Mechanistic insight to acidity effects of Ga/HZSM-5 on its activity for propane aromatization

Abstract

Ga-modified HZSM-5 precursors, containing 1 wt% Ga, were first prepared using the incipient wetness impregnation method, and then subjected to one or three-times consecutive treatment by cycles of reduction in hydrogen and re-oxidation in air. The resulting Ga/HZSM-5 catalysts were characterized by N₂ physical adsorption, ICP-AES, DRIFT, Py-FTIR, NH₃-TPD, H₂-TPR, XPS, DRIFT-TPSR and MS-TPSR techniques to obtain clear mechanistic details that the acidity of these Ga/HZSM-5 catalysts affected their activity for propane aromatization. The characterization data suggested that the impregnated introduction of Ga to ZSM-5 zeolite and subsequent reduction–oxidation treatment led to a great decrease in the number of Brønsted acid sites (BAS) and promoted the formation of strong Lewis acid sites (LAS) attributed to highly dispersed Ga species. The formed strong LAS specifically promoted the dehydrogenation steps during propane aromatization, whereas the original BAS were responsible for the whole aromatization process. The TPSR results suggested that propane was converted to propylene through the dehydrogenation on BAS and strong LAS, and simultaneously converted to ethylene through the β-scission on BAS at a low temperature. With the elevation of temperature, the generated propylene and ethylene on the strong BAS and the strong LAS were further converted into BTX aromatics accompanied by hydrogen release. It was plausible that the highly efficient synergy between BAS and strong LAS could result in lower aromatization temperatures and more product of benzene (β state) over the Ga/HZSM-5 catalysts than that over the Ga-modified HZSM-5 precursors.

---

1. Introduction

The worldwide supply of liquefied petroleum gas (LPG) was about 140 million metric tonnes in 1990 and rose to about 275 million metric tonnes by year 2014.¹ Moreover, urgent demand of BTX (benzene, toluene and xylene) has drawn the attention of academic and industrial sectors to deeply investigate the conversion of lower alkanes into these aromatics. Therefore, propane (light alkane) aromatization has been extensively studied in recent years. Previous studies^2–17 focused on the bifunctional Ga modified ZSM-5 catalysts for aromatization because of their superior catalytic performance compared to Zn and Pt modified ZSM-5 catalysts.^4,8 It was found that incorporation of Ga to ZSM-5, which was prepared using different methods, such as hydrothermal in situ synthesis and post synthesis treatment (ion-exchange, impregnation, chemical vapour deposition [CVD] and reduction–oxidation treatment), could adjust the acidity of the Ga/ZSM-5 catalysts. Moreover, the varied acidity, in turn, had a significant effect on the propane aromatization performance of the catalysts.^4,6 Choudhary et al.¹⁸ found that the Si/Ga and Si/Al ratios of H-gallosilicate (MFI) prepared using the in situ hydrothermal synthesis method strongly affected its acidity/acid strength distribution, therefore inducing a dramatic difference of catalytic activity/selectivity in propane aromatization. The highest propane conversion and selectivity of aromatics over these HMFI catalysts came to about 75% and 80%, respectively. Moreover, Al-Yassir et al.¹⁶ further reported that, through de-galliation from the ZSM-5 framework in hydrothermal synthesis process, Lewis acid sites (LAS) associated with highly dispersed and reducible extra-framework Ga₂O₃ were generated on the catalyst surface, which finally resulted in high activity of propane aromatization. The optimal propane conversion and selectivity of BTX achieved was 42.0% and 52.0%, respectively. El-Malki et al.¹⁹ prepared a highly efficient Ga/ZSM-5 catalyst for aromatization using a CVD method together with subsequent water treatment. They found that the subsequent treatment of the catalyst with water could create secondary strong Brønsted acid sites (BAS), contributing to the aromatization activity of Ga/ZSM-5 catalyst. Abdul Hamid et al.²⁰ revealed that the reductive atmosphere could promote the migration of Ga species and formation of active Ga species, which further affected the acidity of Ga/HZSM-5 catalysts. These active Ga species and BAS over ZSM-5 had a synergistic effect on the catalytic performance of the Ga/HZSM-5 catalysts. The highest propane conversion and selectivity of BTX were not more than 10% and 40%, respectively. Nowak et al.³ investigated the effect of H₂–O₂ cycles on the activity of Ga/HZSM-5 prepared using the ion-exchange method. They found the synergistic effect between the reduced state Ga species generated by H₂–O₂ treatment and BAS over the ZSM-5 zeolites contributed to the highest selectivity of aromatics, which was 19.1% at a high conversion (97–99%) for the aromatization of propane at 823 K using Ga/(0.5) as catalysts. Through the structure–activity correlations between the acidity of Ga/ZSM-5 catalyst and its activity, Rodrigues et al.^3,10,11,20 thought that the reduction–oxidation treatment promoted the formation of highly dispersed oxidic gallium species, which were anchored to ion-exchange sites of the zeolite framework, and contributed to the strong LAS. Moreover, it was also found that the synergy between LAS and BAS led to alkane aromatization activity rather than exclusive effect of LAS or BAS in the reaction. It was believed that the reducibility of strong Lewis acid species facilitated the formation of reduced gallium species, which was the actual active site in alkane activation. As observed, initial rates of propane conversion were at least 20 times higher for the Ga-containing catalysts obtained by them than the respective zeolite, but were almost independent of gallium content. Considering the aromatization activities alone, gallium impregnation led to catalysts that were approximately 200 times more active in the case of the TZ35 series because of synergy between LAS and BAS.

In our previous study,¹³ a highly efficient Ga/ZSM-5 catalyst was prepared by formic acid impregnation and in situ treatment for propane aromatization. This novel treatment process promoted the formation of strong LAS attributed to highly dispersed (GaO)⁺ species through decomposition of intermediate species and the interaction between Ga₂O₃ and water steam. Based on the investigation of the catalytic system, it was thought that the high catalytic activity of this catalyst was attributed to the synergistic effect between strong LAS and BAS. The highest propane conversion and selectivity of BTX on the Ga/HZSM-5 catalyst we prepared by this novel method were 53.6% and 58.0%, respectively. It is obvious that the activities of the catalysts prepared using different methods are different. Considering the distinction between test conditions, we are reluctant to make a certain conclusions on which preparation method is preferential to obtain good catalytic performance. However, it is widely acknowledged that the presence of highly dispersed Ga on HZSM-5, which is obtained through special treatment process such as reduction–oxidation operation, is favorable for propane aromatization. This is evidenced from our present study. In the present study, the preparation process of the Ga/HZSM-5 and the formation process of active Ga species are different from the previous study.¹³ Ga-modified HZSM-5 precursors were first prepared using an incipient wetness impregnation method, and then subjected to one or three-time consecutive treatment by cycles of reduction in hydrogen and re-oxidation in air. The reduction–oxidation treatment could promote the migration of these Ga species into the zeolite and the formation of highly dispersed (GaO)⁺ species is inferred from our techniques and previous research.^3,11 It was apparent that these two different preparation processes (previous treatment and present treatment) promoted the formation of (GaO)⁺ species. Wherein, propane conversion and selectivity of the Ga/HZSM-5 catalyst prepared using reduction–oxidation method showed a significant increase of 15.2–22.3% and 9.5–10.9%, respectively, in contrast to that of the Ga/HZSM-5 catalyst prepared using traditional impregnation methods at the same test conditions. The superior activity of the Ga/HZSM-5 catalyst prepared using the reduction–oxidation method was attributed to the synergistic effect between the strong LAS generated by the (GaO)⁺ species and the BAS. Conclusively, it is apparent that the reduction–oxidation treatment improves the activity of Ga containing HZSM-5 zeolites greatly. Simultaneously, it is obvious that, if one looks forward to a high activity of Ga/ZSM-5 catalyst for propane aromatization, obtaining a suitable amount of BAS and LAS related to Ga species through adopting some various/special preparation and treatment methods is necessary.

