Mordenite is a zeolite that has an orthorhombic structure with the dimensions of a unit cell [1-5] of: a ¼ 18:13) b ¼ 29:49) c ¼ 7:52). As mordenite has a Si/Al molar ratio ≥5, it is very resistant to severe thermal and chemical treatments. At a Si/Al molar ratio equal to 5 the completely hydrated sodium form has the ideal composition of Na₈Al₈Si₄₀O₉₆.24H₂O [1-10]. The mordenite structure consists of two channel systems: an elliptical pore channel (12MR 6.7×7.0 Å), which runs parallel to the c-axis and another channel that runs parallel to the b-axis (8MR 2.6×5.7 Å). Due to its high thermal and acid stability, mordenite has been used as a catalyst for important reactions, such as hydrocracking, hydroisomerization, alkylation, reforming, dewaxing and the production of dimethylamines [11-24]. Mordenite has also been used in the adsorptive separation of gas or liquid mixtures [25-30]. In addition, mordenite has been considered for use in semiconductors, chemical sensors and nonlinear optics [30-34].

The HZSM-5 zeolite is active in several reactions such as alkylation, isomerization, disproportionation and cracking. Acidity is one of the most important characteristics of zeolites that make them catalytically important. HZSM-5 contains both Lewis and Bronsted acid sites. In order to accurately evaluate the catalyst and the product selectivity pattern, it is essential to estimate not only the total acidity but also to quantify the number of Bronsted and Lewis sites and their strength. The relative acidity of the reactant molecule as well as the quantitative distribution of the type of different acid sites will determine the catalytic activity. There are many reports on the relationship between the acidity of zeolites and their catalytic activity [1-19, 35-40]. Various methods such as titration, gravimetric adsorption, temperature programmed desorption of bases, calorimetric measurement and the infrared spectroscopy of adsorbed species have been used to measure the acidity of zeolites [1-17]. In the modern petrochemical industry, propylene serves as an essential feedstock for the production of various polypropylene plastics, acrylonitrile, propylene oxide and many other commodity chemicals that can be used as substitutes for non-plastic materials such as paper, steel and wood. So far its market demand has been increasing steadily. Propylene, however, has always been produced as the byproduct of steam cracking and catalytic cracking of crude oil. Without exception such routes rely on the catalytic transformation of petroleum. Because of the petroleum crisis, finding a new way to produce propylene has attracted much interest. Numerous studies have examined the promoters of the catalysts for the MTO reaction and have found that impregnation with lanthanum and silver can improve the C2–C4 alkene selectivity from 38.9% to 56.0% because of a reduction in the apparent pore size of the zeolite channels [43-49]. Previous work [50-64] revealed that phosphorus modification led to a significant decrease in activity and increased the yield of light alkenes from methanol. It has been reported that entrained phosphate is an impediment to the sterically demanding reactions of methanol conversion [65-70]. Other studies [71-72] investigated the deactivation and product distribution of the MTO process over a phosphorus-modified HZSM-5 catalyst, on which propylene selectivity reached about 40%. However, there are few reports concerning the MTP process [22-32]. Little research exists on the effect of phosphorus on the selective conversion of methanol to a specific olefin, in particular, propylene. This work shows the catalytic effects of HMOR and HZSM-5 in the conversion of methanol to propylene and their selectivity and the results are compared with those reported in the literature.

Mordenite synthesis by standard procedure

A gel with the composition 6Na₂O:Al₂O₃:30.SiO₂:780. H₂O was prepared. First, 4.75 g NaOH were dissolved in 10 g water. To this solution 3.575 g of sodium aluminate were added and the mixture was stirred (350 rpm) until dissolution. Thereafter, 161.25 g of H₂O were added. Finally, 24.55 g of SiO₂ were added; the mixture was stirred (350 rpm) for 30 min. The gel was transferred to Teflonlined stainless steel autoclaves (60 mL capacity). The crystallization temperature was 170°C. The influence of crystallization time (12, 24 and 36 h) and seed addition was studied. A commercial mordenite (Zeolyts, SiO₂/Al₂O₃=35) acquired from Conshohocken, PA 19428- 2240, USA was used as the seed (5 mg of seed for 204,125 g of gel).

