Brazilian Journal of Chemical Engineering SYNTHESIS OF 2 , 6-DIMETHYLNAPHTHALENE OVER SAPO-11 , SAPO-5 AND MORDENITE MOLECULAR SIEVES

Shape-selective methylation of naphthalene over SAPO-11, SAPO-5 and mordenite molecular sieves were carried out in a fixed-bed flow reactor under atmospheric pressure. Methanol and mesitylene were used as methylation and solvent agents. The experiment results showed that SAPO-11 exhibited higher stability, higher selectivity of 2,6-dimethylnaphthalene (2,6-DMN) and higher 2,6-/2,7-DMN ratio than SAPO-5 and mordenite molecular sieves. The catalytic performances for the methylation of naphthalene were mainly related to the pore structure of the catalysts. The comparison of the spent SAPO11 with fresh SAPO11 suggested that structure collapse of the SAPO-11 by dealumination was occurred during the methylation of naphthalene with methanol, which may have been caused by high temperature steam from water produced in the reaction or by high temperature methanol vapor.


INTRODUCTION
2,6-Dialkylnaphthalene (2,6-DKN) is an important product in industry, which can be oxidized to 2,6-naphthalene dicarboxylic acid (2,6-NDA) and then to prepare the commercially valuable polymer of poly(ethylene naphthalene dicarboxylate) (PEN).Compared with poly(ethylene terephthalate) (PET), in addition to its excellent properties as a gas barrier, PEN has good thermal and chemical stability as well as mechanical properties, leading to its wide application in electronics components, insulation material, food containers, aviation, and so on (Song et al., 1993;Tsutsui et al., 2004;Wu et al., 2010;Wu et al., 2015).Highly regioselective synthesis of 2,6-DKN can be achieved over different molecular sieve catalysts using various alkylating agents and solvents (Brzozowski et al., 2012;Liu et al., 1997;Smith et al., 2000;Smith et al., 2003;Smith et al., 2012;Song et al., 2000;Sugi et al., 2008).Liu et al. (1997) reported that the alkylation of naphthalene with tbutanol should be both practical and attractive not only because of the high activity, but also due to the easy separation of the desired product from the reaction mixtures by crystallization.Smith et al. (2000) surveyed the dialkylation of naphthalene over a dealuminated mordenite molecular sieve using tertbutanol as an alkylating agent.The experimental results showed that a 60% yield of 2,6-di-tertbutylnaphthalene was obtained with a 2,6/2,7-ratio of over 50.Sugi et al. (2008) reported that the channels of molecular sieves with three dimensional poresystem were too large to differentiate 2,6-DAN from other β,β-DAN by restriction of their transition states in the channels even with bulky alkylating agents of 1-butene and 2-methylpropene.
Among 2,6-DKNs, 2,6-dimethylnaphthalene (2,6-DMN) is regarded as the most suitable raw material for the synthesis of PEN because no carbon atoms of 2,6-DMN are lost in the oxidation reaction and 2,6-DMN can be oxidized easily compared with other 2,6-DKNs (Pu et al., 1996).At present, 2,6-DMN is mainly produced by BP-Amoco through a four-step process from o-xylene and butadiene, but the process is quite complex (Lillwitz et al., 2001).In order to establish a simple and "green" synthesis route of 2,6-DMN with fewer reaction steps and less reaction waste, it is desirable to methylate naphthalene to 2,6-DMN directly over a shape-selective molecular sieve catalyst.
