Zeolites in Pechmann condensation: Impact of the framework topology and type of acid sites

Journal Pre-proof Zeolites in pechmann condensation: Impact of the framework topology and type of acid sites ´ sova´ Ondˇrej Vesel´y, Mariya Shamzhy, Michal Mazur, Pavla Eliaˇ

PII:

S0920-5861(19)30556-5

DOI:

https://doi.org/10.1016/j.cattod.2019.10.003

Reference:

CATTOD 12508

To appear in:

Catalysis Today

Received Date:

9 July 2019

Revised Date:

26 September 2019

Accepted Date:

7 October 2019

´ sova´ P, Zeolites in pechmann Please cite this article as: Vesel´y O, Shamzhy M, Mazur M, Eliaˇ condensation: Impact of the framework topology and type of acid sites, Catalysis Today (2019), doi: https://doi.org/10.1016/j.cattod.2019.10.003

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.

ZEOLITES IN PECHMANN CONDENSATION: IMPACT OF THE FRAMEWORK TOPOLOGY AND TYPE OF ACID SITES Ondřej Veselý, Mariya Shamzhy, Michal Mazur, Pavla Eliášová Department of Physical and Macromolecular Chemistry, Faculty of Science, Charles University, Hlavova 8, 128 43 Prague 2, Czech Republic

Hierarchical MFI and MTW show higher conversions than their bulk counterparts Additional mesoporosity in *BEA does not enhance its performance significantly Conversions decrease in order: *BEA>MTW>MFI (independently on their form) MFI is not suitable catalyst for Pechmann condensation due to diffusion limitation Catalytic performance of aluminosilicates and gallosilicates is nearly comparable

Jo

    

ur

Highlights:

na

lP

re

-p

ro of

Graphical abstract

Abstract In this contribution, zeolites with different pore sizes and connectivity (MFI 10-10-10R, MTW 12R, Beta 12-12-12R) were prepared in bulk and hierarchical “nanosponge” form and with different

ro of

composition (aluminosilicate, gallosilicate). Properties of the samples were characterized by X-ray diffraction, argon adsorption, transmission electron microscopy and adsorption of pyridine followed by FTIR spectroscopy. Catalytic performance (conversion of reactants, yields of products and turnover frequencies) of prepared materials was investigated in the Pechmann condensation of resorcinol with ethyl acetoacetate. The lowest conversion (less than 15 %) over the MFI samples suggests that its 10R channels are too narrow for the products to escape from the framework and thus the reaction proceeds mainly on the crystal surface. Conversion of reactants over MTW and Beta zeolites is higher (10 %, 35 %, 71 % and 69 % for bulk MTW, nanosponge MTW, bulk Beta and nanosponge Beta, respectively), however, undesired side-products are formed. The hierarchical form of MFI and MTW gives higher conversions and yields of products than their bulk counterparts. Interestingly, the difference between bulk and hierarchical Beta is negligible. The overall performance of aluminosilicates and gallosilicates in the reaction is comparable. Catalytic activity of the samples was determined mainly by their textural properties. Keywords: Hierarchical zeolites; Nanosponge zeolites; Mesoporous materials; Gallosilicates; Pechmann condensation 1. Introduction

ur

na

lP

re

-p

Coumarin and its derivatives are important intermediates for syntheses of fine chemicals. Synthetic coumarin derivatives are used e.g. in pharmaceutical, agrochemical and fragrance industries. Some of their applications include enzyme inhibition, inflammation reduction, and the use as insecticides and antioxidants. [1] Methods for coumarin preparation include Perkin reaction [2], Wittig reaction [3] or Pechmann condensation. [4] In Pechmann condensation, coumarin is prepared in single step by acid-catalysed reaction of activated phenol with β-ketoester. Despite that the mechanism of this reaction has not been disclosed yet, assumingly it consists of three steps; electrophilic aromatic substitution, transesterification and dehydration. [4] Various inorganic and organic acids, such as: H2SO4, AlCl3, and CF3COOH, have been used as homogenous catalysts for this reaction. [5] However, toxicity, corrosive properties of those acids and difficulties of separation from the products limits their large scale application. [6, 7] On the contrary, heterogeneous solid catalysts such as heteropolyacids, zeolites or functionalised mesoporous silicas possess many advantages compared to the mineral acids. Most importantly, heterogeneous catalysts are easier to handle, separate from the reaction mixtures and reusable. [8-10] Zeolites were reported to give reasonable conversions at lower quantities and to work efficiently under milder conditions limiting unwanted consecutive reactions of the product. [11-13]

Jo

A unique property of zeolites is their shape-selectivity; the ability to allow or reject molecules to enter the micropores depending on their size. [14] Moreover, the composition of zeolites can be altered to tune their acidic properties. It is realized by introduction of various tetrahedrally coordinated elements to the framework (e.g. gallium, tin, titanium, boron, or iron). [15-20] Shapeselectivity of zeolites is a clear advantage; however the microporosity usually generates a significant drawback - the diffusion limitations. In order to overcome this issue large and extra-large pore zeolites may be used as the catalysts. [11] Other possibility is the application of hierarchical zeolites. These materials, besides the micropores, contain additional mesopores (2-50 nm in size) which enlarge their external surface and facilitate mass transfer of the reactants/products to the active centres. [21] Countless methods for preparation of hierarchical zeolites have been described;

