High-Impact Promotional Effect of Mo Impregnation on Aluminum-Rich and Alkali-Treated Hierarchical Zeolite Catalysts on Methanol Aromatization.

A systematic change of HZSM-5 (HZ5) as a catalyst of the methanol to aromatics (MTA) reaction was undertaken by employing a fixed-bed tubular-type reactor under ambient pressure, applying a weight hourly space velocity (WHSV) of 2 h-1 at 375 °C, as the first report on the application of low-Si/Al-ratio alkaline-[Mo,Na]-HZSM-5 in the MTA process. To characterize the surface and textural properties of the catalysts, powder X-ray diffraction (PXRD), nitrogen adsorption/desorption, temperature-programmed desorption of ammonia (NH3-TPD), pyridine-infrared spectroscopy (Py-IR), thermogravimetric analysis (TGA), and energy-dispersive X-ray (EDX) methods were employed. Gas chromatography (GC) and gas chromatography-mass spectrometry (GC-MS) measurements demonstrated a selectivity of up to 86 wt % (65.7 wt % for benzene, toluene, and xylene (BTX)) over 2[Mo]HZ5. NH3-TPD and Py-IR results indicated a sensible decrease of strong acid sites on the impregnated samples, while the surface analyses revealed the highest Lewis acid sites (LAS) together with the largest mesopore surface area for 2[Mo]alk-HZ5, supporting the migration of Mo species to the bulk of the catalysts. Mo impregnation had a minor effect on the observed coke formation in the promoted catalyst.

Introduction

Among the aromatic hydrocarbons, benzene, toluene, and xylene (BTX) are essential chemical starting materials for the production of important organic intermediates for many chemical industry processes, namely, medicines, electronic devices, petrochemicals, automotive vehicles, polymers, vehicle tires, detergents, solvents, plastics, rubbers, dyes, cosmetics, perfumes, aspirin, home care products, and many other industries.[1−9] Currently, aromatics production is regularly governed by the petroleum resources, derived from catalytic reforming, alkylation, and cracking applying crude oil.[4,10−12] Thus, exploring alternative renewable energy resources as substitutes for crude oil to produce BTX and introducing new processes for the aromatics production are current scientific challenges worldwide.[2,13]

As an industrially feasible and sustainable high-efficiency process, the methanol to aromatics (MTA) process is considered as a platform for the production of basic aromatics on a large scale,[10] since methanol is a relatively low-priced starting material along with its simple production from different resources, namely, coal, biomass, and natural and/or shale gas.[1,5−7,14,15] Therefore, the development of new and efficient catalysts for the MTA process is a significant prerequisite for their industrial applications.[16] Meanwhile, scientists have explored several zeolites as promising MTA catalysts.[17] Among them, ZSM-5 zeolites have fascinated researchers, due to their typical three-dimensional network structure, outstanding stability, suitable pore sizes, shape selectivity, texture, extremely high thermal stability, adjustable acidic properties, and ability to act as a model in MTA mechanistic studies.[5,16−24] Basically, several aspects of ZSM-5 zeolites, namely, their morphology and the effect of the applied promoters and the Si/Al ratio on the stability, activity, and selectivity of the resultant catalyst,[18,25] are decisive topics in the MTA process.

On the one hand, the stability of zeolites is systematically linked to their acidic properties. Gao et al. reported on lowering the energy barrier along with increasing the rate of the aromatization reaction over a zeolite catalyst by employing highly acidic conditions.[16] On the other hand, the activity of ZSM-5 is determined by the strength and density of the acid sites and occupancy of the available Brønsted acid sites (BASs) over the surface together with the content of the tetrahedrally coordinated Al in the framework of the zeolite. For instance, low-Si/Al-ratio ZSM-5 zeolites normally enjoy from a great amount of acid sites plus high activity, originating from their high Al content.[26] Moreover, the acidity of ZSM-5 has been shown to be a key factor influencing the nature of the products, commonly adjusted by zeolite modification by changing the Si/Al ratio.[27] As an example, several scientists described the remarkable presence of olefins on ZSM-5 catalysts with a high Si/Al ratio, which resulted in boosted selectivity via hydrogen transfer in the methanol to olefins (MTO) process; however, the catalysts with low-Si/Al-ratio ZSM-5 were found to be adequate candidates for aromatics production in the methanol to hydrocarbons (MTH) reaction.[28−30]

But the rate of coke creation is enhanced over a zeolite with high Al content.[26] To lower the rate of coke formation, the presence of mesopores with a narrow pore size distribution (PSD) on the ZSM-5 surface is a basic requisite to support the required molecular diffusion inside the zeolite pores.

Alkaline desilication with NaOH solution is a well-known substitute for the conventional methods to create intracrystalline secondary pores on the surface of ZSM-5 zeolite, acting as a selective method to remove Si atoms from the framework.[2,10,31,32] Besides, ZSM-5 modification by the impregnation of metal ions is another effective alternative to reach higher selectivity for aromatics, namely, BTX, owing to the existence of three-dimensional 10-ring channel morphologies within ZSM-5, having a micropore size of ∼0.55 nm, very close to the kinetic diameter of BTX products.[11,33] Several earlier reports deal with the effect of modification of ZSM-5 with various metal species, such as Ga,[10,17,33−41] Zn,[3,4,6,9,11,18,42−49] Ag,[11,39,50] La,[51] Cu,[39,52] Ni,[11,39,53] Re,[33] Sn,[9] and Cd,[6] on improving the aromatics selectivity in the MTA reaction.