In recent years, methylation of naphthalene over molecular sieve catalysts to synthesize 2,6-DMN was reported by several research groups.As a result, there were great differences in the catalytic activity and selectivity among these molecular sieves (Fraenkel et al., 1986;Park et al., 2002;Jin et al., 2006;Yoo et al., 2002;Jin et al., 2006;Wu et al., 2010).For example, Fraenkel et al. (1986) first investigated the methylation of naphthalene with methanol over ZSM-5, mordenite and HY molecular sieves, and found that the medium pore size ZSM-5 showed a high product selectivity and moderate activity, while the large pore size molecular sieves, HY and mordenite, providing enough space for multiple reactions, would lead to a lower product selectivity and a higher activity.Park et al. (2002) reported that the selectivity for 2,6+2,7-DMN was only 16% using MCM-22 as catalyst after 1 hour time-on-stream (TOS).Yoo et al. (2002) reported that the high stability of ZSM-12 molecular sieves during the methylation was because ZSM-12 possessed one dimensional non-interpenetrating channels which behaved as "perfect tubes" and did not lead to the accumulation of coke precursors.Jin et al. (2006) reported that Zr/(Al)ZSM-5 could improve the selectivity of 2,6-DMN and the stability of the catalyst.They attributed it to the weakening of the acid strength and the enlargement of the pore dimensions because of the part incorporation of Zr in the framework instead of Al.Wu et al. (2010) concluded that the small dimensions of zeolites crystals and the advanced mesoporosity were the general reasons for the remarkable catalytic action of MTW zeolites in naphthalene methylation.Among these molecular sieves, SAPO-11 molecular sieve has been proved to show high catalytic performances in the methylation of naphthalene with methanol (Liu et al., 2013;Wang et al., 2012;Wang et al., 2013;Wang et al., 2015).First of all, SAPO-11 is a one-dimensional pore zeolite with pore opening of 0.39 nm × 0.64 nm (Liu et al., 2015;Wu et al., 2015;Zhang et al., 2007), which is between the pore size of large-pore molecular sieves and the pore size of medium-pore molecular sieves, making it an interesting material for catalytic cracking, reforming and alkylation (Subramanian et al., 1997;Prakash et al., 1996;Han et al., 2014;Liu et al., 1991;Rabaev et al., 2015;Tian et al., 2014;Wang et al., 2015).Similar to ZSM-12, SAPO-11 also possesses one dimensional and non-intersecting channels (Zhang et al., 2010), which shows excellent resistance to deactivation by carbonaceous deposition.
Based on these reports, it is concluded that the catalytic activity for the methylation of naphthalene has relation with the acidity and the structure of molecular sieves.However, the effect of the acidity and the structure of molecular sieves on the naphthalene methylation with methanol have not been understood clearly, which hinders the development of a commercial catalyst for naphthalene methylation.In addition, the research on deactivation of SAPO-11 molecular sieve during methylation of naphthalene with methanol has not been reported.Meanwhile, there are few studies about the methylation of naphthalene with methanol over SAPO-5 molecular sieve so far.
In the present work, SAPO-11, SAPO-5 and mordenite were evaluated for the methylation of naphthalene with methanol, respectively.And they were investigated in detail by using ammonia temperature programmed desorption (NH3-TPD), infrared spectroscopy with pyridine adsorption (Py-IR) and N2 adsorption techniques in order to study the impact of the pore structure and acidity of molecular sieves on catalytic performance in the methylation of naphthalene with methanol.In addition, the fresh and the spent SAPO-11 molecular sieves have been characterized by different characterization methods to analyze the cause of deactivation of the SAPO-11.