The previous study focused on a novel proposed preparation method for catalyst and the probable formation route of highly active GaO⁺ species,¹³ whereas in the present study, we concentrated on the systematic investigation of how the acidity of Ga/HZSM-5 catalyst affected the evolution of reactants and intermediates (such as β-scission of propane to methane and ethylene and dehydrogenation of propane to propylene) on specific acid sites (BAS and LAS) of Ga/HZSM-5 as well as the sorption/desorption process of reactants and intermediates on these acid sites. In other words, we paid more attention to the relationship between the acid sites (LAS and BAS) and specific reactions as well as the sorption/desorption of reactants and intermediates, as the systematic investigation on this research was not reported to us. Therefore, in this study, we in-detail rationalized the results in terms of correlations between the acidity (Lewis acidity and Brønsted acidity) of Ga/HZSM-5 catalyst and its reactivity (specific reactions as well as the sorption/desorption process of reactants and intermediates) in propane aromatization by various characterizations of the catalysts such as N₂ physical adsorption, ICP-AES, DRIFT, Py-FTIR, NH₃-TPD, H₂-TPR, XPS, DRIFT-TPSR and MS-TPSR techniques. Furthermore, the mechanistic details that the acidity of the Ga/HZSM-5 catalysts affected their activity for propane aromatization were deeply illustrated.

2. Experimental section

2.1 Catalyst preparation

HZSM-5 zeolites (silica to aluminum mole ratio = 19) were bought from Nankai University. Ga was incorporated into this zeolite by incipient wetness impregnation with gallium nitrate solution (1 wt% Ga loading). The impregnated sample was calcined at 550 °C for 5 h under static air. Thus, the Ga/IM catalyst was gained accordingly. Subsequently, 0.15 g of Ga/IM catalyst was placed in a micro fixed-bed reactor and then subjected to the reduction–oxidation treatment. Herein, reduction was performed under a hydrogen flow of 15 mL min⁻¹ at 540 °C for 1 h and re-oxidation was carried out under an air flow of 15 mL min⁻¹ for 1 h. The resulting catalysts were denoted as Ga/IMRO1C and Ga/IMRO3C based on one and three times reduction–oxidation treatment to Ga/IM, respectively.

2.2 Catalyst characterization

The textural properties of catalyst were measured by N₂ physical sorption at 77 K using a Tristar 3000 machine. Surface areas were calculated by the BET method and micro-, meso-, and macro-pore volumes were calculated by the t-plot method.

The powder X-ray diffraction (XRD) patterns were obtained on a Rigaku MiniFlex II X-ray diffractometer using CuKα radiation. The anode was operated at 40 kV and 40 mA. The 2θ angles were scanned from 5° to 60°.

TEM micrographs were obtained by a JEOL-JEM-2100F transmission electron microscope operating at 200 kV.

Diffuse reflectance infrared Fourier transform spectra (DRIFT) measurements were obtained using a Bruker Tensor 27 instrument with a MCT detector (64 scans, 4 cm⁻¹). 20 mg of catalyst was weighed and placed in an infrared cell with KBr windows for in situ treatments. The DRIFT spectra were obtained after treating the catalyst at 300 °C for 1 h with a flow of argon.

Pyridine-adsorbed Fourier transform infrared (Py-FTIR) spectroscopy was used to determine the amount of BAS and LAS using Bruker Tensor 27 equipment. 20 mg of catalyst was pressed into a regular wafer (R = 1.3 cm⁻¹) and placed in an infrared cell. The infrared spectrum was obtained after sample treatment at 400 °C for 2 h under vacuum, and this spectrum was used as background for the adsorbed pyridine experiments. Pyridine was then adsorbed to a 5.0 × 10⁻² Pa equilibrium pressure at 40 °C. FTIR spectra were obtained after consecutive evacuation at 150, 250, 350 and 450 °C.

The temperature-programmed desorption of ammonia (NH₃-TPD) was used to test the amount and strength of the acid sites of the as-prepared catalysts in TP-5080 chemisorption instrument. The catalyst (100 mg) was pre-treated at 500 °C under a flow of N₂ (30 mL min⁻¹) for 2 h and then cooled to 100 °C. Then, NH₃ was introduced into the flow system. The TPD spectra were obtained at a ramp rate of 10 °C min⁻¹ from 100 °C to 700 °C.

The H₂-temperature-programmed reduction (H₂-TPR) was carried out to study reducibility of the catalysts with chemisorption instrument (TP-5080). The catalyst (100 mg) was pre-treated at 500 °C under a flow of N₂ (32 mL min⁻¹) for 1 h and then cooled to 100 °C, then the flow was changed to a H₂/N₂ = 0.09 mixture (35 mL min⁻¹). The temperature-programmed reduction was performed between 100 °C and 900 °C with a ramping rate of 10 °C min⁻¹.

X-ray photoelectron spectroscopy (XPS) was used to analyze the change of surface composition measured by AXIS ULTRA DLD equipment. The following photoelectron lines were obtained: Ga2p_3/2, Ga2p_1/2, Ga3d, Si2p, Al2p, O1s, O2s and C1s. The binding energy values were corrected for charging effect by referring to the adventitious C1s line at 284.5 eV.

Diffuse reflectance infrared spectra (DRIFTS) investigation of temperature-programmed surface reaction (TPSR) was carried out in an in situ reaction cell. The catalyst (20 mg) was pre-treated at 400 °C under a flow of argon (30 mL min⁻¹) for 1 h and then cooled to 50 °C. Furthermore, propane flow was introduced to the reaction cell for absorption for 0.5 h and then switched to the argon flow (30 mL min⁻¹) to remove the physical adsorption of propane (blowing for 0.5 h). The above-treated sample was heated from 50 °C to 450 °C at a ramping rate of 3 °C min⁻¹; moreover, its diffuse reflectance infrared spectra were obtained.

Mass spectrum investigation of temperature programmed surface reaction (MS-TPSR) was carried out on a chemisorption instrument (TP-5080) and OMNI star. The catalyst (100 mg) was first pre-treated at 500 °C for 1 h under a flow of argon (30 mL min⁻¹) and then cooled to 50 °C. After that, propane was introduced into the flow system until saturation sorption (about 0.5 h). The sample was heated again from 50 °C to 650 °C with a ramping rate of 3 °C min⁻¹ and the effluents from the reactor were analysed by an on-line mass spectrometer synchronously.