Wei et al.[54] showed that the desilication and dealumination of ZSM-5 led to an increase of aromatic selectivity and lifetime and a decrease of the number of BASs, owing to the creation of more mesopores on the MTA catalyst. Lai et al.[36] also reported that desilication on Ga-doped ZSM-5 can broaden both pore size and pore distribution of the methanol aromatization catalyst. Yang et al.[43] confirmed the above-mentioned effects of desilication on Zn-HZSM-5 applied in methanol aromatization and reported an improved catalyst aromatic yield due to better mass diffusion, coke formation, and coke tolerance.

Recently, we carried out a comparative study on both alkali treatment and Fe–Zn postmodification of low-Si/Al HZSM-5 zeolite, employed in aromatics production in the MTA reaction.[55] The result was successful in terms of the achievement of novel aromatics selectivity of [0.2Fe,0.3Zn]-alk-HZSM-5. As an unstudied case in the MTA reaction, it convinced us to undertake a comparative study of Mo impregnation as a new single-component post-treatment on low-Si/Al HZSM-5.

Although Mo has been explored as an effective promoter on ZSM-5 zeolite for methane aromatization,[56−67] to the best of our knowledge, there is no report on the effect of Mo impregnation on low-Si/Al-ratio HZSM-5 in the MTA process. In this paper, we prepare alkali-treated, aluminum-rich hierarchical [Mo,Na]-HZSM-5 zeolites by hydrothermal synthesis, followed by a metal-ion impregnation method. We also investigate the loading effect of Mo over the parent HZSM-5 in the MTA reaction, and the subsequent effect of this loading on the yield and selectivity of BTX in the MTA process is also surveyed. Due to its remarkable resistance to the coke deactivation process, which is a prominent fault in the MTA process,[9] the resulting catalyst deactivation is also explored.

Results and Discussion

Catalyst Characterization

Figure S1 displays Fourier transform infrared (FT-IR) analysis of HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 catalysts in the range 400–4000 cm–1. The FT-IR spectra showed the bending vibrations of the primary SiO4 and AlO4 tetrahedra at ∼445 cm–1. The additional band at ∼549 cm–1 was attributed to the five-membered ring vibration in the tetrahedral units of the ZSM-5 structure. Also, the FT-IR bands in the 750–850 and 1050–1200 cm–1 regions were assigned to the characteristic asymmetric and symmetric stretching modes of SiOSi, respectively. Another band at 1620 cm–1 was assigned to the OH bending vibration, escorted with a broad band at 3449 cm–1, owing to the absorbed water and the connecting OH groups. Besides, the observed bands at 3500–3800 cm–1 were assigned to OH groups, indicating the presence of Si–OH groups. As represented in Figure S1, all of the samples showed a typical band at ∼3600 cm–1 for BASs accompanied by another band at 3740 cm–1 for the silanol groups on the catalyst.[32,68−72]

The X-ray diffraction (XRD) patterns of the samples are presented in Figure .

Figure 1

XRD patterns of the HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 catalysts in the 2θ ranges of (a) 5–80°, (b) 7–10°, and (c) 22–25°.

Figure a displays the XRD pattern of HZ5 (2θ = 8.1, 8.9, 23.1, 23.3, and 23.8°), observable in all of the orthorhombic ZSM-5 catalysts. The observation of representative ZSM-5 peaks in the absence of any amorphous baseline in the corresponding XRD patterns signifies the occurrence of a high degree of crystallinity for the catalyst samples, even in the impregnated samples with the metal ions. Since these patterns did not show any other diffraction peaks pertinent to Mo, high dispersions of the metal species on the Mo-impregnated samples were practically anticipated. Figure also shows some intense diffraction peaks at 2θ = 22–25°, specifying the retained crystal structures of the ZSM-5 samples after being loaded with the Mo salt. Moreover, in the absence of any XRD pattern pertaining to Mo oxides, good dispersions of small-size metallic particles over the surface of the impregnated catalyst are expected.

From the practical point of view, small-angle X-ray diffraction patterns recorded for microporous materials can be more sensitive than the rest of the patterns to show any penetrated species into the micropore channels, indicating that both the d-spacing and intensity of these diffraction peaks are more likely susceptible to morphological changes. The effects of metal impregnation on the X-ray diffraction patterns of the ZSM-5 samples for low-angle diffraction peaks between 2θ = 7 and 10° are presented in Figure b,c. As can be seen in these figures, fairly large 2θ shifts to somewhat lower intensities are observable in the corresponding diffraction peaks, after loading of HZ5 with Mo. For instance, although the diffraction patterns at 2θ = 8.1 and 9° for 1[Mo]HZ5 experience no remarkable shift with respect to the parent HZ5, the other x[Mo]HZ5 (x = 2 and 4) catalysts exhibit significant changes, suggesting that the thermal diffusion of MoO3 inside the ZSM-5 micropores may have occurred by anchoring to the BAS on the catalyst. Accordingly, Figure b represents the following order of Mo penetration in the micropore channels in terms of the above diffraction peak alternation at 2θ = 8–9°Figure b also shows the significantly lower intensity of the diffraction patterns at 2θ = 8–9° for the 2[Mo]alk-HZ5 catalyst, compared to those for 2[Mo]HZ5 and the parent catalysts. These observations suggest that the impregnation of Mo on the parent alk-HZMS-5 is a result of better penetration of Mo species into the micropore channels of 2[Mo]alk-HZ5 than the 2[Mo]HZ5 catalyst. Additionally, the diffraction patterns at 2θ = 23.1° (Figure c) for the Mo-modified catalysts embodied slight shifts, assigned to a decrease in the d-spacing expected for smaller pore volumes, giving rise to a proportional quantity of Mo loading into the pores of the parent ZSM-5 sample.[63,73]

Figure 2

Isotherms of nitrogen adsorption/desorption for HZ5, 2[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 catalysts.

