Al2O3 Dehydrogenation Catalysts

Article pubs.acs.org/IECR

Deactivation Studies of the CrOx/Al2O3 Dehydrogenation Catalysts under Cyclic Redox Conditions Vladimir Z. Fridman* and Rong Xing Clariant Corporation, Louisville, Kentucky 40209, United States

ABSTRACT: In this work, the deactivation of a CrOx/Al2O3 dehydrogenation catalyst in an industrial reactor has been systematically studied under cyclic dehydrogenation−regeneration conditions from start up to end of run. The results showed that the CrOx/Al2O3 catalyst deactivation occurred through three major processes. The first process is related to the complex transformation of surface chromium species, the second process refers to the phase transformation and/or sintering of support alumina, and the third process is the interfacial CrOx/Al2O3 transformation into the solid solution of α-(Al, Cr)2O3. In addition, the reduction of the chromium oxide surface area and the migration of Cr3+ into the alumina support resulted in the increase of the surface Cr6+/Cr3+ ratio at the end of the run which increased the specific catalyst acidity and reduced the catalyst selectivity. The main causes of irreversible deactivation of the CrOx/Al2O3 catalyst are the decrease of chromium oxide surface area, the formation of α-(Al, Cr)2O3 in the support, and the increase of surface Cr6+/Cr3+ ratio. fixed-bed CrOx/Al2O3 catalyst nowadays survives from 60 000 to 80 000 redox cycles, equivalent of about four years of continuous operation. In addition, the catalyst bed temperature swings from 560 to 700 °C or higher during the run, which adversely induces the changes of the catalyst properties, causing the decrease of the catalyst activity. In a commercial operation, the loss of the catalyst intrinsic activity is usually compensated by increasing the heat input to the reactor, as a way of maintaining a desired paraffin conversion level. During most run time, the olefin selectivity is not affected by the temperature increase. However, at the end of the run the catalyst selectivity is significantly reduced to a level that makes the operation uneconomical, in which case the catalyst needs to be replaced. The CrOx/Al2O3 catalyst has been the subject of research and development in the field of paraffin dehydrogenation for over 100 years. Previous studies have reported that the potential mechanisms for the deactivation of the CrOx/Al2O3 catalyst include the coke formation;6−13 the irreversible poisoning of the catalyst; the chemical and structural changes of the catalyst, such as CrOx agglomeration; the nature, phase, and structure changes of active sites; and the degradation of the support; etc.14−17,27 The deactivation of the CrOx/Al2O3 catalyst within one dehydrogenation cycle has been mainly attributed to the blockage of active sites by carbonaceous

1. INTRODUCTION The CrOx/Al2O3 catalyst is currently employed in two commercial paraffin dehydrogenation technologies, including the Catofin process1,2 and the fluid bed dehydrogenation process.3,4 In both processes, the catalyst is operated in a cyclic dehydrogenation−regeneration (redox) mode with very short dehydrogenation (7−15 min) and oxidative regeneration (7− 15 min) cycles. The use of short dehydrogenation time in these technologies is to maintain the heat balance of an adiabatic reactor where the heat required for the endothermic dehydrogenation reactions is provided by burning coke and by hot air during regeneration. There is typically no lessening in the catalyst activity after the 10−15 min of dehydrogenation; however, in the case of the Catofin process, the catalyst bed temperature decreases to a level at which the paraffin conversion is too low to maintain an economical operation. In that case, the reactor is switched from the dehydrogenation step to the regeneration one. In the latter step, the catalyst bed temperature is reheated to the desired values because of the exothermic reactions of coke burning and the very hot air coming into the reactor. It is noteworthy that the catalyst bed is further heated during the catalyst reduction which occurs at the start of each redox cycle. It is expected that the mode of cyclic redox operation creates tremendous stress on the CrOx/Al2O3 catalyst, leading to the catalyst deactivation with time on stream. As a result, the lifetime of the first commercial CrOx/ Al2O3 catalysts was only 3−6 months.5 Depending on the severity of reactor operation, the type of paraffin feed, the catalyst formulation, and the catalyst location in a reactor, the © 2017 American Chemical Society

Received: Revised: Accepted: Published: 7937

April 18, 2017 June 19, 2017 June 22, 2017 June 22, 2017 DOI: 10.1021/acs.iecr.7b01638 Ind. Eng. Chem. Res. 2017, 56, 7937−7947

Article

Industrial & Engineering Chemistry Research species deposited on the catalyst surface.9,10,13 Weckhuysen et al. gained insight on the formation and burning of coke deposits on an industrial CrOx/Al2O3 catalyst under reaction conditions using operando ultraviolet−visible (UV−vis) spectroscopy. It was found that the formation of coke during the dehydrogenation step is faster at the top of the reactor as compared to the bottom where coke deposition is a gradual process.8 Airaksinen et al. found that the coke formation was related to the amount of redox Cr3+ on the CrOx/Al2O3 catalyst.10 Sokolov et al. reported that the deactivation rate was determined by both the amount of surface carbon species and their nature.13 However, the coking is not any more responsible for the loss of activity from cycle to cycle for the CrOx/Al2O3 catalysts used in the Catofin and fluid bed dehydrogenation processes, because the coke formed on the catalyst surface is removed under the regeneration step. Besides the coking, the irreversible catalyst poisoning will not be considered as well in this work, because the concentration of halogenates in the feedstock, which can permanently deactivate the catalyst, is usually very low in Catofin and fluid bed dehydrogenation processes. This type of CrOx/Al2O3 catalyst deactivation is not typically observed in the field. To date there have been many articles in the literature discussing catalytic behavior of the CrOx/Al2O3 catalyst; however, only a small number of studies focused on the catalyst deactivation especially induced by the changes of the catalyst properties.9−31 It is well-accepted that the surface chromium species on the CrOx/Al2O3 catalyst exist in two oxidation states as Cr3+ and Cr6+, and the paraffin dehydrogenation activity is mainly associated with the redox and nonredox Cr3+ active sites.11,24,29 According to the literature, the CrOx/ Al2O3 catalyst deactivation in paraffin dehydrogenation has been attributed to the following factors: (1) the loss of Cr3+ active sites by either forming inactive alumina-incorporated Cr3+15 or being converted to inaccessible Cr3+ species,24 (2) the formation of solid solution of α-(Cr, Al)2O3,16,17,19,26 (3) the degradation of alumina phase,20,26 (4) the change of catalyst porous texture,18,26 and (5) the aggregation of Cr2O3.18,19 For example, in the early 1970s, Bremer et al. found that the irreversible deactivation of the CrOx/Al2O3 catalyst within one cycle of reduction and oxidation was due to the formation of a solid solution of α-(Cr, Al)2O3 in the catalyst at a reaction temperature over 650 °C and was independent of the coke deposition.17 Gitis et al. reported that the reduction of the CrOx-Al2O3-K2O catalysts at 700 °C led to two types of deactivation: reversible and irreversible. The former was associated with the decrease of chromium oxide surface area, and the latter was with irreversible changes of the catalyst porous texture as a whole.21 In addition, Sterligov et al. studied the catalytic behavior of the CrOx/Al2O3 catalyst with the additions of potassium, lithium, or niobium oxides in n-butane dehydrogenation at 550 °C for about 1000 redox cycles.19 Some interesting observations were made, including the significant decrease of Cr6+ concentration in the catalyst, the decrease of the Cr2p3/2/Al2s atomic ratio obtained by XPS, little reduction of the total catalyst surface area, and steady reduction of the catalyst activity. On the basis of these results, the authors attributed the catalyst deactivation to the aggregation of the Cr2O3 clusters which caused the reduction of the number of active sites on the catalyst surface. However, 1000 cycles of the catalyst operation in a redox mode is equivalent to only 15 days on stream in the commercial reactor, and the knowledge of the CrOx/Al2O3 catalyst lifetime deactivation is still lacking. Davydov et al. studied two commercial fluid-bed CrOx/Al2O3