templating by carbon nanoparticles, as well as etching with base, acid or HF. [22] Also the “softtemplating” method, which uses specially designed amphiphilic structure directing agents (SDAs) for the synthesis, has been deeply studied. This method enables a direct synthesis of zeolites in a form of aggregated nanocrystals or nanosheets with enlarged external surface areas and good accessibility of the pores. One of these forms is so-called “nanosponge”. [23] The goal of this work is to compare the performance of various zeolites and evaluate the effect of the framework topology, morphology and composition on their catalytic properties. Aluminosilicate and gallosilicate forms of zeolites MFI, MTW and Beta in bulk and hierarchical, nanosponge forms were tested in the Pechmann condensation of resorcinol with ethyl acetoacetate. 2. Experimental 2.1. Synthesis of SDAs

lP

re

-p

ro of

The synthesis of the nanosponge zeolites was carried out using structure directing agents (SDAs) C22H45-N+(CH3)2-C6H12-N+(CH3)2-C6H13](Br-)2 (denoted as C22-6-6 [24, 25]) and + + + + + [C22H45−N (CH3)2−C6H12−N (CH3)2−CH2−(C6H4)−CH2−N (CH3)2−C6H12−N (CH3)2−CH2−(C6H4)−CH2−N (CH3) + 2−C6H12−N (CH3)2−C22H45](Br−)2(Cl−)4 (denoted as C22N6 [23]). The C22-6-6 was prepared by reaction of 1-bromodocosane (98 %, TCI) with six times molar excess of N,N,N’,N’-tetramethyl-1,6diaminohexane (98 %, TCI) in a mixture of toluene and acetonitrile (volume ratio 1:1; 25 ml per 1 g of 1-bromodocosane at 60 °C for 12 h. The solvents were evaporated and the product was washed with diethyl ether and dried at room temperature. Subsequently, the precursor was mixed with three times its molar excess of 1-bromhexane (98 %, Sigma Aldrich) in chloroform (8.5 ml per 1 g of the precursor). The reaction was carried out at 80 °C for 24 h. Afterwards, the solvent was evaporated, the product thoroughly washed with diethyl ether and then dried at room temperature. The purity of the C22-6-6 product was checked with 1H NMR spectroscopy.

Jo

ur

na

The C22N6 was be prepared using the same precursor obtained from the reaction of 1bromodocosane with N,N,N’,N’-tetramethyl-1,6-diaminohexane. The subsequent step was a reaction with ten times molar excess of 1,4-bis(chloromethyl)benzene (98 %, Sigma Aldrich) in a mixture of chloroform and acetonitrile (volume ratio 2:1; 36 ml per 1 g of the precursor). The reaction was carried out at 65 °C for 24 h. After that, the solvents were evaporated, the product thoroughly washed with diethyl ether and acetone and then dried at room temperature. The product was mixed with half its molar amount of N,N,N’,N’-tetramethyl-1,6-diaminohexane and dissolved in chloroform (6.3 ml per 1 g). The reaction was carried out at 85 °C for 24 h. When finished, chloroform was evaporated, the product was washed with diethyl ether and dried at room temperature. Purity of the final product was confirmed by using 1H NMR spectroscopy. 2.2. Synthesis of Zeolites

The synthesis of aluminosilicate bulk MFI zeolite was carried out using tetrapropylammonium bromide as SDA. Potassium hydroxide (89.6 %. Lachner) was dissolved in distilled water and then aluminium nitrate nonahydrate (99.4 %, Lachner) was added. When completely dissolved, TPABr (98 %, Acros organics) was added to the mixture. Finally, tetraethylorthosilicate (98 %, Sigma Aldrich) was added and the mixture was left stirring for 5 h. The final molar composition of the gel was 100 SiO2 : 0.5 Al2O3 : 31 K2O : 37.5 SDA : 24000 H2O. The crystallization was carried out in a Teflon-lined steel autoclave at 175 °C for 2 days with rotation. The product was separated by filtration, washed

with distilled water and dried at 65 °C. The calcination was carried out in a flow of air at 550 °C for 6 h. The gallosilicate bulk MFI was prepared using the same procedure, substituting corresponding amount of gallium nitrate (99.9 %, Alfa Aesar) for the aluminium nitrate.

ro of

Tetraethylammonium hydroxide was used as SDA for the synthesis of aluminosilicate bulk MTW zeolite. Sodium aluminate (80-90 %, Riedel-de Haen) was mixed with distilled water and when completely dissolved, TEA−OH (25 % in H2O, Acros Organics) was added to the mixture. In a separate vessel colloidal silica (Ludox HS-40, Sigma Aldrich) was diluted to 30% solution with distilled water. Both solutions were mixed together and stirred until completely homogenous gel. The final molar composition of the gel was 100 SiO2 : 1 Al2O3 : 1.46 Na2O : 25 SDA : 1330 H2O. The crystallization was carried out in a Teflon-lined steel autoclave at 160 °C for 6 days under static conditions. The product was separated by filtration, washed with distilled water and dried at 65 °C. The calcination was carried out in a flow of air at 550 °C for 6 h. Gallosilicate bulk MTW was prepared using the same procedure replacing the sodium aluminate with gallium nitrate. Additional sodium hydroxide (99.2 %, Lachner) was added to adjust the pH. The calcination was carried out in a flow of air at 250 °C for 6 h followed by 450 °C for 2 h.

lP

re

-p

Aluminosilicate nanosponge MFI zeolite was prepared using the C22-6-6 as SDA. The SDA was dissolved in distilled water and 55 % solution of sodium silicate (55 % in H2O, Sigma Aldrich) was added. In a separate container aluminium sulphate (95 %, Fluka analytical) was dissolved in an equal volume of water, both solutions were mixed together, shaken vigorously and left stirring for 2 h. Further, 12.2 % sulphuric acid (98 %, Lachema) was added to the mixture dropwise and aging was carried out at 60 °C for 20 h. The final molar composition of the gel was 100 SiO2 : 0.5 Al2O3 : 28 Na2O : 7.5 SDA : 16.6 H2SO4 : 6000 H2O. The crystallization was carried out in a Teflon-lined steel autoclave at 150 °C for 6 days with rotation. The product was filtered, washed with distilled water and dried at 65 °C. The calcination was carried out under a flow of air at 580 °C for 6 h. Gallosilicate nanosponge MFI was prepared by the same procedure replacing the aluminium sulphate with the corresponding amount of gallium nitrate.