The nitrogen adsorption–desorption isotherms for HZ5, alk-HZ5, 2[Mo]HZ5, and 2[Mo]alk-HZ5 in Table indicate their textural properties including the surface area and pore volume, according to the Brunauer–Emmett–Teller (BET) isotherms.

Table 1

Textural Properties of HZ5, alk-HZ5, 2[Mo]HZ5, and 2[Mo]alk-HZ5 Catalysts

	characteristic
sample name	SBET (m2 g–1)a	SMicro (m2 g–1)b	SMeso (m2 g–1)c	VTotal (cm3 g–1)d	VMicro (cm3 g–1)e	VMeso (cm3 g–1)f
HZ5	229.89	209.16	20.73	0.1235	0.0965	0.0270
alk-HZ5	349.57	291.02	58.55	0.1679	0.1056	0.0523
2[Mo]HZ5	140.54	123.15	17.39	0.1223	0.0718	0.0505
2[Mo]alk-HZ5	190.97	158.42	32.55	0.1273	0.0773	0.0500

Calculated total surface areas at P/P0 = 0.05–0.25 were determined using adsorption data obtained from the BET method.

t-plot method was applied to assess the micropore surface area.

SBET – SMicro were employed to calculate the mesopore contribution of the surface area.

Total pore volumes were estimated from the amount of the adsorbed catalyst at P/P0 = 0.99.

Micropore volume calculated by the t-plot method.

Mesopore volume calculated using VTotal – VMicro.

Table demonstrates the total pore volumes and external (SMicro) and BET surface areas (SBET) of the HZ5 catalyst, wherein the alkaline treatment results in a remarkable increase. Accordingly, the mesopores surface area (SMeso) and volume (VMeso) for alk-HZ5 exhibited twofold higher values than those for the parent HZ5. Moreover, the Mo impregnation on alk-HZ5 significantly reduced the corresponding surface area parameters (SBET and VTotal), attributed to the pore blockage by metallic species in earlier reports.[46,74] Thus, analogous to the XRD results described for the samples in Figure , SBET, SMicro, and SMeso parameters of 2[Mo]HZ5 were noticeably less than those of the parent HZ5 (Table ). Likewise, the comparison made between two samples, namely, 2[Mo]HZ5 and 2[Mo]alk-HZ5, revealed that the alkali treatment on the catalyst samples resulted in higher surface areas and total pore volumes.

The aforesaid differences were also confirmed by the nitrogen adsorption–desorption experiments of the samples (Figure ).

Figure presents two types of isotherms for the catalysts, namely, type I and type IV. While the parent HZ5 exhibits a type I isotherm consistent with a typical microporous structure, 2[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 catalysts show type IV isotherms pertaining to the presence of mesoporosity in their morphologies according to the IUPAC classification.[75] Moreover, Figure displays the isotherms for the 2[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 catalysts, denoting a hysteresis loop of type IV isotherms mainly at a relative pressure of 0.4–1, including both types of microporous and mesoporous structures. Furthermore, the alk-HZ5 sample indicated a hysteresis loop at a relatively higher P/P0 (Figure ) due to the mesoporous structure mainly formed by the alkaline treatment. In comparison with the mesoporous alk-HZ5, the conventional HZ5 exhibited type I isotherms with a small hysteresis loop due to the occurrence of intracrystalline spaces between the small crystalline particles. Besides, the adsorption isotherms of alk-HZ5, 2[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts indicated larger hysteresis, owing to the presence of restricted mesopores into which nitrogen can only have limited access, meaning no direct admission through the crystal surface.[76,77]

Figure 3

PSD according to Barrett–Joyner–Halenda (BJH) curves of the HZ5, 2[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 catalysts; the presentation of the PSD at (a) the complete range, (b) <10 nm, and (c) 20–50 nm.

The BJH method was characteristically employed to calculate the mesopore size distribution.[27]Figure exhibits the PSD curves of HZ5, 2[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 catalysts, derived from the adsorption isotherms using the BJH model.

Figure also demonstrates a bimodal mesoporous size distribution from 1 to 4 and 20 to 50 nm (Figure b,c) for the catalysts. Incidentally, the PSD in Figure c revealed that a prominent mesopore size distribution, concentrated at 20–50 nm, was observable for 2[Mo]HZ5, alk-HZ5, and 2[Mo]alk-HZ5 samples. The mesopore distribution within the said interval was previously reported, owing to the formation of intergranular mesopores, created by the aggregation of small grains.[22] Thus, Figure c indicates the formation of mesopores with various diameters in 2[Mo]HZ5 as a result of the Mo impregnation process.

Figure 4

Temperature-programmed desorption of ammonia (NH3-TPD) analysis of HZ5, alk-HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts.

The NH3-TPD experiments together with FT-IR spectra of pyridine adsorption were employed to assess the Mo loading effect on the acidic properties of the parent HZ5 catalyst. The total number of acid sites on the catalysts and acid site strength were also evaluated by the NH3-TPD technique.[63]Figure displays the corresponding NH3-TPD curves for HZ5, alk-HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts.

