Hydrothermal stability is one of the most important properties that influence the application of zeolitic catalysts. It is well-known that the hydrothermal stability of a zeolitic material depends mainly on its topological structure and framework silica-to-alumina ratio (SAR). However, hydrothermal stability is also significantly influenced by zeolite crystal size. Typically, large-pore zeolites possessing low SARs and nano-particle sizes have poor hydrothermal stability. Conversely, small- and medium-pore zeolites having high SARs and large crystal sizes have improved hydrothermal stability. Although the unique medium-pore MFI structure and high-silica framework composition gives ZSM-5 zeolites excellent hydrothermal stability, efforts to further enhance ZSM-5 hydrothermal stability are certainly meaningful for practical uses, especially when subjecting nano-sized ZSM-5 to harsh hydrothermal conditions to applications such as fluidized catalytic cracking (FCC), where the catalyst is required to withstand steaming treatment at >700 ℃. Several techniques have been developed hitherto for enhancing zeolite hydrothermal stability. For example, rare earth ion exchange [1-3], dealumination and silicon-addition by (NH4)2SiF6 modification [4, 5], and high-temperature steam calcination [6-8] are applied to low silica type zeolites such as X and Y. Additionally, phosphorus modification by H3PO4, (NH4)2HPO4 or NH4H2PO4 impregnation has been proposed for high SAR zeolites, such as β and ZSM-5 [9-13]. Multiple strategies are available for improving the hydrothermal stability of low SAR zeolites. However, few approaches are known for high SAR zeolites. As for ZSM-5 zeolite, phosphorus modification with H3PO4, (NH4)2HPO4 or NH4H2PO4 does not always result in satisfactory improvement to hydrothermal stability, and there are few published reports detailing hydrothermal stability enhancement of nano-sized ZSM-5 zeolites.
Gasoline obtained from FCC processes account for a major proportion of the Chinese gasoline pool and generally contains high sulfide and olefin content. To conform to international gasoline standards, the China V gasoline standard has been issued to meet the necessary specifications and will be implemented in 2017. Stringent threshold levels governed by the China V gasoline standard state that the sulfur content should be ≤ 10 μg/g, and olefin content ≤ 24 vol.%. Therefore, China has a long-term requirement to upgrade FCC gasoline by decreasing sulfide and olefin content. Currently, the 'S-zorb' process configured within a fluidized bed reactor has been accepted by many large-scale refineries for FCC gasoline desulfurization. In this process, the sulfides in FCC gasoline are removed by selective adsorption over a selective adsorbent. Thereafter, the adsorbent is regenerated to recover the adsorption capacity. The 'S-zorb' process is efficient for the deep desulfurization of FCC gasoline. However, it is unaffordable for smaller refineries. Therefore, hydrodesulfurization will continue to be an important alternative to the 'S-zorb' process. There already exist several well-known hydrodesulfurization processes for FCC gasoline, such as 'SCANfining' and 'Prime G+'. Although these processes are able to simultaneously reduce sulfide and olefin content, hydrodesulfurization usually suffers from relatively big research octane number (RON) loss owing to the hydrogenation saturation of olefins. This problem is of concern in the case of deep hydrodesulfurization. Additionally, the removal of extra olefins from FCC gasoline via hydrogenation is too expensive because of considerable H2 consumption.
In previous studies [14-16], we introduced a fixed-bed olefin-to-aromatics (OTA) reaction using modified nano-sized ZSM-5 zeolites to reduce olefin content in full-range FCC gasoline. The transformation of olefins into aromatics via the OTA reaction was considered as an alternative solution to the olefin problem of FCC gasoline, as aromatics have a higher RON rating than olefins, and the process does not require the consumption of H2 if the FCC-OTA gasoline is not hydrodesulfurized down-stream. However, the economic viability of the OTA reaction being implemented within industry is limited because of low OTA conversion and fast catalyst deactivation when set-up in a fixed-bed operation mode. Further potential restrictions will arise if future legislated threshold levels of aromatics in gasoline are lowered. For these reasons, we are now interested in studying the viability of olefin catalytic cracking (OCC) of full-range FCC gasoline with nano-sized ZSM-5 zeolites when configured within a fluidized-bed. The objective is to efficiently transform the extra olefins in FCC gasoline into value-added short-chain olefins like ethylene, propylene and butenes, and to direct the FCC-OCC gasoline for further hydrodesulfurization treatment. Employing nano-sized ZSM-5 zeolites as the catalyst is likely to benefit OCC of full-range FCC gasoline by enhanced performance, however, it is important to improve the hydrothermal stability of nano-sized ZSM-5 zeolites, suitably for fluidized-bed configuration, as if the crystal size is ultrafine the inherent stability of the ZSM-5 structure will weaken.