2.3 Catalyst evaluation

The catalyst test, as our previous illustration,¹³ was carried out in a horizontal quartz tube fixed-bed reactor at T = 540 °C, P = 100 kPa, WHSV = 6000 mL (g⁻¹ h⁻¹) and with a N₂/C₃H₈ molar ratio of 2. Outlet gas was analysed after 0.5–4.5 h reaction.

The products were analysed by an online gas chromatograph (HUAAI GC 9560) equipped with a flame ionization detector (FID) with a HP-INNOWAX capillary column, an off-line gas chromatograph (HUAAI GC 9560) equipped with a FID with an Al₂O₃ packed column, and an offline gas chromatograph (East & West GC 4000A) equipped with a thermal conductivity detector (TCD) with a carbon molecular sieves packed column, respectively. The propane conversion and product selectivity were calculated using the following eqn (1) and (2).

where X_propane is the conversion of propane, S_i is the selectivity of target product i (i = BTX and CH₄), N_propane,in and N_propane,out are the numbers of moles of propane in the inlet and outlet gas phases, respectively, and N_i,out is the total number of moles of product i.

3. Results and discussion

3.1 Catalyst composition and textural properties

The pore volume and BET surface areas (S_BET) of different catalysts are obtained based on the sorption isotherms of N₂ condensation at 77 K and the results are listed in Table 1. As observed, the micropore surface area and micropore volume of the Ga-containing ZSM-5 zeolites decreases compared to that of HZSM-5, especially for Ga/IMRO1C and Ga/IMRO3C treated by reduction–oxidation process, the decrease is more severe. It is thought that the decrease of micropore surface area and micropore volume is probably due to the pore blockage resulting from the deposition of extra-framework gallium and aluminum species within the micropores.^11,13 In addition, compared with surface areas and porous volumes of meso and macropores of Ga/IM, those of Ga/IMRO1C and Ga/IMRO3C show a slight increase. This indicates that the reduction–oxidation treatment somewhat has a potential effect on the expansion of the micropores. TEM characterization of these catalysts is also used to visualize the pore structure. As shown in Fig. S1a–d,† Ga/IMRO1C and Ga/IMRO3C show regular hexagonal prisms such as Ga/IM and HZSM-5. It suggests that impregnation and reduction–oxidation treatment slightly affect the morphologies of ZSM-5 zeolites. At the same time, it is found that no irregular pore structures as well as messy species are created after impregnation and reduction–oxidation treatment, as observed in Fig. S1a–d.† In combination of BET data, it is thought that the pore structures of the zeolites are slightly affected and most of them could be preserved during catalyst preparation and pre-treatment.

Table 1 Physical parameters of different catalysts

Sample	SBET^a (m^2 g^−1)	Smicro^a (m^2 g^−1)	Sext (m^2 g^−1)	Vtotal^b (m^2 g^−1)	Vmicro^b (m^2 g^−1)	Vext (m^2 g^−1)
a Determined by t-method.b Determined by v adsorbed at p/p^0 = 0.97.
HZSM-5	420	324	95.8	0.185	0.129	0.056
Ga/IM	400	301	99.1	0.180	0.121	0.059
Ga/IMRO1C	384	275	109	0.177	0.111	0.066
Ga/IMRO3C	381	270	111	0.177	0.109	0.068

---

The XRD patterns of the as-prepared samples reveal the typical characteristics of HZSM-5 (Fig. S2†). However, no XRD reflections ascribed to segregated bulk Ga₂O₃ or framework gallium species can be found on these catalyst.²¹ It is probably due to low Ga content, highly dispersed ions and/or extra-crystalline oxide-type domain with sizes below 4.0 nm on the external surface.¹⁶ The crystallinities of these materials are determined on the intensity of characteristic peaks (6.9–9.1° and 22.6–24.8°) in XRD patterns. The peak intensity of parent HZSM-5 is highest and its crystallinity is defined as 100%. The crystallinities of other materials depend on the ratios of their strength value of characteristic peaks divided by that of parent HZSM-5. Based on that, crystallinities of the Ga/IM (55.8%), Ga/IMRO1C (66.4%) and Ga/IMRO3C (72.4%) can be obtained. It is found that crystallinity of the Ga/IM dramatically decreases to about 55.8% from original 100% of ZSM-5. However, the crystal can be recovered partly through the reduction–oxidation treatment as observed from 66.4% crystallinity of Ga/IMRO1C and 72.4% crystallinity of Ga/IMRO3C, respectively. It is believed that the reduction–oxidation treatment is beneficial for preserving the pore structures of the zeolites.

Table 2 shows the bulk and surface compositions of as-prepared catalysts based on measurements of ICP-AES and XPS. As observed, the surface Si/Al ratios of these catalysts decrease and are lower than that of the bulk with the introduction of Ga, indicating the enrichment of aluminum on catalyst surface. The increase of aluminum implies that the framework of the Ga modified ZSM-5 zeolites suffers from de-alumination caused by introduction of Ga and the subsequent reduction–oxidation treatment. Therefore, the de-alumination of Ga/IMRO1C and Ga/IMRO3C is more severe. Furthermore, it can be observed that the ratios of Si/Ga of the catalysts' surface are lower than that of its bulk. This indicates that the Ga species are enriched on the surface of these catalysts, whereas higher surface Si/Ga ratios of Ga/IMRO1C and Ga/IMRO3C than that of Ga/IM indicates that the reduction–oxidation treatment is in favor of promoting the entry of the surface Ga into the intra-crystalline region/channels of the zeolite.

Table 2 Bulk and surface compositions of as-prepared catalysts

Sample	Si/Al ratio		Si/Ga ratio
	Bulk^a	Surface^b	Bulk^a	Surface^b
a Determined by ICP-AES method.b Determined by XPS.
HZSM-5	19.1	20.9	—	—
Ga/IM	18.9	17.2	150	21.2
Ga/IMRO1C	19.0	15.1	151	98.0
Ga/IMRO3C	19.0	12.6	151	137

---

3.2 DRIFT

To observe the OH stretching change of special groups that exist on the catalyst surface, the as-prepared catalysts are investigated by means of DRIFT, and the results are shown in Fig. 1. In the region from 3400 cm⁻¹ to 4000 cm⁻¹, four bands are observed on four samples.^3,6 Wherein, (i) the band at 3601 cm⁻¹ is ascribed to the stretching vibration of Si(OH)Al bridged bonds, which is associated with the zeolite BAS;^6,11 (ii) the band at about 3650 cm⁻¹ is ascribed to the extra-framework aluminium OH group, which is associated with zeolite LAS;² (iii) the band at 3671 cm⁻¹ is assigned to the extra-framework gallium OH group, which is associated with weak LAS;^2,22 and (iv) the band at 3737 cm⁻¹ is attributed to the terminal Si–OH group. As observed in Fig. 1, the intensity of the peak at 3601 cm⁻¹ on Ga/IM decreases with Ga incorporation to HZSM-5, suggesting that the number of BAS decreases. It is plausible that the decrease results from exchange of the partially acidic protons with ionic gallium species or de-alumination of ZSM-5 zeolites.⁷ Moreover, a further reduction in intensity of the 3601 cm⁻¹ peak occurs on Ga/IMRO1C and Ga/IMRO3C, implying that the reduction–oxidation treatment also leads to the decrease of BAS, whereas the intensity of the peak close to 3650 cm⁻¹ on Ga-containing ZSM-5 zeolites, especially for Ga/IMRO1C and Ga/IMRO3C, increases in contrast with that on HZSM-5; however, the spectra indicate that de-alumination has occurred with Ga incorporation to HZSM-5 and is aggravated by subsequent reduction–oxidation treatment, as previously verified by the results of XPS in Table 2. Furthermore, an increase in intensity of the peak at 3671 cm⁻¹ on Ga-containing ZSM-5 zeolites is not obvious, probably because part of the peak is concealed by the peak of Al–OH and more surface Ga enters into the intra-crystalline region/channels of zeolites by reduction–oxidation treatment.