EXPERIMENTAL Catalyst Preparation
Mordenite was obtained from the Catalyst Plant of Nankai University.The acidic form of the molecu-Brazilian Journal of Chemical Engineering Vol. 34, No. 01, pp. 295 -306, January -March, 2017 lar sieve was prepared by ion-exchange with aqueous NH4NO3 solution and then calcination at 550 °C for 2 h in air.
SAPO-5 was synthesized by a hydrothermal method from a gel molar composition of 1.0Al2O3: 1.1P2O5:0.66SiO2:1.6DPA:40H2O.The gel mixture was transferred into a stainless steel autoclave, heated in an oven at 200 °C for 6 h, and subsequently at 300 °C for 9 h.The product thus obtained was washed, dried at 120 °C for 6 h, and calcined in air at 550 °C for 24 h to remove the template completely.

Catalyst Characterization
X-ray powder diffraction (XRD) analysis was performed on RigakuD/maxrB X-ray diffractometer.Diffraction patterns were recorded with Cu Ka radiation at 40 kV and 100 mA in the scan range between 5 o and 50 o to identify the phase structure of the catalyst.
The morphologies of the samples were examined by Hitachi S-4800 scanning electron microscopy (SEM).
The textural properties of the samples were derived from N2 adsorption-desorption measurement on Micromeritics Tristar 3000.In each case, the sample was outgassed under vacuum at 300 °C for 3 h before N2 adsorption.
The acidity of the samples was examined by temperature-programmed desorption of ammonia (NH3-TPD) techniques carried out by a flow system with a thermal conductivity detector.All samples were preheated from room temperature to 500 °C in an argon flow and kept at 500 °C for 1 h, which was followed by NH3 saturation in a flowing NH3/Ar stream at 40 °C for 5 min.Evacuation at 40 °C for 40 min was carried out to remove physically adsorbed NH3 then the catalyst was heated to 600 °C at a linear rate of 10 °C•min -1 , and the detector signal of NH3 was recorded.
The acid sites of the catalyst samples were characterized by FT-IR (EQUINOX55) spectroscopy with chemisorbed pyridine.
The thermogravimetric analysis and differential thermal analysis (TG-DTA) were recorded on a Rigaku Thermo plus Evo TG 8120 instrument at a heating rate of 10 °C•min −1 in air with a flow rate of 30 ml•min −1 .
The total (bulk) Si and Al contents in the catalysts were determined by an Inductive Couple Plasma (ICP) emisson spectrometer (Thermo ICAP6300).The SAPO-11 catalyst was subjected to themogravimetric analysis both before reaction (fresh) and after reaction (spent).

Catalytic Evaluation
The experiments were performed in a fixed-bed continuous-flow reactor equipped with 20 mm diameter and 600 mm length stainless steel tube.2.5 g of 20-40 mesh molecular sieve catalysts were loaded in the reaction tube.After preheating, the reaction mixture including naphthalene, methanol and mesitylene (solvent) in a molar ratio of 1:5:3.5 was fed into the reactor by a quantity measuring pump under the pressure kept by N2.The weight hourly space velocity (WHSV) of naphthalene was 0.19 h -1 in all experiments, and the reactive temperature was 350 °C.Reaction products were analyzed by gas chromatography (GC9560) with a Beta-Dex120 capillary column.
The naphthalene conversion was calculated as follows: where nN,0 and nN are the molar percentage of naphthalene before and after the reaction.Product distribution includes the corresponding molar percentages of ethylnaphthalene (EN), methylnaphthalene (MN), dimethylnaphthalene (DMN) and trimethylnaphthalene (TMN) in the total product mixture.

