1 Introduction
Light olefins (LOs, including ethylene (C2 = ), propylene (C3 = ), and butylene (C4 = )) are important platform chemicals for producing cosmetics, lubricants, rubbers, detergents and polymers [1], and the current demand for LOs is predicted to reach $329.30 billion by 2028, growing at a rate of 4.5% from 2020 [2,3]. However, the supply of LOs is limited by traditional production routes such as steam cracking. Therefore, enhancing the production of LOs via optimising traditional technologies (such as catalytic cracking) is regarded as the alternative to meet the high market demand of LOs globally. One of the conventional technologies to produce LOs worldwide is the fluidized catalytic cracking (FCC) process using naphtha which allows for 22% of global LOs production [4]. Yet, the development of LOs-selective FCC catalysts is needed to increase the FCC capacity for LOs production.
Zeolites, especially Zeolite Y (ZY) and ZSM-5, are key components in FCC particles since 1960 [5], whose formulation is based on several components including zeolites, clay, binder and filler. In the matrix in FCC particles ZY is employed to increase liquefied petroleum gas, olefins, and gasoline by primary cracking of heavy liquid paraffin through the 12 membered rings (MR) in its large-pore framework. Comparatively, addition of ZSM-5 is recognized to be able to increase the yield of LOs because its medium 10 MR pore size favors the monomolecular mechanism to promote the selectivity to LOs [6].
Zeolites have microporous frameworks, providing the essential acid active sites and shape selectivity needed for high performance in hydrocarbon cracking reactions [7]. However, the micropores can impose diffusion limitation for bulky reactants, reducing catalytic effectiveness and causing deactivation [8]. To address this issue, strategies focusing on engineering hierarchical meso-micro-porous zeolites and/or composites were developed, exemplified by hierarchical ZY by different post-synthetic treatments (such as chemical treatment using chelating agents [9]) and ZSM-5/MCM-41 composite [10]. Another possible strategy is physically mixing microporous zeolites with mesoporous silicas, such as KIT-6, MCM-41, and SBA-15, to prepare the composite catalysts, in which the presence of mesoporous silicas in the FCC particles can enhance the mass transfer within the FCC particles and/or pre-crack the reactant molecules, then zeolites can do the further cracking with the improved selectivity. The latter approach is flexible allowing independent tuning of the properties of each component to optimize the activity/selectivity of the resulting composite catalyst.
Here, we explore the effect of composite catalyst formulation on the performance of catalytic cracking with the aim of improving the production of LOs of C2 = , C3 =, and C4 = . The composites catalysts were formulated using two zeolites (of ZY and ZSM-5) and three ordered mesoporous silicas (of KIT-6, SBA-15, and MCM-41) by varying the weight percent of the zeolites and maintaining the weight percent of silicas at 60 wt %. The prepared composite catalysts were characterized comparatively using various techniques to understand their physicochemical properties and assessed using catalytic cracking of n-heptane (n-C7) (as the model compound of naphtha). In addition, a pulsed-field gradient nuclear magnetic resonance (PFG-NMR) study of the selected composite catalysts was conducted to know the molecular diffusion properties, which was correlated with the catalytic performance.
2 Experimental
2.1 Materials and Chemicals
ZSM-5 (CBV 2314) and ZY (CBV 712) zeolites, both in NH3 form (with the silicon to aluminum ratio (Si/Al) of 11.5 and 6, respectively) were purchased from Zeolyst International, and the mesoporous MCM-41 was purchased from ACS Materials. The zeolites were first calcined at 550 °C for 5 h, heating from ambient at 2 °C·min–1, to convert to the H-form counterparts. HCl (12 mol·L–1), tetraethoxysilane (TEOS, 99%) and n-butanol (n-BuOH, 99.5%) were purchased from Fisher Scientific. Poly (ethylene glycol)-block-poly (propylene glycol)-block-poly (ethylene glycol) PEO20-PPO70-PEO20 (P123 with MW = 5800, Aldrich) was used as the direct-structuring agent for synthesizing the mesoporous silicas of SBA-15 and KIT-6.
2.2 Synthesis of mesoporous SBA-15 and KIT-6 silica
Mesoporous SBA-15 was synthesized using the procedure reported elsewhere [11]. In detail, 10 g of P123 was dissolved in the HCl solution (containing 74.5 mL of deionized water and 291.5 mL of 12 mol·L–1 HCl) under stirring (at 300 r·min–1) at 35 °C for 15 min (the temperature was maintained by an oil bath). Then, 20 mL of TEOS was added dropwise for 5 min under 300 r·min–1 at 35 °C to form a homogeneous liquid, and the mixture was stirred continuously at 35 °C and 300 r·min–1 for 20 h in the sealed flask. The resulting mixture was transferred to a sealed 1000 mL round-bottomed flask with a condenser attached and placed in an oil bath for 24 h at 80 °C under static conditions. After the synthesis, the solid product was filtered and dried at 80 °C for 12 h without washing, followed by calcination at 500 °C for 6 h (ramp rate = 1 °C·min–1).
