Producing pyridines via thermo-catalytic conversion and ammonization of glycerol over nano-sized HZSM-5

Abstract

In this study, nano-sized HZSM-5 catalysts with different Si/Al ratios were synthesized and employed for producing pyridines from glycerol via a thermo-catalytic conversion and ammonization (TCC-A) process. The catalytic performance of micro-sized HZSM-5 and nano-sized HZSM-5 was studied firstly. The nano-sized HZSM-5 showed better catalytic performance in pyridine production in the TCC-A process due to its smaller particle size. When the nano-sized HZSM-5 (Si/Al = 25) served as the catalyst, and the reaction temperature was about 550 °C with the weight hourly space velocity of glycerol to catalyst at 1 h⁻¹ and the ammonia to glycerol ratio at 12 : 1, the highest yield of pyridines was up to 42.1%, which was much higher than that when using micro-sized HZSM-5 (35.6%) reported before. In addition, nano-sized HZSM-5 also showed a better catalytic performance than micro-sized HZSM-5 in the catalytic conversion of bio-derived furans to produce indoles. After five reaction/regeneration cycles, the catalytic performance of nano-sized HZSM-5 slightly decreased compared with the first run, but was still higher than that of micro-sized HZSM-5.

---

Introduction

Pyridines have been widely used as intermediates in the synthesis of many agrochemicals, pharmaceuticals and other fine chemicals.^1–4 Pyridines are produced from acyclic molecules (formaldehyde, acetaldehyde and acrolein) and ammonia over zeolites (HZSM-5, HY and H-β) via Chichibabin condensation in industry.^5–9 However, the raw materials for the synthesis of pyridines, such as formaldehyde and acrolein, are very toxic, carcinogenic, unstable, and non-renewable.¹⁰ Therefore, it would be highly desirable to find a renewable, stable, and environmentally friendly feedstock to produce sustainable pyridine compounds. Recently, we reported that pyridines can be synthesized directly from renewable glycerol or waste polylactic acid (PLA) in a fixed bed reactor over HZSM-5 under ammonia atmosphere via a thermo-catalytic conversion and ammonization (TCC-A) process.^11,12 In the process from glycerol to pyridines, glycerol initially dehydrates to acrolein and some by-products such as acetaldehyde, acetol, acetone, etc. Then acrolein, a mixture of acrolein and acetaldehyde, or other by-products react with ammonia to form imines and finally pyridines over HZSM-5.¹¹ PLA undergoes a similar reaction to form pyridines. PLA initially thermally decomposes to form lactic acid and some by-products such as acetaldehyde, acetone, etc., and then lactic acid, a mixture of acetaldehyde and acetone, or other by-products react with ammonia to form imines which finally undergo complicated reactions to form pyridines.¹² Luo and co-workers also reported a similar result about producing pyridines from glycerol and ammonia in a series-connected two-stage fixed-bed reactor.¹³ In addition to glycerol, they also synthesized pyridine or 3-picoline from other renewable feedstocks, such as acrolein dimethyl acetal or acrolein diethyl acetal and acrolein, which are glycerol-derived chemicals.^14–16

According to the literature on the synthesis of pyridines, HZSM-5 is one of the most studied catalysts, which possesses many advantages such as good shape selectivity, strong hydrothermal stability and variable acidity. However, there are still some shortcomings of HZSM-5 in the production process of pyridines, such as low pyridine yield, coking, and low stability, which were also evident in our previous study on TCC-A of glycerol to produce pyridines. In our previous study, the catalyst showed a slight deactivation after five reaction/regeneration cycles.¹¹ Recently, much effort has been made to address these problems. To extend the lifetime of HZSM-5, the alkaline treatment method was employed in many studies. The alkaline treatment could enlarge the pore structure and reduce the acidic strength of the zeolite catalyst, which could reduce the coke and extend the lifetime of the catalyst. Jin and co-workers employed alkali-treated HZSM-5 to catalyze the reaction of formaldehyde, acetaldehyde and ammonia to synthesize pyridines and found the stability of alkali-treated HZSM-5 was much better than that of the untreated HZSM-5.^17,18 However, the changes of acidity and pore structure of the catalyst impaired the catalytic performance of HZSM-5. Therefore, a modified alkaline-acid sequential treatment method was applied to improve HZSM-5 by Luo and co-workers. Compared to the alkali-treated HZSM-5, the alkaline-acid sequentially treated HZSM-5 showed better catalytic performance in the reaction of glycerol to produce pyridines.¹⁵ In addition, modification of HZSM-5 catalyst with unique combination of metals is another option to improve the catalyst system with more acceptable channel structure and architecture. Zhang and co-workers prepared Cu/HZSM-5 for catalytic conversion of glycerol to produce pyridines. They observed that Cu/HZSM-5 increased the carbon yield of pyridines to 42.8%, and the lifetime of Cu/HZSM-5 could last about 16 hours.¹⁹ Li and co-workers screened Ga/MFI and Fe/MFI for catalyzing the reaction of formaldehyde, acetaldehyde and ammonia to produce pyridine and 3-picoline, and found that Ga/MFI could increase the pyridine selectivity. Zhang and co-workers reported that ZnO/HZSM-5 and SO₄²⁻/ZrO₂–Fe–HZSM-5 were effective catalysts to produce pyridines.¹⁵ Besides, Co–HZSM-5, Pb–ZSM-5, Co/Pb–HZSM-5 were also screened to catalytically synthesize pyridines from formaldehyde, acetaldehyde and ammonia.²⁰