Characterization of Catalysts
XRD patterns of as-synthesized SAPO-5 and SAPO-11 samples presented in Figure 1 show the presence of highly crystalline SAPO phases and no impurity phases are detected (see Figure 1).The characteristic peaks of the SAPO-11 phase (i.e.2θ = 8. 15°, 9. 40°, 13. 20°, 15. 57°) are observed and are identical to those reported for SAPO-11 in the literature (Lok et al., 1984).Also the positions of the lines of the SAPO-5 are identical to those reported for SAPO-5 (Wei et al., 2015;Zhu et al., 2016).It is a well-crystallized sphere-shaped material as determined by the SEM studies (see Figure 2).
According to the reference materials (Zhang et al., 2007;Chao et al., 2000;Bandyopadhyay et al., 2002), the physicochemical properties of the investigated molecular sieves are summarized in Table 1.The BET surface area and the pore volume of SAPO-11, SAPO-5 and Mordenite molecular sieves are determined by N2 adsorption-desorption measurement.As observed in Table 1, the surface area and the pore volume of the investigated molecular sieves decrease in the order of Mordenite > SAPO-5 > SAPO-11.The acidic properties are usually evaluated by NH3-TPD, and the profiles can be differentiated both in the integral area of the profiles and in the shift of peak temperature.The former corresponds to the amount of acid sites, and the latter indicates the strength of the acid sites.The NH3-TPD profiles of different catalyst samples illustrated in Figure 3 present three ammonia desorption peaks at ca. 210 °C, 300 °C and 510 °C, corresponding to a weak acid site, a medium strong acid site and a strong acid site.The acidity values of different molecular sieve catalysts are shown in Table 2. From Figure 3 and Table 2, it can been seen that the total acidity of the molecular sieves and the acid strength of their acidic sites both decrease in the order of Mordenite > SAPO-11 > SAPO-5.
Pyridine-IR is used to further evaluate the type (Lewis and Bronsted) of acid sites for all the samples.The Bronsted acid sites of the investigated molecular sieves will protonate pyridine, forming the pyridinium ion, which has a characteristic ring vibrational frequency at 1540 cm -1 .Lewis aluminum, being an electron-pair acceptor, can bind pyridine in a covalent fashion, giving rise to a vibrational band at 1450 cm -1 .The relative amounts of Bronsted and Lewis acid sites can be estimated by integration of these two bands after correction for differences in oscillator strengths (Harris et al., 2016;Nieminen et al., 2004).The concentration of Bronsted and Lewis acid sites after pyridine desorption at 200 °C and 400 °C are presented in Table 3.At any desorption temperature, the density of Bronsted acid sites of these samples decreases in the order of Mordenite > SAPO-11 > SAPO-5, indicating that the strong acid sites of these samples decreases in the same order, which is in agreement with the results of NH3-TPD shown in Figure 3.Meanwhile, Bronsted acid sites are predominant for all the investigated molecular sieves, but the ratio of Bronsted acid sites to Lewis acid sites varies among different molecular sieves.This may be due mainly to the fact that the acidic sites of molecular sieves, i.e., aluminum species of molecular sieves-framework Al and external-framework Al, change with the synthetic methods and conditions (Zhu et al., 2006).

