Shape-selective methylation of naphthalene with methanol over SAPO-11 molecular sieve modified with hydrochloric acid and citric acid

Herein, a series of SAPO-11 molecular sieves were modified by citric acid and hydrochloric acid. They were characterized by ICP, XRD, SEM, N2 adsorption–desorption, NH3-TPD, Si MAS NMR, and Py-IR and evaluated by the methylation of naphthalene with methanol to 2,6-dimethylnaphthalene (2,6-DMN). According to the XRD results, hydrochloric acid at a high concentration not only removed extraframework aluminum but also deleted framework silicon; this resulted in the formation of crystal defects. N2 adsorption–desorption results showed that all the samples possessed micropores and secondary mesopores. The SAPO-11 sample modified with 4 mol L 1 hydrochloric acid for 30 min exhibited the largest secondary mesopore size distributions. NH3-TPD and Si MAS NMR showed that hydrochloric acid at a high concentration could eliminate more acid sites. SAPO-11 modified with 4 mol L 1 hydrochloric acid for 30 min presented high catalytic performance for the methylation of naphthalene; this was mainly attributed to the amount of secondary mesopores in the SAPO-11 molecular sieves.


Introduction
2,6-Dimethylnaphthalene (2,6-DMN) is an important intermediate in the synthesis of polyethylene naphthalene dicarboxylate (PEN). Compared with polyethylene terephthalate (PET), PEN has good thermal and chemical stability as well as mechanical properties in addition to its excellent properties as a gas barrier; this leads to its wide application in electronic components, insulation materials, food containers, aviation, and so on. However, high cost and low availability of 2,6-DMN restrict the large production and commercialization of PEN at present. 1,2 Currently, 2,6-DMN is mainly produced by BP-Amoco through four-step reactions from o-xylene and butadiene. However, this synthetic route is environmentally unfriendly and expensive; this greatly limits its application in industry. To reduce the high cost of production of 2,6-DMN, a process of preparation of 2,6-DMN by the methylation of naphthalene or 2methylnaphthalene over molecular sieve is proposed, which is a simple and desirable synthetic route. However, the products of methylation are very complex mixtures, and dimethylnaphthalene (DMN) has 10 different isomers with similar boiling point. It is very difficult to separate 2,6-DMN and 2,7-DMN as the difference between their boiling points is only 0.3 C. Therefore, it is necessary to nd a molecular sieve with high selectivity for 2,6-DMN at acceptable conversion of naphthalene or 2-methylnaphthalene to 2,6-DMN.
Recently, the methylation of naphthalene or 2-methylnaphthalene with methanol has been studied over several molecular sieve catalysts such as ZSM-5, HY, ZSM-12, and SAPO-11. [3][4][5][6] Among these molecular sieves, SAPO-11 has been proven to be effective with high catalytic activity and selectivity for 2,6-DMN in the methylation of NAPH. 7,8 The SAPO-11 molecular sieve is a one-dimensional pore molecular sieve with pore openings of 0.39 nm Â 0.64 nm, 9,10 which are between the pore size of largepore and medium-pore molecular sieves, making it an interesting material for catalytic cracking, reforming, and alkylation. The pore size of SAPO-11 is larger than 0.60 nm; thus, SAPO-11 can effectively sieve the products of naphthalene methylation. In addition, SAPO-11 showed excellent resistance to deactivation by carbonaceous deposits due to its one-dimensional pores. 11 Hence, SAPO-11 is considered to be the most promising catalyst for the selective methylation of naphthalene to 2,6-DMN.
Although SAPO-11 has shown high catalytic performance in the methylation of naphthalene, several modication techniques can further improve the catalytic performance of the SAPO-11 molecular sieve catalyst. Acid dealumination is an important method to improve the catalytic performance of molecular sieves through the modication of pore structure and surface acid properties and leads to high hydrothermal stability. Park et al. reported that HM aer steam treatment at 550 C followed by acid leaching exhibited high and steady conversion of 40%; the ratio of 2,6-/2,7-DMN was 2.4 in 5 h of time on stream. 12 Bai et al. reported that dealumination of HZSM-12 with hydrochloric acid and citric acid has different effects on the pore structure, acidic properties, and catalytic activity. Hydrochloric acid eliminates the strong and weak acid sites simultaneously without destroying the skeleton structure of the molecular sieve and improving the catalytic performance and stability. 13 Najar et al. studied the effect of acid dealumination on the physico-chemical properties of Y molecular sieves and showed that all acids were effective in removing Al from the framework. Moreover, the extra-framework Al extraction from the lattice was dependent on the nature and concentration of the acid. 14 However, the studies about the inuence of acid dealumination on the catalytic performance of SAPO-11 have not been reported to date for the methylation of naphthalene.