Basically, the TPD profiles in Figure for these samples demonstrate two desorption peaks: the first band displays a maximum located at ∼210–280 °C, assigned to weak acid sites, whereas the second band exhibits a maximum at ∼460–490 °C, attributed to strong acid sites.[78] Evidently, the impregnated samples show a sensible decrease in strong acid sites with respect to the parent HZ5. The observed decrease in the concentration of the strong acid sites can be elucidated in terms of the reaction of cationic and/or anionic Mo species with protons[79] and OH[64] groups on the surface of the catalysts.

Besides, Table summarizes the quantitative results of NH3-TPD patterns for HZ5, alk-HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts.

Table 2

NH3-TPD Results for HZ5, alk-HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 Catalysts

	distribution and concentration of acid sites (mmol NH3 g–1)				peak temperature (°C)
catalyst	region I (weak)	region II (strong)	total	strong/weak	Td1	Td2
HZ5	1.103	1.199	2.302	1.087	258.8	474.2
alk-HZ5	0.996	0.541	1.537	0.543	249.1	490.0
1[Mo]HZ5	1.006	1.192	2.198	1.185	248.7	479.7
2[Mo]HZ5	1.305	1.082	2.387	0.829	225.6	464.8
4[Mo]HZ5	1.867	1.165	3.032	0.624	211.9	476.8
2[Mo]alk-HZ5	2.294	0.000	2.294	0.000	273.7

Table suggests the following orders of BASs and Lewis acid sites (LASs) on the catalyst samples:To investigate the acidity of catalysts in more detail, the IR spectra of chemisorbed pyridine on the samples were also studied. On the BAS, pyridine accepts a proton to give a pyridinium ion, whereas on the LAS, pyridine donates a lone pair of electrons to form an adduct.[80] The vibrational bands at 1550 and 1450 cm–1 are characteristic of BAS and LAS, respectively.[4]Figure represents the FT-IR spectra of pyridine-adsorbed samples, namely, HZ5, alk-HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts.

Figure 5

Selected vibrational spectra of adsorbed pyridine on the HZ5, alk-HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts.

BAS

LAS

total acid sites

The IR spectra for pyridine adsorption on the catalyst samples shown in Figure indicate the strength and types of acid sites of the catalysts. In these experiments, 2[Mo]alk-HZ5 exhibited the strongest LAS along with the weakest BAS. NH3-TPD and pyridine-infrared spectroscopy (Py-IR) results suggest that Mo incorporation into the zeolite samples can reduce the strength of the strong acid sites.

In the subsequent experiments, the surface morphologies of HZ5, alk-HZ5, and 2[Mo]HZ5 catalysts were probed by the field-emission scanning electron microscopy (FESEM) method (Figure ).

Figure 6

FESEM images of (a) HZ5, (b) alk-HZ5, and (c) 2[Mo]HZ5 catalysts, as the samples with the best catalytic results.

The FESEM micrographs in Figure exhibit a spongelike microstructure along with individual crystal sizes in the range of 20–40 nm for the HZ5, alk-HZ5, and 2[Mo]HZ5 catalysts. The samples tend to agglomerate into microsized new phases due to their high surface Gibbs energy. On the other hand, the total size of the regular particles seems similar for all catalysts with different metal contents. Evidently, these microsized catalysts simply pile up to arrange into larger stacks, having large spaces among the crystals. In other words, the morphology of ZSM-5 slightly changes after modification with Mo loading and alkaline treatment.[16]

Table presents the Mo loadings on the catalysts, determined by the energy-dispersive X-ray (EDX) technique.

Table 3

Theoretical and Experimental Mo Contents of HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, alk-HZ5 and 2[Mo]alk-HZ5 Catalysts, Determined by the Inductively Coupled Plasma (ICP) Technique

catalyst Mo content (wt %)	HZ5	1[Mo]HZ5	2[Mo]HZ5	4[Mo]HZ5	alk-HZ5	2[Mo]alk-HZ5
theoretical loading	0	1.0	2.0	4.0	0	2.0
experimental loading	0	0.24	0.96	1.57	0	2.01

In the 2[Mo]alk-HZ5 catalyst, the theoretical value for Mo loading (2%) was in complete agreement with the experimental value (2.01% obtained by SEM-EDX). However, according to the EDX results, the nonalkali catalyst samples, i.e., 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, demonstrate lower experimental Mo loading values. Similarly, Abdelsayed et al.[63] reported the elemental surface composition for the Mo/Zn-ZSM-5 sample using EDX analysis, indicating the presence of a lower concentration of 2 wt % Mo in the sample as a result of more Mo diffusing into the bulk of ZSM-5 (the theoretical loading for Mo was 4 wt %). Additionally, Figure S2 clarifies the correlation between Mo and Na loading on the impregnated alkali and nonalkali-treated catalysts.

Elucidating the said relationship, Scheme summarizes the presence of various types of MoO species (shown in Scheme ) on the surface of the aluminosilicate.[64]

Scheme 1

Various Reported Types of Mo Species, Anchored to BASs of the ZSM-5

(a) MoO2+ cluster bridging two neighboring oxygen atoms, (b) MoO22+ cluster bridging two Brønsted acid sites, like H–Y zeolites, and (c) (Mo2O5)2+ cluster bridging two Brønsted acid sites of the zeolite.

Figure shows the EDX dot-mapping analysis on the surface of the catalysts to illustrate the presence and distribution of different elements.

