr2 O3 /Al2 O3 catalyst. No information regarding the effect of chromium oxide loading was given for the ODH of propane. In the present study, the ODH of propane reaction is considered over a series of Cr2 O3 /Al2 O3 and bulk Cr2 O3 samples to understand the effect of loading. The effects of precursor and surface area on the structure-reactivity properties of Cr2 O3 /Al2 O3 catalysts are also considered. Finally, the structure property relationships of alumina supported chromium oxide catalysts are discussed. 2. Experimental 2.1. Catalyst preparation Supported chromia-alumina catalysts were prepared by incipient wetness impregnation technique. The precursor used was chromium nitrate nonahydrate (Cr(NO3 )3 ·9H2 O, Aldrich, 99.98% purity) and alumina (Condea) as support. The support was pretreated with incipient volumes of distilled water and then calcined at 600 ◦ C for 6 h. The pretreated support and incipient volumes of solutions containing predetermined amounts of precursor were intimately mixed in order to prepare the catalysts with different loadings of chromium oxide. The mixture was kept in a dessicator overnight, followed by drying at 110 ◦ C for 8 h, and at 250 ◦ C for another 8 h. Finally, the samples were calcined at 600 ◦ C for 6 h. The prepared catalysts were denoted as x% CrAl-600, where x% is the wt.% loading corresponding to Cr2 O3 . The effect of precursor was studied by preparing Cr2 O3 /Al2 O3 catalysts using ammonium dichromate ((NH4 )2 Cr2 O7 ) and adopting the procedure given above. The prepared samples were denoted as x% CrAl-AD. Influence of the surface area on the support-surface interaction was studied by preparing alumina supports possessing different surface areas. The original alumina support was subjected to 900 ◦ C heat treatment and thereafter was repeatedly washed with NH4 OH in order to remove any sodium impurities. The washed alumina supports were pretreated and loaded with chromia by following the above procedure. The prepared samples were denoted as x% CrAl-900. M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 Bulk Cr2 O3 was prepared by decomposing ammonium dichromate at 250 ◦ C. The decomposed sample was calcined at 600 ◦ C for 6 h. XRD analysis reveals the presence of Cr2 O3 crystals and no other phases are present. 2.2. Physical characterization The BET surface area analysis was done by a single point method using N2 adsorption at 77 K in a Micromeritics Pulse Chemisorb 2700 apparatus. X-ray diffraction (XRD) patterns were obtained on a 180 Debye flex-2002 X-ray diffractometer using Ni filtered K␣ radiation from a Cu target (λ = 1.54056 Å); the instrument was equipped with a monochromator. X-ray photoelectron spectroscopy (XPS) studies were done using Mg K␣ or Al K␣ in fixed analyzer transmission (FAT) mode. The catalyst powder was pressed between a stainless steel holder and a polished single crystal silicon wafer; each such sample was then installed in a vacuum chamber of a Model DS 800 XPS surface analysis system. UV–VIS spectroscopy studies were done using a Hitachi 20/120 spectrophotometer using a deuterium lamp for studies between 190 and 340 nm and a tungsten lamp for wavelengths exceeding 340 nm. Barium sulfate was used as the standard. Catalyst pellets were supported on an absorbing background and then placed in the sample holder. The spectra were obtained under ambient conditions. 2.3. Chemical characterization 2.3.1. Temperature programmed reduction (TPR) The TPR studies were carried out in a microreactor containing ca. 0.05 g of catalyst and attached to a Micromeritis Pulse Chemisorb 2705 analyzer. Helium was used as the carrier gas and used also to degas the samples for 0.5 h prior to the reduction experiments. The helium flow was set at 30 ml/min. A 10% H2 /Ar mixture flowing at 40 ml/min was used for reduction and the temperature was ramped from ∼100 ◦ C at a rate of 10 ◦ C/min to 700 ◦ C. The amount of hydrogen consumed was detected using a TCD. 2.3.2. Reactivity of the Cr2 O3 /Al2 O3 The samples were tested for the ODH of propane in a down-flow quartz reactor at atmospheric pressure. 