In this study, a series of SAPO-11 molecular sieves were modied by citric acid and hydrochloric acid to investigate the inuence of acid-dealumination on the physicochemical and catalytic performances of SAPO-11 for the methylation of naphthalene. The physicochemical properties of SAPO-11 molecular sieves and dealuminated SAPO-11 molecular sieves were characterized by ICP, XRD, SEM, FT-IR, N 2 adsorptiondesorption, NH 3 -TPD, Py-IR, and 29 Si MAS NMR, and the catalytic performances of these samples were studied for the methylation of naphthalene with methanol.

Catalyst preparation
SAPO-11 was synthesized by hydrothermally crystallizing a sol-gel mixture with a composition of 1.0Al 2 O 3 : 1.0P 2 O 5 : 0.6SiO 2 : 1.2DPA : 49H 2 O. The nal crystallization temperature of 180 C and crystallization time of 20-28 h were employed. The products were washed with distilled water, then dried at 120 C for 5 h, and calcined at 600 C for 4 h. The obtained sample was SAPO-11.
Dealumination was performed by acid leaching 5 g of SAPO-11 mixed with 150 mL solution of critic acid or hydrochloric acid at the reux temperature. The samples treated with 1 mol L À1 , 4 mol L À1 , and 8 mol L À1 hydrochloric acid for 30 min were denoted as HClS1, HClS2, and HClS3, respectively. Similarly, samples modied with 1 mol L À1 , 4 mol L À1 , and 8 mol L À1 critic acid for 30 min were labeled as CAS1, CAS2, and CAS3, respectively. The obtained molecular sieve samples were washed with water, dried at 120 C for 12 h, and then calcined at 550 C for 4 h.

Catalyst characterization
The elemental compositions of the samples were determined by inductively coupled plasma-optical (ICP) emission spectroscopy (Perkin-Elmer ICP OPTIMA-3000).
X-ray powder diffraction (XRD) analysis was performed using a Rigaku D/max rB X-ray diffractometer. Diffraction patterns were obtained recorded with Cu Ka radiation at 40 kV and 100 mA in the scan range between 5 and 50 to identify the phase structure of the sample.
Scanning electron microscopy (SEM) was performed using a LEO-435VP scanning electron microscope operated at 20 kV and 50 PA.
Fourier transform infrared (FTIR) spectroscopy spectra of the samples were obtained using a Nicolet 380 FTIR spectrometer.
The textural properties of the samples were derived from N 2 adsorption-desorption measurements conducted using a Micromeritics Tristar 3000. In each case, the sample was outgassed under vacuum at 300 C for 3 h before N 2 adsorption. The specic surface area was calculated according to the BET method, and the volume of pores of the sieve was obtained by tplot analysis of the adsorption isotherm.
The acidity was examined by temperature programmed desorption of ammonia (NH 3 -TPD) technique, which was carried out in a ow system with a thermal conductivity detector. All samples were preheated from room temperature to 500 C in an argon ow and kept at 500 C for 1 h; these samples were then saturated with NH 3 by a owing stream of NH 3 /Ar at 40 C for 5 min. Evacuation at 40 C for 40 min was carried out to remove physically adsorbed NH 3 ; then, the catalyst was heated to 600 C at a linear heating rate of 10 C min À1 , and the detector signal of NH 3 was determined.
IR spectra of adsorbed pyridine (Py-IR) were obtained using a Nicolet-510P apparatus. Prior to pyridine adsorption, the calcined samples were pressed into thin wafers and evacuated in situ under vacuum at 300 C for 2 h and then cooled down to room temperature; pyridine was then dosed onto samples until saturation; nally, the system was evacuated at different temperatures, and the adsorbed pyridine was measured by IR spectroscopy. 29 Si MAS NMR experiments were carried out using a Varian innity-plus 300 NMR. All 29 Si MAS NMR spectra were obtained at 79.5 MHz using 2.0 ms pulse with 3.0 s recycle delay and 7120 scans.