Figure 7

FESEM/EDS analysis of (a) HZ5, (b) 1[Mo]HZ5, (c) 2[Mo]HZ5, (d) 4[Mo]HZ5, (e) alk-HZ5, and (f) 2[Mo]alk-HZ5 catalysts.

These figures show the homogeneous dispersion of Mo, Na, Al, and Si elements over the catalyst surfaces.

The above results were also in accordance with the BET observations (Table ), wherein SBET, SMicro, and SMeso of 2[Mo]HZ5 were noticeably less than those of the parent HZ5 (Table ). More importantly, any attempt to obtain 95Mo magic-angle spinning nuclear magnetic resonance (MAS NMR) spectra of the pale blue Mo-containing samples (as an indication of Mo(V) species in the samples) under any probable conditions failed to show any chemical shift in the spectra, suggesting the (at least partially) the incorporation of paramagnetic Mo species in the samples.

To investigate the catalyst morphologies more deeply, transmission electron microscopy (TEM) images of the samples were recorded as presented in Figure .

Figure 8

TEM images of (a) HZ5, (b) alk-HZ5, and (c) 2[Mo]alk-HZ5 catalysts. The marked yellow and red circles signify the observed etched hole and new crystalline phases on the catalysts, respectively.

As shown in Figure , the parent HZ5 zeolite morphology exhibits both intracrystalline and surface etching mesoporosity due to the remarkable particle size changes to afford the alk-HZ5 catalyst. Meanwhile, partial deterioration of the external surface of the alk-HZ5 catalyst gave rise to spreading mesoporous structures within the zeolite morphologies, corresponding to textural changes recorded in Table for alk-HZ5 (SBET = 349.57 m2 g–1 for alk-HZ5 versus SBET = 222.89 m2 g–1 for HZ5). Figure c, however, suggests the particle size retreatment in 2[Mo]alk-HZ5, wherein particle size aggregation processes lead to larger particle sizes, as is also inferred from the FESEM micrographs in Figure a,e,f. Correspondingly, the textural properties of 2[Mo]alk-HZ5 in Table indicate low SBET = 190.97 m2 g–1, which is remarkably lower than the observed value for HZ5, attributable to the formation of new solid phases, i.e., molybdate (highlighted by the red circles in Figure c), on the surface of 2[Mo]alk-HZ5.

MTA Catalytic Performance of the Samples

To determine the performance of the samples as MTA catalysts, gas chromatography (GC) and gas chromatography-mass spectrometry (GC-MS) methods were applied to analyze the compositions of the liquid hydrocarbon products (Table ).

Table 4

Compositional Analysis of the Hydrocarbon Products of the MTA Process at TR = 375 °C over the Catalyst Samples in the Present Work

catalyst component (wt %)	HZ5	1[Mo]HZ5	2[Mo]HZ5	4[Mo]HZ5	alk-HZ5	2[Mo]alk-HZ5
i-C5	1.7	0.9	0.9	0.6	4.5	0.0
C6 saturated	0.3	1.3	0.2	0.4	4.4	0.0
benzene	3.5	1.2	6.7	3.7	0.0	0.0
C7 saturated	0.5	1.0	0.1	0.1	0.2	0.0
toluene	21.2	28.3	32.1	24.2	13.8	0.0
ethyl benzene	1.8	1.5	1.6	1.6	1.8	0.0
o-, m-, and p-XYL	30.2	27.5	26.9	29.9	39.4	13.6
M-E-BZ	2.4	2.0	1.8	2.5	3.6	9.9
3M-BZ	8.2	8.7	8.0	8.9	7.3	16.8
Ar-C10	10.0	7.7	8.9	9.3	9.4	19.5
C10+	10.8	11.6	8.8	11.5	7.3	40.2
others	9.4	8.3	4.0	7.3	8.3	0
sum	100	100	100	100	100	100
BTX	54.9	57	65.7	57.8	53.2	13.6
aromatics (≤C10)	77.3	76.9	86.0	80.1	75.3	59.8

In these experiments, moderate methanol pressure using a weight hourly space velocity (WHSV) of 2 h–1 was applied to achieve high yield of BTX. Under this condition, more poly(methylbenzene) was practically formed, owing to an increase in methanol partial pressure. Thus, coke formation due to oligomerization processes was promoted as a result of inhibited transformation into BTX, hydrogen transfer, etc.[1]

Table signifies the highest level of aromatic hydrocarbons (viz. 86%) for the Mo-impregnated ZSM-5 catalysts. Typically, Table illustrates high level of selectivity for benzene (6.7%), toluene (32.1%) as well as lower xylenes (26.9%) for methanol conversion over 2[Mo]HZ5. Moreover, the highest hydrocarbon products were observed over the 2[Mo]HZ5 catalyst in the MTA reaction as 70.4% alkyl aromatics, i.e., ortho-, meta-, and para-xylenes (o-, m-, and p-XYL), toluene, along with 40.3% alkyl aromatics, 0% benzene, and 0% toluene over the 2[Mo]alk-HZ5 catalyst.

The aromatic selectivity over various catalysts is also depicted in Figure .

Figure 9

Average production of the aromatic components over HZ5, 1[Mo]HZ5, 2[Mo]HZ5, 4[Mo]HZ5, alk-HZ5 and 2[Mo]alk-HZ5 catalysts.

Figure shows the highest selectivity for the aromatic hydrocarbons together with benzene-free specification, determined over the alk-HZ5 catalysts, denoting the production of heavy aromatics over the 2[Mo]alk-HZ5 sample. The average weight percent of total aromatics and BTX products in Table and Figure over all catalyst samples are also summarized in Figures and 11, respectively.