Jo

ur

na

Aluminosilicate nanosponge MTW zeolite was prepared using the C22N6 as SDA. Sodium aluminate was dissolved in water solution of sodium hydroxide. The mixture was heated up to 60 °C and then the SDA was added and stirred until completely dissolved. The mixture was transferred to a polypropylene bottle, tetraethylorthosilicate was added and the whole bottle was shaken intensively to homogenize the gel. Further, an aging was carried out at 60 °C for 20 h. The final molar composition of the gel was 100 SiO2 : 1 Al2O3 : 13 Na2O : 3.333 SDA : 4500 H2O. The crystallization was carried out in a Teflon-lined steel autoclave at 150 °C for 6 days with rotation. The product was filtered, washed with distilled water and dried at 65 °C. The calcination was carried out under a flow of air at 580 °C for 8 h. Gallosilicate nanosponge MTW was prepared using the C22N6 as SDA following the same procedure. Instead of sodium aluminate, gallium nitrate was used as gallium source and additional sodium hydroxide was added to compensate sodium ions in reaction mixture. Nanosponge Beta zeolite was prepared using the C22N6 as SDA following the same procedure as with nanosponge MTW. Only the composition of the gel was altered to result in final molar composition 100 SiO2 : 3.333 Al2O3 : 11 Na2O : 3.333 SDA : 5500 H2O. The crystallization was carried out in a Teflon-lined steel autoclave at 140 °C for 5 days with rotation. The product was filtered, washed with distilled water and dried at 65 °C. The calcination was carried out under a flow of air at

580 °C for 8 h. Commercial Beta zeolite (Zeolyst CP814C, Si/Al = 19) was used as a comparison bulk catalyst for the nanosponge Beta. 2.3. Characterization The structure and crystallinity of the zeolites were determined by X-ray powder diffraction using a Bruker AXS D8 Advance diffractometer equipped with a graphite monochromator and a position sensitive detector LYNXEYE XE-T using CuKα radiation in Bragg–Brentano geometry. TEM imaging was performed using JEOL NEOARM 200 F with a Schottky-type field emission gun at accelerating voltage of 200 kV. Microscope was equipped with TVIPS XF416 CMOS camera. The alignment was performed using standard gold nanoparticles film method. Due to low beam-stability of the sample the dose of electrons was kept below current density of 2 pA/cm2.

re

-p

ro of

Argon adsorption/desorption isotherms were measured on a Micromeritics 3Flex volumetric Surface Area Analyzer at -186 °C to determine surface area, pore volume and pore size distribution. Before the sorption measurements, all samples were degassed in a Micromeritics Smart Vac Prep instrument under vacuum at 250 °C (heating rate 1 °C/min) for 8 h. The specific surface area was evaluated by BET method using adsorption data in the range of a relative pressure from p/p0 = 0.05 to p/p0 = 0.25. The t-plot method was applied to determine the volume of micropores (Vmic). The adsorbed amount at relative pressure p/p0= 0.98 reflects the total adsorption capacity (V tot). The pore size distributions were calculated using the BJH model from the desorption branch of the isotherms.

ur

na

lP

The concentration and type of acid sites were determined by adsorption of pyridine as a probe molecule and observed by FTIR spectroscope Nicolet 6700 AEM equipped with DTGS detector, using the self-supported wafer technique. Prior to adsorption of the probe molecule, self-supported wafers of zeolite samples were activated in-situ by overnight evacuation at temperature 450 °C. Pyridine adsorption proceeded at 150 °C for 20 min at partial pressure 3 Torr, followed by 20-min evacuation at 150 °C. The concentrations of Brønsted and Lewis acid sites in aluminosilicate samples were calculated from integral intensities of individual bands characteristic of pyridine on Brønsted acid sites at 1545 cm–1 and band of pyridine on Lewis acid site at 1455 cm –1 and molar absorption coefficients of ε(B) = 1.67 ± 0.1 cm.µmol–1 and ε(L) = 2.22 ± 0.1 cm.µmol–1, respectively. [26] The spectra were recorded with a resolution of 4 cm-1 by collection 128 scans for single spectrum.

Jo

The purity of prepared organic SDAs was verified by measuring 1H NMR spectra on a Varian Mercury 300 MHz spectrometer. D4 methanol was used as the solvent. 2.4. Pechmann condensation

The catalytic tests were performed in the liquid phase under atmospheric pressure at 110 °C in a multi-experiment workstation StarFishTM. Prior to the experiment, 200 mg of the catalyst were activated at 450 °C for 90 min with a rate of 10 °C/min. Then 0.5 g of n-dodecane (99 %, Acros organics), 8.5 mmol of resorcinol (99 %, Sigma Aldrich), 10 ml of nitrobenzene (99 %, Alfa Aesar) and the catalyst were placed in a two-necked vessel equipped with a condenser and a thermometer and heated to the reaction temperature. When the temperature was achieved sample in time zero was taken and then 10 mmol of ethyl acetoacetate were added into the vessel to start the reaction. Samples of the reaction mixture were taken in 10 min, 20 min, 30 min, 40 min, 1h, 2h, 3h, 4h, 5h and