Catalyst evaluation
Experiments were performed in a continuous ow xed-bed reactor equipped with a stainless-steel tube (with 20 mm diameter and 600 mm length). Then, 2.5 g of 20-40 mesh molecular sieve catalysts were loaded onto this reaction tube. The reaction mixture was fed into the reactor by a quantity measuring pump, and the pressure was maintained by N 2 . The weight hourly space velocity (WHSV) of naphthalene was 0.19 h À1 in all the experiments. The reaction temperature was 400 C, and the liquid reactants naphthalene, methanol, and mesitylene (solvent) in a molar ratio of 1 : 5 : 3.5 were preheated before passing them to the reactor. The reaction products were analyzed by gas chromatography (GC9560) using a Beta-Dex 120 capillary column. Conversion of naphthalene was calculated as follows: where n N,0 and n N are the molar percentage of naphthalene before and aer the reaction. The product distribution includes the corresponding molar percentages of ethylnaphthalene (EN), methylnaphthalene (MN), dimethylnaphthalene (DMN), and trimethylnaphthalene (TMN) in the total product mixture. The selectivity of 2,6-DMN is the corresponding molar percentage in the sum of all DMN isomers. 2,6-/2,7-DMN stands for the molar ratio of 2,6-DMN to 2,7-DMN. 2,6-DMN yield ¼ (naphthalene conversion Â 2,6-DMN distribution)/100%.

Results and discussion
3.1 Characterization of catalysts 3.1.1 XRD and ICP. The XRD patterns of untreated and acid-treated SAPO-11 samples are shown in Fig. 1. The bulk composition and relative crystallinity of the samples are shown in Table 1. As shown in Fig. 1, the characteristic peaks of the SAPO-11 phase (2q ¼ 8. 15 , 9.40 , 13.20 , and 15.57 ) are observed for all the acid-dealuminated SAPO-11 samples, which are identical to those of the pure SAPO-11 material. 15 This suggests that acid dealumination does not destroy the AEL topology of the SAPO-11 molecular sieves. Table 1 clearly shows that the relative crystallinity of SAPO-11 molecular sieves decreases as the concentration of HCl and CA increases because of the formation of amorphous phases during dealumination that do not diffuse-off from the channels. In addition, the Si/Al ratio of SAPO-11 modied by CA increases as the concentration of citric acid increases. This is probably due to the removal of extra-framework aluminum of pure SAPO-11. The Si/Al ratios in HClS1 and HClS2 samples are also larger than those in SAPO-11, but the Si/Al ratio in the HClS3 samples is smaller than that in HClS2. This is because the acid concentration of HClS3 is very high such that it not only removes extra-framework aluminum but also deletes framework silicon; this results in the formation of crystal defects. It has been reported in literature that hydrolysis of silicon bonds occurs in the case of 6 mol L À1 nitric acid. 16 Thus, our experimental results are consistent with ndings reported in the literature.
3.1.2 SEM. SEM results of untreated and acid-treated SAPO-11 samples are presented in Fig. 2(a-g). From Fig. 2, it is evident that the bulk of acid-treated SAPO-11 samples is slightly modied as compared to that of the untreated SAPO-11; moreover, HClS3 is corroded and cracked due to severe dealumination accompanied by desilication.
3.1.3 FT-IR. Fig. 3 shows the FTIR spectra of untreated and acid-treated SAPO-11 samples. According to the literature, 17 all peaks are assigned as follows: 1121 cm À1 arising from the asymmetric stretching vibration of inner tetrahedra; 706 cm À1 ascribed to the symmetric stretching vibration of inner tetrahedra; 630 cm À1 and 556 cm À1 due to deformation vibration of the four-membered ring and the deformation vibration of sixmembered ring, respectively. The 417 cm À1 corresponds to the deformation vibration of inner tetrahedra. From here we can see that the positions of peaks of all acid-treated SAPO-11 samples are in good agreement with those of SAPO-11.
3.1.4 N 2 adsorption-desorption characterization. Table 2 presents the pore structure parameters of untreated and acidtreated SAPO-11 samples. As seen in Table 2, the mesopore volume of acid-treated SAPO-11 samples (except for HClS3) is  larger than that of SAPO-11 as HCl and CA can remove aluminum oxide species that block the mesopores without the destruction of the SAPO-11 framework. Table 2 shows that the mesopore volume of HClS2 is the largest of all untreated and acid-treated SAPO-11 samples. In addition, the surface area and pore volume of the acid-treated SAPO-11 samples (except for HClS3) is larger than that of SAPO-11. This is because framework alumina and extra-framework alumina can escape from the channels during the acid treatment process. However, compared with the case of SAPO-11, the surface area, pore volume, and mesopore volume of HClS3 markedly decreased because the amorphous silica in SAPO-11 formed nonframework materials during the acid treatment process that blocked the micropore channels and led to the decreased surface area, pore volume, and mesopore volume. The N 2 adsorption-desorption isotherms of SAPO-11, HClS2, HClS3, and CAS3 are shown in Fig. 4. The typical N 2 adsorptiondesorption isotherms of as-synthesized samples are of the IV isotherm type according to the IUPAC classication. High adsorption of N 2 occurred in the low relative pressure range, and obvious hystereses are detected, suggesting the existence of micropores and secondary mesopores in untreated and acidtreated SAPO-11 samples. Hysteresis between adsorption and desorption branches can be observed at medium relative pressure (0.3-1.0) for all samples; this demonstrates the presence of a large number of secondary mesopores. Fig. 4 shows that the secondary mesopore size distributions decreases in the following order: HClS2 > CAS3 > SAPO-11 > HClS3.
