Junwen Wanga,
Xiaofeng Gaob,
Guoliang Chena and
Chuanmin Ding*a
aCollege of Chemistry & Chemical Engineering, Taiyuan University of Technology, Taiyuan 030024, P. R. China. E-mail: dingchuanmin@tyut.edu.cn; Fax: +86 0351 6014 498; Tel: +86 0351 6014 498
bSchool of Chemistry and Chemical Engineering, Liaocheng University, Liaocheng 252000, P. R. China
First published on 10th September 2019
For an industrial-scale catalytic process with a fixed or packed bed reactor, powder catalysts are not suitable because they may block the reaction pipe and increase the pressure of the reactor. Therefore, catalyst molding is essential for the industrial application of a catalyst. During the catalyst molding, binders are employed as indispensable additives that can achieve the mechanical strength requirements for industrial applications. However, the addition of binders may cover the activity sites of the catalyst and suppress the mass transfer of the reactants and products. So, traditional processes of catalyst molding significantly affect the catalytic performance. In this study, we proposed a vapor-phase-treatment to synthesize a pure shaped ZSM-5 zeolite with the re-crystallization of the binder incorporated silica sol and aluminum nitrate, which were converted into a part of ZSM-5 on a commercial H-ZSM-5 zeolite substrate. Subsequently, the shaped ZSM-5 catalyst was evaluated using the catalytic conversion of methanol to an aromatic (MTA reaction). The results showed that compared to the EPHZ catalyst, the SPHZ catalyst exhibited a long lifetime with a relatively high shape selectivity for methanol and aromatics. To rationalize these results and establish a structure–activity relationship, the zeolite catalysts were thoroughly characterized by XRD, NH3-TPD, FT-IR, N2 adsorption, TG, SEM, TEM, ICP and Al MAS-NMR. The results demonstrated that an interesting intra-particle pore structure was formed within the monoliths of the SPHZ catalyst. Moreover, the superior catalytic performance obtained for SPHZ may have also been due to the broad acid strength distribution and the conversion of the silicon aluminum adhesive agent to zeolite crystals.
However, all the above efforts mainly focused on zeolite powder rather than the shaped catalysts. For industrial scale applications, extrusion is the necessary step for the transition from powder to a technical body. A few academic studies that are related to the extrusion process have been widely neglected up to date.9–12 Although the vast majority of this knowledge comes from pure zeolite powder, the effect of shaping is still unknown. The extrusion process will add matrix materials that usually include natural clay, alumina or/and silica to gratify the mechanical, attrition resistance and thermal stability. They can exhibit beneficial or deleterious impacts on the catalyst reaction. However, the binders and other matrix materials are inert components for the catalyst in our subconscious mind. The physical and chemical effects of the extrusion process are the following: the modification of mass transfer characteristics, modification of porosity characteristics and improvement in the attrition strength. Their role is worthy of greater consideration than has hitherto been the case.13,14
Nevertheless, the process of compaction has a positive or adverse effect on the given system. When the catalyst was extruded with alumina binder, the ion-exchange ability and acid-catalytic activity of high-silica ZSM-5 were significantly enhanced as reported by Chang.15 Both, the external and intracrystalline acidity of the catalyst decreased due to the addition of a silica binder at a concentration of 10% or 50%. The increased mass transfer efficiency needs to weaken the influence from the added silica binder and exert a positive impact on the acid catalytic activity of the added aluminum binder. It is a challenge to kill two birds with one stone: one challenge is to achieve the molding purpose, and the other is to eliminate the binder residue on the catalyst surface. Recently, Zhou et al.16 only utilized silica sol as binders for a successful, organic template, steam-assisted recrystallization after the extrusion process. It was discovered that a superior activity, higher propylene selectivity, and a significantly longer life-span in the MTH reaction as well as the catalytic cracking of the C4 olefin propylene production were achieved. There existed a problem that the crystallization of the silica species into pure silica zeolite eliminated the effect of the residue matter on the mass transfer efficiency. However, the pure silica species did not affect the acidity. It is worth investigating how to transform silica species into a zeolite phase with activated species during the recrystallization process. In this present study, we achieved the conversion from the binder phase to the crystalline zeolite phase for activated sites on the surface of the catalyst by adopting added silica sol and a small amount of alumina nitrate as the binder. The finished products were employed in the MTA (methanol to aromatics) technology. The catalytic performance and catalytic lifetime are discussed in detail.
Scanning electron microscopy (SEM) images were recorded on a HITACHI SU-8010 microscope at 10 kV for investigating the morphology and crystal size of the samples. Transmission electron microscopy (TEM) images were recorded using a Tecnai G2 F20 instrument operating at 200 kV.