HZSM5 synthesis by hydrothermal method

Synthesis of zeolite ZSM-5: ZSM-5 zeolite was synthesized by simultaneously adding appropriate amounts of an alkaline sodium aluminate solution and an acid solution to the solution of Ludox under agitation. The first solution was prepared by dissolving sodium aluminate with sodium hydroxide in distilled water and the second one by dissolving sulphuric acid and tetrapropylammonium bromide in distilled water. The mixture obtained had the following molar composition: 0.012 Al₂O₃, 0.27 Na₂O, 0.1 TPABr, 0.2 H₂SO₄, 1 SiO₂ and 54 H₂O [23]. The gel formed was stirred vigorously at room temperature for 2h and put in a PTFE-lined stainless steel autoclave and treated at 150^oC for 24 h under autogenous pressure, as shown in Figure 1. After synthesis,the product was filtered, washed with deionized water and dried at 80^oC for 12h. In order to prepare the exchanged forms of ZSM-5 zeolite, the synthesized material was heated to 550^oC for 5h under atmospheric conditions.

Figure 1: Equipment for synthesis of zeolites.

Preparation of H-ZSM-5 zeolite: The H-form of ZSM-5 zeolite was prepared as follows the procedure reported in literature [11]: ZSM-5 zeolite was stirred in a 1 mol L^-1 aqueous ammonium chloride solution for 24 h at room temperature. The NH₄-ZSM-5 obtained was recovered by filtration, washed with distilled water and dried at 80^oC for 12 h. After that, the solid was converted to the H-form at high temperature (550^oC) for 12 h under atmospheric conditions.

Methods of characterization

The powder XRD patterns were obtained with CuKa 1 radiation (k = 1.5406 Å) on a RIGAKU Miniflex model benchtop diffractometer. Thermal analysis (TGA, DTA) was performed in air using a NETZSCH STA 409PC DTA–TGA analyser. All samples were heated to 1000°C at a heating rate of 5°C.min^-1. The morphology and size of the crystals were determined by scanning electron microscopy using a JEOL scanning electron microscope JSM-6610LV model. The Si/Al ratio of the crystals was determined by energy-dispersive X-ray (EDAX) analysis in the SEM chamber.

Nitrogen adsorption/desorption isotherms were determined by QUANTHACHROME 3B at 77 K. All samples were outgassed at 300°C under vacuum for 12 h. The specific surface area was determined by the BET equation.

Infrared transmission spectra (Ft-IR) of synthetic zeolites were obtained by the KBr wafer technique. The spectra were recorded on a FTIR-Spectrometer (Nicolet) in the region of 370-4000 cm^-1 (middle range IR).

TPD study

In order to study the acid and basic strength, ammonia was adsorbed on these zeolites. The NH₃ temperature programmed desorption (TPD) study was carried out using a heating rate of 10°C/min. Prior to TPD, 50 mg of the sample was placed in a pulse microreactor and calcined at 540°C. Then a series of 0.5 mL NH₃ pulses were given at 100°C. These pulses were continued until no more uptake of NH₃ was observed. Helium was then passed over the sample at a flow rate of 30 ml/min followed by the start of temperature programming. The desorbed NH₃ was detected by a thermal conductivity detector and the TPD profile was drawn.

Differential adsorption heats of zeolites

The heat of adsorption was measured in a heat flow microcalorimeter of the Tian-Calvet type (local construction, Figure 2) linked to a volumetric line that permitted the introduction of successive small pulses of ammonia gas. Before the calorimetric experiments a known amount of sample was heated in vacuo (1•10-3 Pa) for 6 h at 673 K.

Figure 2: Adsorption Microcalorimeter: 1,2 Vacuum and gas injection. 3. Tube connecting. 4. Calibration volume. 5,6. Transducter pressure. 7-16. Valves presicion. 17,18 Cells calorimetrics.