Mesoporous KIT-6 was prepared following the procedure reported in the literature [12,13]. The synthesis first involves dissolving the surfactant P123 (~10 g) in the n-BuOH-HCl mixture containing 366 mL of deionized water, 12.3 mL of n-BuOH and 16.3 mL of 12 mol·L–1 HCl in a polypropylene bottle. The mixture was stirred at 300 r·min–1 and kept at 35 °C for 15 min. Then TEOS (20.0 mL) was added dropwise for 5 min under 300 r·min–1 at 35 °C to form a homogeneous liquid, and the mixture was stirred continuously at 300 r·min–1 for 20 h at 35 °C in a sealed 500 mL flask that was placed in an oil bath. The resulting suspension was transferred to a sealed 1000 mL round-bottomed flask (same as above in the synthesis of SBA-15) for static hydrothermal synthesis for 24 h at 80 °C. Finally, the resulting solid product was filtered, and dried at 80 °C for 12 h without washing, followed by calcination at 500 °C for 6 h (ramp rate = 1 °C·min–1).
2.3 Formulation of the composite catalysts
The physical mixtures of ZY/ZSM-5/X, where X = KIT-6, SBA-15 or MCM-41, were prepared by stirring the three components in deionized water for 3 h at ambient temperature. The weight percent of the components are detailed in Tab.1, and the ratio of deionized water to the mixture was 1.5:1 w/w. After the 3 h mixing, the resulting slurry was dried in an oven at 120 °C overnight, then the resulting dry solid was calcined at 550 °C for 6 h (ramp rate = 2 °C·min–1). Finally, the composites were pelletized (with the particle sizes of 100 to 450 µm).
Tab.1 Proportion of components in the formulated composite catalysts |
| Composite catalyst | Weight percent of the component/wt % | ||
|---|---|---|---|
| ZY | ZSM-5 | Mesoporous silica | |
| ZY/ZSM-5/KIT-6(35:5:60) | 35 | 5 | 60 |
| ZY/ZSM-5/KIT-6(30:10:60) | 30 | 10 | 60 |
| ZY/ZSM-5KIT-6(20:20:60) | 20 | 20 | 60 |
| ZY/ZSM-5/SBA-15(35:5:60) | 35 | 5 | 60 |
| ZY/ZSM-5/SBA-15(30:10:60) | 30 | 10 | 60 |
| ZY/ZSM-5/SBA-15(20:20:60) | 20 | 20 | 60 |
| ZY/ZSM-5/MCM-41(35:5:60) | 35 | 5 | 60 |
| ZY/ZSM-5/MCM-41(30:10:60) | 30 | 10 | 60 |
| ZY/ZSM-5/MCM-41(20:20:60) | 20 | 20 | 60 |
2.4 Characterization of materials
Powder X-ray diffraction (PXRD) of the composite catalysts was carried out using a portable desktop XRD instrument (D2 PHASER, Bruker) with Cu kα radiation (λ = 1.5406 Å) at 30 mA and 40 kV. The samples were scanned with a 2θ interval of 5° to 90°, a step size of 0.02° and a step time of 8 s. Low-angle PXRD was performed using a D8 ADVANCE diffractometer (Bruker) instrument using a Cr Ka radiation (λ = 1.5406 Å) at 30 mA and 40 kV. The samples were scanned with a 2θ interval of 0.5° to 5°, a step size of 0.02° and a step time of 8 s. An energy-dispersive X-ray fluorescence (EDXRF) spectrometer was conducted using the PANalytical MiniPal 4 EDXRF instrument for elemental composition analysis operating at 30 kV to calculate the bulk Si/Al ratio of materials. N2 adsorption and desorption isotherms of materials were measured using a Micromeritics ASAP 2060 physisorption analyzer. The sample (about 150 mg) was degassed at 300 °C under a vacuum (0.5 mmHg) for 4 h before the physisorption analysis at the liquid nitrogen temperature of –196 °C (77 K). The specific surface area of the materials was determined using the Brunauer-Emmett-Teller (BET) method, and the pore size distribution (PSD) was estimated by the Barrett-Joyner-Halenda (BJH) method based on the desorption isotherm. Fourier-transform infrared spectroscopy (FTIR) was carried out using a Bruker Vertex 7.0 spectrometer. The spectra were collected with a spectral resolution of 4 cm–1, 32 scans, and over a wavenumber range of 4000–400 cm–1. Scanning electron microscopy (SEM) of the morphology of the composite catalysts was conducted using Tescan Vega 3 at 20 kV and under high vacuum pressure (10−6 to 10−7 torr). The working distance was 10–15 mm. Transmission electron microscopy (TEM) was conducted using an FEI Tecnai G2 20 microscope operated at an accelerating voltage of 200 kV and equipped with an Oxford Instruments X-Max TLE 80 EDX detector and a Gatan Orius CCD camera. Coke deposition on the used catalyst after the catalytic experiments was assessed by thermogravimetric analysis (TGA, Q600 TGA-DSC, TA Instruments). The temperature program and gas atmosphere were (i) under N2 at 40 cm3·min–1 from room temperature to 650 °C, isothermally held for 60 min at 200 °C, 400 °C, and 650 °C, respectively, (ii) at 650 °C switching to air at 40 cm3·min–1 for 3 h.
