The methanol-to-olefin (MTO) process, combined with the transformation of coal or natural gas to methanol, has been demonstrated to be a successful non-petrochemical route to produce ethylene and propylene [1]. Understanding the mechanism of the MTO reaction has drawn considerable interest over the past several decades because of its ability to produce C-C bonds from C1 reactants.
Previous studies have suggested that the MTO reaction occurs through a hydrocarbon pool (HCP) mechanism [2-8]. The reaction takes place over three periods: the induction period, the steady-state reaction period, and the deactivation period [9, 10]. During the induction period, small amounts of polymethylbenzenes (PMBs) and their protonated analogues are formed and accumulated within the cages or channel intersections of the molecular sieves that act as the initial reaction intermediates [8, 11-14]. After the induction period, methanol conversion is observed and increases autocatalytically. Eventually methanol conversion increases dramatically and polycyclic aromatic molecules and large coke fragments form at the same time, which leads to deactivation of the catalyst [8, 15-17]. The time required for the formation of PMBs determines the induction period of the MTO reaction. In the case of ZSM-5 catalyst, we previously investigated reaction behavior and kinetics during the methanol conversion induction period [18]. Our results found that the induction period could be further broken down into three stages: an initial C-C bond formation stage, a HCP species formation stage and an autocatalysis reaction stage. A critical value of HCP species, [HCP]c, that is required for starting the autocatalysis reaction (the third stage) was proposed and measured.
SAPO-34 (CHA type, containing chabazite cages and 8-member ring windows) offers excellent MTO activity and selectivity [19], and has been used for industrial applications with a fluidized-bed reactor and regenerator because of the rapid deactivation of the catalyst [20]. The aromatic-based cycle is the main reaction mechanism over the cage-type SAPO-34 catalyst. The catalytic performance of SAPO-34 is very different from that of ZSM-5. With larger channel sizes (10-member ring channels), the ZSM-5 catalyst shows less ethylene selectivity but better performance in methanol-to-gasoline, methanol-to-aromatics and methanol-to-propylene reactions than SAPO-34 [21-24]. It has been reported that with ZSM-5 catalyst, methanol conversion can proceed according to both the aromatic and the olefin-based cycles during the steady-state stage [25-28]. Based on the different zeolite framework structures, diffusivities, coking behavior and HCP mechanisms between SAPO-34 and ZSM-5 catalysts, it is important to compare the induction stage between SAPO-34 and ZSM-5. This study presents a kinetic investigation of the induction reaction at different temperatures. It is expected to gain more insight into the HCP mechanism.
SAPO-34 ((Al + P)/Si = 16) was synthesized as described elsewhere [10]. The SAPO-34-based catalyst for fluidized reaction was prepared using the spray-drying method with an inert binder. Methanol (AR) was purchased from Xinxi Chemical Reagent Company of Shenyang, China.
Methanol conversion was carried out in a fluidized-bed reactor, which guaranteed spatial uniformity of the catalytic reaction and coke deposition. Prior to the start of the experiment, 10 g of freshly calcinated catalyst was loaded into the reactor and pretreated at 550 °C under He flow (30 mL/min) for 40 min. The temperature was then adjusted to the desired value. By switching a four-port valve, the reactant (40% aqueous methanol solution), which was vaporized by a preheater, entered the reactor bottom through a distributor and contacted the catalyst. The space velocity (WHSV) of methanol was 1.5 h-1. The product was detected using on-line gas chromatography (GC) (Agilent 7890A GC, USA) employing a CPPORAPLOT Q-HT (25 m × 0.32 mm) column and flame ionization detector.