Nitrogen adsorption–desorption isotherms were measured on a Micromeritics ASAP 2020 instrument at 77 K. Prior to the measurement, the samples were degassed at 573 K under high vacuum for 4 h. The specific surface area was calculated using a BET model. The total pore volumes were calculated using the nitrogen adsorbed volume at a relative pressure of 0.99.
Solid state nuclear magnetic resonance with magic angle spinning (MAS NMR) was used to investigate the coordination of the Al atoms in the ZSM-5 framework. The measurements were performed on a Bruker DSX 300 wide-bore. The NMR spectrometer was equipped with two radio frequency (rf) channels and a spin speed of 15 kHz with 2.05 s intervals between the successive accumulations. The 27Al chemical shifts were reported relative to the Al(H2O)63+ solution.
Temperature programmed desorption measurements for NH3 (NH3-TPD) were performed on a micromeritics An auto chem TP5076 instrument was used for monitoring the amount and type. The sample (0.1 g, particle size: 20–40 meshes) was first pretreated at 600 K for 1 h under a flow of He, cooled to 293 K and blown with ammonia for 30 min to saturation. Then, the physically absorbed content was removed by switching the flowing gas to pure He for 2 h. Finally, the temperature programed desorption of chemically absorbed NH3 was conducted by raising the temperature at a rate of 10 K min−1 at 373 K for 873 K.
Brønsted acid sites and Lewis acid sites were determined using a WQF-510 Fourier transform infrared (FT-IR) spectroscope with a resolution of 4 cm−1. For IR analysis, the self-supported wafer was extruded from sample powder. A 1:300 KBr/adsorbed pyridine mass ratio was used at 60 °C for 30 minutes and then 200 °C for 20 minutes.
Thermogravimetric analysis (TGA) was performed on a Setaram Labsys TG-DTA/DSC 1600 instrument to detect the amount of coke formed on the used catalyst surface. The sample was raised from 300 K to 973 K with a speed of 10 K min−1 in the air. The weight loss of the catalyst was detected with the increase in temperature.
Scanning electron microscopy (SEM) was utilized to further demonstrate the structure and morphology of the above catalyst. As shown in the display in Fig. 2(a) and (b), the commercially-available parent ZSM-5 has a sharp surface with a typical uniform hexagonal shape. After the extrusion process, the ordered structure of zeolite disappeared and ZSM-5 was surrounded by numerous binders to form a block in Fig. 2(c) and (d). The silicate species of the binders restrained the access of the raw material to the zeolite active sites in that the shaped catalyst was almost covered by the binders. Similar crystallographic texture morphologies in Fig. 2(e) and (f) were observed. At the same time, the adhesive particles almost completely disappeared and only the zeolite crystals still existed, which suggested that the majority of the aluminosilicate species were translated into zeolite crystals.18
Fig. 3 shows evident structural differences among EPHZ and SPHZ by TEM characterization. The entire uniform crystallinity of the EPHZ sample (Fig. 3(a)) was covered by the binders. This seriously affected the mass transfer efficiency and changed the acid distribution, which was in accordance with the N2 adsorption and the NH3-TPD consequence. At the same time, there was no evident large spot observed for the EPHZ sample. For the SPHZ sample (Fig. 3(b)), no additional aluminum nitrate particles were observed, indicating that aluminum nitrate was successfully transformed to zeolite in the course of the recrystallization process. As we can see, there are numerous white dots around the crystal lattice, which indicated that the removal of the silicon species in the skeleton by alkali vapor led to lattice defects that resulted in the mesoporous or even hollow structure.19
As shown in the 27Al solid-state MAS NMR spectra in Fig. 4, the small peak detected at 0 ppm was attributed to the octahedral aluminum species of HZSM-5 and SPHZ. EPHZ exhibited a relatively high resonance at 0 ppm, which demonstrated that the majority of aluminum nitrate was in the form of non-framework aluminum species. After the steam-assisted process, some of the extra aluminum framework was transformed into the aluminum skeleton. The three samples indicated a relatively sharp resonance at 56 ppm, which was assigned to tetrahedral aluminum species in the zeolite framework. The above characterization results confirmed that all of the zeolite had a highly crystalline aluminosilicate structure with MFI frameworks. The resonances of the three samples at 56 ppm were in the order of SPHZ > HZSM-5 > EPHZ. The EPHZ framework had the lowest aluminum framework peak, which indicated that the decrease in the catalyst proportion affected the peak height of the aluminum skeleton. After the steam-assisted process, there were two main factors that affected the peak for SPHZ: the conversion of the binder increasing the proportion of the catalyst and that the aluminosilicate species transformed into a crystalline phase, which increased the skeleton of aluminum.