23 The reactor was a single piece of quartz with an inlet of 10 mm i.d. and 15 cm length and an outlet of 5 mm i.d. and 15 cm length. The two sections were tapered and the catalyst bed was placed just above the tapered region on quartz wool. The reactor tube was mounted vertically in a tubular furnace. The temperature of the reactor and the catalyst bed was measured by a thermocouple located inside the reactor tube just above the catalyst bed and was controlled by a PID temperature controller (FUJI Micro-controller X Model PXZ 4). The product gases were sent for online analysis to a gas chromatograph (AIMIL-NUCON 5765) equipped with a methanizer. The carbon oxides and hydrocarbons are analyzed in an FID mode using an activated alumina column. The propane flow rate was adjusted through an electronic flow controller (Bronkhost Hi-Tec, El-Flow Mass flow controller) and the air flow rate was adjusted through a rotameter (Eureka, Model SRS/MG-5) to maintain a 3:1 propane to oxygen ratio. A physical mixture of 0.1 g of the catalyst and amount of quartz glass powder required to form a bed height of 1 cm was loaded into the reactor. Runs were performed at different temperatures starting from 300 to 450 ◦ C with a constant total flow rate of 42.7 sccm. The experiments were performed in a differential mode and the conversions are limited to below 5%. Based on the inlet and outlet concentrations and assuming differential reactor conditions one can calculate the conversion, selectivity, yield and TOF as follows: nc conversion (%) = 100 nf nhc Nhc selectivity (%) = 100 nc Np Nhc nhc yield (%) = 100 nf Np TOF (s−1 ) = FA0 XA nCr where nc and nf are the number of moles of propane consumed and fed, respectively, nhc the number of moles of products (eg. propene, ethene, carbon oxides, etc.) formed, Nhc and Np the number of carbon atoms present in the products and propane, respectively, FA0 the molar flow rate, moles of propane fed per second, 24 M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 XA the conversion of propane per gram of the catalyst, and nCr the moles of Cr per gram of the catalyst. Since propene is the desired product, the propene selectivity is reported. The by-products are CO and CO2 and the by-product selectivity can be obtained by difference. Blank reactor runs were conducted and no significant conversions were observed under the experimental conditions. For each catalyst, several runs were taken at a particular temperature and the average value is reported. constant. Others have observed similar variations in surface area [23,29,30]. The initial decrease in surface area might be attributed to the plugging of pores. In contrast, the surface area for pure Cr2 O3 sample is 27 m2 /g. The surface area of the Al2 O3 -900 sample is 134 m2 /g. The third column in Table 1 gives the surface concentration of chromium oxide, mol Cr/m2 , based on the wt.% Cr2 O3 and the surface area of the sample. 3.2. XRD 3. Results Several Cr2 O3 /Al2 O3 catalysts were prepared as a function of loading, precursor used and surface area and were analyzed by XRD, XPS and UV–VIS spectroscopic techniques. The characterized samples were then studied by TPR and ODH of propane. Initially, the results of the CrAl-600 samples are presented, followed by the results related to the effect of precursor used and surface area. 3.1. BET The BET surface areas of pure alumina (Al2 O3 -600 and -900) and the Cr2 O3 /Al2 O3 samples were obtained and are tabulated as the second column in Table 1. The surface area values for CrAl-600 samples are between 140 and 185 m2 /g. The surface area initially decreased with loading and then remained relatively Table 1 BET surface area and TPR data of CrAl-600 and bulk Cr2 O3 samples Samples Surface area (m2 /g catalyst) Concentration (mol Cr/m2 ) Tmax (◦ C) H/Cr Al2 O3 -600 Al2 O3 -900 5% CrAl-600 7.5% CrAl-600 10% CrAl-600 15% CrAl-600 20% CrAl-600 Bulk Cr2 O3 201 134 185 170 156 141 155 27 – – 3.6 5.8 8.4 14.0 17.0 18.0a n.d. n.d. 388 374 373 377 – – 1.9 1.6 1.5 1.2 n.d.: not detected. a Based on a value 11 Cr/nm2 [9]. b Not shown. b b 292 0.7 The XRD patterns of pure Al2 O3 , 5, 10, 15, 20% CrAl-600 samples and pure Cr2 O3 samples are shown as Fig. 1. For all the CrAl-600 samples the major peaks are observed at 46.1 and 67.8◦ with an additional broad peak between 36 and 38◦ . These peaks are due to the ␥-Al2 O3 support. The broadness of the peak suggests that this alumina support is amorphous. For the 20% CrAl-600 sample additional peaks appear at 24.8, 32.8 and 54.6◦ , corresponding to peaks due to Cr2 O3 crystals. Thus, from the XRD it appears that the monolayer coverage is exceeded for the 20% Cr2 O3 /Al2 O3 sample and that an XRD amorphous chromium oxide species is present for lower loadings. Similar changes in XRD patterns are reported in the literature [31,32]. 