catalysts (containing 18% Cr2O3) operated for one or two years.20 It was found that the concentration of α-Al2O3 in the catalyst increased from zero in the fresh catalyst to 5.6% in the one-year aged sample and to 12.5% in the two-year aged sample, while the α-Cr2O3 concentration decreased from 7.9% to 2.0% to 0.9% accordingly. There was no change of crystallite size of α-Cr2O3 observed during the aging. More recently, Weckhuysen and Puurunen investigated the deactivation mechanism of the model CrOx/Al2O3 catalyst (4% Cr) thermally treated from 600 to 1200 °C using multiple spectroscopic techniques including diffuse reflectance spectroscopy (DRS), X-ray photoelectron spectroscopy (XPS), and electron spin resonance spectroscopy (ESRS). They found that the formation of catalytically inactive alumina-incorporated Cr3+ species was responsible for the catalyst irreversible deactivation. The alumina-incorporated Cr3+ species were formed by (1) entrapment of Cr3+ ions inside the alumina support during sintering of the alumina support (e.g., collapse of the small pores) and (2) migration of Cr3+ ions into the alumina support.15 Although interesting observations were reported in their work, it is not clear whether these conclusions can be applied to the industrial CrOx/Al2O3 catalysts for two reasons: (1) The Cr loading in their work was only 4%, which is much lower than that in a commercial catalyst (∼13.2% and above). (2) The catalyst aging in their work was performed by an oxidative treatment of the catalyst in a temperature range of 900−1200 °C for 24−96 h. The aging temperature and time are quite different as the industrial aging of the catalyst which is operated under cyclic oxidizing and reducing atmospheres at a much lower temperature of 550−750 °C for much longer time of over two years. We believe all these differences would affect the CrOx/Al2O3 catalyst deactivation mechanism. In our previous work, we identified the surface chromium species of an industrial-like CrOx/Al2O3 catalyst and evaluated its relative activity and short-term stability in isobutane dehydrogenation.28,29 In the present study, the deactivation of the CrOx/Al2O3 catalyst in an industrial dehydrogenation reactor was systematically studied from start up to end of run. Using combined analytic techniques including chemical analysis and spectroscopic and electroscopic analysis, the catalyst characteristics at different aging levels were investigated to understand the mechanistic origins governing the catalyst stability under cyclic dehydrogenation−regeneration conditions and to develop a model for the CrOx/Al2O3 catalyst deactivation. It is expected that the knowledge obtained from this work is useful for further improvement of the catalyst activity, selectivity, and lifetime.

2. EXPERIMENTAL SECTION 2.1. Catalyst Preparation. The model catalyst of fresh AlCr-13.2 with Cr loading of 13.2%, equivalent of 19% Cr2O3, was prepared by the impregnation of alumina carrier with the CrO3 aqueous solution in the presence of small amounts of alkali metal ions, followed by the calcination at 760 °C for 4 h using a 15% O2/85% N2 gas stream. The detailed physical− chemical properties of the fresh Al-Cr-13.2 have been reported previously.28,29 It is noteworthy that the catalyst studied in this work is a replica of an old-version Catofin catalyst that was made obsolete and substituted by a much more stable one about 10 years ago.2 The study of this type of catalyst allows us to observe relatively quick changes of the catalyst physical− chemical properties within about two years of operation. To obtain aged catalysts, a mixture of fresh Al-Cr-13.2 catalyst and 7938

DOI: 10.1021/acs.iecr.7b01638 Ind. Eng. Chem. Res. 2017, 56, 7937−7947

Article

Industrial & Engineering Chemistry Research alumina chips was first loaded into a metal basket, then the basket was placed at the top layer of a commercial Catofin reactor and aged for different times, i.e., 55, 180, 350, and 645 days on stream (DOS). The resulting catalysts were denoted as the 55DOS-Al-Cr-13.2, 180DOS-Al-Cr-13.2, 350DOS-Al-Cr13.2, and 645DOS-Al-Cr-13.2, respectively. It should be noted that the catalyst aging behavior obtained in this work represents only that of the top layer catalysts in the commercial reactor. The top-layer catalysts account for only a small portion of the total catalysts in the reactor and do not represent the properties of the whole catalyst bed. The top-layer catalysts are typically operated under harsher conditions than the middle- and bottom-layer catalysts. The 5DOS-Al-Cr-13.2 sample was made by aging the fresh Al-Cr-13.2 in a pilot plant reactor under cyclic redox conditions at 600 °C for 300 cycles, which is equivalent to about 5 DOS operation in a commercial reactor. 2.2. Dehydrogenation Activity Test. The dehydrogenation activity tests were conducted in a testing unit that simulates the commercial cyclic dehydrogenation−regeneration process and is operated in successive cycles of reduction, dehydrogenation, and regeneration. The detailed testing procedure was described previously.32 A plug reactor with an outer diameter of 25.4 mm was used. To minimize the internal diffusion limitation, all samples with an average particle size of 0.25−0.50 mm were used. The activity of the fresh and aged catalysts was tested in isobutane dehydrogenation at a liquid hourly space velocity (LHSV) of 2 h−1 and a temperature range of 540−600 °C. Each sample was tested three times, and the average values of conversion, selectivity, and yield were used in this paper. A full testing cycle includes the following steps: first evacuation, catalyst reduction by H2, second evacuation, dehydrogenation (9 min), third evacuation, and regeneration (9 min). The reaction products were analyzed online using an Agilent gas chromatograph equipped with thermal conductivity and flame ionization detectors. 2.3. Catalyst Characterizations. In this work, the catalysts were extensively characterized using the following analytic techniques: N2 adsorption, O2 chemisorption, temperatureprogrammed desorption of NH3 (TPD-NH3), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), and chemical analysis. It should be noted that because of the limitations of in situ sample analysis, all samples were analyzed by an ex situ approach in this work, which may not truly represent the changes of an active catalyst under the real reaction conditions. However, it is known that in situ results for the CrOx/Al2O3 catalyst generally support those received by ex situ methods, indicating that our method is still very useful for the investigation of this catalyst. We believe that the information obtained from this work helps to elucidate the long-term deactivation mechanism of the catalyst. 2.3.1. Concentration of Cr2+, Cr3+, Cr6+, and the Total Cr. To determine the total Cr and Cr6+ concentration, a titration method was used where the sample was digested with a mixture of sulfuric and phosphoric acid as reported previously.33 The difference between the total Cr concentration and the total Cr6+ concentration in the catalyst was used to determine the Cr3+ concentration. An Autochem (II) apparatus (Micromeritics) equipped with a mass spectrometer was applied to determine the “potential” amount of Cr2+ that can be formed in the catalyst using the method reported previously.28 2.3.2. N2 Adsorption. The Brunauer−Emmett−Teller (BET) surface area of the catalysts was measured by nitrogen