2.5 PFG-NMR analysis
To study the diffusion of n-C7 in the composite catalysts, PFG-NMR measurements were performed. In detail, the dried catalyst sample (~100 mg, dried at 120 °C for 3 h) saturated using n-C7 for 48 h according to the reported procedure [14]. To remove the excess liquid from the sample, the saturated catalyst was gently dried using the filter paper. Then the catalyst was transferred to an NMR tube (of 5 mm diameter) to a depth of 25 mm, to fit a similar height of the radiofrequency coil in the probe of the spectrometer. The tube was sealed with polytetrafluoroethylene tape and left for 1 h to achieve thermal equilibrium before the measurement. PFG-NMR measurements were conducted at a higher magnetic field strength employing a Bruker spectrometer at 500 MHz with a diffusion probe to measure the relaxation times of T1 and T2 of the probe molecules in the sample. Measurements were conducted at ambient temperature (20 °C), employing stimulated echo using bipolar gradient pulse pair sequences for diffusion measurements, a 15 μs 1H 90° pulse, a 1.4 ms sinusoidal gradient pulse width (δ), varying gradient pulse amplitudes ranging from 2% to 15% of 15 Tm–1 across 16 increments at diffusion intervals (Δ) of 100 ms.
2.6 Catalytic cracking
Catalytic cracking using different catalysts was performed using a fixed bed reactor. The reactor is a stainless-steel tube (inner diameter of 9 mm and outer diameter of 14 mm) packed with ~0.6 g catalyst (pelletized, 100–450 µm), and quartz wool and glass beads were used to support the catalyst bed. The catalyst was activated in situ by purged with air (50 mL·min–1) and heating the reactor to 550 °C at 10 °C·min–1 for 30 min, followed by N2 (70 mL·min–1), at atmospheric pressure for 1 h. n-C7 was then introduced to the reactor using N2 as the carrier gas (passing through three bubblers in series in a chiller at –2 °C) to initiate the reaction. The schematic of the experimental rig is shown in Fig.1. The cracking reactions were typically conducted for 4 h at atmospheric pressure with a weight hourly space velocity (WHSV) of 0.41 h–1, and the stability test was conducted for 12 h.
Fig.1 Schematic of the experimental rig for performing the catalytic cracking reaction. 1. pressure regulators for N2 and air; 2. valves; 3. mass flow controllers; 4. bubblers; 5. chiller bath; 6. pressure gauge; 7. pressure relief valve; 8. reactor tube; 9. furnace; 10. glass wool; 11. catalyst bed; 12. glass beads; 13. GC; 14. data collector. |
Gas samples from the outlet of the reactor were taken every hour for gas chromatography (GC) analysis using a Varian 3800 GC equipped with a Restek PLOT Al2O3 KCl capillary column (50 m × 5 mm × 0.32 µm) and a flame ionization detector. The temperature program of the oven is: isothermal at 70 °C for 5 min, ramp to 225 °C at 10 °C·min–1 then isothermal at 225 °C for 30 min. The GC was calibrated using a standard gas mixture with a concentration of 1% mol of C1–C5 (BOC Linde Group). Based on the peak area from a GC chromatogram, the conversion (), selectivity (), and yield () were calculated using Eqs. (1–3).
where and are the mole of n-C7 in the inlet and outlet stream of the reactor, respectively, CxHy is the total mole of LOs (C2 = , C3 =, and C4 = ) produced.
3 Results and discussion
3.1 Physicochemical properties of the materials under investigation
Small-angle X-ray patterns of the commercial MCM-41 and the as-prepared SBA-15 and KIT-6 mesoporous silicas (Fig.2(a)) confirm their well-ordered frameworks with uniform mesopores as evidenced by the relevant characteristic diffraction peaks. In details, for MCM-41, the peaks at 2θ = 2.3° (100), 3.8° (110), and 4.4° (200) reflections, corresponding to the reflections of the 2D hexagonal (space group p6mm) symmetry [15]. The diffraction pattern of SBA-15 shows the typical reflections in the 2θ range of < 5° with the major peak at about 1.1° (corresponding to the (100) plane) and two insignificant humps at 2θ about 1.8°, and 2.2° (corresponding to the (110) and (220) planes, respectively) [16,17], confirming the hexagonal ordered structure. The low-angle diffraction pattern for KIT-6 shows the distinct peak at 2θ of 1.05°, being assigned to the (211) plane [18], while other relevant diffraction peaks assignable to the (220), (400), and (332) planes are not well resolved.