The N2 adsorption–desorption isotherms also demonstrated the structural and textural properties of the catalysts as shown in Fig. 5. The surface areas and micropore volumes for these catalysts are shown in Table 1. HZSM-5 displayed type I isotherms with a sharp uptake, which indicated that the HZSM-5 micropores were filled with molecular ammonia in the low-pressure region, suggesting that the pristine zeolite microcrystal was a micropore-dominant zeolite.20 The surface area of HZSM-5 was 366 m2 g−1 in total, where 90 m2 g−1 belonged to the external surface. The total pore volume of HZSM-5 was 0.17 mL g−1, which included 0.11 mL g−1 and 0.06 mL g−1 for the micropores and stacking pores, respectively. Compared with the HZSM-5 catalyst, the surface area of the EPHZ catalyst decreased to 251 m2 g−1, which may have been due to the binder covering the surface of the HZSM-5 catalyst. However, the mesopore volume increased to 0.08 cm3 g−1. A large number of mesopores were produced during the operation of catalyst molding. For SPHZ zeolite, the external surface area reached 359 m2 g−1 and the mesopore volume was 0.09 cm3 g−1. The SPHZ sample showed more obvious hysteresis loops, mainly due to alkali desilication. Thus, mesoporous intracrystalline was formed.
Sample | Total Al content (mg L−1) | SBET (m2 g−1) | Smicro (m2 g−1) | Smeso (m2 g−1) | Vtotal (cm3 g−1) | Vmicro (cm3 g−1) | Vmeso (cm3 g−1) | Radial grain crushing strength (N/particle) |
---|---|---|---|---|---|---|---|---|
HZSM-5 | 2.26 | 366 | 276 | 90 | 0.17 | 0.11 | 0.06 | 44.6 |
EPHZ | 1.76 | 251 | 172 | 79 | 0.14 | 0.06 | 0.08 | 22.5 |
SPHZ | 2.32 | 359 | 273 | 86 | 0.19 | 0.10 | 0.09 | 43.1 |
The surface amount and strength of the H-form zeolite acid type were determined by the temperature programmed desorption of ammonia (NH3-TPD) characterization, as presented in Fig. 6. All three samples exhibited two well-resolved typical desorption peaks. The temperature regions 440–470 K and 620–670 K were referred to the low temperature and high temperature peaks, respectively.21 However, the low-temperature peak was ascribed to weakly acidic silanol groups that covered the catalyst surface, which were reduced by extending the evaluation time. The weak adsorption sites were inactive for the MTA reaction. Therefore, we mainly concentrated on the change in the high-temperature peak. As can be seen, the intensity of the high temperature desorption peak of EPHZ dramatically decreased after the extrusion process, which indicated that the acid sites on the surface of the catalyst were heavily covered. Obviously, the zeolite recrystallization displayed the highest desorption peak due to the disappearance of the silicate binder coverage over the catalyst surface and the generation of the small grain zeolite increasing the acid sites on the surface of the catalysts. There were huge acidic differences between EPHZ and SPHZ, which indicated that the effect of acidity on the performance could not be ignored between the samples.
The type and concentration of the Brønsted and Lewis acid sites for the HZSM-5, EPHZ and SPHZ samples were determined by Py-IR at a desorption temperature of 200 °C. The result is depicted in Fig. 7. The peaks at 1540 and 1450 cm−1 corresponded to pyridine interacting with the Brønsted and Lewis acid,22 respectively. Moreover, the peak at 1490 cm−1 was ascribed to both Brønsted and Lewis acids. Apparently, the relative amount of the Brønsted acids for the different samples was as follows: SPHZ > HZSM-5 > EPHZ. After the molding process, the total number of acid sites markedly decreased for NH3-TPD, as shown in Fig. 6. The L acid sites and B acid sites exhibited the same trend. In fact, the trend was more likely due to the binder covered or even encapsulated the catalyst that affected the acidity of the catalyst. This result suppressed the diffusion of the pyridine molecules. At the same time, SPHZ exhibited the highest B acid sites, which may have been due to the effects of the ordered structure after the steam-assisted procedure and newly grain zeolite crystals were produced.
Built on the above discussion and in previous studies, we showed the formation of full-crystalline monolithic zeolite in the form of schematics in Fig. 8. First, saturated SDA steam condensed into a liquid phase at the top of the reactor and the liquid then dispersed around the catalyst. Silica sol and aluminum nitrate were dissolved by the binder. Meanwhile, the catalyst formed mesopores under alkali vapor. With the dissolved silica sol as the silicon source, the dissolved aluminum nitrate as the aluminum source as well as the condensed SDA as the basic source and structural guide agent, tiny grains catalyst began to form. Finally, the regular catalyst formed around the near-no binder with a large number of intercrystalline mesopores. At the same time, the catalyst crystal increased to some extent after the recrystallization process. This was consistent with the results of transmission electron microscopy.