Despite some of the advances mentioned above, the particle size of ZSM-5 should also be considered in the pyridine production process. The particle size is another important property of ZSM-5, which can facilitate the adsorption–desorption and diffusion behavior of molecules. The reaction over ZSM-5 is limited by the diffusion rate of reactant molecules within zeolite crystals rather than by intrinsic reaction rates. Recently, some studies have revealed the added advantages of nano-sized zeolites. Compared with the conventional zeolites, the nano-sized zeolites have a large external surface and a better accessibility to the internal micropores.^21–23 These properties may make the nano-sized zeolites possess higher activities in some catalysis and adsorption applications. Firoozi and co-workers compared the catalytic performance of micro-sized and nano-sized HZSM-5 in the methanol to propylene reaction, in which the nano-sized HZSM-5 showed higher activity and stability and also had a higher selectivity to propylene compared with micro-sized HZSM-5.²⁴ Zheng and co-workers investigated the effect of particle size of ZSM-5 on the catalytic fast pyrolysis of biomass for producing aromatics, and found small-sized HZSM-5 (200 nm) exhibited better catalytic performance than micro-sized HZSM-5.²⁵ Ni and co-workers synthesized nano-sized H[Zn, Al]ZSM-5 zeolite to catalyze the reaction of methanol to aromatics and improved the BTX yield and catalytic stability.²⁶ Besides, Viswanadham and co-workers synthesized nano-sized HZSM-5 to catalyze the reaction of ethanol to gasoline, and the nano-sized HZSM-5 showed higher gasoline yields.²⁷ Rownaghi and co-workers investigated the influence of crystal size and acid sites of nano- and micro-sized HZSM-5 on catalytic cracking of hexane for production of light olefins, and found nano-sized HZSM-5 was more attractive than micro-sized HZSM-5.²⁸ Although many studies on nano-sized zeolites for catalytic production of propylene, aromatics and gasoline have been reported, there is no study on nano-sized zeolites for catalytic production of pyridines from glycerol or acyclic molecules.

In this study, nano-sized HZSM-5 with different Si/Al ratios were synthesized and used to catalyze the production of pyridines from glycerol and indoles from bio-derived furans via TCC-A of glycerol.

Experimental

Materials

Ethanol, glycerol, benzene, toluene, xylene, NH₄NO₃, and NaOH were purchased from Sinopharm Chemical Reagent Co. Ltd. Pyridine, 2-methylpyridine, 3-methylpyridine, 4-methylpyridine, tetrapropylammonium hydroxide (TPAOH), NaAlO₂ and bi-cyclohexane were purchased from Aladdin Chemical Reagent Co. Ltd. Fumed silica was obtained from Qingdao Haiyang Chemical Co. Ltd. NH₃ (≥99.995), N₂ (99.999%), Ar (99.999%), He (99.999%) and standard gases (H₂, CO, CH₄, C₂H₄, C₃H₆) were purchased from Nanjing Special Gases Factory. All these chemicals and gases were used as received without further treatment.

Catalyst preparation

The nano-sized HZSM-5 catalysts were prepared following the synthesis procedure reported in a previous study.²⁸ The molar ratio of SiO₂ : Al₂O₃ was in the range of 25 to 200. Fumed SiO₂ was used as the silica source. NaAlO₂ was used as the Al source. The preparation procedure was as follows: an appropriate amount of NaOH solution was added to fumed silica under stirring at room temperature. Afterward, the mixture solutions of NaAlO₂ and TPAOH were individually added to the solution with vigorous stirring. The gel was stirred for 1 h and then transferred into a Teflon-lined steel autoclave, and crystallized under autogenous pressure and static condition at 175 °C for 48 h. After crystallization, the product was filtered, rinsed with deionized water, dried, and calcined to remove the template. The resultant nano-sized Na–ZSM-5 samples were ion-exchanged with a 1 M solution of NH₄NO₃ four times to obtain nano-sized NH₄–ZSM-5 samples. After drying, the nano-sized NH₄–ZSM-5 samples were calcined in air at 550 °C for 6 h to afford the nano-sized HZSM-5 samples.

To compare the catalyst performances, conventional HZSM-5 (Si/Al = 25) was purchased from the Catalyst Plant of Nankai University. The particle size of the catalysts was about 40 meshes.