24h. Directly after taking each sample was diluted to three times of its mass with nitrobenzene, centrifuged to remove the catalyst and analysed by gas chromatography. Analyses were performed using gas chromatograph Agilent 7890B GC equipped with HP-5 column (length 30 m, diameter 0.320 mm, and film thickness 0.25 μm) and flame ionization detector. Analysis of the products was carried out using the Thermo Scientific Trace 1310 gas chromatograph equipped with TG-SOC column (length 15 m, diameter 0.25 mm, and film thickness 0.25 μm) and connected to the Thermo Scientific ISQ LT single quadrupole mass spectrometer. 3. Results and discussion

ro of

Samples with MFI, MTW, and Beta zeolite topologies prepared in bulk and nanosponge form were characterised and tested in Pechmann condensation of resorcinol. Each zeolite topology is described separately for clear comparison of the influence of textural properties and composition on the catalytic performance. At the end of this chapter all samples are compared together based on the turnover frequencies. 3.1. Characterization of the MFI zeolite samples

lP

re

-p

MFI zeolite samples were prepared in the aluminosilicate (Al) and gallosilicate (Ga) forms, both as bulk (bMFI) and nanosponge (nsMFI). Crystallinity of samples was determined by the X-ray diffraction (Figure 1). Broadening of the diffraction peaks of the samples in the nanosponge form indicates small and thin crystallites. Closer analysis of the background of the diffraction patterns reveals a presence of secondary amorphous phase, most notably in case of the nsMFI (Al) sample. A broad peak around 4.5° in the diffraction pattern of the bMFI (Al) indicates the secondary phase may be partially ordered.

na

Textural properties of the samples were investigated by argon adsorption method (Figure 2, Table 1). The bulk samples give clear type-one isotherms with plateau reached at relative pressure 0.05, reflecting their microporous nature. On the other hand, the isotherms of both nanosponge samples show the type-two profile, which indicates the presence of mesopores as can be seen from the pore size distribution. Relatively thin hysteresis loop of the nanosponge samples suggests that majority of the mesopores is directly accessible from the external surface. [27]

Jo

ur

Figure 3 shows the TEM images of the aluminosilicate MFI samples. The difference between the bulk and the nanosponge samples is clearly visible. The bulk zeolite (A, C) is composed of 7 to 10 µm monocrystals, whereas, the nanosponge (B, D) is made of randomly aggregated layer-like small crystallites with the thickness of several nanometers. Gallosilicate MFI samples are displayed in the Figure 4. The gallosilicate bulk MFI (A, C) contains large monocrystals with well defined edge. On the other hand, the gallosilicate nanosponge (B, D) is formed of several-nanometers thick, crystalline, aggregated lamelas. The FTIR measurement of the MFI samples after adsorption of pyridine was used to investigate the concentration of Brønsted and Lewis acid sites (Figure 5). Each spectrum contains a peak at 3744 cm-1 which belongs to the vibration of the silanol groups on the crystal surface. The peak at 3610 cm 1 belongs to the bridging Si-(OH)-Al or Si-(OH)-Ga groups. After the adsorption of pyridine, this peak disappears as a result of the interaction of the Al or Ga site with the protonated pyridine ring. The band at 1545 cm-1 corresponds to a pyridine adsorbed on Brønsted sites. Another peaks at 1455 cm -1

and 1445 cm-1 belong to a pyridine adsorbed on Lewis acid sites of different strength. [28] Total concentration of acid sites in both gallosilicate MFI zeolites is comparable (0.09 mmol/g and 0.08 mmol/g for the bMFI(Ga) and nsMFI(Ga), respectively). It is noticeably lower than respective acid site concentration of their aluminosilicate counterparts (0.27 mmol/g and 0.17 mmol/g for the bMFI(Al) and nsMFI(Al), respectively), Table 2. This is a direct result of less favourable incorporation of gallium into the structure during the synthesis. Despite the synthesis mixture having the same molar composition, the final zeolite contains less gallium, and thereby, lower amount of acid sites compared to the aluminosilicate. It is noteworthy to point out, that the peak of weak Lewis sites at 1445 cm-1 is dominant for both gallosilicate samples, whereas in aluminosilicates the stronger Lewis sites (1455 cm-1) are more common.

ro of

3.2. Characterization of the MTW zeolite samples

lP

re

-p

MTW zeolite samples were prepared as aluminosilicate and gallosilicate, both in the bulk and nanosponge form (denoted in the same manner as MFI). The diffraction patterns of the samples in Figure 6Fig matches well with the simulated powder patterns in the database. In the diffraction pattern of the bMTW (Ga) we can observe diffraction peaks of some secondary phase. The undesired phase was identified as MFI. Due to lower stability of the gallosilicate MTW the formation of MFI is difficult to avoid. Nevertheless, as we discuss in the section 3.4, the MFI is virtually inactive in the reaction and therefore its presence should not affect the resulting yields or selectivity significantly. Broadening of the diffraction peaks of both nanosponge samples is a result of their small crystallite morphology.

na

The adsorption isotherms of the bulk and nanosponge MTW samples (Figure 7) show isotherm types one and two, respectively. Textural properties (Table 3) point to clear difference between two morphologies, especially in their external surface areas (57 m 2/g and 266 m2/g for the bMTW(Al) and nsMTW(Al) accordingly).