3.3. XPS The CrAl-600 and -900 samples were analyzed using XPS. The XPS results reveal the presence of only Cr, Al, O and C on the samples. Trace amounts of F, Cl, and Sn were also detected. From the XPS data the surface atomic concentration of Cr and Al were calculated and the Cr/(Cr+Al)surface ratio was determined. From the knowledge of chromia in the sample the Cr/(Cr + Al)bulk ratio was also calculated. The Cr/(Cr+Al)surface ratio is plotted versus the Cr/(Cr + Al)bulk ratio in Fig. 2. Fig. 2 shows that the Cr/(Cr+Al)surface ratio increases with Cr/(Cr+Al)bulk ratio up to a Cr/(Cr + Al)bulk value of 0.09 after which the Cr/(Cr + Al)surface ratio remains constant. A Cr/(Cr + Al)bulk value of 0.09 corresponds to ∼12% Cr2 O3 loading. Plateauing of such type of plots has been attributed to multilayer formation [33]. Similar results have been observed by others [34,35]. M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 25 Fig. 1. X-ray diffractograms of Al2 O3 , x% CrAl-600 samples and bulk Cr2 O3 . 3.4. UV–VIS The spectra obtained in the diffuse reflectance modes of the supported chromium oxide samples and of pure Cr2 O3 physically mixed with Al2 O3 are shown in Fig. 3. For the 5 and 10% loading, two peaks centered at 276 and 370 nm are observed that correspond to Cr6+ . These peaks of lesser intensity are also present in the 15% CrAl-600 samples. As the loading increases above 15% Cr2 O3 loading, new broad peaks centered at 460 and 600 nm, similar to those of bulk Cr2 O3 (top spectrum) are observed. Thus, from UV–VIS also it appears that the monolayer coverage is exceeded for the 20% CrAl-600 sample. Furthermore, UV–VIS spectra reveal the presence of Cr6+ for low loadings (up to 15%) and gradual development of Cr3+ octahedral symmetry transitions due to Cr2 O3 with increase in loading. The studies are similar to the results reported in the literature for Cr2 O3 /Al2 O3 catalysts [36,37]. 3.5. Temperature programmed reduction TPR studies were performed for the prepared Cr2 O3 /Al2 O3 samples and for pure Cr2 O3 . The TPR profile of pure Al2 O3 sample does not reveal any features in the temperature range considered. The TPR profiles for 5, 7.5, 10 and 15% CrAl-600 are shown in Fig. 4, where the TCD signal is plotted versus the temperature. A single Tmax is observed for all samples; the value is between 373 and 388 ◦ C. From the peak area and the calibration amounts of hydrogen, the hydrogen consumption can be calculated; the value is then used to calculate the H/Cr ratio. Based on the data in Fig. 4, the Tmax temperature and H/Cr ratios are tabulated in the fourth and fifth columns of Table 1 for the CrAl-600 catalysts. From the data 26 M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 Fig. 2. (Cr/Cr + Al)XPS vs. (Cr/Cr + Al)bulk for x% CrAl-600 samples. presented in Table 1, one may observe that the Tmax temperature appears to be constant within experimental error. Furthermore, the H/Cr ratio appears to decrease somewhat with loading; however, this variation is within experimental error. For bulk Cr2 O3 , the Tmax and H/Cr ratio are 292 ◦ C and 0.7, respectively; these have been included in Table 1 for reference. Similar TPR results showing the difference between Cr2 O3 and Cr2 O3 /Al2 O3 have also been observed before [23]. 3.6. ODH of propane The ODH of propane reaction over the 5, 7.5, 10, 15 and 20% CrAl-600 samples, and that of pure Cr2 O3 and Al2 O3 were performed at different temperatures ranging from 380 to 500 ◦ C. The conversion and selectivity data at a reaction temperature of 400 ◦ C are plotted in Fig. 5 as a function of loading. The pure Al2 O3 support did not show any activity towards ODH under the present operating conditions. As the chromia loading is increased, the propane conversion increases, reaches a maximum at 15% loading, and de- creases at higher loadings. The maximum conversion for 15% CrAl-600 at 400 ◦ C is 3.1%. The selectivity is observed to follow the same trend as that of conversion with loading. The selectivity increases with loading up to 15% and decreases