adsorption at 77 K with an Autochem II automated adsorption instrument (Micromeritics). The surface areas were determined from the adsorption values at five relative pressure values (P/ P0) ranging from 0.05 to 0.2 using the BET method. 2.3.3. O2 Chemisorption. The O2 chemisorption experiments were conducted by the pulse technique using an AutoChem (II) apparatus (Micromeritics) at −78 °C. The obtained data were used to calculate the surface area and the average size of chromium oxide particles using the previous approach.34 2.3.4. Surface Acidity. The surface acidity was measured by the temperature-programmed desorption of NH3 using an Autochem (II) apparatus coupled with a Thermo-Star Quadrupole Mass Spectrometer (Micromeritics). Before the measurement, each sample was reduced by H2 at 600 °C for 1 h, followed by cooling the sample to 25 °C and introducing a flow of NH3 for adsorption. Then, the sample was purged by Helium for 30 min to remove the physically adsorbed NH3. The total catalyst surface acidity was defined as the total molar amount of adsorbed NH3 per gram catalyst, while the specific surface acidity was defined as the adsorbed molar amount of NH3 per catalyst surface area. 2.3.5. Powder X-ray Diffraction. The phase composition of the catalysts was studied by XRD. Diffractograms were measured by a Siemens D5000 X-ray diffractometer, and the phases were identified with database of Siemens Diffrac AT program (@ 1993). Diffractometer was calibrated to determine the α-Cr2O3 concentration and the solid solution α-(Cr, Al)2O3 concentration if present. The α-Cr2O3 concentration was calculated based on the intensity of the bands at d spaces of 1.67 and 2.18, and the α-(Cr, Al)2O3 concentration was calculated based on the intensity of the bands at d spaces of 1.61, 1.74, 2.1, and 2.56. 2.3.6. X-ray Photoelectron Spectroscopy. The XPS measurements were performed using a VG Thermo MultiLab 3000 surface analysis system inside a multichamber ultrahigh vacuum (UHV) surface science facility (VG Scientific/RHK Technology) equipped with a 150 mm radius CLAM 4 hemispherical analyzer. The base chamber pressure was ∼3 × 10−9 Torr. All the measurements were performed using Mg Ka X-rays with characteristic energy of 1253.6 eV. High-resolution Al2s, O1s, and Cr2p3/2 spectra were collected and analyzed for each sample. Elemental quantification was performed using XPS MultiQuant software. The background curves were calculated using Shirley model, and the cross sections were calculated using the Scofield model. Atomic sensitivity factors (ASF) were used to account for the sensitivity difference between these lines. 2.3.7. Transmission Electron Microscopy. TEM was performed using a field-emission gun TECNAI F20 microscope operated at 200 kV. The bulk compositions of chromium oxide particles were identified by scanning transmission electron microscopy (STEM) and energy dispersive X-ray (EDX) mapping analysis using a beam nanoprobe with a diameter of ∼1 nm. Typically, each sample was prepared by scrapping the catalyst surface using a clean razor blade and crushing it between two glass plates. The resulting fine powder was then dispersed on Lacey carbon film supported by a 300 mesh copper grid.

3. RESULTS AND DISCUSSION As described in the Experimental Section, all the samples were aged at the top layer of the catalyst bed in the commercial 7939


Article

Industrial & Engineering Chemistry Research

Figure 1. Normalized isobutane conversion and isobutylene selectivity of the fresh and aged CrOx/Al2O3 catalyst tested at 540 °C and LHSV of 2 h−1 in isobutane dehydrogenation reactions.

reactor except for the fresh and 5DOS-Al-Cr-13.2 samples. Unfortunately, the exact operation temperatures for the samples taken from the commercial reactor were not known, because there was no thermocouple installed at the top layer in the commercial reactor. Also, the catalyst bed temperature varies with time on stream. For example, the inlet temperature of hot air entering the reactor is initially about 650 °C and is usually increased to a maximum of 680 °C after a couple of months of operation. After several months of operation, the plant starts to increase the air temperature by burning extra fuel gas to heat inlet air, and the amount of fuel gas is usually steadily increased to provide sufficiently high air temperature to keep paraffin conversion constant during the run. Based on a heat balance calculation, the inlet temperature of hot air is typically in the range of 630−700 °C. Considering the top layer temperature is 20−30 °C lower than the inlet temperature of hot air, the top layer temperature is thus estimated to be in the range of 610−670 °C. 3.1. Isobutane Dehydrogenation. Figure 1 compares the normalized activity for isobutane conversion and the normalized isobutylene selectivity of the series of CrOx/ Al2O3 catalysts tested at 540 °C in isobutane dehydrogenation reactions. It is seen that after a 5 day operation of the fresh catalyst the isobutane conversion was slightly reduced, and the isobutylene selectivity was slightly increased. It should be noted that the decrease of the initial activity is not considered as a catalyst deactivation, but a catalyst surface stabilization in which the chromium species originally stable only on the air-calcined catalyst surface experienced restructuring to form stable chromium species under redox conditions. After the surface stabilization, the resulting catalyst is called an equilibrated catalyst.28,29 After a 55 day operation, the resulting catalyst of 55DOS-AlCr-13.2 had little change of the catalytic performance as compared to the 5DOS-Al-Cr-13.2. This confirms that the 5DOS-Al-Cr-13.2 reached stable performance after the initial operation. At 180 DOS, the resulting catalyst showed ∼26% lower isobutane conversion as compared to the fresh one; at 350 DOS and 645 DOS, the aged catalysts showed ∼37.3% and ∼83.6% lower isobutane conversion than the fresh one, respectively. In terms of the catalyst selectivity, the catalysts showed almost constant isobutylene selectivity from the start of the run to 350 DOS of operation. At 645 DOS, the resulting catalyst showed ∼16.4% lower isobutylene selectivity as compared to other samples. 3.2. Nitrogen Adsorption, O2 Chemisorption, and Pore Volume Measurements. Figure 2 shows the changes of the normalized BET surface area and the normalized chromium

Figure 2. Changes of the normalized BET surface area and chromium oxide surface of the CrOx/Al2O3 catalysts as a function of time on stream.

oxide surface area of the series of CrOx/Al2O3 catalysts as a function of time on stream. After the first 5 days of initial operation, the BET surface area of the 5DOS-Al-Cr-13.2 decreased by only ∼5.6%, while the chromium oxide surface area decreased dramatically by 45.4%. This supports the previous conclusion that significant restructuring of surface chromium species in the fresh sample occurred during the initial operation.28,29 According to them, the restructuring of surface chromium species involved the complete loss of “isolated” mono-/poly aluminum chromate(s) species, the transformation from Cr6+ to Cr3+, the agglomeration of Cr2O3 clusters, and the migration of crystalline α-Cr2O3 from initially being covered by amorphous Cr2O3 to being exposed on the catalyst surface. At 55DOS, the catalyst showed slight decreases of the BET surface area by ∼2.4% (relative) and the chromium oxide surface area by ∼2.5% (relative) as compared to the 5DOS-Al-Cr-13.2. This shows that the CrOx/Al2O3 catalyst reached a relatively stable state within 5 days of operation, which is in agreement with the measured activity and selectivity (Figure 1). At 180 DOS, the measured BET surface area and chromium oxide surface area of the catalyst dropped by 18.3% (relative) and 53.8% (relative) as compared to the 55DOS-AlCr-13.2 catalyst, respectively. At 645 DOS, the catalyst lost 80.9% of the total surface area and 93.3% of the chromium oxide surface area as compared to the fresh one, showing that the Al-Cr-13.2 was already severely aged with significant changes of the pore texture and active sites of the catalyst after 7940


Article


Figure 3. Normalized pore volume fraction of relative small pores in the range of 50−150 Å changes with time on stream.