Fig.2(b) and Fig.2(c) show the wide-angle XRD patterns of the formulated composite catalysts, in which the characteristic diffraction peaks of the ZSM-5 MFI crystal structure (indicated by the star symbols) and that of the ZY FAU crystal structure (indicated by the nabla symbols), respectively) can be identified clearly. The results suggest that after the formulation synthesis the frameworks of ZY and ZSM-5 remained intact. In additional, the amorphous humps at around 2θ of 20°–30° were identified in the composites, confirming the presence of KIT-6, SBA-15, and MCM-41 in them [19]. Interestingly, the diffraction peak intensities of ZY and ZSM-5 do not correspond to the variation in their weight percent in the composite catalysts, which could be attributed to the heterogeneous distribution of the components in the samples used for XRD characterization.
The textural properties of the mesoporous silicas and composite catalysts were examined by N2 physisorption, as shown in Fig. S1 (cf. Electronic Supplementary Material, ESM) and Tab.2. N2 adsorption-desorption isotherms and BJH PSDs of the mesoporous silicas (KIT-6, MCM-41, and SBA-15) are shown in Fig. S1, displaying the typical IV isotherms with hysteresis loops, which proves the presence of uniform mesopores with narrow PSD. MCM-41 shows a narrow PSD centring at about 1.6 nm, the KIT-6 synthesized in this study shown a larger mesopore size of about 2.4 nm. Regarding the synthesized SBA-15, its PSD is widest ranging from 2 to 4 nm. All three mesoporous silicas show the high specific external surface areas (Sext) and mesopore volumes (Vmeso). After the formulation with different proportions of ZY and ZSM-5 (Fig. S2, cf. ESM), the total specific external surface areas (SBET) and mesopore volumes of the resulting composites decreased by 10%–20%, while the micropore surface areas (Smicro) and micropore volumes (Vmicro) were enhanced by 50%–55%, confirming the improved microporosity provided by the zeolites. The findings demonstrate that the formulated composites consist of both well-developed mesoporous and microporous properties, which could potentially benefit their performance in catalysis, outperforming the individual materials
Tab.2 Textural properties of the mesoporous silicas and the composite catalysts |
| Materials | Specific surface areaa)/(m2·g–1) | Specific surface areab)/(cm3·g–1) | Si/Alc) | |||||
|---|---|---|---|---|---|---|---|---|
| SBETa) | Sextb) | Smicroc) | Vtotald) | Vmicroe) | Vmesof) | |||
| KIT-6 | 755 | 559 | 202 | 0.34 | 0.11 | 0.23 | – | |
| MCM-41 | 810 | 767 | 42 | 0.33 | 0.02 | 0.31 | – | |
| SBA-15 | 755 | 442 | 312 | 0.66 | 0.16 | 0.5 | – | |
| ZY/ZSM-5/KIT-6(35:5:60) | 601 | 418 | 183 | 0.30 | 0.1 | 0.2 | 48.4 | |
| ZY/ZSM-5/KIT-6(30:10:60) | 585 | 351 | 234 | 0.27 | 0.12 | 0.15 | 47.5 | |
| ZY/ZSM-5/KIT-6(20:20:60) | 541 | 331 | 211 | 0.25 | 0.11 | 0.14 | 42.1 | |
| ZY/ZSM-5/MCM-41(35:5:60) | 726 | 564 | 161 | 0.31 | 0.08 | 0.23 | 36.7 | |
| ZY/ZSM-5/MCM-41(30:10:60) | 705 | 535 | 169 | 0.31 | 0.09 | 0.22 | 36.2 | |
| ZY/ZSM-5/MCM-41(20:20:60) | 718 | 583 | 135 | 0.31 | 0.07 | 0.24 | 36.3 | |
| ZY/ZSM-5/SBA-15(35:5:60) | 604 | 391 | 212 | 0.30 | 0.11 | 0.19 | 16 | |
| ZY/ZSM-5/SBA-15(30:10:60) | 535 | 301 | 234 | 0.25 | 0.12 | 0.13 | 20.4 | |
| ZY/ZSM-5/SBA-15(20:20:60) | 577 | 368 | 210 | 0.30 | 0.11 | 0.19 | 26.7 | |
a) Total surface area (SBET) by the BET method, external surface area (Sext) = SBET – Smicro, microspore surface area by the t-plot method; b) total pore volume at P/P0 (Vtotal) = maximum pore volume, microspores volume (Vmicro), mesoporous volume (Vmeso); c) by XRF; d) by the BJH method. |
The N2 physisorption isotherms of the formulated composites are shown in Fig. S2, and that of the three selected samples is shown in Fig.3. The Type IV isotherms (Fig.3(a)) suggest the overall mesoporous structure since the main components in them are mesoporous silica. The BJH PSDs (Fig.3(b)) show that the ZY/ZSM-5/KIT-6 (20:20:60) and ZY/ZSM-5/SBA-15 (20:20:60) composites possess similar pore sizes centered at about 2.4 nm, while that for ZY/ZSM-5/MCM-41 (20:20:60) is narrower centered at about 1.6 nm. The results suggest that these composite catalysts contain the mixed mesoporosity and microporosity, potentially benefiting the accessibility of the active sites (in zeolites) to reactant molecules and mass transport within them. For the same category composites, the PSDs (Fig. S2) are similar without being affected significantly by the variation in the weight percent of the two zeolites.