slightly at higher loadings. To observe the difference in reactivity over alumina supported chromia oxide catalysts and bulk Cr2 O3 , the conversion and selectivity during ODH of propane of 15% CrAl-600 and bulk chromia between 380 and 500 ◦ C is shown in Fig. 6. The conversion of both catalysts increases with temperature. However, the conversions of the supported chromium oxide catalysts are greater than the conversions of bulk Cr2 O3 at all temperatures. For bulk Cr2 O3 catalysts, the selectivity decreases with temperature, however, for 15% CrAl-600 catalyst it reaches a maximum and remains constant at higher temperatures. Conversion trends similar to the 15% CrAl-600 catalysts shown in Fig. 6 are observed for the other CrAl-600 catalysts. The yield of propene is calculated based on the data given in Fig. 5 and values are tabulated in Table 2. The turn over frequency (TOF, s−1 ) calculated based on the total moles of Cr present in sample is also tabulated. M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 27 Fig. 3. UV–VIS spectra of x% CrAl-600 samples, and Cr 2 O3 + Al2 O3 physical mixture. Table 2 shows that the propene yield increases with loading up to 15% loading and then decreases. The TOF, however, is observed to remain constant below 15% loading and then to decrease. The activation energies of the Cr2 O3 /Al2 O3 samples based on a first-order reaction and differential conditions were also calculated and varied between 90 and 110 kJ/mol, Table 2 Reactivity data for CrAl-600 and Cr2 O3 catalysts at 400 ◦ C Samples Activity (× 106 mol C3 H8 /(g s)) Yield (%) TOF (× 103 , s−1 )a 5% CrAl-600 7.5% CrAl-600 10% CrAl-600 15% CrAl-600 20% CrAl-600 Bulk Cr2 O3 1.2 1.7 2.3 3.2 1.7 0.7 0.8 1.2 1.7 2.4 1.2 0.5 1.9 1.8 1.7 1.6 0.7 0.2 a Related to chromium concentration given in Table 1. which are similar to values reported earlier [38,39]. However, the activation energy for pure Cr2 O3 was ∼40 kJ/mol. 3.7. Effect of the precursor The influence of the precursor on the structure and activity of chromia species was studied by performing BET, TPR, XPS, Raman and activity studies on 5 and 10 wt.% of Cr2 O3 on Al2 O3 catalysts. These Cr2 O3 /Al2 O3 samples are prepared using chromium nitrate and ammonium dichromate as the different precursors. Comparisons of the results of the above studies are summarized in Table 3. It is observed that the highest chromium–oxygen peak positions in the ambient Raman spectra and the XPS surface to bulk atomic ratio are similar. The H/Cr ratio from TPR data shows a small variation, however, the Tmax positions for the samples prepared from the two precursors were similar. The surface area of the Cr2 O3 /Al2 O3 28 M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 in Table 3. The difference in their values is within 5%. 3.8. Effect of surface area Fig. 4. TPR profile for x% CrAl-600 catalysts. catalysts prepared from ammonium dichromate had a lower surface area, 165 and 145 m2 /g, compared to the samples prepared from chromium nitrate. The reactivity properties of Cr2 O3 /Al2 O3 samples prepared using both chromium nitrate and ammonium dichromate were also studied for the ODH of propane and the activity, selectivity and TOF are tabulated The effect of surface area on the structure and activity properties of Cr2 O3 /Al2 O3 catalysts were studied by using alumina supports possessing different surface areas. The BET surface areas for CrAl-900 samples also show a behavior similar to that of the CrAl-600 samples. The surface area for the CrAl-900 samples varied between 92 and 134 m2 /g. The XRD patterns for the CrAl-900 samples reveal strong peaks corresponding to crystals of θ -Al2 O3 , for which the major peaks are present at 2θ values at 46 and 67.7◦ . For CrAl-900 samples, crystals of Cr2 O3 were not observed by XRD. XPS analysis of these samples reveals the absence of impurities. UV–VIS, Raman, TPR and the ODH of propane studies were used to analyze the structure and reactivity properties of the CrAl-900 samples. The results of these studies are summarized in Table 4. The Raman band of the highest chromium–oxygen vibration is at 882–890 cm−1 , corresponding to monochromate and polychromate species. The 10% CrAl-900 samples had a Raman band of lesser intensity at 550 cm−1 corresponding to a small amount of Cr2 O3 crystals. The spectroscopy studies on the CrAl-900 samples reveal the presence of Cr6+ peak at ∼278 and ∼372 nm. The Raman and UV–VIS data for CrAl-900 samples are similar to those for CrAl-600 samples. The H/Cr and Tmax values from the TPR experiments reveal similar values for the CrAl-900 samples and the CrAl-600 samples, except for the H/Cr value for the 10% CrAl-900 sample. For this sample, small amounts of Cr2 O3 crystals were detected in the Raman, which Table 3 Effect of precursor used on the properties of Cr2 O3 /Al2 O3 catalyst Precursors (CrNO3 )3 ·9H2 O (NH4 )2 Cr2 O7 Wt.