about two years of operation. It was found that the decrease of the catalyst total surface area and chromium oxide surface area was in part caused by the change of catalyst pore structure. Figure 3 compares the normalized volume fraction of pores in the range of 50−150 Å among the samples. It is seen that there was no change of small pore volume fraction during the initial operation. Compared to the fresh sample, the number of small pores decreased by 2.0% (which is close to the error of the measurement), 15%, 40%, and 95% at the time on stream of 55, 180, 350, and 645 DOS, respectively. The loss of small pores was likely caused by the catalyst carrier sintering due to the high temperature cyclic redox operation, which explains the loss of BET surface area during the same period. It is likely that the loss of chromium oxide surface after 55 DOS and beyond was caused by the collapse of small pores in the catalyst that entrapped some chromium oxide particles into alumina.15 In addition to the entrapment of chromium oxide particles into alumina, the formation of solid solution of α-(Al, Cr)2O3 was also attributed to the loss of both catalyst total surface area and chromium oxide surface area in this period (Figure 2). It is likely that the entrapped chromium oxide particles participated in the formation of solid solution during the phase conversion of transitional alumina to α-Al2O3. 3.3. Average Size of Chromium Oxide Particles. Figure 4 shows the changes of the calculated average chromium oxide particle size as a function of time. It is seen that the average chromium oxide particle size increased from 180 to 375 Å after 5 days on stream. The increase of chromium oxide particle size can be explained by the agglomeration of Cr2O3 clusters on the catalyst surface under cyclic redox conditions.28 The average size of chromium oxide particles on the catalyst kept almost constant in the period of 5−55 DOS. After 55 DOS, the size of chromium oxide particles showed gradual increase with time. For example, the calculated average particle size of chromium oxide increased from 180 to 1680 to 3200 Å as the time increased from 180 to 350 to 645 DOS. Two representative elemental Cr mapping images in Figure 4 reveal the particle size and morphological difference between the fresh Al-Cr-13.2 and the 645DOS-Al-Cr-13.2. It is seen that the fresh Al-Cr-13.2 catalyst contains relatively small chromium oxide clusters with a size of a few tens of nanometers which are well dispersed on the catalyst surface, while the 645DOS-Al-Cr-13.2 comprises large chromium oxide particles with a size over 500 nm. The combined O2 chemisorption measurement and TEM/elemen-

Figure 4. Changes of the calculated average chromium oxide particle size as a function of time on steam. The inset pictures show elemental Cr mapping images of the fresh Al-Cr-13.2 and the 645DOS-Al-Cr13.2.

tal mapping thus demonstrated the growth of the chromium oxide particles with time on stream. 3.4. XRD Analysis. The XRD method was used to determine the concentration of crystalline α-Cr2O3 and the solid solution of α-(Al-Cr)2O3 in the catalysts. Figure 5 shows the changes of measured α-Cr2O3 and α-(Al-Cr)2O3 concentration in the CrOx/Al2O3 catalysts as a function of time on stream. The α-(Al, Cr)2O3 is typically formed at high temperature by the replacement of Al3+ with Cr3+ in α-Al2O3, because α-Al2O3 and α-Cr2O3 are sesquioxides having the same corundum crystal structure (approximately hexagonal closepacked oxide ions with the Al3+ and Cr3+ ions occupying twothirds of the available octahedral interstitial sites).35 As can be seen from Figure 5, the fresh Al-Cr-13.2 contained ∼7.8 wt % α-Cr2O3, indicating that ∼40% chromium oxide is present in the form of an amorphous state. It should be noted that in addition to α-Cr2O3 and (Al, Cr)2O3, the Cr3+ species can be present in several forms. Using combined diffuse reflectance spectroscopy (DRS) and electron spin resonance spectroscopy (ESR) techniques, Weckhuysen and Puurunen reported three types of Cr3+ species including Cr3+ exposed to the surface (“isolated” Cr3+ and Cr3+ clusters), Cr3+ entrapped into alumina support, and Cr3+ buried inside chromia crystallite.15 To make 7941


Article


Figure 6. Correlation of the solid solution α-(Al, Cr) 2 O 3 concentration with the percentage of small pore loss in the catalyst.

Figure 5. Concentration of α-Cr2O3 and solid solution of α-(Al, Cr)2O3 present in the CrOx/Al2O3 catalysts varies with time on stream. The inset pictures are the photographs of fresh Al-Cr-13.2 and 645DOS-Al-Cr-13.2 catalysts.

the catalyst. As discussed above, the loss of small pores was due to the pore collapse in the course of the carrier sintering. It is possible that some chromium oxide particles were entrapped inside the small pores during the pore collapse. Thus, the strong correlation between the α-(Al, Cr)2O3 concentration and the small pore loss may suggest that the entrapment of chromium oxide particles inside the small pores promoted the formation of α-(Al, Cr)2O3. 3.5. Total Cr6+ Concentration and TPD_NH3. Figure 7 shows the changes of measured total Cr6+ concentration and

it simple, all noncrystalline Cr3+ species were denoted as amorphous Cr2O3 in this work. The initial 5 day operation caused the decrease of α-Cr2O3 concentration to 6.7 wt %. After that, the α-Cr2O3 concentration in the catalyst showed fast decrease with time. For instance, the concentration of αCr2O3 was reduced to 4.2 wt % at 55 DOS; at 180 DOS, the concentration of α-Cr2O3 phase was further reduced to 2.7%; at 350 DOS, there was no α-Cr2O3 detected by XRD. In contrast to α-Cr2O3, there was no solid solution of α-(Al, Cr)2O3 detected from the start of the run to 55 DOS. After that, the solid solution concentration showed fast increase with time. For example, the measured α-(Al, Cr)2O3 concentration increased from 12.3% to 34% to ∼73% as the time increased from 180 to 350 to 645 DOS, respectively. In the period of 55−180 DOS, the formation of α-(Al, Cr)2O3 was likely associated with the solid phase reaction between α-Cr2O3 and α-Al2O3, which resulted in some loss of α-Cr2O3 in dehydrogenation reactions. Also, the formation of α-(Al, Cr)2O3 suggests that some transitional alumina started to degrade to form α-Al2O3.13 Because α-Cr2O3 is the most active species among all surface Cr species,29 the loss of α-Cr2O3 by 35.7% (relative) caused the overall activity for isobutane conversion to decrease by 16.1% (relative). From 180 to 350 DOS, the α-Cr2O3 concentration in the catalyst decreased from 2.7% to zero, while the α-(Al, Cr)2O3 concentration showed an almost three times increase from 12.3% to 34%. From 350 to 645 DOS, the α-(Al, Cr)2O3 concentration in the catalyst increased double from 350 to 645 DOS in the absence of α-Cr2O3. All these results suggest that amorphous Cr2O3 also participated in the formation of solid solution of α-(Al, Cr)2O3. Two photographs in Figure 5 reveal the difference in catalyst color between the fresh Al-Cr-13.2 and the 645DOS-Al-Cr-13.2. The fresh catalyst appears greenish because it has ∼19 wt % of Cr2O3, while the sample aged 645 days shows a bright pink color because it has a high concentration of α-(Al, Cr)2O3. To gain more information on the solid solution formation, a plot between the α-(Al, Cr)2O3 concentration and the percentage of small pore loss was made as shown in Figure 6. It is seen that the formation of α-(Al, Cr)2O3 correlates well with the percentage of small pore loss in

Figure 7. Changes of the total Cr6+ concentration and the total surface acidity of the CrOx/Al2O3 catalysts as a function of time on stream.