Catalyst characterizations

The Si/Al ratio of zeolites was measured by XRF (Shimadzu Corporation, Japan). The data were analyzed using the semi-quantitative program UniQuant. XRD patterns were obtained with a Rigaku D/MAX-2500 diffractometer in a 2θ range of 5–80° (with a 2θ step of 0.02°) using Cu Ka radiation (k = 1.5406 Å). The N₂ adsorption/desorption isotherms of the catalysts were measured at −196 °C using a Coulter SA 3100 analyzer. The surface area of the catalyst was calculated by the Brunauer–Emmett–Teller (BET) method. The surface morphologies and particle sizes of catalysts were investigated by scanning electron microscopy (SEM) with a Sirion 200 instrument. To prepare the samples for SEM observation, a drop of dilute colloidal solution of nano-sized HZSM-5 was placed onto the SEM sample stud surface and then dried at room temperature. Shortly before an SEM image was obtained, the samples were coated with gold. For the NH₃-TPD tests, about 200 mg of sample was put in a reactor and pre-treated in situ for 1 h at 500 °C in a flow of argon. After cooling to 90 °C, ammonia adsorption was performed by feeding pulses of ammonia to the reactor using a flow of dry argon of 60 mL min⁻¹. After the catalyst surface became saturated, the sample was kept at 90 °C for 2 h to remove the base excess. Ammonia was thermally desorbed by increasing the temperature with a linear heating rate of approximately 8 °C min⁻¹ from 90 °C to 700 °C. The amount of NH₃ desorbed was measured by a gas chromatograph (GC-SP6890, Shandong Lu-nan Ruihong Chemical Instrument Co. Ltd, Tengzhou, China) with a thermal conductivity detector (TCD).

Apparatus for thermo-catalytic conversion of glycerol over zeolites to pyridines

The apparatus in this study was a bench-top continuous flow reactor, the same as the apparatus reported in previous studies.¹¹ It consists of a quartz tube reactor heated by a furnace and a condensation tube bathed in liquid nitrogen. The catalyst bed supported by quartz wool was built up in the heating zone of the reactor. The liquid feed was fed into the reactor with a peristaltic pump under a certain flow rate and purged with ammonia. Volatile products were trapped in the condensation tube cooled by liquid N₂. The apparatus for TCC-A of furan, pyrrole, 2-methylfuran and furfural over nano-sized HZSM-5 for producing indoles was the same as that for TCC-A of glycerol over nano-HZSM-5.

Product analysis

The liquid produced was analyzed by GC-MS (Thermo Trace GC Ultra with an ISQ i mass spectrometer) equipped with a TR-35 MS capillary column (30 m × 0.25 mm × 0.25 mm), which was also the same as in the previous study.¹¹ The GC heating ramp was: (1) hold at 40 °C for 3 min, (2) heat to 180 °C at 5 °C min⁻¹, (3) heat to 280 °C at 10 °C min⁻¹, and (4) hold at 280 °C for 5 min. Split injection was performed at a split ratio of 50 using helium as carrier gas.

The total amount of liquid products was determined by the weight difference of the condensation tube before and after the experiment. The coke at the end of the run was measured after the reaction by weighing the solid and deducting the weight of the catalyst. The carbon yield of coke was further determined by elemental analysis. The glycerol and major liquid products were quantified by gas chromatography (GC 1690, Kexiao, China) employing a 30 m × 0.25 mm × 0.25 μm fused-silica capillary column (HP-Innowax, Agilent). The liquid sample was mixed with bi-cyclohexane as the internal standard and diluted by ethanol. The GC operating conditions were as follows: carrier gas: nitrogen; injection port: 250 °C in a split mode; flame ionization detector (FID): 250 °C; column temperature: 40 °C; oven temperature program: heating up to 250 °C at a rate of 10 °C min⁻¹, and holding at a final temperature for 5.0 min. The gas products were analyzed by collecting the gas in gas bags. The gas bags were weighed before and after reaction and their contents were analyzed using gas chromatography (GC-SP6890, Shandong Lu-nan Ruihong Chemical Instrument Co. Ltd, Tengzhou, China) with two detectors, a TCD for analysis of H₂, CO, CH₄, and CO₂ separated on a TDX-01 column, and a FID for gas hydrocarbons (C₂H₄, C₃H₆) separated on a Porapak Q column. The moles of gas products were determined by the normalization method with standard gas. Various equations to calculate the conversion of glycerol, the carbon yield of coke, gases, pyridines and aromatics, and the selectivity of pyridines and gases are given below.