% Cr2 O3 Ambient Raman (cm−1 ) XPS Cr/Al TPR (H/Cr) Tmax (◦ C) TOF (× 103 , s−1 )a Selectivity (%)a Ambient Raman (cm−1 ) XPS Cr/Al TPR (H/Cr) Tmax TOE (× 103 , s−1 )a Selectivity (%)a 5 10 880 882 0.07 0.13 1.9 1.5 351 337 1.9 1.7 63 72 882 887 0.07 0.12 2.11 1.34 354 340 60 71 a Based on propane ODH at 400 ◦ C. 1.7 1.8 M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 29 Fig. 5. Variation of conversion (solid symbol) and selectivity (open symbol) with loading (wt.% Cr2 O3 ) for the CrAl-600 catalysts for ODH of propane at 400 ◦ C. gives a smaller H/Cr value (see Table 1). The selectivity and TOF for CrAl-900 samples for ODH of propane at 400 ◦ C were calculated and are also shown in Table 4. It is observed that the CrAl-900 samples have a slightly higher conversion and selectivity than the CrAl-600 samples. 4. Discussion The Cr2 O3 Cr2 O3 reveal XRD and UV–VIS studies reveal that bulk is formed on the alumina support at 20% loading. Raman spectra on the same samples that crystals of Cr2 O3 are present for the Table 4 Effect of surface area on the properties of Cr2 O3 /Al2 O3 catalysts CrAl-600 Al2 O3 -201 m2 /g CrAl-900 Al2 O3 -134 m2 /g Ambient Raman UV–VIS TPR Tmax (nm) (H/Cr) (◦ C) TOF (× 103 , s−1 )a Selectivity Ambient (%)a Raman (cm−1 ) UV–VIS (nm) TPR (H/Cr) Tmax (◦ C) TOF (× 103 , s−1 )a Selectivity (%)a 5 880 280 376 1.92 351 1.9 63 890 276 370 1.7 341 2.8 67 10 882 282 374 1.52 337 1.7 72 862 278 372 1.1 328 2.0 76 Wt.% Cr2 O3 a Based on propane ODH at 400 ◦ C. 30 M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 Fig. 6. Variation of conversion (solid symbols) and selectivity (open symbols) with temperature for 15% CrAl-600 and bulk Cr2 O3 catalysts for ODH of propane. synthesized 15% CrAl-600 sample [40]. Consequently, Raman spectroscopy is more sensitive than XRD and UV–VIS for the detection of Cr2 O3 crystals. Data from XPS suggest that monolayer coverages, defined as the coverages where crystals of Cr2 O3 are first observed, is achieved at 12% Cr2 O3 loading. Based on the surface area of the 12% Cr2 O3 /Al2 O3 sample, this corresponds to ∼9 mol Cr/m2 . Similar values of monolayer coverage have also been observed before [41]. Prior to bulk Cr2 O3 formation, the chromium oxide species is molecularly dispersed. The presence of dispersed chromium oxide is observed as Cr6+ species in the UV–VIS spectra and as mono- and polychromate species in the Raman spectra obtained on the same samples under ambient condition [40]. Surface chromium oxide in the +6 oxidation state is the predominant species present below monolayer coverages [23,42]. Minor amounts of Cr5+ have also been proposed [43]. When the surface moisture is removed the surface chromium oxide species on alumina is present as tetrahedral monomeric and polymeric species different from those present under ambient conditions [44]. Furthermore, the ratio of the two species does not change as a function of surface coverage below monolayer loadings. Since XPS reveals that no impurities are present in the Cr2 O3 /Al2 O3 samples, similar results can be expected. Indeed Raman spectra of 15% CrAl-600 under dehydrated conditions also reveals bands at 866 and 1002 cm−1 corresponding to monomeric and polymeric tetrahedral species [45]. Thus, below monolayer coverage, which in the present study is achieved at ∼12% Cr2 O3 loading, the surface chromium oxide species is predominantly present as Cr6+ species monomeric and polymeric species. M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 The reactivity studies reveal that the activity and selectivity of the Cr2 O3 /Al2 O3 catalysts depend upon the type of the chromium oxide species present. The maximum conversion is observed for the 15% Cr/Al-600 catalyst. At this loading, the physical characterization