total surface acidity of the CrOx/Al2O3 catalyst with time. It should be noted that according to our in situ XPS study and the literature,7,36−39 the Cr6+ or Cr5+ is present only on the catalyst surface of the CrOx/Al2O3 catalyst after the oxidation step, not that after the catalyst reduction and dehydrogenation because most of the Cr6+ or Cr5+ is reduced to Cr3+. Thus, the Cr6+ species mentioned in our paper refers to the precursors of the active Cr3+ species formed after the reduction or during the dehydrogenation. It is seen that the total Cr6+ concentration in the catalyst decreased significantly from an initial 1.1% to 0.6% 7942


Article

Industrial & Engineering Chemistry Research down by 45.5% (relative) after the first 5 days of operation. At the same time, the total catalyst surface acidity decreased from 2.76 to 2.2 mmol NH3/g catalyst, down by 27.5% (relative). This correlation suggests that the Cr6+ species is responsible for the formation of certain types of acidic sites on the catalyst surface which are active in the side reactions of hydrocracking and coking.29 Thus, the improvement of the catalyst selectivity after the initial operation can be attributed to the reduction of the total Cr6+ concentration and the total catalyst acidity (Figure 1). In the period of 5−55 DOS, the total Cr6+ concentration decreased from 0.6% to 0.5%, while the total surface acidity remained almost constant. Further aging of the catalyst caused simultaneous decrease of the total Cr6+ concentration and the total surface acidity, but their decline rates were much lower than those during the initial 5 day operation. At 645 DOS, the catalyst contained only 0.19% of Cr6+, which is about 6 times lower than that in the fresh catalyst. The total surface acidity was reduced to 1.17 mmol NH3/g catalyst, most of which was likely attributed to the conversion of transitional alumina to α-Al2O3 and the formation the solid solution of α-(Al, Cr)2O3. In spite of the reduction of the total catalyst acidity, the 645DOS-Al-Cr-13.2 catalyst was not able to retain as high an isobutylene selectivity as other samples at the targeted isobutane conversion (40− 45%) because of a significant loss of chromium oxide surface area (Figure 2). In practice, the targeted conversion was achieved by further increasing the operation temperature in a commercial reactor, which further decreased the catalyst selectivity because an increase in the reaction temperature kinetically favors side reactions like hydrocracking and coking reactions.40 In addition, the reduction in isobutylene selectivity at the end of the catalyst life was also related to the changes of surface Cr6+ and Cr3+ distribution, which will be discussed later. Figure 8 compares the measured total surface acidity with the calculated specific surface acidity of the CrOx/Al2O3 catalysts at

catalyst acidity, some additional changes in the distribution of the active sites for dehydrogenation and the active sites for hydrocracking and coking may occur. 3.6. XPS, Cr6+, and Cr3+ Chemical Analysis. Figure 9 shows the examples of Cr2p XPS spectra of the 5DOS-Al-Cr-

Figure 9. Representative Cr 2p XPS spectra of 5DOS-Al-Cr-13.2, 180DOS-Al-Cr-13.2, and 645DOS-Al-Cr-13.2 samples.

13.2, 180DOS-Al-Cr-13.2, and 645DOS-Al-Cr-13.2 samples. All samples show two peaks centered at 576.4 and 585.8 eV, respectively. The two peaks can be attributed to the Cr2p1/2 and Cr2p3/2 levels of the Cr3+ species.41 The deconvolution of the two peaks revealed an additional two small peaks centered at 580.4 and 589.7 eV, respectively, which were attributed to the Cr6+ species on the CrOx/Al2O3 catalyst surface.17,28 It should be noted that there was no Cr2+ or Cr5+ detected by XPS for all the samples. With increase of the time on stream, all Cr3+ and Cr6+ peaks shifted to slightly lower binding energy likely because of the changes of the Cr environment on the catalyst surfaces.27 It should be noted that XPS spectrum of the catalyst aged for 645 DOS had much higher relative intensity of the Cr6+ peak than those of the fresh and 180 DOS catalysts (Figure 9). Figure 10 shows the changes of the measured surface Cr6+/Cr3+ ratio and Cr/Al atomic ratio (determined by XPS) for the CrOx/Al2O3 catalysts at different aging levels. For direct comparison, the bulk Cr6+/Cr3+ ratios determined by chemical analysis are also given in the figure. As can be seen, from the start of run to 180 DOS all three ratios demonstrated similar behavior, showing gradual decrease with the time on stream. The simultaneous decrease of surface and bulk Cr6+/ Cr3+ ratios was caused by the significant loss of Cr6+ in the catalyst during that period (Figure 7), which is also in agreement with previous work.28 For the fresh catalyst, the measured surface Cr6+/Cr3+ ratio was about three times higher than the bulk Cr6+/Cr3+ ratio. This big difference, if true, suggests that either some Cr3+ species migrated into alumina lattice after calcination or they were located at the inner layer of large size Cr2O3 clusters. It should be noted that some difference may be also introduced from using two different analytical techniques. During the initial operation, it was

Figure 8. Variations of the total surface acidity and specific surface acidity of the CrOx/Al2O3 catalysts with the time on stream.

different aging levels. The specific surface acidity was defined as the molar amount of adsorbed NH3 per total catalyst surface area. To our interest, the change of calculated specific surface acidity was opposite to that of the total surface acidity as the time increased. In contrast to the total surface acidity, the specific surface acidity of the catalyst increased with the time and reached the maximum value at the end of run. This observation indicates that besides the reduction of the total 7943


Article


Cr6+ concentration due to a relatively high operation temperature used at the later stage of operation. Among all the samples, the 645DOS-Al-Cr-13.2 had the largest Cr6+/Cr3+ ratio, which explains its lowest catalyst selectivity because the Cr6+ species favored side reactions leading to coke and light gases.29 After 180 DOS, the surface Cr/Al ratio also showed gradual increase with time, which can be explained by (1) the CrOx migration on the catalyst surface leading to the formation of large clusters and/or (2) the replacement of Al3+ by Cr3+ due to the formation of solid solution of α-(Al, Cr)2O3. 3.7. CrOx/Al2O3 Catalyst Deactivation Model. To develop a CrOx/Al2O3 catalyst deactivation model, the results obtained in our previous work and the current study will be applied. Our previous work has elucidated the evolution of Cr species on the Al-Cr-13.2 catalyst surface during the initial operation and provided information on the short-term stability of the chromium species of the CrOx/Al2O3 catalysts under cyclic dehydrogenation−regeneration conditions.28,29 It was reported that a total of five types of surface Cr species are present in the fresh CrOx/Al2O3 catalyst, including the “isolated” redox mono/poly-aluminum chromate(s), redox chromium chromate, redox aluminum-chromium-chromate(s), nonredox Cr3+ on amorphous Cr2O3, and nonredox Cr3+ on crystalline α-Cr2O3 (Table 1). After the initial operation, four types of surface Cr species survived and stayed on the equilibrated catalyst surface except for the “isolated” mono/ poly-aluminum chromate(s) which was converted into nonredox Cr3+ ions or very small Cr2O3 clusters.28 Two parallel changes of the surface Cr species were observed during the initial operation. The first change is the conversion of the “isolated” mono/poly-aluminum chromate(s) into small Cr2O3 clusters during the reduction step. As a result of the change, the oxidation state of the Cr species changed from +6 (chromate) to +3 (Cr2O3), likely through the breakage and/or weakening of chemical bonds between the chromates with alumina.29 In the following regeneration step, the newly formed Cr2O3 clusters were reoxidized into the redox chromium-chromate (its amount is negligible), and the aluminum-chromiumchromate(s) coupled with Cr2O3 at the interface of Al2O3 and Cr2O3 clusters. Compared to the fresh catalyst, about 50% loss of initial Cr6+ in the 5DOS-Al-Cr-13.2 (Figure 7) can be attributed to the conversion of “isolated” mono/poly-aluminum chromates to Cr2O3. The second change is the coalescence of small Cr2O3 clusters on the alumina support to form large and stable agglomerates, which explains another 50% loss of initial chromium oxide surface area after the initial operation (Figure 2). The presence of steam, either formed during the reduction of Cr6+ or used during the steam purge (i.e., after dehydrogenation), might also help to mobilize the Cr2O3 clusters on the