Fig. 1 DRIFT spectra in the OH stretching region of the as-prepared catalysts. (a) HZSM-5; (b) Ga/IM; (c) Ga/IMRO1C; and (d) Ga/IMRO3C.

3.3 Py-FTIR

FTIR spectra of pyridine adsorption on different catalysts are obtained to identify the BAS and LAS, and the quantified results of different types of acid sites are shown in Fig. 2. It is noted that there are three adsorption peaks of pyridine at ∼1450 cm⁻¹, ∼1490 cm⁻¹ and ∼1540 cm⁻¹, which are assigned to the characteristic bands of LAS, the co-contribution of LAS and BAS, and BAS, respectively (as shown in Fig. 2A).¹¹ As observed in Fig. 2B, the number of BAS decreases with the elevated evacuation treatment temperature for all as-prepared samples. It implies that both weak and strong BAS are present on these catalysts. Compared with HZSM-5, Ga/IM shows a slight decrease in the number of BAS. Furthermore, the decrease in BAS is greatly deteriorated when Ga/IM is treated with reduction–oxidation as that of Ga/IMRO1C and Ga/IMRO3C in Fig. 2B. The results are consistent with the previous infrared results in the OH stretching region. The concentration of pyridine adsorption on LAS of the catalysts as a function of evacuation temperature is shown in Fig. 2C. The figure shows that pyridine adsorbed on LAS of HZSM-5 disappears substantially after evacuation at 250 °C. However, as for Ga/IMRO1C and Ga/IMRO3C, pyridine absorption is still present even at temperatures as high as 450 °C, indicating that strong LAS are generated as a result of Ga incorporation via the subsequent reduction–oxidation treatment. Previous study^11,23–26 reported that the incorporation of highly dispersed Ga species to ZSM-5 zeolites could promote the formation of strong LAS, which improved dehydrogenation ability of the catalysts. That is to say, the reduction–oxidation treatment to Ga/IM enhances the formation of highly dispersed Ga species contributing to strong LAS. Moreover, the numbers of BAS and the numbers of LAS over the Ga/HZSM-5 catalysts prepared by reduction–oxidation treatment are higher than the Ga/HZSM-5 catalysts prepared by previous treatment (formic acid impregnation and in situ treatment).¹³ That is because the reduction–oxidation treatment leads to less framework damage to the catalysts. From Fig. 2D, it can be observed that pure HZSM-5 has the largest ratio of Brønsted acid concentration/Lewis acid concentration, indicating that large amounts of BAS are present on its surface. However, the ratio decreases with Ga incorporation (such as Ga/IM). Furthermore, the decrease is very dramatic after the subsequent reduction–oxidation treatment as Ga/IMRO1C and Ga/IMRO3C show. Previous research^4,27 suggested that BAS of HZSM-5 are exclusive active sites for propane aromatization during the whole reaction, whereas the LAS associated with extra-framework aluminium OH do not contribute to the propane aromatization. However, our prepared Ga/IMRO1C and Ga/IMRO3C exhibited higher conversion of propane and selectivity of BTX than Ga/IM and HZSM-5, as shown in Fig. 3. In combination with other^4,11 and our previous reports,¹³ it can be deduced that the formed strong LAS by Ga modification to HZSM-5 is equally important for propane aromatization. It can be concluded, apparently, that the synergy effect between strong LAS is related to highly dispersed Ga species and BAS on Ga/IMRO1C and Ga/IMRO3C contributes to propane aromatization activity rather than the exclusive effect of BAS.

Fig. 2 FTIR spectra of the pyridine adsorption on the catalysts and the quantified results of different acid sites. (A) Spectrum of HZSM-5 (a), Ga/IM (b), Ga/IMRO1C (c) and Ga/IMRO3C (d) after evacuation at 450 °C; (B) Brønsted acid site concentration of samples at different temperatures; (C) Lewis acid concentration of samples at different temperatures; (D) Brønsted acid concentration/Lewis acid concentration ratios of samples at different temperatures.

Fig. 3 Propane conversion (a) and selectivity of BTX (b) as a function of time on stream for the as-prepared catalysts. Reaction conditions: P = 100 kPa, T = 540 °C, WHSV = 6000 mL (g⁻¹ h⁻¹) and N₂/C₃H₈ molar ratio = 2.

3.4 NH₃-TPD

Shown in Fig. 4 are the NH₃-TPD profiles obtained over the as-prepared catalysts. As observed, all catalysts show two desorption peaks of NH₃ (centred at 150–340 °C and 360–570 °C), which are associated with weak acid sites and strong acid sites, respectively. Compared with two desorption peaks on HZSM-5, the Ga-containing ZSM-5 zeolites don't show an obvious shift indicating that the incorporation of Ga to HZSM-5 induces little change of acid strength. On the other hand, one can note that the intensity of desorption peaks at 150–340 °C on the Ga-containing samples show a slight decrease in comparison with that on HZSM-5, whereas the intensity of the desorption peak at 360–590 °C on Ga/IMRO1C and Ga/IMRO3C dramatically decreases. It also can be further confirmed based on the data of the quantified acid sites as listed in Table S1.† The abovementioned result suggests that the present method of preparing catalyst can obviously reduce the amount of total acid sites, and most of the decrease is caused by the removal of partial strong acid sites, which is actually associated with the BAS of the catalysts. Thus, it is realized that the decrease of the strong acid sites ascribed to BAS is mainly due to the de-alumination of ZSM-5 zeolite and the exchange of BAS′ acidic proton with Ga species.⁷ The results are consistent with previous Py-IR results. Moreover, the numbers of strong acid sites over the Ga/HZSM-5 catalysts treated by reduction–oxidation are more than that over the Ga/HZSM-5 catalysts prepared by the previous treatment,¹³ because the reduction–oxidation causes less de-alumination of the framework and decrease in BAS.

Fig. 4 NH₃-TPD profiles of different catalysts. (a) HZSM-5; (b) Ga/IM; (c) Ga/IMRO1C; and (d) Ga/IMRO3C.