studies suggest that the maximum amount of surface chromia species is present and a small amount of bulk Cr2 O3 is formed. Below 15% loading, the alumina surface is also exposed to the reactant and product molecules. It is this exposed alumina surface below monolayer coverages that is partially responsible for the decrease in selectivity. As the loading increases above 15%, the presence of bulk Cr2 O3 crystals of the sample increases, which results in a decrease in propane conversion and propene selectivity and consequently in the yield. These results suggest that the bulk Cr2 O3 species that is observed by XRD, Raman and UV–VIS spectroscopies is less active than the two-dimensional surface phase that is present below monolayer loadings. The difference in propane reactivity between the surface chromium oxide species and bulk Cr2 O3 is clearly observed in Fig. 6, where the variation of conversion and selectivity with temperature were significantly different. Furthermore, the smaller activation energy for bulk Cr2 O3 suggests that a different rate-determining step is involved. The difference between the surface chromium oxide phase and bulk Cr2 O3 is also observed in the TPR experiment. For bulk Cr2 O3 the Tmax is at 292 ◦ C, whereas the surface chromium oxide species possesses a Tmax temperature between 373 and 388 ◦ C. The H/Cr ratio is also different, since the value for bulk Cr2 O3 is 0.7 and for the surface chromium oxide species it is 1.3–1.9. All of these factors clearly suggest that the surface chromium oxide species that are present below monolayer coverages are more active and behave differently than bulk Cr2 O3 towards the ODH of propane. The TPR experiments reveal that the Tmax temperature and the H/Cr ratio are relatively insensitive to the loading for the monolayer and sub-monolayer catalysts. Furthermore, the H/Cr ratio is less than 3. An H/Cr value of 3 corresponds to the reduction of the surface chromium oxide species from +6 to +3 oxidation state. In situ Raman spectroscopy studies reveal that the polymeric surface chromium oxide species is reducible and the monomeric species is not [46]. Correlating the H/Cr ratio and the in situ Raman studies suggests that only a fraction of the surface chromium 31 oxide species is reducible and, consequently, the H/Cr ratio is less than 3. This reducible fraction is given by the polymeric chromium oxide species. Furthermore, since the ratio of monomeric and polymeric species is independent of coverage, a constant H/Cr ratio is observed for the monolayer and sub-monolayer catalysts. To understand the effect of coverage of the two-dimensional surface chromium oxide species, we calculated the TOF based on the total chromium oxide loading (see Table 2). The TOF is relatively constant for the monolayer Cr2 O3 /Al2 O3 catalysts. A constant TOF value is also observed for supported vanadium oxide catalysts for the ODH of propane [47,48] and ethane [49]. If one assumes that a redox mechanism occurs, it appears that reducible polymeric species are active for the ODH of propane. Consequently, the TOF is independent of coverages, since the polymeric chromium oxide species form a constant fraction of the total surface chromium oxide sites. Thus, a constant fraction of the total surface chromium oxide phase represented by the polymeric chromium oxide species is active for the ODH of propane, which gives rise to a TOF value that is independent of coverage. Several mechanisms have been proposed for the ODH of propane [1,50,51]. One of the mechanisms proposed for supported metal oxide catalysts involves three routes [51]. In this mechanism, propane is directly converted to propene (k1 ) and carbon oxides (k2 ), and a secondary reaction in which propene is converted to carbon oxides (k3 ). It was suggested that, as loading increases, the selectivity increases due to the formation of large domains of poly vanadates or poly molybdates, which decrease the k2 /k1 ratio. In the present study, however, the ratio of monochromates to polychromates does not change with respect to loading [44]. Therefore, it appears that the increase in selectivity with loading for supported chromium oxide catalysts is not controlled by the decrease in k2 /k1 ratio associated with the polychromates. Moreover, the selectivity is improved with an increase in