Jo

ur

The Figure 8 shows the TEM images of the aluminosilicate MTW samples. Unlike the MFI, the particles of bulk MTW (A, C) are polycrystalline (approximately 1 µm in diameter), made of aggregates of smaller crystallites. Due to this, the surface of bulk MTW is not as well defined as the MFI, which is the cause of the sample’s higher external surface area (6 m 2/g compared to 57 m2/g for the bMFI and bMTW, respectively). The nanosponge MTW (B, D) consists of many aggregated small (10 – 30 nm) crystallites. The morphology of gallosilicate bulk MTW (Figure 9 A, C) is similar to its aluminosilicate counterpart, although its crystals are slightly smaller (around 0.8 µm in diameter).

The FTIR spectra of the MTW samples (Figure 10) contain the same peaks that were described for the MFI samples. The difference in intensities of the peak of silanol groups at 3744 cm -1 between the nanosponge MTW sample and the bulk MTW (57 m2/g) is more striking than in case of the MFI samples despite the differences between their textural properties are comparable. The peak of the

bridging hydroxyl groups at 3610 cm-1 disappears after the adsorption of pyridine, indicating that all active sites are well accessible. The band at 1545 cm -1 corresponds to a pyridine adsorbed on Brønsted sites and the bands at 1455 cm-1 and 1445 cm-1 to a pyridine adsorbed on strong and weak Lewis acid sites. The aluminosilicate bulk MTW is predominantly Brønsted acidic (Table 4) with 0.11 mmol/g compared to 0.03 mmol/g of Lewis acid sites. On the other hand, the nanosponge form contains slightly more Lewis sites (0.06 mmol/g) than Brønsted sites (0.05 mmol/g). The concentration of acid sites in the gallosilicate nanosponge MTW is lower (0.04 and 0.06 mmol/g for the Brønsted and Lewis sites, respectively) due to more difficult incorporation of gallium into the structure.

3.3. Characterization of the Beta zeolite samples

-p

ro of

The sample of nanosponge Beta was prepared in the aluminosilicate form and commercial bulk Beta zeolite (Zeolyst CP814C, Si/Al = 19) was used for comparison. The direct hydrothermal synthesis of the gallosilicate samples turned out to be very difficult despite wide range of synthesis conditions we applied. The probably reason is lower hydrothermal stability of gallosilicate Beta forms. The diffraction patterns of the commercial bulk and nanosponge Beta (Figure 11) are in a good agreement with the simulated powder patterns in the database. The broad diffraction peaks at 7.6° and 22.4° in both patterns are a result of the disordered nature of the Beta structure.

lP

re

The adsorption isotherms (Figure 12, Table 5) of the Beta samples confirm the almost purely microporous nature of the bulk sample and presence of mesoporosity in the nanosponge Beta. The hysteresis loop of the isotherm is relatively thin, indicating well accessible mesopores.

ur

na

Figure 13 shows TEM images of Beta zeolite samples. Similar to MTW, the bulk sample (A, C) contains relatively large polycrystals (approximately 1 µm in diameter) made of smaller crystalline domains, which results in its increased external surface area (170 m 2/g) compared to the monocrystalline MFI (6 m2/g) and polycrystalline MTW (57 m2/g). The nanosponge (B, D) is formed of aggregated small crystallites (10 – 30 nm in size) with large mesopores and interparticle volume contributing largely to high total pore volume 1.387 cm3/g.

Jo

The concentrations of Brønsted and Lewis acid sites in the Beta samples (Figure 14, Table 6) differ significantly between the nanosponge and bulk sample. The majority of the acid sites in the bulk Beta sample is Brønsted acidic (0.30 mmol/g), whereas in the nanosponge Beta Lewis sites are dominant (0.22 mmol/g). 3.4. Catalysis over the MFI zeolite samples

The catalytic properties of the MFI samples were investigated in the Pechmann condensation of resorcinol with ethyl acetoacetate. The reaction was performed in an excess of ethyl acetoacetate. The Figure 15 shows the conversions of the resorcinol up to 24 hours of the reaction time, as well as yields of observable products for the same time period. In the means of conversion of reactant and yields of products the bulk MFI samples appear virtually inactive. The approximate kinetic diameter of resorcinol is 8.2 Å [29] and therefore it cannot penetrate the 10-ring MFI channels (5.1 x 5.5 Å, 5.3

ro of

x 5.6 Å). Thus, the reaction is expected to proceed mostly on the acid sites located on the external surface, which is very low in both aluminium and gallium bulk MFI form (6 and 18 m2/g for bMFI(Al) and bMFI(Ga), respectively). Only the aluminosilicate nanosponge MFI was active. We can see a slow increase in the conversion which reaches 15 % after 24 hours. Simultaneously, corresponding growth of the yield of the product can be seen as well. Formation of any side-products has not been detected during the experiment. The improved performance of the nsMFI(Al) sample is most probably caused by the presence of the additional mesoporosity. Nevertheless, due to the narrow diameter of the micropores it is more likely that in case of the nanosponge MFI the reaction is catalysed by acid centres located on the external surface of the zeolite or at its micropore openings. The negligible conversion over the gallosilicate nanosponge MFI is ascribed to its lower total concentration of acid sites (0.08 mmol/g) compared to its aluminosilicate counterpart with 0.17 mmol/g.