Adsorption was carried out by introducing successive doses of a known amount of ammonia to the sample. For every adsorbed amount the equilibrium pressure was measured by means of a differential pressure gauge (Pfeiffer). Identical doses were sent repeatedly to the sample until a final equilibrium pressure of 0.5 Torr was obtained. The calorimetric and volumetric data were stored and analyzed by microcomputer processing. The adsorption temperature was maintained at 393 K to avoid physical adsorption.

Catalytic conversion of methanol

The zeolites synthesized were tested in the methanol conversion following the procedure described in [24]. The methanol conversion was carried out in a fixed bed reactor at 450^oC under atmospheric pressure. Prior to each reaction, the samples (0.5 g) were pretreated in he flow at 550^oC for 2 h and cooled to the reaction temperature. Methanol (Sigma-Aldrich, C99.9%) was fed into the reactor by a liquid mass flow controller (Bronkhorst High-Tech, LIQUID-FLOW series L10/L20) and the weight hourly space velocity (WHSV) was 2.88 h^-1. A homogeneous mixture of MeOH (10%) and He (90%) was obtained using a pre-heater to vaporize the methanol. All products were passed through a heated transfer line to a gas chromatograph with a thermal conductivity detector and a flame ionization detector (column: HPPLOT Q, Agilent) in series.

Mordenite and HZMS-5

The diffractograms of materials prepared using the standard procedures for mordenite (HMOR) with seeds are presented in Figure 3a. The solid phase obtained was mordenite, according to comparison with a standard diffractogram [20-27]. The use of seeds for synthesis of HMOR produced materials with high crystallinity; the XRD powder pattern of the synthesized sample is shown in Figure 3b. Such a pattern is characteristic of ZSM-5 and is consistent with earlier work [23-30]. Specifically, the zone between 10 and 30 degrees has the typical mordenite peaks, show that a zeolite ZSM5 has been obtained.

Figure 3a: XRD for Zeolite HMOR.

In the scanning electron micrograph of HMOR shown in Figure 4, crystals with two different morphologies can be seen: one being an acicular form, which is characteristic of mordenite and another being a “ball of string” form that is a characteristic of these zeolites.

Figure 3b: XRD for Zeolite HZSM-5.

Figure 4: Scanning electron microscope of mordenite.

Figure 5 shows SEM micrographs of a highly crystalline material. The H-ZSM-5 sample displayed a hexagonal prismatic crystal morphology with different crystal sizes (three main crystal sizes of about 2µm x 5µm x 10µm were observed). No impurities were detected.

Figure 5: Scanning electron microscope of HZSM-5.

Physical adsorption

Nitrogen adsorption at 77 K on the standard samples is shown in Figure 6. The adsorption isotherms on HMOR and HZSM5 exhibited a well-defined hysteresis loop of type H1 according to the IUPAC classification [37]. This type of hysteresis loop was observed for the open ended cylindrical mesopores during adsorption and desorption.

The N₂ adsorption–desorption isotherms (cm³/g sample) for the samples are shown in Figure 6. The larger adsorbed volume per gram sample is a result of the different type of zeolite synthesized. The micropore volume per gram sample increased from HMOR to HZSM5, as shown in Figure 6 and Table 1. Thus HMOR has a much higher mesopore volume compared to HZSM5. Up to the relative pressure at which hysteresis begins, the difference between the two isotherms in the volume adsorbed was less than 70 cc/g. However, this difference abruptly increased as the mesopores were filled, giving differences as large as 120 cc/g; the sample of HMOR had a higher pore volume than HZSM5.

Figure 6: Adsorption isotherm of nitrogen at 77K on HMOR and HZSM-5.

Table 1: Textural characteristic of HMOR and HZSM-5and properties of infrared.

Table 2: MTP reaction over unmodified and mesopore-modified HZSM-5 catalysts ^c.

The structural properties of the standard samples were obtained from both XRD and physical adsorption analysis and are shown in Table 1.

It can be seen that the HMOR sample had larger pore sizes, surface area, micropore volume and total pore volume than HZSM5. The wall thickness calculation based on the adsorption branch could not be obtained for HMOR since the pore diameter was higher than the a-parameter. The desorption branch finally closed for sample HMOR but the closure point was at low relative pressure when the normal point was around 0.45 relative pressure. This might be due to experimental error during the sorption isotherm measurement.