3.5 H₂-TPR

H₂-TPR profiles of the as-prepared samples are depicted in Fig. 5, which show the Ga species evolution for each sample. The pure Ga₂O₃ exhibits two reduction peaks at about 507 °C and 771 °C, which are ascribed to small Ga₂O₃ particles and segregated bulk Ga₂O₃, respectively,³ whereas Ga/IM shows three reduction peaks centred at about 528 °C, 671 °C and 810 °C, which are assigned to the small Ga₂O₃ particles, highly dispersed gallyl ion species probably (GaO)⁺ and segregated bulk Ga₂O₃ according to previous research.^3,9,21,28,29 In the case of Ga/IMRO1C and Ga/IMRO3C, there are only two reduction peaks observed (centered at about 514–522 °C and 633–692 °C) and the peak ascribed to the segregated bulk Ga₂O₃ tends to disappear. It is thought that the segregated bulk Ga₂O₃ can be disassembled into smaller Ga₂O₃ particles on the ZSM-5 zeolite host as well as formation of well-dispersion gallyl ion species via the reduction–oxidation treatment. Through fitting the TPR profiles and calculating their peak areas (Table S2†), it is found that the (GaO)⁺ content of Ga/IMRO1C and Ga/IMRO3C are higher than that of the Ga/IM, implying that the reduction–oxidation treatment promotes the formation of highly dispersed (GaO)⁺ species. This is further verified by XPS results. From the characterization of XPS (Fig. S3†), it is noted that the Ga2p_3/2 binding energy (BEs) of Ga/IM (1119.1 eV) is higher than that for Ga₂O₃ (1117.9 eV), indicating a strong interaction between the ionized Ga species and ZSM-5 zeolite. However, Ga2p_3/2 BEs for the Ga/IMRO3C ((1118.6 eV)) and Ga/IMRO1C ((1118.8 eV)) are lower than that for Ga/IM. This is because the reduction–oxidation treatment to Ga/IM promotes the formation of (GaO)⁺ species easily and induces strong covalent-character bonding between the highly dispersed (GaO)⁺ species and the framework of the ZSM-5.³ It has also been proved that highly dispersed (GaO)⁺ species could exchange acidic protons of BAS of the zeolite framework, contributing to the strong LAS.¹¹ The results are concordant with the Py-FTIR results that there are more strong LAS on Ga/IMRO1C and Ga/IMRO3C.

Fig. 5 H₂-TPR profiles of different samples. (a) Ga₂O₃; (b) Ga/IM; (c) Ga/IMRO1C; and (d) Ga/IMRO3C.

3.6 TPSR experiment

TPSR is an important technique to study the surface reaction mechanisms of supported metal catalysts,^30–36 because it can monitor the evolution of species instantaneously and obtain lots of visualized information during reactions directly. Therefore, in this research, DRIFTS-TPSR and MS-TPSR are used to investigate the mechanistic details in the propane aromatization reaction.

In combination with these DRIFT, FTIR, XPS, and NH₃-TPD data, we pay more attention to study systematically how acidity of Ga/HZSM-5 catalysts affect the evolution of reactants, intermediates and products on the surface of Ga/HZSM-5 as well as the sorption/desorption process by means of DRIFTS-TPSR and MS-TPSR.

DRIFT-TPSR. The DRIFT spectra of different catalysts during TPSR experiment are shown in Fig. 6b, d, f and h, and DRIFT spectra of TPSR experiment without introduction of propane are also shown for comparison in Fig. 6a, c, e and g. As observed in Fig. 6b, d, f and h, multiple absorption peaks (at about 2886 cm⁻¹, 2904 cm⁻¹, 2922 cm⁻¹, 2967 cm⁻¹ and 2982 cm⁻¹) over HZSM-5, Ga/IM, Ga/IMRO1C and Ga/IMRO3C appear in a wide region within 2800–3000 cm⁻¹ when the propane is introduced to the in situ infrared cell in which the catalyst is placed. These peaks are assigned to the adsorbed alkyl species, which are characteristic of ethyl and/or methyl groups bonded to Ga.^37,38 On the other hand, other bands such as 2819 cm⁻¹, 2844 cm⁻¹ and 2856 cm⁻¹ are probably assigned to the C–H stretching of those alkanes with three or more carbon atoms such as propyl and butyl.^37,38 Therefore, it is plausible that the bands in the 2800–3000 cm⁻¹ region should be assigned to the C–H stretching of propane as well as the intermediates such as methane and ethane. Simultaneously, it can be observed that the intensity of these peaks (at about 2886 cm⁻¹, 2904 cm⁻¹, 2922 cm⁻¹, 2967 cm⁻¹ and 2982 cm⁻¹) over HZSM-5, Ga/IM, Ga/IMRO1C and Ga/IMRO3C first increases with enhanced temperature until about 125 °C and then decreases as a function of temperature (this is also observed based on the profiles of the peak intensity as a function of temperature in Fig. S4†). Within the reaction temperature of 50–125 °C, the dehydrogenation and β-scission of propane occur mainly with the increase of temperature, accordingly forming the intermediates with ethyl and/or methyl groups such as methane, ethane, and ethylene. However, when the reaction temperature is further elevated (>125 °C), the formed intermediate species are quickly and easily converted to higher carbon products such as butane and BTX aromatics. This suggests that propane aromatization is strongly affected by temperature. In another peak region: around 3500–3800 cm⁻¹, it is also found that there are three hydroxyl (–OH) vibration peaks centred at 3601 cm⁻¹, 3650 cm⁻¹ and 3737 cm⁻¹ from Fig. 6a, c, e and g, which are ascribed to the stretch of Si(OH)Al group, extra-framework Al–OH group and extra-framework Si–OH group, respectively. The intensity of the peaks increases with the increase of temperature, which indicates an enhanced thermal vibration of chemical bond as a function of temperature. In addition, a reduction in peak intensity of Si(OH)Al group can be observed with Ga introduction to HZSM-5 zeolite, indicating decreased amounts of BAS.⁷ The peak intensity of the Al–OH group is enhanced as well as the appearance of the Ga–OH group with Ga incorporation. These results are consistent with the previous DRIFT results. However, when propane was introduced into the catalytic system, the peak intensities of the Si(OH)Al group associated with BAS and the Al–OH group associated with LAS for these catalysts decrease to different extents, as shown in Fig. 6b, d, f and h. Especially for the Si(OH)Al group, the decrease is very severe. Yet the Si–OH groups with little acidity show no decrease in peak intensity. The results imply that propane and intermediates are more facilely absorbed on the acid sites, especially on the strong acid sites.^39–41 From Fig. 6, it is observed that peak intensity of these Si(OH)Al, Al–OH, and Si–OH groups over the as-prepared catalysts change with the increase of temperature accordingly. As is well known, the peak intensity could reflect the relative amount of the groups on catalyst surface, which is useful to reveal the acidic properties of catalyst. However, it is very difficult to observe the obvious difference of the peak intensity with various temperatures. To calculate accurately the intensity of each peak, the DRIFT spectra of these samples from 3500 cm⁻¹ to 3800 cm⁻¹ are de-convoluted. Different hydroxyl (–OH) stretching groups (Si(OH)Al, Al–OH, Ga–OH and Si–OH) are identified from four peaks with the following wavenumbers: 3601 cm⁻¹, 3650 cm⁻¹, 3671 cm⁻¹ and 3737 cm⁻¹, respectively, by deconvolution of the experimental spectra with Gaussian peaks (R² > 0.99) after deducting the base line. The peak intensities of these bands (at 3601 cm⁻¹, 3650 cm⁻¹, 3671 cm⁻¹ and 3737 cm⁻¹) over the catalysts as a function of temperature are plotted and the profiles are displayed in Fig. S5.† As observed, the intensities of Si(OH)Al peaks over these Ga-containing ZSM-5 catalysts (Fig. S5d, S5f and S5h†) are lower than that over HZSM-5 (Fig. S5b†), whereas the intensities of Al–OH peaks are somewhat higher. Moreover, with the introduction of propane to this catalytic system (S5b, S5d, S5f and S5h†), the intensities of Si(OH)Al peaks and Al–OH peaks over these catalysts decrease especially for the Si(OH)Al peaks, whereas the intensities of Si–OH peaks show little increase. It is probable that the decrease is caused by adsorption of propane and intermediates on the acid sites ascribed to Si(OH)Al and Al–OH groups. This is in accordance with the conclusion obtained from Fig. 6. The intensity of Ga–OH peak at 3671 cm⁻¹ as a function of temperature is also investigated. It can be observed from Fig. S5† that the intensity of Ga–OH group on the Ga modified ZSM-5 zeolites is similar to Al–OH group, implying that Ga–OH group, which shows weak Lewis acidity is probably also the sorption center for propane and intermediates.²²