3.5. Catalysis over the MTW zeolite samples

na

lP

re

-p

The conversions of bulk MTW samples (Figure 16) after 24 hours reached 11 and 7 % over the aluminosilicate and gallosilicate MTW, respectively. Yields of the product correlate well with the conversions and no side-products were detected. The product (product 1) was identified by the GCMS as the desired product of the reaction; 4-methyl-7-hydroxycoumarin. In contrast to the MFI zeolites, the improved performance of the bulk MTW samples implies that its 12-ring pores (5.6 x 6.0 Å) are wide enough for the reactant and product molecules to enter and leave the framework. Moreover, their smaller crystal size and different crystal morphology may enhance the reaction rate as well. As a result of the enhanced external surface area the nanosponge samples give higher conversions than their bulk counterparts; 40 % over the aluminosilicate MTW and 36 % over the gallosilicate MTW after 24 hours. However, formation of second product (product 2) after 5 hours of the reaction has been detected. This compound has been recognised as a product of subsequent Pechmann condensation on the 4-methyl-7-hydroxycoumarin. The yield of the second product reaches approximately 5 % after 24 hours. The large external surface of the nanosponge facilitates the molecular transport, ergo increasing the conversion and yield of product 1 (4-methyl-7hydroxycoumarin), nevertheless, at the same time enables bulkier molecules to react on its surface and thereby to form the undesired side-product.

ur

3.6. Catalysis over the Beta zeolite samples

Jo

The conversions of the resorcinol over the Beta samples as well as yields of observable products are depicted in Figure 17. Both values increase, compared to the reaction over the MTW samples as a result of the larger pore size (6.6 x 7.7 Å) of the Beta as well as three-dimensional connectivity of micropores within its framework. Unexpectedly, the difference in conversion after 24 hours over the bulk and nanosponge Beta samples is rather small despite the vast differences in textural properties and acidity. More interestingly the bulk Beta sample gives higher conversion (72 %) than the nanosponge (69 %). It implies that for Beta structure the limitation by diffusion is not as prominent and ergo the additional mesoporosity does not augment the reaction rate significantly. Instead, the rate of the reaction is governed by the difference in acidity of the two samples (Table 6). Another observation distinct from reactions over the MFI and MTW samples is the decrease of the yield of the product 1 (4-methyl-7-hydroxycoumarin) after 5 hours from 38 % to 28 % over bulk Beta.

Simultaneously the formation of the second product (product 2) can be observed. This confirms a presence of subsequent reaction where the first product, the coumarin, reacts to form the secondary product. The secondary products observed in both reactions over the MTW and Beta were identified as the same compound.

ro of

To compare all the samples together conversions of resorcinol at 3 hours of the reaction were used to calculate the turn over frequencies (Figure 18). In all cases the turnover frequency of the nanosponge is higher than the turnover frequency of the bulk sample. The additional mesoporosity facilitates the diffusion of reactants and products, makes the active sites more easily accessible and thereby increasing the turnover frequency. The difference is significant in case of MFI (10-10-10R) and MTW (12R) samples but nearly negligible for Beta (12-12-12R). It can be deduced that the effect of the additional mesoporosity becomes less significant with the increasing pore size and connectivity of the channel system. Due to the low conversions and low concentrations of acid sites, the results of gallosilicate MFI zeolites are inevitably affected by high relative error. From comparison of the turn over frequencies of the MTW zeolites it can be concluded that the activity of aluminosilicates and gallosilicates in this reaction is comparable. 4. Conclusions

re

-p

Zeolites MFI, MTW and Beta were prepared in the bulk and nanosponge forms by direct hydrothermal synthesis. MFI and MTW were prepared both as aluminosilicates and gallosilicates. The lower stability of gallosilicate Beta compromised the synthesis attempts and therefore only the aluminosilicate forms were investigated.

ur

na

lP

The catalysts were tested in Pechmann condensation of resorcinol with ethyl acetoacetate. Both bulk MFI samples are inactive as a result of their low external surface area and too small pore diameter. The aluminosilicate nanosponge MFI gave slightly higher conversion of 15 %, due to the additional mesoporosity and higher external surface area. Similarly, high conversions (11 % and 7 %) were observed for both bulk MTW samples. The use of nanosponge MTW samples led to higher conversions (40 % and 36 % for the aluminosilicate and gallosilicate, respectively). Nevertheless, the formation of secondary product can be observed after 5 hours of the reaction time. The conversions over bulk and nanosponge Beta are very similar (72 % and 69 %, respectively) which indicates that additional mesoporosity is not very significant in case of this structure. Again, formation of second, subsequent product is observed after 5 hours.

Jo

In the summary, conversion of reactants over individual samples increases with the increasing diameter and connectivity of the channels. Correspondingly, the differences in activity (in the means of turn-over frequencies) between the bulk and hierarchical zeolites become less significant with the increase of micropore size and conectivity (MFI > MTW > Beta). On top of that, the selectivity decreases following the same trend, which results in formation of subsequent, secondary products. The catalystic results for MTW zeolite indicate that the activity of aluminosilicates and gallosilicates in the Pechmann condensation is comparable. Acknowledgement

OV acknowledges the Czech Science Foundation for the ExPro project (19-27551X). MM and MS would like to acknowledge the OP VVV "Excellent Research Teams", project No. CZ.02.1.01/ 0.0/0.0/15_003/0000417 – CUCAM. References