Adsorption heats of HZSM-5 and HMOR

Figure 7 shows the differential heats of adsorption of the samples at 393 K versus the adsorbed volume. It shows that there are a few acid sites that are able to strongly chemisorb ammonia at heats greater than 80 kJ.mol^-1 in HZSM5 compared to HMOR. Ammonia adsorption heats of above 80 kJ.mol^-1 may be either due to Bronsted or Lewis acid sites. The inflection point of the differential heat curves shows that the HZSM5 sample contained about 200 µmol.g^-1 of Lewis acid sites and the HMOR sample contained 390 µmol.g^-1. The remaining sites were assigned to Bronsted acid sites.

Figure 7: Differentials heats for HMOR and HZSM-5 were made 393 K.

FT-IR of HZSM-5 and HMOR

Similar two-band spectra have already appeared in the literature. It thus seems that the characteristic vibration of framework OH groups from Bronsted acid sites in H-ZSM-5 is around 3600 cm^-1. Since the latter band has been assigned to a well-defined vibration in the lattice, it follows that the environment for the framework hydroxyl groups in H-ZSM-5 is far less homogeneous. The existence of only a single framework OH vibration could mean that the groups are vibrating inside the small cages formed at the channel intersections. The absence of a band at 3720 cm^-l is indicative of highly crystalline material with negligible amounts of terminal silanols (due to the large crystal size) and without gel or other, not fully cross-linked, impurities.

HZSM5 shows a single OH band around 3600 cm^-1. Its rather large half-bandwidth indicates that the environment in which it was synthesized is less homogeneous given the very high degree of crystallinity of HZSM5 (as shown in SEM). This can be related to the A1 zoning reported for this zeolite [12, 22-34]. Presently, the catalytic consequences of the presence and/or absence of the 3720 cm^-1 band in proton-catalyzed reactions are under investigation.

Figure 8 shows that the FT-IR spectra of synthesized HMOR are very similar to those of HZSM-5. In the FTIR of HZSM5 bands near 1080 cm^-1 (internal asymmetric stretch), 790 cm^-1 (external symmetric stretch), 542 cm^-1 (double ring vibration) and 450 cm^-1 (Si-O band) were apparent in all samples, due to the formation of the HZSM-5 phase. Additional evidence for the HZSM-5 zeolite was the asymmetric stretch vibration of the band at 1225 cm^-1, which has been assigned to external linkages (between the SiO₄ tetrahedral) and is a structure-sensitive IR band of HZSM-5 zeolite [20-24]. In the FT-IR spectra (Figure 8) an absorption signal corresponding to the MFI phase skeletal vibration is apparent at 542 cm^-1. Figure 8 shows the FT-IR spectra of HMOR obtained using kaolin. The spectra show the hydroxyls of HMOR after synthesis. There are two main peaks at 3749 and 3608 cm1. According to previously reported results [20-32, 40-43] these two peaks should correspond to terminal silanols and framework Si–OH–Al groups, respectively. A small band at 3660 cm^-1 is also apparent. This peak has previously been observed [24-27, 30- 35, 43-45] and can be attributed to the hydroxyls associated with the ill-defined extra-framework alumina. The peaks in HMOR are more intense and defined.

Figure 8: FTIR of HMOR and HZSM-5.

The mordenite framework (Si/Al ratio=5) is characterized by the presence of straight channels running along the [001] crystallographic direction, which are accessible through twelve-membered (elliptical) silicon oxygen rings having dimensions of 0.65x0.70 nm (and hence slightly larger than those of ZSM-5, which are delimited by ten-membered rings with [0.51]0.55 nm radii). A second family of smaller channels intersects the previous one in the perpendicular [001] direction. These channels are delimited by eight-membered rings. Unlike ZSM-5, this second family of channels does not allow the full passage of molecules (even of small dimensions) from one straight channel to another. In fact the two halves of this second family of channels are slightly stacked, giving rise to narrow-necked obstructions. Summarizing, the mordenite structure shows only a set of permeable parallel straight channels with regularly disposed side pockets. The Si/Al ratio in mordenite (ca. 5) is lower than in ZSM-5 and, as a consequence, a larger number of counter cations, balancing the negative charge of the framework, are present. It is worth noticing that, with respect to H-ZSM-5 with the lowest Si/Al ratio (ca. 14); the proton density in H-Mord is larger by at least a factor of 3, which implies substantially lower distances between the Bronsted sites.