3.1.5 NH 3 -TPD and Py-IR. NH 3 -TPD is frequently used to determine the number and strength of the acid sites in catalysts. The NH 3 -TPD results of SAPO-11, HClS2, HClS3, and CAS3 are presented in Fig. 5. There is only one distinct region at 230 C for all the samples, which indicates that weak acid sites are dominant in SAPO-11, HClS2, HClS3, and CAS3. Fig. 5 shows that the acid strength of all the samples is qualitatively similar. The total acid number of the samples decreases in the following order: CAS3 > HClS2 > SAPO-11 > HClS3. HClS2 has a larger number of acid sites than SAPO-11 because hydrochloric acid at a low concentration removes the aluminum oxide species; this blocks the micropores; therefore, more acid sites are exposed. The number of acid sites of HClS3 are lower as compared to those of SAPO-11; this indicates that hydrochloric acid at a high concentration eliminates more acid sites. Compared to other samples, CAS3 has most number of acid sites; this is related to the concentration of critic acid.
The surface acidity of SAPO-11, HClS2, HClS3, and CAS3 was determined by in site Py-IR technique. Usually, the bands at approximately 1450 cm À1 and 1621 cm À1 are assigned to Lewis acid sites, whereas the bands at around 1540 cm À1 and 1639 cm À1 are assigned to Brönsted acid sites. In addition, the   band at around 1490 cm À1 can be associated with pyridine adsorbed on both Lewis and Brönsted acid sites. 18 The Py-IR spectra of SAPO-11, HClS2, HClS3, and CAS3 obtained at 200, 300, and 400 C are presented in Fig. 6. Both Lewis acid sites and Brönsted acid sites decrease in the following order: CAS3 > HClS2 > SAPO-11 > HClS3, which is in agreement with NH 3 -TPD results shown in Fig. 5. In addition, Lewis acid sites and Brönsted acid sites decrease in line with the increasing temperature as the amount of chemisorption is less at high temperatures than that at low temperatures. 3.1.6 29 Si MAS NMR. 29 Si MAS NMR spectra of SAPO-11, HClS2, HClS3, and CAS3 are presented in Fig. 7. It is evident that the Si environments have signicant effect on the acidity of SAPO molecular sieve, which is enhanced in the order: Si(4Si) < Si(4Al) < Si(3Al, 1Si) < Si(2Al, 2Si) < Si(Al, 3Si). 19 As shown in Fig. 7, there are ve peaks at À90, À100, À105, À108, and À112 ppm, induced by the Si(4Al), Si(3Al, Si), Si(2Al, 2Si), Si(Al, 3Si), and Si(4Si) environments, respectively. HClS3 shows an intensive peak at À112 ppm assigned to Si(4Si) environments, which indicates that the number of acid sites of HClS3 is the least. CAS3 has the most abundant acid sites because of its intensive Si(Al, 3Si) environments (À108 ppm).