Therefore, this study is two-fold: first, enhancing the hydrothermal stability of nano-sized ZSM-5 zeolites by a novel phosphorus modification route featuring trimethyl phosphate and, second, the evaluation OCC of full-range FCC gasoline when employing the stabilized zeolite catalysts under fluidized-bed configuration.
Nano-sized NaZSM-5 powder (20-50 nm, SiO2/Al2O3 = 26) was supplied by Dalian Ligong Qiwangda Chemical Technology. First, the received zeolite was calcined in a muffle furnace in the presence of static air to remove the organic template. Thereafter, the template-free zeolite was treated by a conventional ammonia ion exchange method to obtain HZSM-5. Further modification of the zeolite was performed by introducing 0.2 wt.%, 0.6 wt.%, 1.0 wt.%, 2.0 wt.%, and 3.0 wt.% phosphorus into the HZSM-5 zeolite by incipient wet impregnation via an aqueous solution of trimethyl phosphate ((CH3)3PO4) at room temperature. For the purpose of comparison, the same phosphorus modifications were repeated with ammonium dihydrogen phosphate (NH4H2PO4). The modified catalysts were denoted as 0.2P(Z)-3.0P(Z) for (CH3)3PO4-modified zeolites, and 0.2P(Y)-3.0P(Y) for NH4H2PO4-modified analogues. To investigate changes to hydrothermal stability of HZSM-5 zeolites as a function of phosphorus content and type, the phosphorus-modified catalysts were further steam treated at 800 ℃ for 4 h (100% steam). The steamed samples were given a “-st” postfix to the original catalyst codes. All catalysts were pressed, crushed and sieved to 40-60 mesh sizes for reaction purpose.
Chemical composition of the catalyst samples was analyzed using a Bruker SRS 3400 X-ray fluorescence (XRF) spectrometer. X-ray diffraction (XRD) was performed using a Rigaku D/MAX-2004 diffractometer with Cu Kɑ radiation (40 kV, 100 mA) at a scanning rate of 0.02°/min (2θ). The relative crystallinity (RC) of the modified samples was obtained by comparing the sum of intensities of its five characteristic peaks at 2θ = 8°-9° and 23°-25° with that of the precursor. Nitrogen adsorption measurements were conducted at -196 ℃ employing a Micrometrics ASAP 2020 instrument. Before measurements, the samples were subjected to vacuum treatment at 350 ℃ for 6 h. The specific surface area was calculated using the classical Brunauer-Emmett-Teller (BET) model, while the pore volume was calculated using the t-plot method. Ammonia temperature-programmed desorption (NH3-TPD) was performed on a CHEMBET-3000 (Quantachrome) apparatus. Typically, 0.15 g sample was used for each measurement. First, the sample was purged with He at 600 ℃ for 1 h. Thereafter, the pretreated sample was subjected to NH3 adsorption at 150 ℃ until saturation, followed by a He sample sweep at 150 ℃ for at least 30 min to expel physically adsorbed NH3. Finally, the NH3 desorption profile was recorded during sample heating from 150 to 600 ℃ at a rate of 15 ℃/min in a He flow 80 ml/min. The relative acid concentrations of the modified samples were estimated from the areas of their desorption profiles, in reference to a precursor. Weak acid concentrations were determined by the desorption profile area in the temperature range of 150-400 ℃. Correspondingly, strong acid concentrations were determined by the desorption profile area in the temperature range of 400-600 ℃. IR spectra were recorded on a Nicolet 6700 FT-IR spectrometer at room temperature, the spectral resolution was 4 cm−1. To obtain the spectra, the catalysts were pressed into self-supporting thin wafers and activated in a quartz IR cell equipped with CaF2 windows. Activation was performed at 400 ℃ for 4 h under vacuum (10−3 Pa). Ammonia adsorption was performed at room temperature for 30 min at a flow rate of 30 ml/min. After adsorption, samples were evacuated at 150 ℃ for 30 min to remove physically adsorbed ammonia.