Glycerol conversion (%) = moles of glycerol reacted/moles of glycerol fed × 100% (1)

Coke yield (C%) = moles of carbon in solid residue/moles of carbon in glycerol feed × 100% (2)

Gases yield (C%) = moles of carbon in gases/moles of carbon in glycerol feed × 100% (3)

Pyridines yield (C%) = moles of carbon in pyridines/moles of carbon in glycerol feed × 100% (4)

Aromatics yield (C%) = moles of carbon in aromatics/moles of carbon in glycerol feed × 100% (5)

Pyridines selectivity (%) = moles of carbon in specific pyridine/total moles of carbon in all pyridines identified × 100% (6)

Gas selectivity (%) = moles of carbon in specific gas/total moles of carbon in gases identified × 100% (7)

Results and discussion

Catalyst characterization

Five kinds of nano-sized HZSM-5 with different Si/Al ratios (25, 50, 100, 135 and 165) and micro-sized HZSM-5-25 were chosen in this study. The XRD patterns of micro-sized HZSM-5 and nano-sized HZSM-5 samples (Fig. 1) exhibited a doublet at 2θ = 5–10° along with a triplet at 2θ = 5–10° representing the ZSM-5 framework structure as reported before. The high intensities of all peaks and low background lines indicated good crystallinity of these catalysts. The SEM images of micro-sized HZSM-5 and nano-sized HZSM-5-25 (Fig. 2) exhibited the particle size, morphology and aggregation of the ZSM-5 catalysts. The mean particle size of micro-sized HZSM-5-25 and nano-sized HZSM-5-25 was about 2 μm and 300 nm, respectively. The morphology of nano-sized HZSM-5 was more regular than that of micro-sized HZSM-5.

Fig. 1 XRD patterns of micro-sized HZSM-5-25 and nano-sized HZSM-5 with different Si/Al ratios (25, 50, 100, 135, and 165).

Fig. 2 SEM images of micro-sized HZSM-5-25 and nano-sized HZSM-5-25: (a) SEM image of micro-sized HZSM-25; (b) SEM image of nano-sized HZSM-5-25.

The typical properties of the catalysts, such as real Si/Al ratio, BET surface area, pore volume, and total acid amounts, are listed in Table 1. The BET surface area of nano-sized HZSM-5-25 was 339 m² g⁻¹, which was slightly less than that of micro-sized HZSM-5-25 (370 m² g⁻¹). However, the external surface area of nano-sized HZSM-5 (110 m² g⁻¹) was much more than that of micro-sized HZSM-5-25 (84 m² g⁻¹). With the Si/Al ratio increasing, the BET surface area and external surface area increased from 339 m² g⁻¹ and 110 m² g⁻¹ to 383 m² g⁻¹ and 123 m² g⁻¹, respectively. The micro volume of nano-sized HZSM-5 (0.109 cm³ g⁻¹) was also more than that of micro-sized HZSM-5 (0.104 cm³ g⁻¹). The total acid amounts of nano-sized HZSM-5 with different ratios measured by NH₃-TPD are also shown in Table 1. The total acid amount of nano-sized HZSM-5-25 was about 595.3 μmol g⁻¹, which was similar to that of the micro-sized HZSM-5-25 (580.6 μmol g⁻¹). The NH₃-TPD patterns of the nano-sized HZSM-5-25 and micro-sized HZSM-5-25 samples (Fig. 3a) reveal the comparable acidity patterns of these two samples which differ in particle size but have similar Si/Al ratio. Thus, the acid amounts and acidity did not change a lot with the particle size of HZSM-5 decreasing from micro level to nano level. In addition, the total acid amounts of nano-sized HZSM-5 with different Si/Al ratios were also measured by NH₃-TPD (Fig. 3b). With the Si/Al ratio increasing from 25 to 165, the total acid amounts of nano-sized HZSM-5 decreased from 595.3 μmol g⁻¹ to 78.6 μmol g⁻¹.

Table 1 Typical properties of the catalysts

Catalyst	Si/Al^a	BET surface area (m^2 g^−1)	External surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Micro volume (cm^3 g^−1)	Total acid (μmol g^−1)
a Si/Al ratio: the ratio of SiO2 to Al2O3.
Micro-HZSM-5-25	25.0	370	84	0.200	0.104	580.6
Nano-HZSM-5-25	26.8	339	110	0.189	0.109	595.3
Nano-HZSM-5-50	53.0	333	108	0.188	0.111	315.5
Nano-HZSM-5-100	98.9	350	116	0.199	0.115	192.9
Nano-HZSM-5-135	135.4	360	123	0.203	0.116	102.7
Nano-HZSM-5-165	164.0	383	120	0.209	0.119	78.6

---

Fig. 3 (a) NH₃-TPD patterns of micro-sized HZSM-5-25 and nano-sized HZSM-5-25. (b) NH₃-TPD patterns of the nano-sized HZSM-5-25 with different Si/Al ratios.