Jo

ur

na

lP

re

-p

ro of

[1] J. Oyamada, C.G. Jia, Y. Fujiwara, T. Kitamura, Direct synthesis of coumarins by Pd(II)-catalyzed reaction of alkoxyphenols and alkynoates, Chemistry Letters, (2002) 380-381. [2] J.R. Johnson, The Perkin Reaction and Related Reactions, Organic Reactions, (2011). [3] I. Yavari, R. Hekmat-Shoar, A. Zonouzi, A new and efficient route to 4-carboxymethylcoumarins mediated by vinyltriphenylphosphonium salt, Tetrahedron Letters, 39 (1998) 2391-2392. [4] S. Tyndall, K.F. Wong, M.A. VanAlstine-Parris, Insight into the Mechanism of the Pechmann Condensation Reaction Using NMR, The Journal of Organic Chemistry, 80 (2015) 8951-8953. [5] A.K.D. Gupta, R.M. Chatterje, K.R. Das, B. Green, Coumarins and related compounds. Part IV. Aluminium chloride-catalysed reaction of phenols with methyl acrylate: a new approach to the synthesis of hydroxycoumarins, Journal of the Chemical Society C: Organic, (1969) 29-33. [6] L.L. Woods, J. Sapp, A NEW 1-STEP SYNTHESIS OF SUBSTITUTED COUMARINS, J. Org. Chem., 27 (1962) 3703-&. [7] J.R.F. Alfred Russell, 2,6-Dihydroxyacetophenone, Organic Syntheses Collective, 3 (1955) 281. [8] E.A. Gunnewegh, A.J. Hoefnagel, H. van Bekkum, Zeolite catalysed synthesis of coumarin derivatives, Journal of Molecular Catalysis A: Chemical, 100 (1995) 87-92. [9] R. Torviso, D. Mansilla, A. Belizán, E. Alesso, G. Moltrasio, P. Vázquez, L. Pizzio, M. Blanco, C. Cáceres, Catalytic activity of Keggin heteropolycompounds in the Pechmann reaction, Applied Catalysis A: General, 339 (2008) 53-60. [10] B. Tyagi, M.K. Mishra, R.V. Jasra, Synthesis of 7-substituted 4-methyl coumarins by Pechmann reaction using nano-crystalline sulfated-zirconia, Journal of Molecular Catalysis A: Chemical, 276 (2007) 47-56. [11] M. Opanasenko, M. Shamzhy, J. Čejka, Solid Acid Catalysts for Coumarin Synthesis by the Pechmann Reaction: MOFs versus Zeolites, Chemcatchem, 5 (2013) 1024-1031. [12] J.-C. Kim, R. Ryoo, M.V. Opanasenko, M.V. Shamzhy, J. Čejka, Mesoporous MFI Zeolite Nanosponge as a High-Performance Catalyst in the Pechmann Condensation Reaction, ACS Catalysis, 5 (2015) 2596-2604. [13] C. Gutiérrez-Sánchez, V. Calvino-Casilda, E. Pérez-Mayoral, R.M. Martín-Aranda, A.J. LópezPeinado, M. Bejblová, J. Čejka, Coumarins Preparation by Pechmann Reaction Under Ultrasound Irradiation. Synthesis of Hymecromone as Insecticide Intermediate, Catalysis Letters, 128 (2008) 318. [14] C. Baerlocher, L.B. McCusker, D.H. Olson, Atlas of Zeolite Framework Types, Elsevier, Amsterdam, 2007. [15] J. Přech, Catalytic performance of advanced titanosilicate selective oxidation catalysts – a review, Catalysis Reviews, 60 (2018) 71-131. [16] M. Shamzhy, M. Opanasenko, P. Concepción, A. Martínez, New trends in tailoring active sites in zeolite-based catalysts, Chemical Society Reviews, 48 (2019) 1095-1149. [17] R. Fricke, H. Kosslick, G. Lischke, M. Richter, Incorporation of Gallium into Zeolites:  Syntheses, Properties and Catalytic Application, Chemical Reviews, 100 (2000) 2303-2406. [18] M.V. Shamzhy, O.V. Shvets, M.V. Opanasenko, L. Kurfiřtová, D. Kubička, J. Čejka, Extra-LargePore Zeolites with UTL Topology: Control of the Catalytic Activity by Variation in the Nature of the Active Sites, Chemcatchem, 5 (2013) 1891-1898. [19] J.i. C̆ejka, A. Vondrová, B. Wichterlová, G. Vorbeck, R. Fricke, The effect of AI, Fe, and in strength, Zeolites, 14 (1994) 147-153. [20] H. Berndt, A. Martin, H. Kosslick, B. Lücke, Comparison of the acidic properties of ZSM-5 zeolites isomorphously substituted by Ga, In, B and Fe, Microporous Materials, 2 (1994) 197-204.

Jo

ur

na

lP

re

-p

ro of

[21] J. Přech, P. Pizarro, D.P. Serrano, J. Čejka, From 3D to 2D zeolite catalytic materials, Chemical Society Reviews, 47 (2018) 8263-8306. [22] J. Přech, K.N. Bozhilov, J. El Fallah, N. Barrier, V. Valtchev, Fluoride etching opens the structure and strengthens the active sites of the layered ZSM-5 zeolite, Microporous and Mesoporous Materials, 280 (2019) 297-305. [23] W. Kim, J.-C. Kim, J. Kim, Y. Seo, R. Ryoo, External Surface Catalytic Sites of Surfactant-Tailored Nanomorphic Zeolites for Benzene Isopropylation to Cumene, ACS Catalysis, 3 (2013) 192-195. [24] C. Jo, K. Cho, J. Kim, R. Ryoo, MFI zeolite nanosponges possessing uniform mesopores generated by bulk crystal seeding in the hierarchical surfactant-directed synthesis, Chemical Communications, 50 (2014) 4175-4177. [25] C. Jo, R. Ryoo, N. Žilková, D. Vitvarová, J. Čejka, The effect of MFI zeolite lamellar and related mesostructures on toluene disproportionation and alkylation, Catalysis Science & Technology, 3 (2013) 2119-2129. [26] H.S. Shin, M. Opanasenko, C.P. Cabello, R. Ryoo, J. Čejka, Surfactant-directed mesoporous zeolites with enhanced catalytic activity in tetrahydropyranylation of alcohols: Effect of framework type and morphology, Applied Catalysis A: General, 537 (2017) 24-32. [27] M. Thommes, Physical Adsorption Characterization of Nanoporous Materials, Chemie Ingenieur Technik, 82 (2010) 1059-1073. [28] J.A. Lercher, A. Jentys, Chapter 13 - Infrared and Raman Spectroscopy for Characterizing Zeolites, in: J. Čejka, H. van Bekkum, A. Corma, F. Schüth (Eds.) Studies in Surface Science and Catalysis, Elsevier, 2007, pp. 435-476. [29] B. Koubaissy, J. Toufaily, M. El-Murr, T. Jean Daou, H. Hafez, G. Joly, P. Magnoux, T. Hamieh, Adsorption kinetics and equilibrium of phenol drifts on three zeolites, Central European Journal of Engineering, 2 (2012) 435-444.