Figure 10. Changes of the measured surface Cr/Al atomic ratio, surface Cr6+/Cr3+ atomic ratio (determined by XPS), and bulk Cr6+/ Cr3+ atomic ratio (determined by chemical analysis) in the CrOx/ Al2O3 catalysts as a function of time on stream.

observed that both the surface and bulk Cr6+/Cr3+ ratios decreased with time, which was likely related to the restructuring of CrOx species on the catalyst surface.29 During the stable operation period of 5−180 DOS, both the bulk and surface Cr6+/Cr3+ ratios showed slight decrease with time. For the Cr/Al ratio, its decease during the initial operation can be explained by (1) the decreased of CrOx surface area (Figure 2) and (2) the migration of Cr3+ into the alumina lattice.36 As discussed previously,28 this observation provides additional evidence that significant restructuring of surface CrOx species occurred during the cyclic redox mode of aging, and large size Cr2O3 clusters were formed by collision and agglomeration of small clusters. The Cr/Al ratio remained almost constant during the operation period of 5−180 DOS, indicating that the phase composition and cluster sizes of surface Cr species reached a relatively stable state after the initial operation. After 180 DOS, significant difference was observed between the surface and bulk Cr6+/Cr3+ ratios. The surface Cr6+/Cr3+ ratio showed a gradual increase from 0.19 to 0.41 to 0.45 as the time increased from 180 to 350 to 645 DOS, while the bulk Cr6+/ Cr3+ ratio kept almost constant at ∼0.3. It is evident that the increase of surface Cr6+/Cr3+ ratio was mainly attributed to the loss of Cr3+ from the catalyst surface due to its migration into the lattice of α-Al2O3 to form the solid solution of α-(Al, Cr)2O3 (Figure 5). Also, it is plausible that the increase of surface Cr6+/Cr3+ ratio was caused by the increase of surface

Table 1. Surface Chromium Species Present in the Series of CrOx/Al2O3 Catalyst Cr6+ fresh Al-Cr-13.2 5DOS-Al-Cr-13.2 and 55DOS-Al-Cr-13.2 180DOS-Al-Cr-13.2 360DOS-Al-Cr-13.2 and 645DOS-Al-Cr13.2

“isolated” mono/poly-aluminum chromate(s), chromium-chromate(s), aluminum-chromiumchromate(s) chromium-chromate(s), aluminum-chromiumchromate(s) chromium-chromate(s), aluminum-chromiumchromate(s) chromium-chromate, aluminum-chromiumchromate

Cr3+ nonredox Cr on amorphous Cr2O3, nonredox Cr3+ on α-Cr2O3 (small clusters with an average size of ∼180 Å) 3+

“isolated” nonredox Cr3+ ions or small amorphous Cr2O3 clusters, nonredox Cr3+ on amorphous Cr2O3, nonredox Cr3+ on α-Cr2O3 (large clusters with an average size of ∼375 Å) nonredox Cr3+ on α-(Al, Cr)2O3, nonredox Cr3+ on amorphous or defective Cr2O3, nonredox Cr3+ on α-Cr2O3 (large clusters with an average size of 790 Å) nonredox Cr3+ on α-(Al, Cr)2O3, nonredox Cr3+ on amorphous or defective Cr2O3 (large clusters with an average size >1700 Å)

7944


Article


Figure 11. Schematic representation of the evolution of surface Cr species and the alumina support of the CrOx/Al2O3 catalysts at different aging levels.

alumina support. In addition, it is possible that crystalline αCr2O3 was busied inside amorphous Cr2O3 clusters on the fresh Al-Cr-13.2 catalyst surface, which excluded its participation in the initial catalyst activity.28 The initial operation caused the migration of amorphous Cr2O3 clusters from the surface of αCr2O3, resulting in freeing crystalline α-Cr2O3 to be accessible to the dehydrogenation reactions.29 Regarding the catalytic activity change in this period, the small decrease of isobutane conversion (Figure 1) was not considered as the catalyst deactivation, but as a surface stabilization where unstable chromium species were converted into the thermodynamically stable ones under the redox conditions. In the period of 5−55 DOS, most of the surface Cr species remained relatively stable as indicated by small changes of the catalyst activity, the total Cr6+ concentration, the Cr6+/Cr3+ and Cr/Al atomic ratios, and the average size of chromium oxide clusters. For example, compared to ∼50% loss of Cr6+ after the initial operation, only ∼10% decrease of the Cr6+ concentration was observed in the following 50 days or 3300 cycles of the run. The Cr6+ concentration in the catalyst continued to decrease until the end of operation (Figure 7). Meanwhile, the α-Cr2O3 concentration showed significant decrease from 6.7% to 4.2%, down by 37.3% (relative). After 55 DOS, the solid solution of α-(Al, Cr)2O3 started to form. At 180 DOS, about 12.3 wt % of the solid solution was detected by XRD. The α-Cr2O3 concentration in the catalyst further dropped to ∼2.7%. The total Cr6+ concentration dropped to 0.33%. At 360 and 645 DOS, there was no α-Cr2O3 detected in the catalysts, while the solid solution concentration reached ∼34 and ∼73 wt %, respectively. This suggests that the formation of solid solution should not be considered as a simple solid-phase reaction between α-Al2O3 and α-Cr2O3, but as a gradual replacement of Al3+ in α-Al2O3 by migrated Cr3+ ions starting at the interface between Cr2O3 and α-Al2O3. The total Cr6+ concentration further dropped to 0.25% at 360 DOS and 0.19% at 645 DOS, respectively. The possible types of Cr species present in both fresh and aged catalysts are summarized in Table 1. In addition, the comparison of the total catalyst acidity and the total Cr6+ concentration suggests that the reduction of Cr6+ helped to reduce the total catalyst acidity. In contrast to the total catalyst acidity, the specific acidity showed a different trend due to significant loss of chromium surface area. After 180 DOS, the continuous increase of the surface Cr6+/Cr3+ ratio was observed, which can be explained by the incorporation of Cr3+ inside the alumina support,14 the formation of α-(Al, Cr)2O3 phase, or the increase of air and catalyst bed temperatures during the regeneration at the end run, or all of them. As a result of the large increase of the Cr6+/Cr3+ ratio, the isobutylene selectivity was noticeably reduced at the end of the catalyst life.