Fig. 6 (a–h) DRIFT spectra of the as-prepared samples during TPSR experiments from 50 to 450 °C in the region of 2800–3800 cm⁻¹. (a) HZSM-5 (without introduction of propane); (b) HZSM-5; (c) Ga/IM (without introduction of propane); (d) Ga/IM; (e) Ga/IMRO1C (without introduction of propane); (f) Ga/IMRO1C; (g) Ga/IMRO3C (without introduction of propane); and (h) Ga/IMRO3C.

In addition, it is difficult to detect the characteristic peak of GaO⁺ species by infrared spectroscopy even using probe molecules such as carbon monoxide;¹⁰ however, the scientific recognition that highly dispersed (GaO)⁺ species are present on Ga/IMRO1C and Ga/IMRO3C has been verified as a result of the impregnated incorporation of Ga and subsequent reduction–oxidation treatment, inferred from previous Py-FTIR, H₂-TPR and XPS results. As thought,¹¹ the (GaO)⁺ species can exchange with acidic protons of BAS, generating strong LAS. We deduce that the strong LAS should be the sorption centers for propane and intermediates based on the strong sorption of propane and intermediates on the strong acid sites as shown above.

Conclusively, the above results indicate that the sorption of propane and intermediates mainly occurs on acid sites, especially on strong acid sites. Moreover, the generated strong LAS related to highly dispersed Ga species and BAS on ZSM-5 zeolites have a synergistic effect on sorption/desorption processes of propane and its intermediates.

MS-TPSR. Propane-TPSR profiles for the as-prepared catalysts are shown in Fig. 7. These profiles are divided into three temperature regions: 150–270 °C (temperature region I), 270–400 °C (temperature region II) and 400–550 °C (temperature region III). In every temperature region, the detailed information (formation temperatures and peak intensity) of mainly the desired reactant and products (propane, methane, ethylene, propylene, hydrogen, benzene, toluene and xylene) are collected and listed in Table S3–S6.† As observed in Fig. 7a, the desorption peaks of propane, methane, ethylene, propylene and hydrogen over HZSM-5 appear in the region I. According to the literature,^4,27 it could be understood that propylene was generated from the dehydrogenation of propane with release of hydrogen, whereas methane was produced from β-scission of propane accompanied by ethylene release at the primary stage of propane aromatization. Herein, dehydrogenation and β-scission of propane almost occur instantaneously based on the close starting formation temperatures of methane and propylene. As for region II and III in Fig. 7a, BTX can be observed, indicating the occurrence of an aromatization reaction. It has been previously accepted that propylene and ethylene are primary intermediates for producing BTX aromatics. Therefore, it is deduced that in the present TPSR experiment, a part of the formed propylene and ethylene in region I desorbs from the catalyst surface while another part is further converted into benzene, toluene and xylene with the elevated temperature by a series of processes such as dehydrogenation, cyclization and aromatization. It is apparent that no less than two active sites for propane aromatization exist on the pure HZSM-5. Wherein, one type of active site contributes to the generation of propylene and ethylene only, and another type of active site turn to contribute to two processes containing the generation of propylene and ethylene and their aromatization to BTX. As for Ga-containing ZSM-5 zeolites (Fig. 7b–d), the distribution and adscription of the peaks are similar to that of HZSM-5, but the peak intensity and formation temperature of the products, such as hydrogen, ethylene, and BTX, are tremendously different. The methane and ethylene both come from the β-scission of propane over Ga/IM, Ga/IMRO1C and Ga/IMRO3C as HZSM-5 shows. However, only trace amount of ethylene can be observed on Ga/IM, Ga/IMRO1C and Ga/IMRO3C in region II and III (as shown in Fig. 7b–d). Because most of the formed intermediate ethylene obtained by the β-scission of propane over these Ga-containing ZSM-5 catalysts can be converted to BTX product rapidly considering that ethylene and propylene aromatization is quite an easy process. Therefore, the result that the amount of ethylene over HZSM-5 is more than that over other Ga-containing ZSM-5 catalysts is due to the lower activity of HZSM-5 in aromatization. Compared with the peak intensity of hydrogen over HZSM-5, the same over the Ga/IM is obviously enhanced in region II and III (as shown in Fig. 7b and Table S4†), indicating that the incorporation of Ga promotes the dehydrogenation of propane, and intermediates, to form BTX. In the case of Ga/IMRO1C and Ga/IMRO3C, the peak intensity of hydrogen in region II and III is also enhanced similar to that of Ga/IM. In fact, through curve-fitting results of the detected hydrogen and benzene profiles for the catalysts (Fig. S6†), it is notable that Ga/IMRO1C and Ga/IMRO3C show an additional dehydrogenation peak (defined as θ peak) besides those α, β and γ dehydrogenation peaks, as shown in Fig. S6a.† It is plausible that the appearance of the θ peak caused by the reduction–oxidation treatment is closely related with the improved dehydrogenation behavior of Ga/IMRO1C and Ga/IMRO3C. Based on a slight shift of the peaks of benzene (α state), toluene (α state) and xylene (α state) to higher temperatures over the Ga/IM than that over the HZSM-5, it is thought that the impregnated incorporation of Ga can lead to an increased aromatization temperature. As observed in Fig. 7c, the peaks of products, such as benzene (α state), toluene (α state) and xylene (α state), over the Ga/IMRO1C shift to lower temperatures by 25 °C, 25 °C and 18 °C (Table S5†), respectively, compared to the same over Ga/IM. This suggests that the benzene (α state), toluene (α state) and xylene (α state) are more facilely generated on Ga/IMRO1C at lower temperatures than Ga/IM. Moreover, it can also be found that the peak intensity of benzene on Ga/IMRO1C increases in region III in contrast with that on Ga/IM (Fig. 7c and S6b†), which implies that more benzene (β state) are generated on Ga/IMRO1C. The case of Ga/IMRO3C catalyst is similar to Ga/IMRO1C (as shown in Fig. 7d and Table S6†), except for a slight difference on temperature shift.

Fig. 7 (a–d) Propane-TPSR profiles for the as-prepared catalysts. (a) HZSM-5; (b) Ga/IM; (c) Ga/IM RO1C; and (d) Ga/IMRO3C.