Comparison study on the catalytic performance of micro-sized HZSM-5 and nano-sized HZSM-5 in pyridine production

Based on our previous study, HZSM-5 with Si/Al ratio of 25 was the optimum catalyst for pyridine production in the TCC-A process.¹¹ As shown in Entry 1 and Entry 2 in Table 2, we firstly tested the catalytic performance of micro-sized HZSM-5-25 and nano-sized HZSM-5-25 in pyridine production. The experiments were carried out at 500 °C with weight hourly space velocity (WHSV; glycerol to catalyst) and NH₃ to glycerol molar ratio fixed at 1 h⁻¹ and 8 : 1, respectively. The main products from thermo-catalytic conversion of glycerol were coke, gases, pyridines and a small amount of aromatics. The detected gases were CO, CH₄, C₂H₄ and C₃H₆. When the glycerol conversion was catalyzed by micro-sized HZSM-5-25, the overall carbon yield of coke, gases, pyridines and aromatics was 13.2%, 47.5%, 26.0% and 0.9%, respectively. When the nano-sized HZSM-5 was used as the catalyst, the carbon yield of coke and gases decreased to 11.8% and 43.2%, respectively. The carbon yield of pyridines increased by 30% and reached 34.1%. The selectivity of pyridines catalyzed by nano-sized HZSM-5-25 was similar to that of micro-sized HZSM-5. The nano-sized HZSM-5 showed better catalytic performance in pyridine production in the TCC-A process. Since the Si/Al ratio and the total acid amounts of micro-sized HZSM-5-25 and nano-sized HZSM-5-25 were similar, the better catalytic performance of nano-sized HZSM-5-25 could result from the smaller particle size. Therefore, the pyridine production in the TCC-A process was a function of the particle size of HZSM-5 catalyst. Although the catalytic performance of nano-sized HZSM-5 was better than that of micro-sized HZSM-5, the carbon yield of pyridines under this condition (34.1%) was still less than the carbon yield under the optimized condition (35.6%) catalyzed by micro-sized HZSM-5 reported in our previous study.¹¹ Thus, more systematic experiments should be done to investigate the effect of parameters on the TCC-A of glycerol for producing pyridines over nano-sized HZSM-5 to improve the carbon yield of pyridines.

Table 2 Thermo-catalytic conversion of glycerol with ammonia over nano-sized HZSM-5^a

a Reaction conditions: reaction temperature, 500 °C; catalyst: 1 g; WHSV, 1 h^−1; NH3 to glycerol molar ratio, 8 : 1; all glycerol conversion was 100%.b Carbon yield: carbon mol%.
Entry	1	2	3	4	5	6
Catalyst	Micro-HZSM-5-25	Nano-HZSM-5-25	Nano-HZSM-5-50	Nano-HZSM-5-100	Nano-HZSM-5-135	Nano-HZSM-5-165
Actual Si/Al ratio	25.0	26.8	53.0	98.9	135.4	164.0
Overall carbon selectivity (C%)^b
Coke	13.2	11.8	12.7	13.2	13.0	14.1
Gases	47.5	43.2	45.3	47.6	50.7	51.5
Pyridines	26.0	34.1	32.3	28.5	26.3	22.1
Aromatics	0.9	1.4	1.7	1.5	1.8	2.2
Pyridine selectivity (%)
Pyridine	67.3	63.4	63.3	63.8	64.0	61.5
2-Methylpyridine	10.0	10.7	9.1	8.3	8.0	8.7
3-Methylpyridine	21.4	21.7	24.8	25.3	24.3	26.9
4-Methylpyridine	1.3	4.2	2.8	2.6	3.6	2.9
Other alkylpyridines	<1	<1	<1	<1	<1	<1

---

Effect of acidity of nano-sized HZSM-5 on pyridine production

Based on our previous studies on producing N-heterocycles via the TCC-A process, acidity is an important factor which could influence the product distribution.^13,14,31,32 HZSM-5-based catalysts are the protonated form of ZSM-5 and have higher acid amounts and better shape selectivity. Higher Si/Al ratio results in lower acid amounts. To further investigate the effect of acidity of nano-sized HZSM-5 on pyridine production, nano-sized HZSM-5 with five different Si/Al ratios from 25 to 165 were tested in this study. As shown in Table 2 (Entry 2 to Entry 6) and Fig. 4, with the Si/Al ratio increasing from 25 to 165, the product distribution changed a lot. The carbon yield of pyridines decreased from 34.1% to 22.1%. Conversely, the carbon yield of coke and gases increased from 11.8% and 43.2% to 14.1% and 51.5%, respectively.

Fig. 4 Effect of Si/Al ratio on glycerol conversion with ammonia over nano-sized HZSM-5 catalyst: (a) overall yield; (b) pyridine selectivity.

As shown in Fig. 4b, the pyridine selectivity did not change a lot with the Si/Al ratio of HZSM-5 increasing from 25 to 165. The selectivity of pyridine, 2-methylpyridine, 3-methylpyridine and 4-methylpyridine was about 63%, 10%, 24%, and 3%, respectively. Therefore, the acidity of nano-sized HZSM-5 could affect the carbon yield of pyridines, gases, and cokes, but could not affect the pyridine selectivity. Therefore, nano-sized HZSM-5 (Si/Al = 25) was selected as the optimal catalyst in the following study.