Catalytic Activity in the Methylation of Naphthalene
The catalytic performances of the investigated molecular sieves with 1h of TOS are listed in Table 4.It can be seen that 1,8-DMN is not detected for all the molecular sieves, probably due to steric repulsion.Among all the molecular sieves in the present study, Mordenite exhibits relatively high naphthalene conversion compared with the other molecular sieves.SAPO-11 exhibits the highest selectivity for DMN, i.e. 52.8% with the very high 2,6-/2,7-DMN ratio of 1.46 for 1 h of TOS.The 2,6-DMN yield decreases in the order of SAPO-11 > Mordenite > SAPO-5.Reaction conditions: Temperature = 350 °C, Pressure = 0.1MPa, WHSV= 0.19h −1 , naphthalene:methanol:mesitylene = 1:5:3.5(molar ratio), TOS = 1h.Product distribution includes the corresponding molar percentages of EN (ethylnaphthalene), MN (methylnaphthalene), DMN (dimethylnaphthalene) and TMN (trimethylnaphthalene) in the total product mixture.
The catalytic conversion of the investigated molecular sieves is compared in Figure 4 shows that mordenite presents the highest conversion of naphthalene (60.2%) at 1h of TOS.In contrast, SAPO-11 shows the highest conversion of naphthalene (60.2%) at 6 h of TOS, and the conversion of naphthalene for SAPO-5 is the lowest in the reaction time.Meanwhile, mordenite and SAPO-5 show initially the highest conversion of naphthalene and the lowest conversion of naphthalene, respectively, which may be related to their acidity.As shown in Figure 3 and Table 2, mordenite and SAPO-5 exhibit the highest acidity and lowest acidity, respectively, leading to the initial activity of mordenite and SAPO-5 molecular sieves.In additon, Table 3 shows that mordenite has the most Bronsted acid sites and SAPO-5 possesses the least Bronsted acid sites, implying that the Bronsted acid sites probably be active sites in the methylation of naphthalene with methanol.However, their relatively low stability may be caused by the deposition of carbonaceous material ("coke") in the pores or the breakdown of the structure of the molecular sieves.The smaller pores of SAPO-11 may restrain the formation of coke, leading to somewhat greater stability.The 2,6-/2,7-DMN ratios for the three molecular sieves are shown in Figure 5.The 2,6-/2,7-DMN ratio is very important in the purification of 2,6-DMN.It is very hard to separate them when the 2,6-/2,7-DMN ratio is less than or equal to 0.7 due to the formation of the eutectic mixture.A higher 2,6-/2,7-DMN ratio can facilitate the subsequent separation of 2,6-and 2,7-DMN mixture.When the 2,6-/2,7-DMN is more than 1.4, 2,6-DMN can be separated more easily from the eutectic mixture.As shown in Figure 5, the ratio of 2,6-/2,7-DMN on the SAPO-11 is always the highest among the investigated molecular sieves (see Table 4).However, the 2,6-/2,7-DMN ratio for the investigated molecular sieves are not associated with the acidity of the molecular sieves shown in Figure 3 and Table 2. Therefore, the 2,6-/2,7-DMN ratio is probably related to the differences in pore structure of the molecular sieves.Fang et al. (2006) calculated that 2,6-DMN is somewhat larger than 2,7-DMN in molecular dimension and suffers Brazilian Journal of Chemical Engineering Vol. 34, No. 01, pp. 295 -306, January -March, 2017 more diffusion resistance than 2,7-DMN does during the diffusion process.Mordenite and SAPO-5, with nominal pore openings of 0.70 nm and 0.73 nm, respectively, are not expected to be very selective in the methylation of naphthalene because their channels are wide enough to accommodate the formation and diffusion of both 2,6-and 2,7-DMN molecules.SAPO-11, however, possesses a pore opening (0.39 nm×0.64 nm), which is comparable with the kinetic diameter of the naphthalene molecule (0.62 nm) and its β-methylation products (at least 0.62 nm), making it more discerning for the formation and diffusion of the 2,6-and 2,7-DMN products.Therefore, the 2,6-/2,7-DMN ratio over SAPO-11 is mainly ascribed to its specific pore structure.Figure 6 shows the 2,6-DMN yield with 1-6 h of TOS on different molecular sieve catalysts.It can be seen that SAPO-11 exhibits the highest 2,6-DMN yield compared with the other molecular sieve catalysts.
All these findings prove that the relatively good catalytic performance of SAPO-11 for the methylation of naphthalene with methanol is mainly related to the catalyst's pore structure and SAPO-11 is the most promising of the catalysts tried for the methylation of naphthalene because of its suitable pore structure, which favors the formation and/or diffusion of the 2,6-DMN product over the 2,7-isomer.

Characterization of the Spent Catalyst
To further analyze the cause of deactivation of SAPO-11, the spent SAPO-11 catalyst (6 hour reaction time) was investigated in comparison with the fresh SAPO-11.

XRD Characterization
The XRD patterns of the fresh and the spent SAPO-11 are shown in Figure 7.The spent SAPO-11 still retains the structure of AEL (aluminophosphate eleven) type material.However, the peak strength of the spent SAPO-11 obviously decreases compared with the fresh SAPO-11 catalyst, which indicates that the crystal structure of the catalyst is severely damaged.

SEM Characterization
The SEM micrographs of the fresh and the spent SAPO-11 catalysts are presented in Figure 8  by irregular particles after the reaction, which is in agreement with the results of the XRD shown in Figure 7.

ICP Characterization
The SiO2/Al2O3 ratio of the fresh and the spent SAPO-11 detected by ICP are shown in Table 5.It shows that the SiO2/Al2O3 ratio of the spent SAPO-11 increases compared with that of the fresh SAPO-11.That would indicate that the content of Al2O3 of SAPO-11 decreases after the end of reaction.
Table 5: ICP results of the fresh and the spent SAPO-11.