Temperature programmed desorption of ammonia (TPD) over HZSM-5 catalysts

Figure 9 shows the TPD profiles of HZSM-5 and HMOR with NH₃. For HZSM-5 with NH₃ adsorbed at 373 K the low temperature peak was observed at 453 K whereas the high temperature peak was observed at 653 K. The low temperature peak indicates weak acidic sites where ammonia molecules are weakly adsorbed. On the other hand the high temperature peak corresponds to strong acid sites present on the catalyst. The total amount of ammonia that was desorbed between 453 K and 553 K was 0.41 mmol of NH₃/g catalyst (HZSM-5). The HMOR catalysts with NH₃ adsorbed both at 523 K and 753 K predominantly show a peak with strong acidic sites only in the TPD profiles. Figure 9 depicts the TPD profile of HZSM-5 with varying amounts of acid sites per gram of catalyst (density of acid sites). HZSM-5 and HMOR catalysts have different TPD profiles; in the case of HMOR, which show some acid sites. The concentration of NH₃ desorbed up to 400 K was considered to indicate the amount of weak acid sites and that coming out above 400 K was considered to indicate the amount of strong acid sites.

Figure 9: Temperature Programmed Desorption profi les of HMOR and HZSM5 and NH₃-TPD profiles of 373 K.

Role of acidic sites

In order to identify the type of sites that are responsible for MTP the following experiments were carried out. NH₃ was adsorbed on HZSM-5 catalyst samples at 373 K. These samples were used as the catalyst at 373 K. The TPD profiles of these catalysts are shown in Figure 8. In the case of HZSM-5 with NH₃ adsorbed at 453 and 653 K the conversion of MTP was only 8.5% with 89% selectivity to hydrocarbons in a 2-h run. Here all the strong acidic sites are blocked by NH₃ as NH₄ ⁺ ions, whereas in the case of zeolite with NH₃ adsorbed at 653 K a fraction of strongly acidic sites are available. MTP takes place on these sites. The HMOR catalyst, which has more strong acidic sites (at 523 and 753 K), generates a high conversion percentage of MTP. This clearly proves that the weakly acidic sites have no activity while the strongly acidic sites are mainly responsible for MTP.

The methanol conversion and the selectivity (product distribution) of HZSM5 and HMOR are presented in Table 2. For the pure HZSM-5 (SiO₂/Al₂O₃ = 80), selectivity to C1–C4 saturated hydrocarbon, ethylene, propylene and butylene was 18.5, 18.0, 22.7 and 8.25 C mol%, respectively. Compared with HZSM-5, the propylene selectivity of the HMOR catalyst was double, while the C1–C4 saturated hydrocarbon selectivity increased and the butylene selectivity decreased. The changes in product distribution must be due to the major acidity of the HMOR catalyst: the propylene and butylene selectivity dramatically increased to 44.2 and 19.4C mol%, respectively, while the C1–C4 saturated hydrocarbon selectivity decreased to 3.62C mol% due to the decrease in strong acidity as well as the weakening of the strong acid strength.

The well-known methanol conversion consists of three main reaction steps. Methanol is dehydrated to dimethylether and the resulting equilibrium mixture, consisting of methanol, dimethylether and water, undergoes further dehydration to produce light olefins. These are subsequently converted to paraffins, aromatics, naphthenes and higher olefins. These results are also associated with Si/Al ratios of zeolites; for over 100 relationships is contrary behavior.

The propyl ene selectivity is affected by the Brønsted acidity of the catalyst. In the case of HZSM-5 without Brønsted acidity, methanol was not converted to hydrocarbon and the main product was dimethylether [16]. The changes in product selectivity with time on stream for HMOR are presented in Figure 10. The results are very similar to those reported in the literature [24] in terms of methanol conversion. However, in our case the higher percentage of conversion was obtained with HMOR. The trend is very similar to that reported by some authors, but our results show a higher variability is probably due to the sensitivity of our used method.