Figure 1. X-ray powder diffraction patterns of MFI samples

ro of

Jo

ur

na

lP

re

-p

Figure 2. Argon adsorption isotherms and pore size distributions of the MFI samples

na

lP

re

-p

ro of

Figure 3. TEM images of aluminosilicate bulk MFI – A, C; and aluminosilicate nanosponge MFI – B, D

Jo

ur

Figure 4. TEM images of gallosilicate bulk MFI – A, C; and gallosilicate nanosponge MFI – B, D

ro of

Jo

ur

na

lP

re

-p

Figure 5. FTIR spectra the MFI samples: left – vibrations of Si-OH and Si-(OH)-T groups before (b) and after (a) the adsorption of pyridine; right – vibrations of pyridine ring adsorbed on the active sites

Figure 6. X-ray powder diffraction patterns of the MTW samples

ro of

Jo

ur

na

lP

re

-p

Figure 7. Argon adsorption isotherms and pore size distributions of the MTW samples

na

lP

re

-p

ro of

Figure 8. TEM images of aluminosilicate bulk MTW – A, C; and aluminosilicate nanosponge MTW – B, D

Jo

ur

Figure 9. TEM images of gallosilicate bulk MTW –A , C; and gallosilicate nanosponge MTW – B, D

ro of

Jo

ur

na

lP

re

-p

Figure 10. FTIR spectra the MTW samples: left – vibrations of Si-OH and Si-(OH)-T groups before (b) and after (a) the adsorption of pyridine; right – vibrations of pyridine ring adsorbed on the active sites

Figure 11. X-Ray powder diffraction patterns of the Beta samples

ro of

Jo

ur

na

lP

re

-p

Figure 12. Argon adsorption isotherms and pore size distributions of the Beta samples

-p

ro of

Figure 13. TEM images of commercial aluminosilicate Beta – A, C; and aluminosilicate nanosponge Beta – B, D

Jo

ur

na

lP

re

Figure 14. FTIR spectra the Beta samples: left – vibrations of Si-OH and Si-(OH)-T groups before (b) and after (a) the adsorption of pyridine; right – vibrations of pyridine ring adsorbed on the active sites

Figure 15. Conversions of resorcinol over MFI samples and yields of the main product

ro of

Jo

ur

na

lP

re

-p

Figure 16. Conversions of resorcinol over MTW samples and yields of the main product

Figure 17. Conversions of resorcinol over Beta samples and yields of the main product

ro of

re

-p

Figure 18. Turn-over frequencies (TOFs) of each catalyst used for Pechmann condensation of resorcinol

na

BET [m²/g] 270 364 385 327

Sext [m²/g] 6 202 18 154

Vtot [cm³/g] 0.112 0.443 0.165 0.374

Vmic [cm³/g] 0.108 0.098 0.160 0.087

ur

Sample bMFI (Al) nsMFI (Al) bMFI (Ga) nsMFI (Ga)

lP

Table 1. Textural properties of the MFI samples based on argon adsorption

Jo

Table 2. Acidic properties of the MFI samples based on pyridine adsorption followed by FTIR cB [mmol/g] cL [mmol/g]

bMFI (Al) 0.14 0.13

nsMFI (Al) 0.10 0.07

bMFI (Ga) 0.05 0.04

nsMFI (Ga) 0.04 0.04

Table 3. Textural properties of the MTW samples based on argon adsorption measurement Sample bMTW (Al) nsMTW (Al) bMTW (Ga)

BET [m²/g] 292 379 223

Sext [m²/g] 57 266 67

Vtot [cm³/g] 0.220 0.729 0.170

Vmic [cm³/g] 0.094 0.038 0.083

nsMTW (Ga)

396

161

0.674

0.099

Table 4. Acidic properties of the MTW samples based on pyridine adsorption followed by FTIR

cB [mmol/g] cL [mmol/g]

bMTW (Al) 0.11 0.03

nsMTW (Al) 0.05 0.06

bMTW (Ga) 0.04 0.06

nsMTW (Ga) 0.04 0.06

Table 5. Textural properties of the Beta samples based on argon adsorption measurement BET [m²/g] 560 712

Sext [m²/g] 170 626

Vtot [cm³/g] 0.301 1.387

Vmic [cm³/g] 0.227 0.069

ro of

Sample cBeta nsBeta

Table 6. Acidic properties of the Beta samples based on pyridine adsorption followed by FTIR nsBeta 0.11 0.22

Jo

ur

na

lP

re

-p

cBeta 0.30 0.10

cB [mmol/g] cL [mmol/g]