A fixed-bed micro-reactor was used to simulate the fluidized-bed reaction configuration for OCC of full-range FCC gasoline. Reactor dimensions were 10 mm inner diameter and 4.8 g catalyst loading capacity. The reaction was conducted at 540 ℃. The full-range FCC gasoline contained 50.80 wt.% olefins, 20.82 wt.% aromatics, 7.69 wt.% n-paraffins, 16.59 wt.% i-paraffins, 3.45% naphthenes and 675 μg/g sulfur. The feed was pumped in a closed-circuit system with a sampling loop. A desired amount of oil was fed into the micro-reactor via the sampling loop when a six-way valve was manually rotated. The catalyst/oil mass ratio was set to 4 in this study. Using N2 as a diluent gas, the contact time of the feed oil with the catalyst was adjusted to ~4 s. The products from the reactor were cooled using an ice-bath and collected as gaseous and liquid fractions. The liquid fraction was analyzed by a Shimadzu GC-14C gas chromatograph (OV-1 capillary column 50 m × 0.2 mm, FID detector), while the gas fraction was analyzed by a Techcomp 7890F gas chromatograph (PLOT Al2O3 capillary column 50 m × 0.53 mm, FID detector). The product yield (Y) and selectivity (S) were calculated on the basis of mass balance as follows:
Additionally, sulfur content of the feed oil and product oils were measured by coulometry. Oil RON was calculated according to a previously published method [17].
Table 1 shows the compositions of the steamed catalysts studied. The catalysts modified with ammonium dihydrogen phosphate or trimethyl phosphate show no significant compositional changes when compared with the steamed HZSM-5 (NZ-st) catalyst.
Fig. 1 indicates that, when the parent nano-sized HZSM-5 zeolite was directly subjected to the harsh steaming treatment (100% steam, 800 ℃ for 4 h), the RC fell to ~70%. Treatment of the parent zeolite by NH4H2PO4 and (CH3)3PO4 impregnation and subsequent steaming under the same conditions resulted in higher RC. These results clearly show the effectiveness of phosphorus modification to enhance hydrothermal stability of nano-sized HZSM-5 zeolites. Furthermore, it can be seen that P loading has a significant influence on the resulting degree of crystallinity. The optimum P loading was estimated to be ~0.6 wt.% for NH4H2PO4, and ~1.0 wt.% for (CH3)3PO4. By comparison, it is easy to see that modification with (CH3)3PO4 results in higher framework integrity than that with NH4H2PO4 and therefore better protects the framework of nano-sized HZSM-5 zeolites from the damage of high-temperature steaming.
Table 2 shows that, both NH4H2PO4, and (CH3)3PO4 modifications reduce the specific surface area of the nano-sized HZSM-5 zeolite even though the P loading was limited to 3 wt.%. Considering the changes of micropore surface area and pore volume, both NH4H2PO4, and (CH3)3PO4 predominately modify the micropores of the nano-sized zeolite. Fig. 2 shows that, as expected, both NH4H2PO4 and (CH3)3PO4 modifications are helpful in stabilizing the pore structures of nano-sized HZSM-5 zeolite. As a result, the modified zeolites had higher retention rates for surface area and pore volume. Complementary to RC, these correlations also indicated the advantage of (CH3)3PO4 over NH4H2PO4 for zeolite modification.
Table 3 indicates that, both NH4H2PO4 and (CH3)3PO4 modifications significantly reduce the concentration of both weak acid and strong acid sites. This suggests interaction of the phosphorous species with framework Al present in the modified zeolites. Fig. 3(a) shows that, (CH3)3PO4 modification resulted in a faster decrease in the total acid concentration when compared with NH4H2PO4 modification when the P loading was below 2.0 wt.%. Conversely, increasing the P loading beyond 2.0 wt.%, NH4H2PO4, not (CH3)3PO4, modification resulted in a faster decrease in the total acid concentration. This phenomenon implies that, for the interaction with framework Al sites, (CH3)3PO4 was more selective and efficient than NH4H2PO4. From Fig. 3(b) it can be seen that both NH4H2PO4 and (CH3)3PO4 modifications enabled far more acid sites in the nano-sized zeolite to be retained after severe high-temperature steaming treatment. Once again, (CH3)3PO4 acts as a more favorable precursor as a modifier as compared with NH4H2PO4.
In addition to the NH3-TPD acidity characterization, FT-IR spectroscopy was also used to determine the nature of the acid sites within different modified catalysts. In Fig. 4, absorbance peaks located at 1457 cm−1 and 1617 cm−1 belong to Brönsted acids and Lewis acids, respectively. It is evident that regardless of the P precursor employed in the modification the resulting acidity of the final catalysts is dominated by Brönsted acid sites.
