| The spent zeolites are traditionally regenerated by several methods, such as thermal treatment, or reaction with oxygen or nitrous oxide-containing mixtures [4,11–14]. Oxidative treatments using oxygen usually operate above 400 ◦C and may result into irreversible chemical modification of the zeolites through hydrothermal process with generated water vapour [4,14,16]. Regeneration by ozone is an attractive low temperature process due to the high oxidizing activity of this compound [17]. This technique is therefore particularly interesting to preserve the structure of the zeolite catalyst and to prevent dealumination. Copperthwaite et al. [16] showed that ozone-enriched oxygen treatment can restore activity of ZSM-5 zeolites used for methanol conversion under mild conditions (150 ◦C), while a temperature of about 450 ◦C was necessary with oxygen only. The main difference observed between the two restored catalysts lay in a higher selectivity towards light gases in the C2–C4 fraction for the ozonized samples (with reduced methane, C5+ and aromatic yields). From FT-IR study of coked HY zeolite regeneration, Mariey et al. [18] also concluded that ozone was effective at 170 ◦C or less (as compared to 500 ◦C for oxygen) and preserved the framework of samples with a low Si/Al ratio. However using coked Y zeolites in the form of 1/16 in. diameter extrudates, Hutchings et al. [19] found that ozone treatment at 200 ◦C only removed coke on the outer surface of the pellets and regeneration obeyed a shrinking core mechanism with pore diffusion control. In the present work, regeneration of coked ZSM-5 extrudates is studied using an ozone-enriched oxygen stream at low temperature, in the range 20–150 ◦C. The objective of this work is to understand the interactions between ozone and spent zeolite and to find the best operating conditions for coke removal and catalyst activity restoration. In the following, the term “coke” will refer to all carbonaceous materials which are located on the zeolite, irrespective of their chemical nature and molecular weight [20]. 2. Materials and methods 2.1. Material ZSM-5 zeolite was provided by Zeolyst International in the form of ca. 3 mm × 10 mm extrudates pelletized with 20 wt.% of aluminium oxide. The pure ZSM-5 phase was given with a SiO2/Al2O3 molar ratio of 23. The extrudates were used as catalyst for the cracking of poly(methyl methacrylate) (PMMA) in a fixed bed reactor operating at 1 bar and 250–300 ◦C. Distribution of coke deposit varied axially along the bed, with a higher amount at the top of the reactor close to the PMMA feed. Nature of the deposit also differed: top particles were coated with a dense, shiny black, layer that eventually stuck them together, while bottom particles of grey colour kept well separate. Particle samples from both locations were collected for regeneration tests. 2.2. Characterization of the zeolites Physicochemical properties of the coked and regenerated catalysts were evaluated by several techniques and compared to those of fresh zeolites. Textural properties, which are essential to evaluate the role of internal mass transfer in the effective burn-off rate, were |
measured by N2 adsorption at 77 K on a Micromeritics ASAP 2010 and by mercury intrusion on a Micromeritics AUTOPORE IV. Prior to N2 sorption, the solid samples were degassed at 200 ◦C (this temperature being selected to avoid coke removal). The specific surface area was calculated from BET plot at relative pressures between 0.01 and 0.2 [21]. The mesoporous and microporous volumes were estimated from gas porosimetry measurements according to Barrett–Joyner–Halenda [22] and Horvath–Kawazoe [23] methods respectively. The macroporous volume (50 nm < pore diameter < 30 m) and skeletal density were evaluated from mercury porosimetry.
Chemical characterization was done using different techniques. Thermogravimetry coupled with infrared analysis of released gases (TGA-IR) was used to qualify the nature of the carbonaceous deposit. These analyses were performed on a Q50 thermobalance (TA Instruments) under controlled atmosphere (nitrogen). The samples were heated at 10 ◦C/min from room temperature to 1000 ◦C, including 60 min plateau at 120 ◦C to remove physisorbed water. The amount of carbon in the aged/regenerated catalysts was determined by a CHN elemental analyzer (flash combustion) using about 20 mg of crushed sample. Particle cross-sections were observed by scanning electron microscopy (Leo 435 VP) coupled with energy dispersive X-ray (Oxford INCA 200) analysis to profile elemental carbon content within the particles. Finally the acidity of zeolite catalysts, which is usually correlated with their activity in C C bond cracking [3], was evaluated by two complementary techniques: -ammonia temperature-programmed desorption (NH3-TPD) carried out on a chemisorption analyzer (Micromeritics AutoChem II 2920) equipped with a thermal conductivity detector, NH3-TPD first provided information about the global acidity of the zeolites. Crushed samples (50 mg) were outgassed under an argon flow (50 ml/min) by heating with a rate of 10 ◦C/min up to 400 ◦C, and cooled back to ambient prior to NH3 saturation. After removal of physisorbed NH3 (at 50 ◦C), a temperature ramp rate of 10 ◦C/min was applied up to 800 ◦C under argon flow (50 ml/min). The free acid sites were evaluated from subtraction of TPD spectra with and without the preliminary saturation of the zeolite with NH3. From deconvolution of the resulting NH3-TPD spectrum, three peaks were also distinguished, each following a gaussian distribution. They corresponded to weak (low desorption temperature, 140–150 ◦C), medium (235–245 ◦C) and strong (high desorption temperature > 450 ◦C) acid sites, respectively [24]. To determine the total amount of acid sites, an amorphous silica–alumina (ASA) which known amount of total acid sites (0.55 mmol g−1 [25]) was used as reference material. Pyridine adsorption/desorption allowed the distinction between Brønsted and Lewis acid sites, due to the narrow character and high extinction coefficient of the typical IR bands [26]. Crushed zeolite samples (10–30 mg) were pressed into thin wafers and pretreated under flowing air at 450 ◦C overnight and then under vacuum. Excess pyridine was adsorbed at 150 ◦C and after a few minutes physisorbed pyridine was removed under vacuum (1 h). The concentration of Brønsted and Lewis acid sites able to retain pyridine at 150 ◦C was estimated from the surface of IR bands at 1545 cm−1 and 1455 cm−1, respectively. Subsequent desorption analyses were performed at |