Figure 10

Total weight percentage of aromatics over different catalysts (applied %Mo): HZ5 (0), 1[Mo]HZ5(1), 2[Mo]HZ5(2), and 4[Mo]HZ5(4).

Figure 11

Average of weight percent for BTX products over different catalyst samples (%Mo): HZ5(0), 1[Mo]HZ5(1), 2[Mo]HZ5(2), and 4[Mo]HZ5(4).

Figures and 11 summarize the average weight percent of total aromatics and BTX products over all catalyst samples, reported in Figure and Table , respectively.

Based on the results of Figures and 11, 2[Mo]HZ5 produced the maximum of total aromatics and BTX products. Although the reason(s) for the observed rise and decline in the product formation, e.g., BTX (Figures –11), is not clear, it can be elucidated in terms of some previous reports.[64,81] In these reports, sensible catalytic influence of Mo impregnation on a low-Si/Al-ratio ZSM-5 catalyst is attributed to either electronic interactions or the deformation of the AlO4– tetrahedra.

Moreover, Figure displays the methanol conversion versus time on stream (TOS) for MTA over the HZ5, alk-HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts.

Figure 12

Methanol conversion as a TOS function over HZ5, alk-HZ5, 2[Mo]HZ5, 4[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts (WHSV = 2 h–1, T = 375 °C).

As can be seen in Figure , 2[Mo]HZ5 and 4[Mo]HZ5 catalyzed significant methanol conversion up to >95% for several hours. Meanwhile, this figure illustrates the declined conversion of methanol over alk-HZ5 and 4[Mo]HZ5 catalysts with time, giving typical conversions of 93 and 89% after 8 h on stream, respectively. While the methanol conversion over alk-HZ5, 4[Mo]HZ5, and 2[Mo]HZ5 catalysts remained constant up to 85% after 8 h, the corresponding conversions over HZ5 and 2[Mo]alk-HZ5 experienced outstanding drops to 75% under the same conditions.

Practically, coke deposition is the key reason for catalyst deactivation in the MTA reaction,[11] resulting from successive reactions between BASs or LASs and light olefins.[5] To compare the extent of coke formation on HZ5 and 2[Mo]alk-HZ5, their thermogravimetric analysis (TGA) results are shown in Figure .

Figure 13

TGA curve of HZ5, 2[Mo]HZ5, and 2[Mo]alk-HZ5 catalysts.

Incidentally, the weight loss in the fresh HZ5 catalyst at 500–600 °C can be due to the elimination of the structure-directing agent, tetrapropylammonium cation (TPA+). Figure also shows different typical dual-step weight losses for these samples in TGA curves: the first step occurs at 50–450 °C, assigned to the adsorbed water and large organic molecules with high molar ratios of H/C (soluble coke), wherein the second step at 450–750 °C can be assigned to the carbonaceous species, namely, insoluble coke,[44] having low H/C mole ratios. Besides, TGA results indicated that the coked catalysts, namely, 2[Mo]HZ5 and 2[Mo]alk-HZ5 after 8 h on stream, exhibit a typical weight loss of ∼9%, while the same reaction over the parent HZ5 resulted in 11% weight loss under the same condition. Regardless of the post-treatments applied on the parent HZ5, the above-mentioned results indicate that neither alkali treatment nor Mo impregnation can cause significant changes in the corresponding coke formation. In a similar example, Zhang et al. observed coke formation on both strong and weak acid sites in butylene aromatization, wherein the catalyst stability poisoning was thoroughly associated with the acid site distribution[1] (see Figure ).

To complete coke formation studies on the catalyst samples, the content of carbon and hydrogen on the spent catalysts was also determined by the carbon and hydrogen (CH) microanalysis method (Table ).

Table 5

Content of Carbon and Hydrogen on Spent Catalysts

catalysts	C (%)a	H (%)a
HZ5	13.80	4.06
alk-HZ5	9.91	2.24
1[Mo]HZ5	15.10	4.48
2[Mo]HZ5	12.70	3.36
4[Mo]HZ5	13.40	3.59
2[Mo]alk-HZ5	6.28	2.88

Determined by CH microanalyses.

Table demonstrates that the spent 2[Mo]alk-HZ5 and alk-HZ5 exhibited the lowest soluble coke content because the impregnation of HZ5 with Mo (as in the other [Mo]HZ5) was less efficient than alkali treatment (in alk-HZ5) in terms of the formation of soluble coke species. The situation for 2[Mo]alk-HZ5 in Table was more pronounced, wherein its C and H contents from CH microanalyses were lower than those of even the alk-HZ5 sample, which is attributed to the occupation of the pores and channels of 2[Mo]alk-HZ5 with large organic molecules. Taking into account the conclusions drawn from TGA measurements (Figure and the subsequent descriptions), a similarity in the total weight loss of the spent HZ5 and 2[Mo]alk-HZ5 catalysts along with their remarkable CH microanalysis difference led us to conclude that the deactivation of the parent HZ5 mostly occurred as a result of pore blocking by large hydrocarbon molecules. However, the primary reason for the deactivation of the 2[Mo]alk-HZ5 catalyst can be the occlusion of pores by nondestructive coke species. These observations were also in accordance with the N2 adsorption/desorption isotherms of the HZ5, 2[Mo]alk-HZ5, and alk-HZ5 catalysts (Figure ). While the parent HZ5 displays a microporous structure, alk-HZ5 and 2[Mo]alk-HZ5 catalysts exhibit a prominent mesopore size distribution in their morphologies, basically due to the alkaline treatment. The observed sharp drop in the catalytic activity of 2[Mo]alk-HZ5 (Figure ) can be attributed to the blockage of the mesopores, viable to the observed selectivity of 2[Mo]alk-HZ5 with respect to HZ5 in the MTA process (Figure ). In conclusion, the lowest coke formation in 2[Mo]alk-HZ5 and alk-HZ5 may happen as a result of the occurrence of the MTA process within their channels and frameworks, while the microporous structure of the parent HZ5 catalyst can lead to methanol conversion on the surface.