The SEM micrographs (Fig. S3, cf. ESM) illustrate the morphology and microstructure of the selected composite catalysts of ZY/ZSM-5/KIT-6 (20:20:60), ZY/ZSM-5/SBA-15 (20:20:60), and ZY/ZSM-5/MCM-41 (20:20:60). Figure S3(a) shows the heterogenous distribution of ZY, ZSM-5, and KIT-6 particles in ZY/ZSM-5/KIT-6 (20:20:60), and the large irregular-shaped crystals of KIT-6 with rough terraces could be identified [20]. Similar heterogenous microstructures were found for the ZY/ZSM-5/SBA-15 (20:20:60) and ZY/ZSM-5/MCM-41 (20:20:60) composites as well (Figs. S3(b) and S3(c)), with the former having the presence of short-rod particles (for SBA-15 [17]) and the latter having the spherical particles smaller than 0.5 µm (for MCM-41 [21]). For the ZY/ZSM-5 crystals, they cannot be identified clearly in the composite catalysts by SEM. TEM characterization of the composite catalysts was conducted. As shown in Fig.4(a), the parallel mesoporous channels of KIT-6 [22] can be clearly seen in ZY/ZSM-5/KIT-6 (20:20:60). In Fig.4(b), the overlapped phases of SBA-15 mesoporous silica and ZSM-5 zeolite were found in ZY/ZSM-5/SBA-15 (20:20:60). The amorphous SBA-15 is visible next to the crystalline ZSM-5 phase. However, the ordered mesoporous structure of SBA-15 is less clear. Like ZY/ZSM-5/SBA-15 (20:20:60), similar microscopic structures were found as well in ZY/ZSM-5/MCM-41 (20:20:60), showing the presence of both crystalline zeolite phases and amorphous MCM-41 phase (Fig.4(c)).
FTIR spectra of the zeolites and mesoporous silicas are shown in Fig. S4(a) (cf. ESM). All mesoporous silicas show the characteristic vibration bands of Si–O–Si, including the asymmetric, symmetric stretching vibrations, bending vibrations at ~1064 cm–1 [23], ~795 cm–1 [22], and ~434 cm–1 [24], respectively. For zeolites, it shows two bands, which are related to the external linkage banding vibration of Al–O4 tetrahedra at ~608 cm–1 [25] and ~529 cm–1 [26,27], depending on the type of zeolite frameworks (i.e., FAU or MFI). The IR bands at ~1243 cm–1, ~811 cm–1, and ~608 cm–1 [26,28], are associated with the internal linkage symmetric stretching and asymmetrical stretching in the zeolite framework, respectively. The IR spectra of the formulated composites are shown in Figs. S4(b–d). The composites show a decrease in the intensity of the characteristic framework bands, indicating the possible interaction between the mesoporous silicas and the zeolites. The decrease in the intensity of the bands related to the zeolitic frameworks (of ~612 cm–1–466 cm–1 and ~1383 cm–1–810 cm–1) in the composites could be due to the dilution by the mesoporous silica (with the weight percents of 60%). In addition, the significant presence of OH stretching bands (e.g., the band at about 3410 cm–1) suggests the water adsorption in the composites after formulation.
3.2 Catalytic cracking of n-Hptane (n-C7)
Catalytic cracking of n-C7 was first performed using the pristine zeolites and mesoporous silicas, and the conversion results as a function of time-on-stream (ToS) are shown in Fig. S5(a) (cf. ESM). Mesoporous silicas are not capable of cracking n-C7, showing comparable insignificant n-C7 conversions (about 2%). Conversely, over the two zeolites the initial n-C7 conversions were considerable at about 95%. The ZSM-5 zeolite (Si/Al = 11.5) was very stable with only 5% loss in activity over 4 h on stream. Comparatively, the ZY zeolite (Si/Al = 6) deactivated continuously with the final n-C7 conversion at 71%. This could be due to the low Si/Al of ZY which contributed to coking during catalytic cracking.
Figure S5(b) shows the total yield of LOs from the system employing different frameworks. Again, the mesoporous silicas showed insignificant yields of LOs. ZY presents an initial higher total LOs yield of 41%, which decreased gradually to about 22% after 4 h on stream. Conversely, ZSM-5 shows a very stable total LOs yield over the catalytic assessment (at about 22.6%). Detailed distributions of the selectivity to different LOs are shown in Fig. S6 (cf. ESM). In detail, ZY shows the initial higher selectivity to C2 = and C3 = , which gradually decreased to about 4.5% and 20%, respectively, while the selectivity of ZY to C4 = was rather stable (at about 6.7%, Fig. S6(a)). Overall, ZY was selective to C3 = and C4 = due to its intrinsic large 12-MR pores [29], which allows heavier hydrocarbons to accumulate, causing the catalyst’s deactivation [30]. Conversely, ZSM-5 was more selective to C2 = (at about 13.5%), followed by C3 = (at about 9%), and produced rather insignificant C4 = , as shown in Fig. S6(b). ZSM-5 can constrain the transition state and increase C2 = and C3 = olefin production due to its three-dimensional pore network and 5.4–5.6 nm ring openings [31]. In addition, compared to the ZY, ZSM-5 was more stable due to the smaller pore sizes and less acidic nature [32]. Under the same condition, catalytic cracking of n-C7 over different composites was conducted comparatively, and the results are shown in Fig.5.