Effect of reaction temperature on pyridine production

The effect of temperature on the glycerol conversion with ammonia was investigated in the temperature range of 400 to 600 °C over nano-sized HZSM-5 catalyst. Fig. 5 shows the overall carbon yield of coke, gases, pyridines, and aromatics (Fig. 5a); the pyridine selectivity (Fig. 5b); and gas selectivity (Fig. 5c). As shown in Fig. 5a, the conversion of glycerol and the carbon yield of coke, gases, pyridines, and aromatics were affected by the reaction temperature. The conversion of glycerol increased and reached 100% at 500 °C. In contrast, the yield of coke decreased from 20.1% to 7.1% with temperature increasing from 400 to 600 °C. Meanwhile, the yield of gases and aromatics increased from 27.3% and 0% to 49.2% and 17.8%, respectively. In addition, the carbon yield of pyridines increased with increasing temperature, and the highest yield of pyridines (38.7%) was obtained at 550 °C. If the temperature further increased to 600 °C, the yield of pyridines decreased to 19.5%. Meanwhile, the yield of aromatics reached 17.8%. These results indicated that a temperature of around 550 °C is favorable for pyridine production.

Fig. 5 Effect of reaction temperature on glycerol conversion with ammonia over nano-sized HZSM-5 catalyst: (a) overall yield; (b) pyridine selectivity; (c) gas selectivity.

Fig. 5b shows the effect of reaction temperature on pyridine selectivity. The pyridine selectivity was influenced by reaction temperature. Four kinds of pyridines, pyridine, 2-methylpyridine, 3-methylpyridine and 4-methyloyridine, were detected in the pyridine products. With temperature increasing from 400 to 600 °C, the selectivity of pyridine and 4-methylpyridine increased from 65.1% and 3.0% to 74.2% and 4.8%, respectively. In contrast, the selectivity of 3-methylpyridine decreased from 24.2% to 11.5%. The selectivity of 2-methylpyridine showed no regularity in the range from 400 to 600 °C.

Fig. 5c shows the effect of temperature on the gas product distribution. The main products detected in the gas phase were CO, CH₄, C₂H₄ and C₃H₆. With the reaction temperature increasing, the gas selectivity changed dramatically. The selectivity of CO decreased from 90.2% to 60.1%. Conversely, CH₄, C₂H₄, and C₃H₆ selectivity increased from 4.7%, 3.8% and 1.3% to 14.8%, 19.9%, and 6.1% with temperature increasing from 400 to 600 °C. At higher temperature, the side reactions such as cracking, decarbonylation, and oligomerization become more apparent. Thus, with the temperature increasing, more olefins and hydrocarbons were produced.

Effect of molar ratio of ammonia to glycerol on pyridine production

The effect of the molar ratio of ammonia to glycerol on the TCC-A of glycerol to pyridines over nano-sized HZSM-5 was also investigated by fixing the reaction temperature at 550 °C, the WHSV of glycerol to catalyst at 1 h⁻¹ and the mass of catalyst at 1 g, while changing the flow rate of ammonia. The molar ratio of ammonia to glycerol was in the range from 4 : 1 to 20 : 1. In the TCC-A process, the ammonia served as both the reactant and the carrier gas. Therefore, a change of the molar ratio of ammonia to glycerol would lead to a change of the residence time of glycerol over the catalyst. Table 3 summarizes the detailed product distributions of TCC-A of glycerol at different molar ratios of ammonia to glycerol. Accordingly, the conversion of glycerol and the product distributions were a function of ammonia to glycerol molar ratio. With the molar ratio of ammonia to glycerol rising from 4 : 1 to 16 : 1, the conversion of glycerol remained at 100%. While, when the molar ratio reached 20 : 1, the conversion of glycerol decreased to 94.6%. At the same time, the coke yield decreased from 10.3% to 6.8%, while the gas yield increased from 41.6% to 52.1%. The carbon yield of aromatics was very low and remained at about 2%. When the molar ratio of ammonia to glycerol was at 12 : 1, the yield of pyridines reached 42.1%. During this process, a higher molar ratio of ammonia to glycerol means a shorter residence time, which could cause an insufficient reaction from glycerol to pyridines, and thus the conversion of glycerol decreased. In contrast, a lower molar ratio of ammonia to glycerol corresponded to a longer residence time, which could cause a longer reaction time of feedstock stream over the catalyst, and led to the further reaction of pyridines, and thus more coke and gases but less pyridines were produced. In addition, with the ammonia to glycerol ratio increasing from 4 : 1 to 20 : 1, the detected pyridines were still pyridine, 2-methylpyridine, 3-methylpyridine, and 4-methylpyridine. The selectivities of pyridine, 2-methylpyridine, 3-methylpyridine and 4-methylpyridine were about 68%, 9%, 18% and 4%, respectively. Therefore, the pyridine selectivity was not a function of ammonia to glycerol ratio.