chromium oxide loading due to the coverage of exposed support sites, which otherwise would degrade propene to carbon oxides. For a more detailed kinetic model for the ODH of propane over supported chromium oxide sites, additional in situ characterization studies are required. The choice of the precursor used to form the surface chromium oxide species appears to be immaterial, since the catalysts prepared by using ammonium 32 M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 dichromate or chromium nitrate do not differ significantly in their physical and chemical characteristics (see Table 3). Different surface areas of the alumina support also do not affect the physical nature of the surface chromium oxide species, since the Raman and UV–VIS spectra are not significantly different. Due to a lower surface area of the support, a small amount of Cr2 O3 crystals are observed for the 10% CrAl-900 sample, since about 9% Cr2 O3 can be accommodated on this alumina. However, minor differences in reactivity between the different surface area Cr2 O3 /Al2 O3 catalysts are observed. These differences in TOF may be related to a slightly higher effectiveness factor of the CrAl-900 samples [52]. The higher effectiveness factor arises from larger pores associated with lower surface areas of the same material. As mentioned before, the selectivity to propene for the CrAl-900 samples is higher, since the amount of exposed support-surface is lower when compared with the CrAl-600 samples. These observations regarding the effect of precursors and surface area are supported by the hypothesis that the preparation method or surface area do not play a significant role in preparing the surface metal oxide phase provided the support matrix remains unchanged [53]. Consequently, the physical and chemical characteristics are similar for both catalysts, since the same surface chromium species is present. In summary, below monolayer coverage, molecularly dispersed surface chromium oxide species are formed independent of the precursor or surface area of the support. These surface species are active for the ODH of propane to propene and show a greater activity and selectivity compared to bulk Cr2 O3 . As the concentration of the surface chromium oxide is increased, the average effectiveness of the chromium oxide species, given by the TOF, is relatively constant and the selectivity to propene increases. The selectivity increases with increase in loading, since the exposed Al2 O3 surface decreases. The independence of the TOF for surface chromium oxide species is related to the constant fraction of the polymeric chromium oxide species present. 5. Conclusions Several Cr2 O3 supported Al2 O3 catalysts were prepared by incipient wetness impregnation method. The samples were analyzed by BET, XRD, TPR, XPS and UV–VIS characterization techniques. The presence of surface chromates was confirmed and the monolayer limit was estimated at ∼9 mol Cr/m2 . The characterization studies suggest that the surface chromia species are predominantly present as Cr6+ species independent of the precursor used and of the surface area of the support. The prepared samples were observed to be chemically active for the ODH of propane. The propane conversion increases with loading up to monolayer coverages and decreases for higher loadings since bulk Cr2 O3 are less active than the surface chromium oxide species. The propene selectivity increases with loading due to the decrease in the amount of exposed support-surfaces. It appears that a constant fraction of surface chromium oxide sites are active for the ODH of propane reaction; such sites give a TOF that is independent of surface coverage. This constant fraction is given by the polymeric chromium oxide species, which form the active sites for the ODH of propane. Acknowledgements The authors gratefully acknowledge Prof. Jitendra Kumar for assisting in obtaining the UV–VIS and Prof. I.E. Wachs for providing the Raman data. This work has been partially supported by the financial assistance of The Department of Science and Technology, India. References [1] H.H. Kung, Adv. Catal. 