Figure 14

XRD patterns of the 2[Mo]alk-HZ5 and coked-2[Mo]alk-HZ5 catalysts in the 2θ ranges of 5.0–70.0° (a), (b) 22.5–25.0°, and (c) 44.5–46.0°.

Practically, the formation of polyaromatic species (also known as coke) under the dehydrogenation and condensation reactions of the unsaturated species advantageously requires strong acid sites.[3]Figure displays the critical differences between the XRD patterns of 2[Mo]alk-HZ5 and coked-2[Mo]alk-HZ5 catalysts.

Figure displays the significant alterations of the XRD pattern after coke formation on the zeolite. Alvarez et al.[82] also reported, for the first time, complete zeolite deactivation by accumulation of the coking species on the zeolite, in terms of changing from the original orthorhombic structure to the tetragonal geometry, determined by a powder diffractometer. A clear change in the diffraction pattern is observed in the region 23–24°. For instance, Figure b represents the doublet corresponding to the lines 23.209 and 23.365° merging to a single line at 23.258°. Moreover, Figure c represents the double bands at (1000) 45.153° and (0100) 45.569°, merged to a single line at 45.358°. Meanwhile, clear intensity changes observed for the coked-2[Mo]alk-HZ5 catalyst were induced due to coke accumulation during the deactivation of the zeolite.[83] In conclusion, in our Mo-impregnated HZ5 samples, the deactivation of the catalysts can be attributed to the coke deposition over the internal pores as well as the surface of the catalyst.

Conclusions

In this paper, we prepared aluminum-rich and alkali-treated hierarchical [Mo,Na]-HZSM-5 catalysts with hydrothermal synthesis and alkali treatment methods, followed by impregnation with Mo. The Mo impregnation of ZSM-5 zeolite resulted in the formation of 70.4% alkyl aromatics, having higher benzene (6.7%), toluene (32.1%), and lower xylene (26.9%) contents for the 2[Mo]HZ5 catalyst. Moreover, the alkali-treated Mo-HZ5 catalyst (2[Mo]alk-HZ5) afforded 40.3% alkyl aromatics, 0% benzene, and 0% toluene. The observed difference in product distribution by variation of Mo content in these catalysts was described in terms of the Mo extra-framework coordination mechanism. The resultant decrease in micropore volume along with the ratio of BAS/LAS in 2[Mo]alk-HZ5 indicates the mesoporous morphology. Among these samples, 2[Mo]alk-HZ5 and alk-HZ5 displaying the lowest coke formation are the best candidates for the MTA process, attributable to their mesoporous morphology. Failing to obtain any 95Mo MAS NMR spectra for the Mo-containing samples indicate the presence of paramagnetic Mo species in the samples. These bifunctional catalysts are practical candidates for methanol aromatization due to their sensible acid and redox activity.

Materials and Methods

Catalyst Preparation Materials

Materials

Tetraethyl orthosilicate (TEOS, 98 wt %), tetrapropylammonium hydroxide (TPAOH, 40 wt % aqueous solution), aluminum isopropoxide (AIP, 97 wt %), HNO3, NaOH, NH4NO3, and (NH4)6Mo7O24·4H2O were purchased from Merck (Germany) and used with no further purification.

Preparation of HZ5 Zeolite

The HZSM-5 (HZ5) zeolite was prepared using a hydrothermal process, consistent with the modified method introduced by Karimi et al.[84] The applied molar ratio for the gel composition was 1Al2O3/22SiO2/5Na2O/2.7TPAOH/2500H2O.

Alkaline Treatment on HZ5

The desilication procedure of HZ5 zeolite was performed using 0.3 M NaOH solution at 80 °C (using 1 g of zeolite/8 mL of NaOH solution) for 2 h. Then, the slurry was chilled before filtering the mixture to remove the residual NaOH and subsequent decantation with diluted HNO3. Afterward, the product was dried at 110 °C for 4 h prior to ion exchanging with 0.8 M NH4NO3 in water (applying 1:8 mass ratio) for 12 h at 80 °C under stirring conditions. Then, the mixture was filtered prior to washing with distilled water to replace the residual protons with Na+ cations. After drying at 110 °C overnight, the NH4ZSM-5 zeolite was calcined for 5 h at 550 °C, forming an alkaline HZSM-5 sample, denoted alk-HZ5. According to EDX analysis, the Na contents of HZ5 and alk-HZ5 samples were determined as 78 and 56%, respectively.