Changes in methanol conversion with reaction time at different temperatures over SAPO-34 are presented in Fig. 1. At 290 °C, initial methanol conversion was low and almost no methanol conversion was observed, indicating that the initial HCP species were difficult to generate at a low reaction temperature. Methanol conversion (> 1%) was only observed after 100 min, after which the conversion increased more significantly. The highest methanol conversion (11.4%) occurred after 222 min and then decreased with gradual deactivation of the catalyst. At the higher temperature of 300 °C, the induction period was greatly shortened and the maximum methanol conversion increased to > 50%. The methanol conversion increased more rapidly during the autocatalytic reaction and was followed by faster deactivation compared with the reaction at 290 °C. Clearly, the MTO induction reaction is sensitive to the reaction temperature over the SAPO-34 catalyst. At higher reaction temperatures (310 and 320 °C), the induction period was further shortened and the maximum methanol conversion continuously increased, followed by more rapid catalyst deactivation. These phenomena are consistent with our previous studies on H-ZSM-5 in a fixed-bed reactor [18].
The MTO reaction is a typical autocatalytic process. The initially accumulated HCP species will promote and accelerate the generation of more HCP compounds. To gain deeper insight into how the fresh catalyst is transformed to the active one, detailed reaction stages that evolve during the induction period should be clearly differentiated. In our previous work, three reaction stages were observed over the MTO; induction period and the generation of initial HCP species were shown to occur during the second stage [18]. By plotting the methanol conversion on a logarithmic scale, as illustrated in Fig. 2, three reaction stages can also be distinguished during the induction period at 290 °C over SAPO-34. According to our previous results [18], and the changes in methanol conversion during different stages, the three stages can also be described as the initial C-C bond formation stage (the initial stage), the HCP species formation stage (the second stage) and the autocatalysis reaction stage (the third stage) [19]. This indicates that there are similar kinetic characteristics during the induction period over HZSM-5 and SAPO-34, regardless of whether the reaction takes place in a fixed- or fluidized-bed.
Identification of the three reaction stages during the induction period is helpful for further proving the importance of the HCP mechanism using the SAPO-34 catalyst. However, because of the complexity of the composition of HCP species and HCP mechanism, it is still very difficult to clarify the relationships among the three stages and the evolution of the reaction mechanism in the absence of kinetic investigations. As such, the apparent activation energies of the three reaction stages were calculated following similar methods to our previous work [18].
It has been reported that the first C-C bond and the initial HCP species (very trace amounts) are generated during the initial stage of the MTO induction period [29, 30]. The formation of the first C-C bond is very likely to be influenced and disturbed by many factors, such as impurities in the feed and catalyst, reaction temperature and the reactor. Consequently, the initial stage is very sensitive to uncontrollable uncertainties that may make calculated results unreliable. Careful investigation of this stage is required. In this study, the use of a fluidized-bed reactor allowed for more catalyst loading (10 g) with a uniform temperature distribution inside the catalyst bed, which decreased uncertainties in the kinetic studies. However, despite the relatively simple reaction processes during this stage, it was still very difficult to observe the initial methanol conversion accurately. As a result, calculation of the apparent activation energy for the initial stage was not performed.
Despite the uncertainty of the initial reaction stage, the existence of the second stage was evident under the conditions used. During the second stage, certain quantities of HCP species are accumulated to trigger the autocatalytic reaction. It has been shown that the generation and accumulation of HCP compounds are influenced by factors like catalyst topology, acidic site density and contact time [31-34]. Because of the different catalyst topologies and acidities of SAPO-34 and ZSM-5, it was essential to clarify which initial active centers are easily generated for each catalyst. Although the detailed reaction mechanisms for the formation of the organic compounds are complex and it is almost impossible to determine the intrinsic activation energy of the whole process, the value of the calculated apparent activation energy is highly representative and can be used for comparisons.
Using a similar principle as described elsewhere [18], the relationship between the reaction rate constant, k2, and the time of the second stage, tc, is
The value of tc in the temperature range 300-320 °C is shown in Fig. 3 and Table 1.