Figure 10: Product selectivity of HMOR as a representative catalyst for MTP reaction as a function of time: (◊) C₃H₆; (■) C₂H₄; (Δ) C₄H₈; (▲) aromatics; (♦) C₁–C₄ saturated hydrocarbons; (□) C₅ and higher hydrocarbons excluding aromatics. Reaction conditions: T=470 °C, WHSV=1h^-1, PC H₃O H=0.5 atm, H₂O:CH₃OH = 1:1.

At the initial stage the selectivity to propylene and butylene increased slightly, whereas the selectivity to ethylene, saturated hydrocarbons and aromatics decreased slightly. After 25 h on stream the reaction reached the steady state. The methanol conversion over the HMOR catalyst after being on stream for 100 his shown in Figure 10. However, the product selectivity’s were similar to those at steady state. For the series catalysts, the propylene selectivity first increased and then decreased with an increase in the Si/Al ratio. The ethylene selectivity decreased gradually from 15.3 to 9.90%, as the Si/Al ratio increased. A decrease in the Brønsted acidity of the catalysts favored the formation of propylene in the reaction and led to an increase in the P/E ratio from 1.75 to 3.66. Meanwhile, the influence of mesopore modification on the product selectivity of the catalyst exceeded that of the acidity. For the HMOR catalyst, the propylene selectivity increased from 37.0 to 42.2% after modification, while the ethylene selectivity decreased from 10.1 to 4.18%, leading to a substantial increase in the P/E ratio from 3.66 to 10.1. It was noted that the selectivity to propylene and P/E ratio a lso improved on the HMOR catalyst. However, the improvement was less striking than that on the HZSM5 catalyst. The more probable reason is that the mesopores or cavities generated inside the zeolite crystals of HMOR catalyst by the method used, improved the diffusion of the gas molecules in the reaction process.

The catalytic stability of HMOR and HZSM-5 zeolites for the MTP reaction was tested at 470^oC and the results are shown in Figure 11. It can be seen that the catalytic stability followed the order of HMOR>HSZM5. As shown in Figure 11, the methanol conversion dropped abruptly after 10 minutes of stream. Coke deposition is the reason for catalyst deactivation [1]. When the coke accumulates to a certain level, it blocks the micropores of HZSM-5 and subsequently results in sharp deactivation of the catalyst. Additionally is should be observed that the catalytic stability that followed the order of HMOR > HSZM5 and that present a low propylene production and a low stability of HZSM-5 zeolite is due to that the Si/Al molar ratio of HZSM-5 used in this our work is low. In general, HZSM-5 zeolites with Si/Al ratio higher than 100. Additionally is should be observed that the catalytic stability that followed the order of HMOR > HSZM5 and that present a low propylene production and a low stability of HZSM-5 zeolite is due to that the Si/Al molar ratio of HZSM-5 used in this our work is low. In general, HZSM-5 zeolite with Si/Al ratio more than 100 exhibits good performance for MTP reaction exhibits good performance for MTP reaction.

Figure 11: Catalytic stability of HMOR and HZSM-5 zeolites for MTP reaction: (□) HMOR; (♦) HZSM-5. Reaction conditions: T = 470°C, WHSV = 1 h⁻¹, PCH₃OH = 0.5 atm, H₂O:CH₃OH = 1:1.

In this work we synthesized HMOR and HZSM5 zeolites and were employed in the MTP process. The study of adsorption by means of nitrogen isotherms, scanning electronic microscopy and XRD demonstrated that the zeolites are synthesized by good percentage of crystalline.

HMOR converts methanol to propylene more effectively than HZSM-5 which associated with the ratio of Si/Al and with the different content of acid groups. This work shows novel results for this type of reaction and higher yields.

The authors would like to thank the Departments of Chemistry at the Universidad Nacional de Colombia and Universidad de Los Andes (Colombia) and the Master Agreement established between these two institutions. Special gratitude goes to the Fondo Especial de Investigaciones de la Facultad de Ciencias de la Universidad de Los Andes (Colombia) for its partial financing of this research.