40 (1995) 1. [2] Y. Iwasawa, Studies in Surface Science and Catalysis, Vol. 101, Elsevier, Amsterdam, 1996. [3] B. Delmon, P. Ruiz, S.R.G. Carrazan, S. Korili, M.A. Vicente Rodriguez, Z. Sobalik, Catalysis in Petroleum Refining and Petrochemical Industries, 1995, p. 1. [4] J. Haber, Handbook of Heterogeneous Catalysis 4 (1997) 2258. [5] C.M. Pradier, F. Rodrigues, P. Marcus, M.V. Landau, M.L. Kaliya, A. Gutman, M. Herskowitz, Appl. Catal. B: Environ. 27 (2000) 73. [6] H. Rajesh, U.S. Ozkan, Ind. Eng. Chem. Res. 32 (8) (1993) 1622. [7] S. Udomsak, R.G. Anthony, Ind. Eng. Chem. Res. 35 (1996) 47. [8] H.J. Lugo, J.H. Lunsford, J. Catal. 91 (1985) 155. M. Cherian et al. / Applied Catalysis A: General 233 (2002) 21–33 [9] S. De Rossi, G. Ferraris, S. Fremiotti, E. Garrone, G. Ghiotti, M.C. Campa, V. Indovina, J. Catal. 148 (1994) 36. [10] O.F. Gorriz, L.E. Cadus, Appl. Catal. A: Gen. 180 (1999) 247. [11] A. Hakuli, A. Kyokivi, A.O.I. Krause, Appl. Catal. A: Gen. 190 (2000) 219. [12] J.K. Sohn, S.G. Ryu, Langmuir 9 (1993) 126. [13] U. Scharf, H. Scneider, A. Baiker, A. Wokuan, J. Catal. 145 (1994) 424. [14] W.K. Jozwiak, I.G.D. Lana, J. Chem. Soc., Faraday Trans. 93 (15) (1997) 2583. [15] M.I. Zuki, M.A. Hasan, N.E. Fouad, Appl. Catal. A: Gen. 171 (1998) 315. [16] F.D. Hardcastle, I.E. Wachs, J. Mol. Catal. 46 (1988) 173. [17] S. Khadder-Zine, A. Ghorbel, C. Naccache, J. Mol. Catal. 150 (1999) 223. [18] A. Rahman, M.H. Mohamed, M. Ahmmed, A.M. Aitani, Appl. Catal. A: Gen. 121 (1995) 203. [19] S.J. Scierka, M. Houalla, A. Proctor, D.M. Hercules, J. Phys. Chem. 99 (1995) 1537. [20] M.A. Vuurman, D.J. Stufkens, A. Oskam, J.A. Moulijn, F. Kapteijn, J. Mol. Catal. 60 (1) (1990) 83. [21] G.C. Bond, Handbook of Heterogenous Catalysis 3 (1997) 762. [22] D.W. Flick, M.C. Huff, Appl. Catal. A: Gen. 187 (1999) 13. [23] B. Grzybowska, J. Sloczynki, R. Grabowski, K. Wcislo, A. Kozlowska, J. Stoch, J. Zielinski, J. Catal. 178 (1998) 687. [24] S.M. Al-Zahrani, N.O. Elbashir, A.E. Abasaeed, M. Abdulwahed, Ind. Eng. Chem. Res. 40 (2001) 781. [25] R. Grabowski, B. Grzybowska, K. Samson, J. Sloczynski, K. Wcislo, React. Kinet. Catal. Lett. 571 (1996) 127. [26] M. Hoang, J.F. Mathews, K.C. Pratt, React. Kinet. Catal. Lett. 61 (1) (1997) 21. [27] S.M. Al-Zahrani, B.Y. Jibril, A.E. Abaseed, Ind. Eng. Chem. Res. 39 (2000) 4070. [28] B. Grzybowska, J. Sloczynki, R. Grabowski, L. Keromnes, K. Wcislo, T. Bobinska, J. Catal. 209 (2001) 279. [29] A. Kytokivi, J.P. Jacobs, A. Hakuli, J. Merilainen, H.H. Brongersma, J. Catal. 162 (1996) 190. [30] V. Amte, M.Tech. Thesis, IIT Kanpur, 1998. [31] S. De Rossi, M.P. Casaletto, G. Ferraris, A. Cimino, G. Minelli, Appl. Catal. A: Gen. 167 (1998) 257. 33 [32] B.D. Raju, K.S.R. Rao, G.S. Selvapathi, P.S.S. Prasad, P.K. Rao, Appl. Catal. A: Gen. 209 (2001) 335. [33] F.P.J.M. Kerkhof, J.A. Moulijn, R. Thomas, J. Catal. 56 (1979) 279. [34] O.F. Gorriz, V.C. Corberan, J.L.G. Fierro, Ind. Eng. Chem. Res. 31 (1992) 2670. [35] Ch. Fountzoula, H.K. Matralis, Ch. Papadopoulou, G.A. Voyiatzis, Ch. Kordulis, J. Catal. 172 (1997) 391. [36] B.M. Weckhuysen, I.E. Wachs, R.A. Schoonheydt, Chem. Rev. 96 (1996) 3327. [37] F. Cavani, M. Koutyrev, F. Trifiro, A. Bartolini, D. Ghisletti, R. Lezzi, A. Santucci, G. Del Piero, J. Catal. 158 (1996) 236. [38] T.C. Watling, G. Deo, K. Seshan, I.E. Wachs, J.A. Lercher, Catal. Today 28 (1996) 136. [39] D. Patel, P.J. Anderson, H.H. Kung, J. Catal. 125 (1990) 132. [40] M. Cherian, M.S. Rao, G. Deo, in: Proceedings of the CHEMINAR 2001, Vol. 159, Bhubaneswar, India. [41] B.M. Weckhuysen, R.A. Schoonheydt, Catal. Today 51 (1999) 223. [42] B.M. Weckhuysen, A.A. Verberchmoes, A.L. Buttiens, R.A. Schoonheydt, J. Phys. Chem. 98 (1994) 579. [43] S. Khaddar-Zine, A. Ghorbel, C. Naccache, J. Mol. Catal. A: Chem. 150 (1999) 223. [44] M.A. Vuurman, I.E. Wachs, J. Phys. Chem. 96 (1992) 5008. [45] M. Cherian, G. Deo, unpublished results. [46] B.M. Weckhuysen, I.E. Wachs, J. Phys. Chem. 100 (1996) 14437. [47] T. Blasco, J.M. Lopez Nieto, Appl. Catal. A: Gen. 157 (1997) 117. [48] E.A. Mamedov, V.C. Corberan, Appl. Catal. A: Gen. 27 (1995) 1. [49] M.A. Banares, M.V. Martinez-Huerta, X. Gao, J.L.G. Fierro, I.E. Wachs, Catal. Today 61 (2000) 295. [50] D. Creaser, B. Andersson, Appl. Catal. A: Gen. 141 (1996) 131. [51] K. Chen, A.T. Bell, E. Iglesia, J. Phys. Chem. B 104 (2000) 1292. [52] J.M. Smith, Chemical Engineering Kinetics, 3rd Edition, McGraw-Hill, New York, 1981. [53] G. Deo, I.E. Wachs, J. Haber, Crit. Rev. Surf. Chem. 4 (1994) 1.
© Copyright 2026 Paperzz