On the basis of the above discussion, the CrOx/Al2O3 catalyst experiences three major processes from the start of operation under the dehydrogenation−regeneration conditions. The first process refers to the complex transformation of chromium species on the catalyst surface throughout the catalyst lifetime. The initial operation causes the conversion of the “isolated” mono/poly-aluminum chromate(s) originally present on the fresh catalyst into small clusters of amorphous and crystalline α-Cr2O3, followed by the surface stabilization of chromium species through their restructuring. After about several months of relatively stable operation, the surface chromium species experience further changes, including the Cr3+ migration into the alumina to form solid solution, the phase transformation from crystalline α-Cr2O to amorphous or defective Cr2O3, and the aggregation of Cr2O3 clusters. All these changes led to the decrease of chromium oxide surface area leading to the catalyst deactivation. The second process refers to the phase transformation and/or sintering of alumina carrier. This process transfers the initial alumina to thermally stable α-Al2O3 with reducing surface area and the amount of small pores. The third process involves the interfacial transformation of CrOx/Al2O3 to form a solid solution of α(Al, Cr)2O3. For all three processes, the operation temperature and number of cycles affect the catalyst deactivation. In terms of the α-(Al, Cr)2O3 formation, there are two possible pathways: (1) the migration of catalyst surface Cr3+ ions into lattice of alumina, which facilitates the conversion of the formed α-Al2O3 into α-(Al, Cr)2O3, and (2) the migration of Cr3+ from chromium oxide clusters entrapped inside the collapsed pores into the surrounding α-Al2O3 to form the solid solution. The strong correlation between the formed α-(Al, Cr) 2 O 3 concentration and the loss of small pores in Figure 6 suggests that most of the α-(Al, Cr)2O3 was likely formed through the latter pathway. Figure 11 illustrates the changes of surface Cr species and the alumina support of the CrOx/Al2O3 catalysts at different aging levels. The changes of surface Cr species and alumina support depicted in the figure suggest that the CrOx/ Al2O3 catalyst deactivation was mainly caused by the loss of chromium oxide surface area due to the increased cluster size and by the formation of the solid solution of α-(Al, Cr)2O3 through the migration of Cr3+ from chromium oxide clusters into the surrounding α-Al2O3 lattice.

4. CONCLUSIONS In this study, we have studied the deactivation of CrOx/Al2O3 catalysts operated in an industrial dehydrogenation reactor under cyclic dehydrogenation−regeneration conditions using combined analytic methods, including XRD, O2 chemical sorption, XPS, N2 adsorption, TPD_NH3, TEM, and chemical analysis. It was found that three major processes occurred on 7945


Article


(7) Sattler, J. J. H. B.; Ruiz-Martinez, J.; Santillan-Jimenez, E.; Weckhuysen, B. M. Catalytic Dehydrogenation of Light Alkanes on Metals and Metal Oxides. Chem. Rev. 2014, 114, 10613. (8) Sattler, J. J. H. B.; González-Jiménez, I. D.; Mens, A. M.; Arias, M.; Visser, T.; Weckhuysen, B. M. Operando UV-Vis spectroscopy of a catalytic solid in a pilot-scale reactor: deactivation of a CrOx/Al2O3 propane dehydrogenation catalyst. Chem. Commun. 2013, 49, 1518. (9) Fang, D.; Zhao, J.; Li, W.; Fang, X.; Yang, X.; Ren, W.; Zhang, H. Investigation of the Characteristics and Deactivation of Catalytic Active Center of Cr-Al2O3 catalysts for Isobutane Dehydrogenation. J. Energy Chem. 2015, 24, 101. (10) Airaksinen, S. M. K.; Kanervo, J. M.; Krause, A. O. I. Deactivation of CrOx/Al2O3 Catalysts in the Dehydrogenation of iButane. Stud. Surf. Sci. Catal. 2001, 136, 153. (11) Hakuli, A.; Kytökivi, A.; Krause, A. O. I.; Suntola, T. Initial Activity of Reduced Chromia/Alumina Catalyst in n-Butane Dehydrogenation Monitored by On-Line FT−IR Gas Analysis. J. Catal. 1996, 161, 393. (12) De Rossi, S.; Ferraris, G.; Fremiotti, S.; Garrone, E.; Ghiotti, G.; Campa, M. C.; Indovina, V. Propane Dehydrogenation on Chromia/ Silica and Chromia/Alumina Catalysts. J. Catal. 1994, 148, 36. (13) Sokolov, S.; Stoyanova, M.; Rodemerck, U.; Linke, D.; Kondratenko, E. V. Comparative Study of Propane Dehydrogenation over V-, Cr-, and Pt-based Catalysts: Time on-stream Behavior and Origins of Deactivation. J. Catal. 2012, 293, 67. (14) Zaki, M. I.; Hasan, M. A.; Fouad, N. E. Stability of Surface Chromate − A Physicochemical Investigation in Relevance to Environmental Reservations about Calcined Chromia Catalysts. Appl. Catal., A 1998, 171, 315. (15) Puurunen, R. L.; Weckhuysen, B. M. Spectroscopic Study on the Irreversible Deactivation of Chromia/Alumina Dehydrogenation Catalysts. J. Catal. 2002, 210, 418. (16) Blasco, V.; Royo, C.; Monzon, A.; Satamaria, J. Catalyst Sintering in Fixed-bed Reactors: Deactivation Rate and Thermal History. AIChE J. 1992, 38, 237. (17) Bremer, H.; Muche, J.; Wilde, M. Proceedings of the 5th International Congress on Catalysis, Hightower, J. W., Ed.; North Holland: Amsterdam, 1973; Vol. 1, pp 195−202. (18) Rombi, E.; Cutrufello, M. G.; Solinas, V.; De Rossi, S.; Ferraris, G.; Pistone, A. Effects of Potassium Addition on the Acidity and Reducibility of Chromia/Alumina Dehydrogenation Catalysts. Appl. Catal., A 2003, 251, 255. (19) Sterligov, O. D.; Gitis, K. M.; Slovetskaya, K. I.; Shpiro, E. S.; Rubinstein, A. M.; Minachev, Kh. M. The Role of Chemical and Structural Changes on the Surface in Deactivation of ChromiaAlumina Catalysts in Dehydrogenation of Paraffinic Hydrocarbons. Stud. Surf. Sci. Catal. 1980, 6, 363. (20) Davydov, E. M.; Lykiyanov, E. N.; Nizamov, A. M.; Ablyakimov, E. I. Reasons of Deactivation Chromium Aluminum Catalyst at Conditions of the Commercial Operation. Nefteperabotka Neftechim (Moscow) 1981, 1, 26. (21) Rozengart, M. I.; Gitis, K. M.; Saltanova, V. P.; Anurov, S. A.; Rashchutkina, Z. A.; Kazanskii, B. A. Thermal Deactivation of Reduced Oxychome Dehydrocyclization Catalysts. Kinet. Katal. 1970, 11, 1446. (22) Kotelnikov, G. P. Technology of Dehydrogenation Catalysts and Problems of Optimization. Applied Chemistry (Russian) 1997, 70, 276. (23) Stitt, E. H.; Jackson, S. D.; Ahern, A.; King, F. Modelling for Design of a Deactivating Non-Isothermal Propane Dehydrogenation Reactor. Stud. Surf. Sci. Catal. 1999, 126, 137. (24) Shee, D.; Sayari, A. Light Alkane Dehydrogenation over Mesoporous Cr2O3/Al2O3 Catalysts. Appl. Catal., A 2010, 389, 155. (25) Fridman, V.; Romaine-Schmidt, E. Reduction of the Al-Cr Catalyst Selectivity As a Result of its Deactivation. Proceedings of 21st NACS Meeting, San Francisco, CA, 2009. (26) Babenko, V. S.; Pakhomov, N. A.; Buyanov, R. A. Investigation of the Thermal Stability of the Chromia−Alumina Catalysts for the Process of the One-Stage Dehydrogenation of n-Butane. Catal. Ind. 2009, 1, 43.