Fig. 9 Time course for the liquid hydrocarbons on HZSM-5, EPHZ and SPHZ. Reaction conditions: 400 °C, 0.1 MPa, WHSV = 1 h−1. |
The coke on the above used HZSM-5, used EPHZ and used SPHZ were analyzed by TG. Since the molecular size of 1,3,5-trimethylbenzene was close to the pore size of ZSM-5 zeolite with a boiling point of 165 °C, it may have been the initial carbon deposition molecule for this series of ZSM-5 catalysts during the aromatic synthesis from methanol.26 As shown in Fig. 10, by careful observation, a slight weight loss existed below 100 °C, which corresponded to the volatilization of the moisture. Meanwhile, the decomposition of dissolved carbon ranged from 165–500 °C and the decomposition of insoluble carbon with a high molar ratio of C/H was in the range of 500 °C to 750 °C. The coke formation rate for used EPHZ was up to 1.7 × 10−3 g gcat−1 h−1, while that for SPHZ was only 1.03 × 10−3 g gcat−1 h−1 and HZSM-5 was 1.68 × 10−3 g gcat−1 h−1. Moreover, the result of the N2 adsorption and XRD for SPHZ indicated that the porous structure of this catalyst was not destroyed during its recrystallization process. As a whole, SPHZ with a longer lifetime showed lower coke formation. Actually, the deposition rate was strongly affected by the structural and textural properties of the catalysts. The large external surface area indicated that the large pore opening was exposed to the outer surface, which was bound to shorten the diffusion path. The coke precursors easily migrated from the channel to the outer surface. For the HZSM-5 formed by tableting, there was no binder to influence the diffusion and the carbon deposits also easily diffused out of the pores.
The most approved model for the hydrocarbon pool mechanism explains the selectivity for a particular shape in methanol during the hydrocarbon reaction. The most important feature of the hydrocarbon route is that methanol reacts with the hydrocarbon pool species through a series of steps to produce olefin and regenerated hydrocarbon over the course of the reaction cycle. We wanted to explore whether olefin was the basic unit of the aromatics. From Fig. 10, we observed that the aromatics increased when olefin was reduced. This also supported the aromatization in the propene experiments for Cu/ZSM-5 zeolite. The formation of the aromatic hydrocarbons using olefin as the basic unit was proven by the methanol feed with propylene.27 As presented in the Table 2, the BTX selectivity of SPHZ increased with time and reached a maximum of 34.84%. HZSM-5 and EPHZ reached a maximum of 32.69% and 31.65%, respectively. SPHZ exhibited a larger C9+ than EPHZ due to the excess alkylation of the xylenes on external surface. It was difficult to generate in the channel because the trimethylbenzene diameter was 0.61 nm and it came directly from low-carbon xylene. Xylene was the main component in the product for EPHZ and SPHZ. The xylene content was similar to the total yield for benzene, toluene and C9+. The poor diffusion capability resulted in the formation of a higher number of C1–C5 hydrocarbons. Due to the steam assisted process, SPHZ exhibited an excellent diffusion capability. Therefore, SPHZ had less C1–C5 hydrocarbons than EPHZ. For HZSM-5 by tableting molding, the aromatic selectivity was higher than EPHZ due to the absence of the binder.
Catalyst | Selectivity of hydrocarbons/% | SBTX/% | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
CH4 | C2b | C03c | C3d | C04 | C4 | C5 | C6–8e | Bf | Tg | Xh | EBi | C9+j | COXk | ||
a Reaction conditions: 0.1 MPa, 400 °C, WHSV = 1 h−1.b C2 hydrocarbons.c C3 alkane.d C3 alkene.e C6–8 aliphatic hydrocarbons.f Benzene.g Toluene.h Xylene.i Ethylbenzene.j C9 and C9+ hydrocarbons.k CO and CO2. | |||||||||||||||
EPHZ | 2.29 | 2.42 | 12.04 | 2.32 | 29 | 3.29 | 2.95 | 3.52 | 2.26 | 10.73 | 18.66 | 1.12 | 8.73 | 0.41 | 31.65 |
HZSM-5 | 2.34 | 2.36 | 11.62 | 2.15 | 28.73 | 3.22 | 2.86 | 3.58 | 2.34 | 11.09 | 19.26 | 1.01 | 8.80 | 0.34 | 32.69 |
SPHZ | 2.18 | 2.34 | 10.86 | 2.33 | 27.79 | 3.20 | 2.65 | 3.41 | 2.45 | 12.33 | 20.06 | 0.93 | 9.57 | 0.28 | 34.84 |
This journal is © The Royal Society of Chemistry 2019 |