Fig. 4 presents lnk2 as a function of 1/T and is a straight line. The activation energy of the second stage of the MTO reaction was 274 kJ/mol. The acidity of HZSM-5 is much stronger than that of SAPO-34 and coke species are formed much more readily over stronger acidic sites. The calculated value in this study was lower than that of HZSM-5 (Si/Al = 19, Ea = 301 kJ/mol). Considering the factors influencing the second stage, the differences in activation energies could be caused by different topologies and diffusivities of the catalysts. HZSM-5 features an MFI topology made up of a 3D network consisting of sinusoidal (5.1×5.5 Å2) and straight (5.3×5.6 Å2) channels defined by 10-rings that results in medium-sized pores. SAPO-34 is a small-pore molecular sieve, featuring a CHA topology with spacious cavities (10×6.7 Å2) that are connected by small (3.8×3.8 Å2) 8-ring windows. HZSM-5 has a much longer lifetime than SAPO-34 during the MTO reaction because of its perfect diffusivity. Because of the limited diffusivity of SAPO-34, the initially formed cyclic compounds and higher olefins are more likely confined in the CHA cages. These retained species would contribute to the generation of active centers and the formation rate would be enhanced because of the spatial confinement effect of the CHA cages. As a result, the accumulation of initial HCP species during the second stage is easier and the apparent activation energy is lower for SAPO-34.
During the third reaction stage, the methanol conversion reaction was initiated and the reaction proceeded in an autocatalytic fashion. To gain more information about the evolution of coke species during this stage, the apparent activation energy was calculated. As shown in Fig. 2, the methanol conversion, if plotted on a logarithmic scale, is a linear function of TOS for the third stage. As such, the autocatalysis reaction rate can be expressed as:
where x is the methanol conversion and k3 is obtained from Fig. 5(a). Fig. 5(b) shows lnk3 as a function of 1/T. The activation energy of the autocatalysis reaction stage was calculated to be 332 kJ/mol.
The kinetic parameters and activation energies during the three stages are summarized in Table 1. With these activation energy data, the evolution of the energy barrier during the MTO induction period for SAPO-34 could be explained according to Scheme 1.
It is reasonable that after the accumulation of some HCP species the autocatalytic reaction would be initiated and the reaction would proceed more easily during the third reaction stage. However, the apparent activation energy was much higher than that of the second stage. This seemed strange and contradictory to previous research at first glance [18]. But, if we focus on the evolution of coke species during the autocatalytic stage, the result seems reasonable. Much work has previously focused on the evolution of retained species during the autocatalysis stage for ZSM-5 and SAPO-34 [35, 36]. For ZSM-5, all of the retained species are active methylbenzenes and their quantities increased continuously as the reaction progressed during the induction period [36]. For SAPO-34, the coking behavior is different. Besides active species like methylbenzenes and methylnaphthalenes, a large amount of inactive methyladamantanes are also generated and accumulated during the autocatalysis stage [35]. It has been shown that the deactivation of SAPO-34 is caused by the accumulation of methyladamantanes under low reaction temperatures [35]. As a result, the evolution of the HCP mechanism is different over SAPO-34 during the autocatalysis stage of the induction period. The MTO reaction proceeds more readily over HZSM-5 because more active centers are continuously accumulated. Over SAPO-34, both the active and inactive species are generated and many of them could be retained in the cages because of the limited diffusivity during the induction period. Active species enhance the activity of the catalyst while inactive species inhibit activity. Consequently, the apparent activity of the SAPO-34 catalyst is based on competition between active and inactive species. Also, the proportion of methyladamantanes continuously increased with the progress of the induction reaction and, as a result, the apparent activation energy of the third stage was higher than that of the second stage.
This study provides new insight into the MTO induction reaction over SAPO-34 catalyst. Three reaction stages were observed that were sensitive to the reaction temperature. Kinetic investigation of the MTO induction reaction was systematically performed and the results were discussed in comparison with those for the ZSM-5 catalyst. For the first time, we suggest that the limited diffusivity and the spatial confinement effect of the CHA cages could help generate initial HCP species. It is also important to note that the apparent activation barrier increased between the second and third stages because of the continuous formation of inactive methyladamantanes.
2016,
Vol. 37 Issue (9): 1496-1501 
PDF
2016,
Vol. 37 Issue (9): 1496-1501 PDF