Propane aromatization belongs to a tandem process and contains a series of steps such as dehydrogenation and β-scission of propane to propylene and ethylene, oligomerization and interconversion of olefins, olefin alkylation, cyclization and aromatization.^4,27 Herein, each step during the reaction is affected by the acidity of catalysts. The DRIFTS-TPSR results have suggested that the sorption of propane and intermediates mainly occurs on acid sites such as BAS related to the Si(OH)Al group and LAS associated with the Al–OH group over ZSM-5 zeolites. Furthermore, previous studies^4,27 also showed that the BAS over ZSM-5 zeolite were the exclusive active sites for propane aromatization on pure HZSM-5, contributing to the whole processes of propane aromatization, whereas LAS over ZSM-5 zeolites were only sorption centers. In the Py-FTIR results in Fig. 2D, we observed that the ratios of Brønsted acid concentration/Lewis acid concentration for HZSM-5 are greater than 6.0 at all evacuation temperatures, while the ratio of weak acid site/strong acid site is 1.2 (Table S1†). This indicates that BAS over ZSM-5 zeolite are comprised of weak and strong BAS, which probably corresponds to two types of active sites as mentioned above from the MS-TPSR results. It is thought that both the dehydrogenation of propane to propylene and the β-scission of propane to methane and ethylene occur on the weak and strong BAS over HZSM-5 at low temperatures. With the elevation of reaction temperature, parts of generated propylene and ethylene on strong BAS are further converted into BTX aromatics. In our present catalytic system, the acidic properties of catalyst actually have a remarkably close relation with Ga incorporation. Introduction of Ga to ZSM-5 zeolite can lead to a modest decrease in the number of BAS, which is caused by the exchange of BAS′ acidic protons with Ga species and de-alumination of ZSM-5 zeolite, and the generation of more LAS related to Ga species. As for Ga/IM, the addition of Ga³⁺ ions into the ZSM-5 inter-crystalline structure is very difficult because of the highly positive electrostatic charge and the bulky size of [Ga(H₂O)₆]³⁺ aqua complexes, and thus Ga³⁺ ions tend to reside predominantly on the external surface of zeolite as single extra-crystalline Ga₂O₃.⁴² It has been verified that the extra-crystalline Ga₂O₃ species have relatively low activity in propane aromatization because their activation is remarkably complicated and difficult. Al-Yassir and Rane et al.^2,10 thought that the active species were formed through two main processes: in situ reduction of extra-crystalline Ga₂O₃ to the highly mobile Ga₂O by H₂ or hydrocarbons in the feed and subsequent interaction with BAS to form a Z–Ga⁺ structure (Z⁻ represents an ion-exchange site at the zeolite framework), as shown in eqn (3) and (4).

Ga₂O₃ + 2H₂ → Ga₂O + 2H₂O (3)

Ga₂O + 2ZH⁺ → 2Z–Ga⁺ + H₂O (4)

Furthermore, the Z–Ga⁺ structure can catalyse propane to propylene as well as subsequent aromatization via dehydrogenation (as shown in eqn (5) and (6)).¹⁰

Z–Ga⁺ + C₃H₈ → Z–Ga⁺H(C₃H₇) (5)

Z–Ga⁺H(C₃H₇) → Z–Ga⁺ + H₂ + C₃H₆ (6)

In the case of Ga/IMRO1C and Ga/IMRO3C, activation of Ga species proceeds relatively facilely. The reduction–oxidation treatment promotes the formation of these highly dispersed Ga species, probably (GaO)⁺. After exchanging with acidic protons of BAS, the (GaO)⁺ species are anchored to ZSM-5 zeolite to form the structures of Z–GaO⁺, which are ascribed to strong LAS. Finally, the Z–Ga⁺ structure is easily obtained via deoxygenation of Z–GaO⁺ in hydrocarbons or H₂ flow (as show in eqn (7)).¹⁰

Z–GaO⁺ + H₂ → Z–Ga⁺ + H₂O (7)

Therefore, it is apparent that the generated strong LAS related to highly dispersed Ga species are active centers for propane and intermediates on Ga/IMRO1C and Ga/IMRO3C for the aromatization reaction. According to a previous study,¹¹ these strong LAS play a key part in the dehydrogenation process, whereas BAS are responsible for the whole aromatization process. Over Ga/IMRO1C and Ga/IMRO3C, dehydrogenation of propane to propylene occurs on the BAS and strong LAS at low temperatures in the propane-TPSR experiment, whereas β-scission of propane to methane and ethylene only occurs on BAS. With elevation of temperature, the generated propylene and ethylene are further converted into BTX aromatics with release of hydrogen on strong BAS and strong LAS. Based on present TPSR results and our discussion, a new and main reaction pathway for propane transformation is well designed, as shown in Scheme 1.

Scheme 1 Main reaction pathway for propane transformation over Ga/HZSM-5. Step (1) dehydrogenation of propane to propylene; step (2) β-scission of propane to methane and ethylene; step (3) conversion of propylene and ethylene to benzene, toluene and xylene.

Apart from that, the incorporation of Ga adds a new dehydrogenation route in propane aromatization and improves dehydrogenation ability of the Ga-containing ZSM-5 zeolites inferred from the above discussions.⁴ This is further verified by the MS-TPSR results that the peak intensity of hydrogen over Ga-containing ZSM-5 zeolites is obviously enhanced in region II and III with Ga incorporation to HZSM-5. Furthermore, the reduction–oxidation treatment to Ga/IM promotes the formation of strong LAS, which are probably responsible for the generation of the additional θ dehydrogenation peak on Ga/IMRO1C and Ga/IMRO3C. Furthermore, the strong LAS could enhance dehydrogenation steps tremendously in the propane aromatization reaction, including the dehydrogenation of propane to propylene and the conversion of propylene and ethylene to BTX aromatics. It is concluded that the synergistic effect between strong LAS and BAS on Ga/IMRO1C and Ga/IMRO3C contribute to dehydrogenation and β-scission of propane as well as the subsequent aromatization reaction process. It is the synergistic reaction between these two types of acid sites that results in lower aromatization temperature and larger amounts of product of benzene (β state) on Ga/IMRO1C and Ga/IMRO3C than that over Ga/IM. That may be the reason why Ga/IMRO1C and Ga/IMRO3C show higher catalytic performance for propane aromatization than HZSM-5 and Ga/IM, as shown in Fig. 3.