As mentioned above, (CH3)3PO4 modification allows the steamed nano-sized HZSM-5 zeolites to retain more crystallinity, micropores and surface acid sites (mainly Brönsted acid sites) on the whole. Therefore, it is of interest to investigate the catalytic performance of the (CH3)3PO4-modified catalysts in OCC of full-range FCC gasoline. For the purpose of comparison, the catalytic performance of the NH4H2PO4-modified catalysts is also included in this section.
Fig. 5 shows that, after hydrothermally treatment at 800 ℃ for 4 h under a 100% steam atmosphere, the parent nano-sized HZSM-5 zeolite has rather low catalytic cracking activity. Consequently, gas and C2-C4 olefin yields over the parent zeolite were only ~25% and 23%, respectively. However, both (CH3)3PO4 and NH4H2PO4 modifications significantly increased the catalytic cracking activity of the nano-sized HZSM-5 zeolite. As a result, the gas and C2-C4 olefin yields over the modified zeolite catalysts notably increased to ~38% and ~35%, respectively, with (CH3)3PO4-modified HZSM-5, and ~34% and ~30%, respectively, with NH4H2PO4-modified HZSM-5. All improved yields reached a maximum at a P loading of 0.6 wt.%. It is evident that the variation in yield trend follows the trend of relative total acid concentration in the catalysts, which is a function of phosphorus loading, therefore the acidity of the catalyst plays a key role in the reaction. Furthermore, the (CH3)3PO4-modified nano-sized HZSM-5 zeolite catalysts have a clear activity advantage over the NH4H2PO4-modified catalysts as a result of increased acid site retention.
Table 4 indicates that, OCC over different modified catalysts gave similar gas products. For example, the gas products contained ~90% C2=-C4=, the content of propylene was approximately 3-5 times that of ethylene and double the concentration of butenes. However, only when OCC of full-range FCC gasoline was performed in the presence of phosphorus-modified catalysts could the olefin content in the product oils be reduced to less than 30%—far below the olefin content of the gasoline feed (Table 5). That means, the phosphorus-modified catalysts had stronger olefin-reduction ability. Undoubtedly, this should be attributed to the enhanced hydrothermal stability of the nano-sized HZSM-5 zeolites as a result of phosphorus modification, which allowed the catalysts to retain a higher degree of Brönsted acid sites after high-temperature steaming. Additionally, a significant increase in the aromatic content in the product oils was observed, which increased from 21.52% in a 100% gasoline feed to ~50% in a 60-70% product oil. This indicates substantial aromatization occurring during the OCC when configured in a fluidized-bed operation mode. Aromatization is thought to proceed via hydrogen transfer by the Brönsted acid-catalyzed carbenium ion-chain mechanism. With respect to RON rating of the product oils, OCC benefits from this concurrent reaction. The calculated data suggests that the product oils of the OCC should have improved anti-knock properties. Additionally, a significant decrease in the sulfide content from 675 μg/g in the gasoline feed to less than 200 μg/g in the product oils can be found when OCC is performed in the presence of phosphorus-modified catalysts, which corresponded to at least an 80% desulfurization rate. Obviously, non-hydrodesulfurization should be an additional advantage of OCC, by significant H2 savings, thus allowing subsequent hydrodesulfurization easier. Furthermore, from Table 6 and Fig. 6 it can be seen that, OCC converted more heavy olefins than light olefins. As a result, the carbon number distribution of olefins in the product oils were shifted towards shorter-chain olefins as compared with that in the full-range FCC gasoline feed. This would facilitate the subsequent hydrodesulfurization as if the product oils from the OCC of full-range FCC gasoline, termed FCC-OCC gasoline, is fractionated into light and heavy cuts, in the following hydrodesulfurization process, the heavy end would contain less olefins and thus high hydrodesulfurization selectivity is expected.
This study shows that (CH3)3PO4 impregnation is an effective phosphorus modification method for enhancing the hydrothermal stability of nano-sized HZSM-5 zeolites. Low phosphorus loading is sufficient for the ultra-fine zeolite to withstand severe steaming treatment and to retain high crystallinity, large specific area and micropore volume, and abundant surface acid sites. It is believed that (CH3)3PO4 modification may open further opportunities for nano-sized HZSM-5 zeolites to act as efficient catalysts for reactions occurring under harsh hydrothermal conditions. Preliminary analysis in OCC of full-range FCC gasoline confirms the significance of the (CH3)3PO4 modification. Furthermore, when the OCC of full-range FCC gasoline is configured in a fluidized bed reaction mode, it may offer an alternative solution for the olefin problem of FCC gasoline upgrade.
2017,
Vol. 38 Issue (1): 138-145 
PDF
2017,
Vol. 38 Issue (1): 138-145 PDF