Catalyst Promotion

The zeolites were promoted by employing the wet impregnation method over the parent HZ5 zeolite (5 g) using 30 mL of aqueous solutions of (NH4)6Mo7O24·4H2O (2.48 × 10–3 M) to achieve 1 wt % Mo loading (denoted 1[Mo]HZ5), while 2[Mo]HZ5 and 4[Mo]HZ5 samples were similarly prepared by applying 60 and 90 mL of the said solution, respectively. Furthermore, the corresponding mixture was stirred for 24 h, prior to filtration. Then, the impregnated samples were dehydrated and heated at 110 °C overnight before calcination at 550 °C for 6 h, respectively. Similarly, the 2[Mo]alk-HZ5 zeolite was prepared by executing the above-mentioned Mo impregnation on alk-HZ5 zeolites by applying the aforesaid aqueous solution of (NH4)6Mo7O24·4H2O to achieve 2 wt % Mo loading with stirring for 24 h. The metal loading on all of the zeolite samples was determined by EDX.

Catalytic Characterization

XRD, X-ray fluorescence (XRF), N2 adsorption–desorption, NH3-TPD, Py-IR, ICP, FT-IR, CH microanalysis, TGA, FESEM, and FESEM/EDX techniques were applied to characterize the zeolites. The measurements were executed on the following instruments.

Powder X-ray Diffraction Studies

The crystallinity of the zeolites was determined according to the XRD patterns, carried out in the 2θ range of 5–80° obtained at room temperature and by employing a D8 Advance Bruker X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å).

Nitrogen Adsorption–Desorption Isotherms

The BET surface areas and PSDs were determined according to nitrogen adsorption–desorption isotherms by a Micromeritics ASAP 2010 instrument at −196 °C. The SBET values of the zeolites were also obtained from the N2 adsorption isotherms by applying the BET method at P/P0 = 0.05–0.25. To determine the sample surface area, 300 mg of the sample was degassed at 300 °C for 2 h prior to the adsorption of N2. VTotal was determined in terms of the adsorbed N2 at P/P0 = 0.99. The mesopore size distribution was also measured using the BJH method. The t-plot method at P/P0 = 0.1–0.4 was applied to measure SMicro and VMicro. VMeso was gauged in terms of the observed difference of VTotal and VMicro.

TPD-NH3 Measurements

TPD-NH3 was executed on a Micromeritics TPD/TPR 2900 chemisorption analyzer. Samples (0.1 g) were pretreated at 500 °C for 2 h in these experiments prior to the NH3 desorption measurement at 100–700 °C.

Py-IR Spectroscopy

Py-IR spectra of the zeolites were measured on an RXI PerkinElmer Fourier transform infrared spectrometer. The sample activation was performed at 150 °C for 3 h under vacuum (5 × 10–3 mmHg) using an Alcatel Adixen 2005SD Pascal vacuum pump. Then, the pyridine adsorption by the samples was followed at 30 °C for 1 h, prior to heating up to 150 °C, by determining their Py-IR spectra.

FESEM Analysis

FESEM images were acquired using a TESCANMIRA3-LMU scanning electron microscope, working at a potential difference of 15 kV. In these experiments, the FESEM chamber was equipped with EDX. Sample preparation was executed by applying the catalyst on a silicon wafer using the dispersion method, before coating with a gold film.

XRF Spectroscopy

The mole ratio of Si/Al in the zeolite samples was also evaluated using an Axios X-ray fluorescence spectrometer.

FT-IR Spectroscopy

Fourier transform infrared spectra were recorded by applying KBr pellets on an RXI PerkinElmer FT-IR spectrophotometer, having a resolution of 4 cm–1 at 400–4000 cm–1, to address the functional groups.

Elemental Analyses

Microanalytical determination of carbon and hydrogen was performed using a Truspec CHNS-Com (Leco).

TGA of the Samples

The thermal decomposition of the samples was evaluated by the TG method, running on an SDT Q600 v20.9 Build 20 TGA instrument at 30–750 °C by executing a heating rate of 10 °C min–1, after heating 5.53 mg of zeolite up to 150 °C, prior to purging the samples with Ar flow (10 L min–1).

95Mo Solid-State NMR

95Mo MAS NMR spectra were obtained on an 800 MHz Bruker spectrometer working at 52.1 MHz (for 95Mo) at room temperature, employing a 3.2 mm probe. The powder samples were packed in rotors and rotated at the magic-angle spinning at 20 kHz. A solid echo pulse sequence (in which half of the echo time was synchronized to the spinning rate) was used.

TEM Studies

Operating at 100 kV, TEM images were recorded on a Zeiss EM900 instrument. The powder samples were sonicated in ethanol for 15 min prior to putting on the carbon-coated Cu grids.

Catalytic Performance

To determine the zeolite performance in the MTA process, a fixed-bed tubular reactor was employed under ambient pressure.[55] The reaction temperature was set at 375 °C in these experiments, while a graduated burette equipped with a microtube pump (prep pump, Chem Tech Co. Ltd.) was used for methanol injection (WHSV of 2 h–1). Meanwhile, methanol stream was heated at 120 °C and mixed with N2 before sending to the setup. The zeolite sample (1.5 g) was placed in the central part of the tubular fixed-bed reactor, made from 380 mm tubular Pyrex (i.d. 9 mm). Cooling down the condenser to −1 °C separated the condensable off-gases accompanied by the product, escaping the reactor. A GC-Varian 3800 was used to analyze the liquid products offline, sampled during 1 h periods, at the end of the experiments. Besides, after every 8 h, the liquid hydrocarbons were collected and analyzed by a gas chromatography mass spectrometer. A Hewlett-Packard 6890 series GC (HP, Palo Alto) fitted with an injector and an HP 5973 mass-selective detector were also used for this purpose.