Table 3 Effect of ammonia to glycerol molar ratio on pyridine production^a

a Reaction conditions: reaction temperature, 550 °C, nano-sized HZSM-5, Si/Al = 25, 1 g; WHSV, 1 h^−1.b Mole conversion: mol%.c Carbon yield: carbon mol%.
Entry	1	2	3	4	5
Ammonia to glycerol molar ratio	4 : 1	8 : 1	12 : 1	16 : 1	20 : 1
Glycerol conversion (%)^b	100	100	100	100	94.6
Overall carbon selectivity (C%)^c
Coke	10.3	7.9	7.4	7.1	6.8
Gases	41.6	44.3	45.2	47.7	52.1
Pyridines	33.9	38.7	42.1	38.4	35.2
Aromatics	3.8	3.2	1.7	2.2	1.8
Pyridine selectivity (%)
Pyridine	71.0	68.3	66.4	67.3	66.3
2-Methylpyridine	8.3	9.0	10.2	9.6	10.2
3-Methylpyridine	17.2	17.8	19.6	18.9	18.3
4-Methylpyridine	3.7	4.9	3.8	4.2	5.2
Other alkylpyridines	<1	<1	<1	<1	<1

---

Production of indoles over nano-sized HZSM-5 catalyst via TCC-A process

To investigate the particle size effect of the catalyst, nano-sized HZSM-5 was also used in the production of indoles from furan, pyrrole, 2-methylfuran and furfural via the TCC-A process. The reaction conditions were the optimized conditions reported in our previous studies.^29,30 When furan, pyrrole, and 2-methylfuran served as the substrates (Entries 1 to 3 in Table 4), the reaction conditions were a reaction temperature of 500 °C, WHSV of 0.5 h⁻¹, and molar ratio of ammonia to furan of 8 : 1. When furfural served as the substrate (Entry 4 in Table 4), the reaction conditions were a reaction temperature of 650 °C, WHSV of 1 h⁻¹, and molar ratio of ammonia to furan of 8 : 1. Table 4 shows the carbon yield of coke, gases, aromatics, and N-containing chemicals and the overall carbon yield of N-containing chemicals from different substrates. When furan served as the feedstock in the TCC-A process over the nano-sized HZSM-5 catalyst, the carbon yield of coke and gases was 17.1% and 16.3%, respectively, which was less than those over normal HZSM-5 (21.0% and 23.4%). In contrast, the carbon yield of N-containing chemicals and indoles was up to 49.9% and 34.4%, which was more than those over normal HZSM-5 (45.7% and 31.7%). When pyrrole, 2-methylfuran and furfural served as the feedstocks, more N-containing chemicals and indoles, and less coke and gases were produced. These results indicated that the catalytic performance of nano-sized HZSM-5 was also better than that of normal micro-sized HZSM-5 for producing indoles via the TCC-A process.

Table 4 The detailed product distributions from TCC-A of different bio-derived furans

a Reaction conditions: WHSV was 0.5 h^−1, the molar ratio of ammonia to furan was 8 : 1.b Reaction conditions: WHSV was 1 h^−1, the molar ratio of ammonia to furan was 8 : 1.c 5-Methylindole, 6-methylindole, 7-methylindole, 2,3-dimethylindole, 2,5-dimethylindole, etc.
Entry	1	2	3	4
Feedstocks	Furan^a	Pyrrole^a	2-Methylfuran^a	Furfural^b
Temperature (°C)	500	500	500	650
Conversion (%)	100	86.4	100	100
Carbon yield of products (C%)
Coke	17.1	16.3	18.7	9.5
Gases	23.5	12.4	23.6	38.9
N-Containing chemicals	49.9	44.1	42.2	37.8
Aromatics	1.9	2.8	2.4	6.3
Overall carbon yield of N-containing chemicals (C%)
Pyridines	2.1	7.2	2.4	7.2
Pyrroles	11.1	—	10.6	4.1
Anilines	2.3	5.3	1.7	3.0
Indoles^c	34.4	31.6	27.5	22.9

---

Catalyst regeneration

To further study the stability of nano-sized HZSM-5 during the TCC-A process of glycerol to pyridines, five reaction–regeneration cycles of the catalyst were conducted at 550 °C. After the reaction, coke was formed on the surface of the catalyst. The spent catalyst was regenerated in an air stream at 550 °C for 3 h to remove the coke. As shown in Fig. 6, the catalyst did not deactivate significantly. After using the catalyst five times, the yield of pyridines decreased from 42.1% to 36.3%, this yield being 30.3% for the micro-sized HZSM-5 as reported in ref. 11. The carbon yield of coke decreased from 7.4% to 5.5%. Meanwhile, the yield of gas increased from 45.2% to 48.9%. The yield of aromatics remained at around 3%. Generally, nano-sized HZSM-5 (Si/Al = 25) is more stable and has better catalytic performance than micro-sized HZSM-5 for the production of pyridines in the TCC-A of glycerol.

Fig. 6 The effect of catalyst recycling on pyridine yields. (Reaction conditions: reaction temperature, 550 °C; catalyst, nano-sized HZSM-5, Si/Al = 25; catalyst usage, 1 g; WHSV, 1 h⁻¹; NH₃ to glycerol molar ratio, 12 : 1; each run, 1 h.)