4. Conclusion

The Ga-modified HZSM-5 precursors (Ga/IM), containing 1 wt% Ga, was prepared by the incipient wetness impregnation method, and then subjected to one or three-times treatment of reduction in hydrogen and re-oxidation in air. The resulting Ga/IMRO1C and Ga/IMRO3C catalysts were characterized by techniques, such as N₂ physical adsorption, ICP-AES, DRIFT, Py-FTIR, NH₃-TPD, H₂-TPR, XPS, DRIFT-TPSR and MS-TPSR, to understand the correlations between acidity of these catalysts and their activity for propane aromatization. The characterization data suggested that the impregnated introduction of Ga to ZSM-5 zeolite and subsequent reduction–oxidation treatment led to a great decrease in the numbers of BAS because of the exchange of BAS′ acidic protons with Ga species as well as the severe de-alumination of the ZSM-5 framework, and thus facilely promoted the formation of strong LAS attributed to highly dispersed Ga species, probably (GaO)⁺, over Ga/IMRO1C and Ga/IMRO3C in comparison with that over Ga/IM. On the other hand, both the BAS and strong LAS were the sorption centers and reaction centers for propane and its intermediates. The strong LAS were experts in promoting the dehydrogenation steps during propane aromatization, whereas the BAS are responsible for the whole aromatization process. According to the TPSR results, the dehydrogenation of propane to propylene occurs on the BAS and strong LAS, whereas β-scission of propane to methane and ethylene only occurs on the BAS at low temperature. With the elevation of temperature, part of the generated propylene and ethylene on the strong BAS and strong LAS were further converted into BTX aromatics with release of hydrogen. It was plausible that the synergistic effect between BAS and strong LAS contributed to the lower aromatization temperature of propane into BTX and more product of benzene (β state) over Ga/IMRO1C and Ga/IMRO3C than that over the Ga/IM.

Acknowledgements

This research was supported by the Chinese Academy of Sciences Knowledge Innovation Project (No. KGCX2-YW-318-1), and the Special-funded Program on National Key Scientific Instruments and Equipment Development of China (No. 2011YQ120039).

Notes and references

1. K. Otto, 22nd World LP Gas Forum, Rio de Janeiro, Brazil, October 8th 2009.
2. N. Al-Yassir, M. N. Akhtar, K. Ogunronbi and S. Al-Khattaf, J. Mol. Catal. A: Chem., 2012, 360, 1–15.
3. I. Nowak, J. Quartararo, E. G. Derouane and J. C. Védrine, Appl. Catal., A, 2003, 251, 107–120.
4. A. Bhan and W. Nicholas Delgass, Catal. Rev.: Sci. Eng., 2008, 50, 19–151.
5. V. R. Choudhary, A. K. Kinage and T. V. Choudhary, Chem. Commun., 1996, 2545–2546.
6. R. Fricke, H. Kosslick, G. Lischke and M. Richter, Chem. Rev., 2000, 100, 2303–2405.
7. M. García-Sánchez, P. C. M. M. Magusin, E. J. M. Hensen, P. C. Thüne, X. Rozanska and R. A. van Santen, J. Catal., 2003, 219, 352–361.
8. G. Krishnamurthy, A. Bhan and W. N. Delgass, J. Catal., 2010, 271, 370–385.
9. B. S. Kwak and W. M. H. Sachtler, J. Catal., 1993, 141, 729–732.
10. N. Rane, A. R. Overweg, V. B. Kazansky, R. A. van Santen and E. J. M. Hensen, J. Catal., 2006, 239, 478–485.
11. V. D. Rodrigues, J. G. Eon and A. C. Faro, J. Phys. Chem. C, 2010, 114, 4557–4567.
12. V. R. Choudhary, K. Mantri and C. Sivadinarayana, Microporous Mesoporous Mater., 2000, 37, 1–8.
13. H. Xiao, J. Zhang, X. Wang, Q. Zhang, H. Xie, Y. Han and Y. Tan, Catal. Sci. Technol., 2015, 5, 4081–4090.
14. T. Choudhary, A. Kinage, S. Banerjee and V. Choudhary, Energy Fuels, 2006, 20, 919–922.
15. A. Ausavasukhi and T. Sooknoi, Appl. Catal., A, 2009, 361, 93–98.
16. N. Al-Yassir, M. N. Akhtar and S. Al-Khattaf, J. Porous Mater., 2011, 19, 943–960.
17. V. R. Choudhary, A. K. Kinage and T. V. Choudhary, Appl. Catal., A, 1997, 162, 239–248.
18. V. R. Choudhary, A. K. Kinage, C. Sivadinarayana, P. Devadas, S. D. Sansare and M. Guisnet, J. Catal., 1996, 158, 34–50.
19. E. M. El-Malki, R. A. van Santen and W. M. H. Sachtler, J. Phys. Chem. B, 1999, 103, 4611–4622.
20. S. B. A. Hamid, E. G. Derouane, P. Meriaudeau and C. Naccache, Catal. Today, 1996, 31, 327–334.
21. B. Zheng, W. Hua, Y. Yue and Z. Gao, J. Catal., 2005, 232, 143–151.
22. C. Otero Areán, A. L. Bellan, M. P. Mentruit, M. R. g. Delgado and G. T. Palomino, Microporous Mesoporous Mater., 2000, 40, 35–42.
23. O. A. Anunziata and L. B. Pierella, Catal. Lett., 1993, 19, 143–151.
24. S. Altwasser, A. Raichle, Y. Traa and J. Weitkamp, Chem. Eng. Technol., 2004, 27, 1262–1265.
25. A. Raichle, S. Moser, Y. Traa, M. Hunger and J. Weitkamp, Catal. Commun., 2001, 2, 23–29.
26. C. Otero Arean, B. Bonelli, G. Turnes Palomino, A. M. Canaleta Safont and E. Garrone, Phys. Chem. Chem. Phys., 2001, 3, 1223–1227.
27. A. Montes and G. Giannetto, Appl. Catal., A, 2000, 197, 31–39.
28. K. M. Dooley, C. Chang and G. L. Price, Appl. Catal., A, 1992, 84, 17–30.
29. G. D. Meitzner, E. Iglesia, J. E. Baumgartner and E. S. Huang, J. Catal., 1993, 140, 209–225.
30. Y. Y. Ji, W. Z. Li, H. Y. Xu and Y. X. Chen, Catal. Lett., 2001, 71, 45–48.
31. R. Lin, M. F. Luo, Q. Xin and G. Q. Sun, Catal. Lett., 2004, 93, 139–144.
32. D. Chen, Z. Qu, W. Zhang, X. Li, Q. Zhao and Y. Shi, Colloids Surf., A, 2011, 379, 136–142.
33. R. Q. Long and R. T. Yang, J. Catal., 2001, 198, 20–28.
34. D. Ma, Y. Shu, M. Cheng, Y. Xu and X. Bao, J. Catal., 2000, 194, 105–114.
35. C. Zhao and I. Wachs, Catal. Today, 2006, 118, 332–343.
36. X. Yang, Y. Wei, Y. Su and L. Zhou, Fuel Process. Technol., 2010, 91, 1168–1173.
37. V. d. O. Rodrigues and A. C. F. Júnior, Appl. Catal., A, 2012, 435, 68–77.
38. A. Kurbakova, L. Leimes, V. Aleksanyan, L. Golubinskaya, E. Zorina and V. Bregadze, J. Struct. Chem., 1974, 15, 961–969.
39. F. Eder, M. Stockenhuber and J. A. Lercher, Stud. Surf. Sci. Catal., 1995, 97, 495–500.
40. F. Eder and J. A. Lercher, J. Phys. Chem. B, 1997, 101, 1273–1278.
41. F. Eder and J. A. Lercher, Zeolites, 1997, 18, 75–81.
42. E. S. Shpiro, D. P. Shevchenko, O. P. Tkachenko and R. V. Dmitriev, Appl. Catal., A, 1994, 107, 147–164.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1–S4; Table S1–S6. See DOI: 10.1039/c5ra15227e

---