It is worth noting that, for a specific silica, variation in the formulation of zeolites caused the significant effect on the catalyst activity of the composite catalysts. Taking the ZY/ZSM-5/KIT-6 composites as the example, an increase in the weight percent of ZSM-5 from 5% to 20% leads to the significant increase in n-C7 conversion, that is, ~35.4% for ZY/ZSM-5/KIT-6 (35:5:60) vs ~82.4% for ZY/ZSM-5/KIT-6 (20:20:60), as well as the total selectivity to LOs, i.e., from ~15.8% to ~27%. Interestingly, similar trends were measured for the other two sets of composite catalysts based on SBA-15 and MCM-41, i.e., the composites with 20 wt % ZSM-5 showed the best catalytic cracking performance. The findings demonstrate the important role played by the ZSM-5 zeolite on the activity of the composite catalysts.
By comparing the catalytic performance of the composites with a fixed formulation of zeolites, the effect of the mesoporous silicas on the catalysis was identified. For the composites with 35 wt % ZY and 5 wt % ZSM-5, the steady-state n-C7 conversion follows the order of ZY/ZSM-5/MCM-41(35:5:60) > ZY/ZSM-5/SBA-15 (35:5:60) > ZY/ZSM-5/KIT-6 (35:5:60). For the composites with 30 wt % ZY and 10 wt % ZSM-5, ZY/ZSM-5/MCM-41 (30:10:60) again shows the highest n-C7 conversion at about 83.4%, conversely, the ZY/ZSM-5/SBA-15 (30:10:60) composite achieved the lowest n-C7 conversion at about 40%. The better performance of the composites containing 5% ZSM-5 and 10% ZSM-5 with MCM-41 is likely a result of the tight range of mesopores size and also the smaller size of the mesopores (typically 0.8–1.2 nm, cf. ESM). The ZSM-5 and the ZY crystals are typically aggregated between 5 to 2500 nm (typically 250–500 nm) and hence will most likely be coating the mesoporous silicas and not inside any pore structure. Therefore performance will be related to the morphology of the individual silicas (e.g., MCM-41, KIT-6, and SBA-15).
For the composites with the equal 20 wt % ZY and ZSM-5 formulation, ZY/ZSM-5/SBA-15 (20:20:60) showed that highest n-C7 conversion at about 92% and total selectivity to LOs, as shown in Fig.5. The addition of 20% ZSM-5 increased activity significantly and the overall conversion of all three composite catalysts was comparable (with conversions in the range 85% ± 5%) suggesting that the support has less influence compared to increase in ZSM-5 content.
The results shown here demonstrate the importance of balancing acidity (from zeolites) and mesopore space (from silicas) to optimize the catalytic cracking performance of the composites. Overall, ZY/ZSM-5/SBA-15 (20:20:60), ZY/ZSM-5/MCM-41 (20:20:60), and ZY/ZSM-5/KIT-6 (20:20:60) were identified as the top three best performed composite catalysts under investigation, which were further assessed.
Tab.3 shows the detailed comparison of the catalytic performance of ZY/ZSM-5/KIT-6 (20:20:60), ZY/ZSM-5/SBA-15 (20:20:60), and ZY/ZSM-5/MCM-41 (20:20:60). ZY/ZSM-5/KIT-6 (20:20:60) and ZY/ZSM-5/SBA-15 (20:20:60) have the similar olefin-to-paraffin ratio (O/P) of about 0.7, while ZY/ZSM-5/MCM-41 (20:20:60) favors LO production based on its higher O/P ratio of 1.1. Among all the catalysts tested, the ZY/ZSM-5/SBA-15 (20:20:60) catalyst shows higher selectivity to LOs (26%). The hydrogen transfer coefficient (HTC) measures a catalyst’s efficiency in facilitating hydrogen atom transfer between molecules, and the HTC of the three catalysts was comparable since the zeolite formulation in them was the same. Overall, results presented in Tab.3 show that the details of product information of the systems employing ZY/ZSM-5/SBA-15 (20:20:60) and ZY/ZSM-5/KIT-6 (20:20:60) are similar, while that of the system over ZY/ZSM-5/MCM-41 (20:20:60) is slightly different.