Catalytic performance for the methylation of naphthalene
The catalytic conversion and stability of SAPO-11, HClS2, HClS3, and CAS3 are compared in Fig. 8. As shown in Fig. 8, the conversion of naphthalene to SAPO-11 is 74.3% and 25.6% with 1 h and 6 h reaction times, respectively. The naphthalene conversion of HClS2 is 79.2% with reaction time of 1 h, but decreases to 45.7% with reaction time of 6 h. Conversion of naphthalene for HClS2 is 71.6% with reaction time of 1 h and 21.6% with reaction time of 6 h. CAS3 has the highest conversion of naphthalene (82.4%) with the reaction time of 1 h, and the conversion of naphthalene is 30.8% with the reaction time   of 6 h. The initial activity of CAS3 is higher than that of the other samples because it has more active centers than others due to its strong acidity. However, conversion of naphthalene does not decrease quickly although its strong acidity can accelerate the coke formation. On the contrary, HClS2 exhibits lower initial naphthalene conversion, and its stability is the highest among all the samples although its acidity is not the weakest. These results indicate that there is no correlation between naphthalene conversion and acidity of samples. As seen in Table 2 and Fig. 4, the secondary mesopore size distributions decreases in the following order: HClS2 > CAS3 > SAPO-11 > HClS3. Thus, an increase in second mesoporosity can lead to the introduction of more reactants into the major channels or can facilitate the products desorption out of the major channels, thus resulting in enhancement of conversion of naphthalene by HClS2. Therefore, the conversion of naphthalene is probably related to the differences in the pore structure of SAPO-11 samples, especially relating to the amount of secondary mesopores in the SAPO-11 molecular sieves. More number of secondary mesoporses in SAPO-11 molecular sieves result in the higher conversion of naphthalene over SAPO-11 samples. Moreover, it can be seen from Fig. 8 that the naphthalene conversion over SAPO-11, HClS2, HClS3, and CAS3 decreases with increased reaction time. It is possible that because some reactant molecules or intermediates are formed in the pore channels during the synthesis process, which block the pores of channels of SAPO-11 samples and prevent the further diffusion of the reactant and the product molecules, 20 thus leading to lower naphthalene conversion at longer reaction times. Based on the abovementioned reason, the authors think that the naphthalene conversion increases with time because the pore size of SAPO-11 molecular sieves can be increased by modifying the synthetic method or modifying the SAPO-11 molecular sieves. [21][22][23] The catalytic performance of SAPO-11, HClS2, HClS3, and CAS3 with 1 h and 6 h time on stream (TOS) are listed in Table 3. Table 3 shows that the main products are methylnaphthalene (MN), dimethylnaphthalene (DMN), and trimethylnaphthalene (TMN). The 2,6-DMN yield over the investigated molecular sieves decreases in the order: HClS2 > CAS3 > SAPO-11 > HClS3. The selectivity of 2,6-DMN and 2,6-/2,7-DMN ratio decrease in the order: HClS2 > CAS3 > SAPO-11 > HClS3 with the reaction time, which is not associated with the acidity of the samples. Fang et al. have calculated that 2,6-DMN is somewhat larger than 2,7-DMN in molecular dimension, 24 with their molecular dimensions in length, thickness, and cylindrical diameter being 10.06Å, 2.76Å, 6.44Å for 2,6-DMN and 9.73Å, 2.76Å, 6.03Å for 2,7-DMN. Therefore, 2,6-DMN suffers more diffusion resistance than 2,7-DMN, diffusing out from the pore channel of SAPO-11. HClS2 exhibits broader pore size distributions than the other SAPO-11 samples (see Table 2 and Fig. 4). The broad pore size distributions are helpful for the diffusion of 2,6-DMN, particularly, for the generation of the secondary mesopores over HClS2; this makes them more favorable for the diffusion of 2,6-DMN, thus nally leading to higher selectivity of 2,6-DMN and 2,6-/2,7-DMN ratio. Thus, SAPO-11 molecular sieves with more amount of secondary mesoporses present higher selectivity of 2,6-DMN and 2,6-/2,7-DMN ratio than other SAPO-11 samples.

Conclusions
SAPO-11 molecular sieves were modied by hydrochloric acid and citric acid. The inuence of modication of SAPO-11 by the two acids on the acidity, pore structure, and catalytic performance of the modied SAPO-11 molecular sieve-supported catalysts for the methylation of naphthalene was investigated. The XRD results showed that the modication of SAPO-11 by hydrochloric acid and citric acid decreased the relative crystallinity of the sample without destroying the AEL topology of SAPO-11. N 2 adsorption-desorption results showed that micropores and secondary mesopores were present in all samples. Moreover, NH 3 -TPD and 29 Si MAS NMR measurements showed that hydrochloric acid at a high concentration could eliminate more acid sites.
The treatment of SAPO-11 with 8 mol L À1 hydrochloric acid for 30 min (HClS3) not only removed some extra-framework aluminum of SAPO-11 but also deleted some framework silicon of SAPO-11; this resulted in lower degree of the secondary mesopores size, the total acid number, the conversion of naphthalene, and 2,6-DMN yield. SAPO-11 modied by 4 mol L À1 hydrochloric acid for 30 min (HClS2) showed the highest crystallinity, the broadest pore size distributions, and the highest catalytic performances among all the SAPO-11 samples modied in this study.