Pore Structure Characterization
The pore characteristics of the fresh and the spent SAPO-11 catalysts were studied by the N2 adsorption-desorption technique, and the results of SBET and Vpore results are listed in Table 6.The SBET of 215 m 2 g −1 and the Vpore of 0.145 cm 3 g −1 on the fresh SAPO-11 sample decrease to 121 m 2 g −1 and 0.091 cm 3 g −1 on the spent SAPO-11 sample after the reaction, respectively.The N2 adsorption-desorption isotherms of the fresh and the spent SAPO-11 catalysts are shown in Figure 9.The typical N2 adsorption-desorption isotherms of the SAPO-11 samples are of the type IV isotherm according to the IUPAC classification (Sing et al., 1985).A hysteresis between adsorption and desorption branches can be observed at medium relative pressure (0.45-1.0) for all the samples, which demonstrates the presence of a large number of secondary mesopores.However, the N2 adsorptiondesorption isotherms of the spent SAPO-11 catalyst is smaller than that of the fresh SAPO-11 catalyst.Figure 10 shows the micropores and secondary mesopores size distributions of the fresh and the spent SAPO-11 catalysts.Generally speaking, the secondary mesopores probably result from SAPO-11 microcrystals piled up during the synthesis process.As seen in Figure 10, the fresh SAPO-11 catalyst has broader micropores size distributions and secondary mesopores size distributions than the spent SAPO-11 catalyst.From the above characterization results, we speculate that deactivation of the SAPO-11 is mainly due to the breakdown of the structure caused by dealumination (Lucas et al., 2009).

Thermogravimetric (TG) Analysis
The TG analysis was performed to investigate the coke deposition on the fresh and the spent SAPO-11 catalyst.The weight loss in 50-250 °C shown in Figure 11 and Figure 12 is attributed to the physically adsorbed water in the porous materials (Garces et al., 1988).The weight loss in 250-480 °C shown in Figure 11 is attributed to the volatilization of residual template.The weight loss in 480-800 °C in Figure 11 and the weight loss in 250-800 °C in Figure 12 are attributed to the coke deposition, which may block the active sites in the catalyst (Wang et al., 2011).The amount of coke deposition over the fresh SAPO-11 is 0.613% (see Figure 12), and the amount of coke deposition over the spent SAPO-11 only slightly increases to 1.64% (see Figure 12), which indicates that there is not much coke deposition caused by the acidity of SAPO-11 catalyst during the reaction process.

Acidity Characterization
The differences of the acid amount and the acid strength between the fresh and the spent SAPO-11 catalysts can be shown in Figure 13 by the NH3-TPD profiles.Compared with the fresh SAPO-11 catalyst, the acid amount and the acid strength of the spent SAPO-11 catalyst decrease slightly.

Brazilian Journal of Chemical Engineering
Based on the thermal gravimetric analysis and the acidity characterization, it can be inferred that the deactivation of SAPO-11 molecular sieves is not because of the coke deposition.According to the value of n(SiO2/Al2O3) by ICP characterization, the deactivation of SAPO-11 is possibly because water is generated during the methylation of naphthalene with methanol, leading to the structure collapse by dealumination of the catalyst by high temperature steam or even by the methanol itself.

CONCLUSION
In summary, the catalytic activities for the methylation of naphthalene were mainly related to the pore structure of the molecular sieves catalysts and the experimental results showed that SAPO-11 exhibited higher catalytic performance than Mordenite and SAPO-5.In addition, the spent SAPO-11 was investigated in comparison with the fresh SAPO-11, which indicated that the structure of SAPO-11 had collapsed due to dealumination, probably by high temperature steam resulting from water formed in the reaction or by high temperature methanol vapor, thus resulting in the deactivation of the catalyst.
. Compared with the fresh SAPO-11, the SEM micrograph of the spent SAPO-11 changes dramatically.It shows that the spherical particles disappear and are replaced Brazilian Journal of Chemical Engineering

Figure 8 :
Figure 8: SEM images of the fresh and the spent SAPO-11.

Figure 10 :
Figure 10: Pore size distributions of the fresh and the spent SAPO-11.

Figure 11 :Figure 12 :
Figure 11: TG curve of weight loss of the fresh SAPO-11.

Figure 13 :
Figure 13: NH3-TPD results of the fresh and the spent SAPO-11.

Table 3 : Bronsted acid sites and Lewis acid sites of molecular sieves spectra of absorbed pyridine a .
Brazilian Journal of Chemical Engineering