Tab.3 Averaged activity and selectivity data of n-C7 cracking over the selected composite catalysts at 550 °C, 0.4 h–1 and atmospheric pressure |
| Composite catalyst | Conversion | C2=–C4=a) | C2=/C3= | C2=/C2 | C3=/C3 | C4=/C4 | O/Pb) | HTCc) |
|---|---|---|---|---|---|---|---|---|
| % | – | |||||||
| ZY/ZSM-5/SBA-15(20:20:60) | 84 | 26.2 | 0.8 | 2.5 | 0.6 | 0.4 | 0.7 | 0.7 |
| ZY/ZSM-5/KIT-6(20:20:60) | 94 | 26.6 | 0.6 | 2.5 | 0.7 | 0.5 | 0.7 | 0.7 |
| ZY/ZSM-5/MCM-41(20:20:60) | 82 | 25.5 | 0.5 | 2.4 | 0.7 | 0.5 | 1.1 | 0.6 |
a) C2=–C4= : averaged total yield of LOs; b) O/P: olefin-to-paraffin ratio; c) HTC = (n-C4 + i-C4)/C4=). |
Fig.6(a) shows the n-C7 conversion from the stability tests of the three catalysts (as a function of ToS). Again, the ZY/ZSM-5/SBA-15 (20:20:60) composite catalyst showed the best ability to crack n-C7 with the stable conversions at 95%, followed by ZY/ZSM-5/KIT-6 (20:20:60). Conversely, the ZY/ZSM-5/MCM-41 (20:20:60) composite catalyst showed the initial decrease in n-C7 conversion from 82% (within the first hour of ToS) the stabilized at about 79%. The three catalysts show rather stable total yields of LOs (Fig.6(b)). ZY/ZSM-5/SBA-15 (20:20:60) produced the highest LOs yield at about 25%, while ZY/ZSM-5/MCM-41 (20:20:60) only achieved the LOs yield of about 21%. Detailed product distribution is shown in Fig.6(c) and 6(d), showing that the three composites are selective to propylene, and ZY/ZSM-5/SBA-15 (20:20:60) and ZY/ZSM-5/KIT-6 (20:20:60) are also capable of producing ethylene and the aromatic compounds benzene, toluene and xylene (BTX). Overall, The MCM-41 mesoporous silica is less effective in the formation of composite catalysts for catalytic cracking, possibly due to the relatively smaller pore size of MCM-41 compared to SAB-15 and KIT-6, being less capable of improving diffusion and accessibility to the zeolites.
TGA was conducted to investigate the coke deposition of the used ZY/ZSM-5/KIT-6 (20:20:60), ZY/ZSM-5/MCM-41 (20:20:60), and ZY/ZSM-5/SBA-15 (20:20:60) composite catalysts after the longevity tests. The TGA was divided into 4 stages for a total of 12 h on stream, as shown in Fig.7. The first stage heated the catalyst under N2 from room temperature up to 200 °C (at 10 °C·min–1) and held for 1 h to remove the absorbed moisture from the composites. The second step was from 200 °C to 400 °C and held at 400 °C for 1 h to ensure the decomposition of the light hydrocarbons of C1–C5, which may be that possibly exist in the composite. The third step was from 400 °C to 650 °C and held at 650 °C for 1 h to remove heavy hydrocarbons. The final stage performed at 650 °C isothermally under air for 1 h to burn the remaining carbon residue from the catalysts. The results show that the three used catalysts of ZY/ZSM-5/KIT-6 (20:20:60), ZY/ZSM-5/MCM-41 (20:20:60), and ZY/ZSM-5/SBA-15 (20:20:60) displayed weight losses of 0.4 wt %, 1.4 wt %, and 1.5 wt %, respectively, at the final stage, suggesting that they are relatively resistant to coking during the reaction.
3.3 PFG-NMR study of self-diffusion of n-C7 within the selected composite catalysts
To study the structure-diffusion-performance relationship of the selected composite catalysts, PFG-NMR characterization was conducted to investigate the diffusion of n-C7 within the composites. We first performed the measurement using the neat n-C7 at room temperature to estimate the self-diffusion coefficient (Dself). Figure S7 (cf. ESM) shows the log attenuation plot for the free bulk liquid n-C7 sample, which is linear as expected, and the calculated Dself value for n-C7 is 32.4 × 10−10 m2∙s–1, being consistent with the reported value [33]. The log10 attenuation plots for the n-C7 saturated composites are shown in Fig.8, which are nonlinear attenuation curves, showing the distinct presence of two linear regions. The first region represents the self-diffusivity of n-C7 (D1) in the intercrystallite regimes of the composites, which could be influenced by the size and shape of the zeolite/silica particles, while the second indicate the self-diffusivity within the frameworks of the components, i.e., the intracrystalline diffusivity (D2) [34].
The calculated values for D1 and D2 of the n-C7 saturated composite catalysts are presented in Tab.4. Comparatively, ZY/ZSM-5/KIT-6 (20:20:60) has the largest D1 value of 11.2 ± 0.2 × 10−10 m2·s–1 among the three composites, followed by ZY/ZSM-5/SBA-6 (20:20:60) with D1 = 9.7 ± 0.1 × 10−10 m2·s–1, indicating that the effective diffusion of n-C7 molecules through the intercrystallite region of the two composites. Conversely, the D1 value of ZY/ZSM-5/MCM-41(20:20:60) is the lowest at 6.9 ± 0.1 × 10−10 m2·s–1, which corresponds to the conversion data (Fig.6(a)), indicating that catalytic conversion of n-C7 within the composites was affected by the intercrystallite diffusion of reactant molecules.