the CrOx/Al2O3 catalysts that caused the catalyst deactivation. The first process refers to the complex transformation and restructuring of surface chromium species. The phase transformation involves many changes, e.g., the conversion of “isolated” mono/poly-aluminum chromate(s) into “isolated” nonredox Cr3+ ions and/or Cr2O3, the conversion from αCr2O3 to amorphous Cr2O3, etc. The restructuring started at the beginning of the run and caused the coalescence of small chromium oxide clusters to form large ones, thereby reducing the chromium oxide surface area and the amount of active sites for dehydrogenation. The second process refers to the phase transformation and/or sintering of alumina. In this process, the phase transition and sintering of the alumina support take place during the catalyst operation at high temperatures. The third process is the interfacial phase transformation of CrOx-Al2O3 to form the solid solution of α-(Al, Cr)2O3. In this process, the formation of α-Al2O3 in the alumina support facilitates the migration of Cr3+ into the lattice of α-Al2O3 to form the inactive solid solution of α-(Al, Cr)2O3. The solid solution formation was observed at the middle stage of the run. In addition, the reduction of the chromium oxide surface area and the migration of Cr3+ into the alumina support resulted in the increase of the surface Cr6+/Cr3+ ratio, as a result increasing the specific catalyst acidity and reducing the catalyst selectivity. The main causes of irreversible deactivation of the CrOx/Al2O3 catalyst are thus the decrease of chromium oxide surface area, the transformation of the alumina carrier, the formation of solid solutions α-(Al, Cr)2O3 in the support, and the change of surface Cr6+/Cr3+ ratio at the end of the run.

■

AUTHOR INFORMATION

Corresponding Author

*600 Hill Street, Louisville, KY 40210. Tel: (502) 634-7423. Email: [email protected]. ORCID

Vladimir Z. Fridman: 0000-0001-6048-2804 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by Clariant Corporation and was performed in Louisville, KY. The authors thank Mrs. Sally Davis and Mr. Jim Howard at Clariant Corporation and Dr. Jacek B. Jasinski at University of Louisville for analytical support of this project.

■

REFERENCES

(1) Arora, V. K. Propylene via CATOFIN® Propane Dehydrogenation Technology. In Handbook of Petrochemicals Production Processes, Meyer, R. A., Ed.; McGraw-Hill: New York, 2005; Chapter 10.5. (2) Oviol, L.; Bruns, M.; Fridman, V.; Meriam, J.; Urbancic, M. Mind the Gap. Hydrocarbon Engineering, September 2012. (3) Kotelnikov, G. R.; Komarov, S. M.; Bespalov, V. P.; Sanfilippo, D.; Miracca, I. Application of FBD processes for C3-C4 olefin production from light paraffins. Stud. Surf. Sci. Catal. 2004, 147, 67. (4) Sanfilippo, D.; Buonomo, F.; Fusco, G.; Miracca, I. Paraffins activation through Fluidized Bed Dehydrogenation: the answer to light olefins demand increase. Stud. Surf. Sci. Catal. 1998, 119, 919. (5) Hornaday, G. F.; Ferrell, F. M. Advances in Petroleum Chemistry and Refining; Interscience: Paris, 1961; Vol. 4, p 451. (6) Weckhuysen, B. M.; Wachs, I. E.; Schoonheydt, R. A. Surface Chemistry and Spectroscopy of Chromium in Inorganic Oxides. Chem. Rev. 1996, 96, 3327. 7946


Article

Industrial & Engineering Chemistry Research (27) Grünert, W.; Saffert, W.; Feidhaus, R.; Anders, K. Reduction and Aromatization Activity of Chromia-Alumina Catalysts. J. Catal. 1986, 99, 149. (28) Fridman, V.; Xing, R.; Severance, M. Investigating the CrOx/ Al2O3 Dehydrogenation Catalyst Model: I. Identification and Stability Evaluation of the Cr Species on the Fresh and Equilibrated Catalysts. Appl. Catal., A 2016, 523, 39. (29) Fridman, V.; Xing, R. Investigating the CrOx/Al 2 O 3 Dehydrogenation Catalyst Model: II. Relative Activity of the Chromium Species on the Catalyst Surface. Appl. Catal., A 2017, 530, 154. (30) Cao, S.; Monnier, J. R.; Williams, C. T.; Diao, W.; Regalbuto, J. R. Rational nanoparticle synthesis to determine the effects of size, support, and K dopant on Ru activity for levulinic acid hydrogenation to γ-valerolactone. J. Catal. 2015, 326, 69. (31) Ellison, A.; Sing, K. S. W. Magnetic and Optical Studies of Chromium Oxides. Part 3.Calcination of Coprecipitated Chromium and Aluminium Hydroxide Gels. J. Chem. Soc., Faraday Trans. 1 1978, 74, 2807. (32) Fridman, V. Z. Pathways of light compounds formation during propane and isobutane dehydrogenation on Al-Cr catalysts. Appl. Catal., A 2010, 382, 139. (33) Cavani, F.; Koutyrev, M.; Trifino, F.; Bartolini, A.; Ghisletti, D.; Lezzi, R.; Santucci, A.; Del Piero, G. Chemical and Physical Characterization of Alumina-Supported Chromia-Based Catalysts and Their Activity in Dehydrogenation of Isobutane. J. Catal. 1996, 158, 236. (34) Mathur, J.; Bakhshi, N. N.; Mathews, J. F. Study of Chromia Alumia Catalysts. Can. J. Chem. Eng. 1980, 58, 601. (35) Bondioli, F.; Ferrari, A. M.; Leonelli, C.; Manfredini, T. Reaction Mechanism in Alumina/Chromia (Al2O3-Cr2O3) Solid Solutions Obtained by Coprecipitation. J. Am. Ceram. Soc. 2000, 83, 2036. (36) Weckhuysen, B. M.; Schoonheydt, R. A.; Jehng, J.-M.; Wachs, I. E.; Cho, S. J.; Ryoo, R.; Kijlstra, S.; Poels, E. Combined DRS-RSEXAFS-XANES-TPR study of supported chromium catalysts. J. Chem. Soc., Faraday Trans. 1995, 91 (18), 3245. (37) Sattler, J. J. H. B.; Mens, A. M.; Weckhuysen, B. M. Real-time quantitative operando Raman spectroscopy of a CrOx/Al2O3 propane dehydrogenation catalyst in a pilot-scale reactor. ChemCatChem 2014, 6, 3139. (38) Okamoto, Y.; Fujii, M.; Imanaka, T.; Teranishi, S. X-Ray Photoelectron Spectroscopic Studies of Catalysts: Chromia-Alumina Catalysts. Bull. Chem. Soc. Jpn. 1976, 49 (4), 859. (39) Grünert, W.; Shpiro, E. S.; Antoshin, G. V.; Feldhous, R.; Anders, K.; Minachev, H. M. Formation of the metallic chromium on the surface of aluminum/chromium catalysts. In Reports of academy of science, 1985; pp 372−378. (40) Turyaev, I. Y. Book “Theoretical Fundamentals of Butadiene and Isoprene Production by Dehydrogenation Methods”, Kiev. Science 1973, 271. (41) Pervaiz, E.; Gul, I. H. Low temperature synthesis and enhanced electrical properties by substitution of Al3+ and Cr3+ in Co−Ni nanoferrites. J. Magn. Magn. Mater. 2013, 343, 194.

7947