Tab.4 Relevant diffusion coefficients and calculated tortuosity values by PFG-NMR |
| Composite | Dself of n-C7 | D1 (10–10 m2·s–1) | D2 | τ1 | τ2 |
|---|---|---|---|---|---|
| ZY/ZSM-5/KIT-6 (20:20:60) | 32.4 | 11.2 ± 0.2 | 2.0 ± 0.2 | 2.9 | 16.3 |
| ZY/ZSM-5/SBA-15 (20:20:60) | 9.7 ± 0.1 | 2.4 ± 0.2 | 3.4 | 13.7 | |
| ZY/ZSM-5/MCM-41 (20:20:60) | 6.9 ± 0.1 | 1.7 ± 0.1 | 4.7 | 19.3 |
Compared to the intercrystallite diffusivity (D1), the intracrystalline diffusivity (D2) was about half a magnitude lower, suggesting the higher diffusion resistance within the frameworks of zeolites/silicas than that in the space between their particles. Since the proportion of the three components of the composite catalysts was the same at 20 wt % for ZY, 20 wt % for ZSM-5, and 60 wt % for silica, then the D2 values of the composites under investigation could be correlated to the averaged pore diameters of the mesoporous silicas employed. As shown above (Fig. S1), the pore diameters of the silica follow the order of SBA-15 > KIT-6 > MCM-41, corresponding well to the calculated D2 values of 2.4 ± 0.2 × 10−10 m2·s–1 > 2.0 ± 0.2 × 10−10 m2·s–1 > 1.7 ± 0.1 × 10−10 m2·s–1.
The intercrystallite and intracrystalline tortuosity (τ1 and τ2) values are calculated by comparing the bulk self-diffusivity (Dself) to D1 and D2, as presented in Tab.4, and a higher τ value suggests more restrictive diffusion within the considered domain [35,36]. The values of τ2 are larger than that of τ1, which is reasonable since intracrystalline diffusion resistance is higher than that of the intercrystallite diffusion due to the significantly reduced diffusion length scale within the framework materials. Tab.4 shows that ZY/ZSM-5/SBA-15 (20:20:60) has the smallest τ2 value of 13.7 and moderate τ1 value of 3.4.
Compared to the other two composites, ZY/ZSM-5/MCM-41 (20:20:60) has the highest τ value of 4.70, indicating the presence of a intercrystalline pore network with significant diffusion resistance. This is in line with its lower self-diffusion coefficient compared to ZY/ZSM-5/KIT-6 (20:20:60) and ZY/ZSM-5/SAB-15 (20:20:60), which have tortuosity of 2.89 and 3.35, respectively. Findings here suggest that by engineering the heterogeneous meso-microporosity, diffusion within a composite catalyst can be tuned to enhance the cracking performance.
4 Conclusions
In the study, the successfully synthesized of formulated catalysts was confirmed through detailed XRD and BET analyses. An XRD analysis confirmed the integrity of ZY and ZSM-5 frameworks within the composites. At the same time, BET results indicated that external surface areas and mesopore volumes had decreased by 10%–20%, while microporosity had increased by 50%–55%, suggesting a well-balanced micro-mesopore structure conducive to catalysis. Based on the results of the zeolite composites with different silica supports, pore size is an important factor in catalytic performance. Additionally, the PFG-NMR study indicates that the structure of composite catalysts influences n-C7 self-diffusion. In the intercrystallite region, there is less resistance than in the intercrystalline regions, indicating higher intercrystallite diffusivity (D1). The composite ZY/ZSM-5/KIT-6 (20:20:60) shows the highest diffusion coefficient. In the case of ZY/ZSM-5/MCM-41 (20:20:60), the tortuosity values indicate that intercrystalline diffusion faces greater resistance. In general, LO selectivity and conversion of n-C7 are affected by the diffusion characteristics of catalysts, and meso-microporosity can be adjusted to optimize. Due to the narrower and smaller mesopores range in MCM-41, the composite ZY/ZSM-5/MCM-41 (20:20:60) converted 79% n-C7. With larger mesopores in KIT-6, the ZY/ZSM-5/KIT-6 composite (20:20:60) achieved an 84% conversion rate. ZY/ZSM-5/MCM-41 (20:20:60) exhibited the lowest yield of LOs (21%), followed by ZY/ZSM-5/KIT-6 (20:20:60) (22%), and ZY/ZSM-5/SBA-15 (20:20:60) (24%), demonstrating the importance of mesoporous structures for diffusion and accessibility of reactants. Overall, this research suggests the potential of zeolite-mesoporous silica composite catalysts in catalysis and separation applications. It is possible to develop highly efficient and selective catalysts for industrial applications by combining microporous and mesoporous structures within these composites.