Conclusions

This study demonstrated that smaller particle size of the catalyst benefits the TCC-A process for heterocycle production. Nano-sized HZSM-5 with different Si/Al ratios were synthesized, and employed for producing N-heterocycles. Nano-sized HZSM-5 (Si/Al = 25) showed better catalytic performance than micro-sized HZSM-5 for catalytic production of pyridines from glycerol, and indoles from bio-derived furans via the TCC-A process due to its favorable particle size and acidity. Temperature, WHSV and ammonia to glycerol ratio displayed a significant effect on the product distribution. The optimal conditions for producing pyridines were achieved with nano-sized HZSM-5 (Si/Al = 25) at 550 °C with WHSV of glycerol to catalyst of 1 h⁻¹ and ammonia to glycerol ratio of 12 : 1. The highest yield of pyridines was up to 42.1%, which was higher than that when using micro-sized HZSM-5 (35.6%). In addition, five reaction/regeneration cycles demonstrated that the catalytic activity of nano-sized HZSM-5 slightly decreased compared with the first run, but was higher than that of micro-sized HZSM-5. All in all, it is necessary to make more effort to further develop more efficient and stable catalysts for heterocycle production via the TCC-A process.

Conflict of interest

The authors declare no competing financial interests.

Acknowledgements

The authors are grateful to the National Natural Science Foundation of China (21572213, 21325208), the National Basic Research Program of China (2013CB228103), Anhui Provincial Natural Science Foundation (1408085MKL04), and the Fundamental Research Funds for the Central Universities for financial support.

References

1. G. D. Henry, De novo synthesis of substituted pyridines, Tetrahedron, 2004, 60, 6043–6061.
2. M. Movassaghi, M. D. Hill and O. K. Ahmad, Direct synthesis of pyridine derivatives, J. Am. Chem. Soc., 2007, 129, 10096–10097.
3. K. S. K. Reddy, I. Sreedhar and K. V. Raghavan, Interrelationship of process parameters in vapor phase pyridine synthesis, Appl. Catal., A, 2008, 339, 15–20.
4. K. S. K. Reddy, C. Srinivasakannan and K. V. Raghavan, Catalytic vapor phase pyridine synthesis: a process review, Catal. Surv. Asia, 2012, 16, 28–35.
5. S. Shimizu, N. Abe, A. Iguchi, M. Dohba, H. Sato and K. I. Hirose, Synthesis of pyridine bases on zeolite catalyst, Microporous Mesoporous Mater., 1998, 21, 447–451.
6. S. Shimizu, N. Abe, A. Iguchi and H. Sato, Synthesis of pyridine bases: general methods and recent advances in gas phase synthesis over ZSM-5 zeolite, Catal. Surv. Asia, 1998, 2, 71–76.
7. Y. Higashio and T. Shoji, Heterocyclic compounds such as pyrrole, pyridines, pyrrolidine, piperidine, indole, imidazol and pyrazines, Appl. Catal., A, 2004, 260, 251–259.
8. J. R. Calvin, R. D. Davis and C. H. McAteer, Mechanistic investigation of the catalyzed vapor-phase formation of pyridine and quinoline bases using ¹³CH₂O, ¹³CH₃OH, and deuterium-labeled aldehydes, Appl. Catal., A, 2005, 285, 1–23.
9. F. Jin, Y. Cui and Y. Li, Effect of alkaline and atom-planting treatment on the catalytic performance of ZSM-5 catalyst in pyridine and picolines synthesis, Appl. Catal., A, 2008, 350, 71–78.
10. D. E. Webster and I. M. Rouse, Production of formaldehyde, US 4208353 A, 1997.
11. L. Xu, Z. Han, Q. Yao, J. Deng, Y. Zhang, Y. Fu and Q. Guo, Towards the sustainable production of pyridines via thermo-catalytic conversion of glycerol with ammonia over zeolite catalysts, Green Chem., 2015, 17, 2426–2435.
12. L. Xu, Q. Yao, Z. Han, Y. Zhang and Y. Fu, Producing pyridines via thermocatalytic conversion and ammonization of waste polylactic acid over zeolites, ACS Sustainable Chem. Eng., 2016, 4, 1115–1122.
13. C. Luo, C. Huang, A. Li, W. Yi, X. Feng, Z. Xu and Z. Chao, Influence of Reaction parameters on the catalytic performance of alkaline-treated zeolites in the novel synthesis of pyridine bases from glycerol and ammonia, Ind. Eng. Chem. Res., 2016, 55, 893–911.
14. C. Luo, A. Li, J. An, X. Feng, X. Zhang, D. Feng, Z. Xu and Z. Chao, The synthesis of pyridine and 3-picoline from gas-phase acrolein diethyl acetal with ammonia over ZnO/HZSM-5, Chem. Eng. J., 2015, 273, 7–18.
15. X. Zhang, C. Luo, C. Huang, B. Chen, D. Huang, J. Pan and Z. Chao, Synthesis of 3-picoline from acrolein and ammonia through a liquid-phase reaction pathway using SO₄²⁻/ZrO₂–FeZSM-5 as catalyst, Chem. Eng. J., 2014, 253, 544–553.
16. X. Zhang, Z. Wu, W. Liu and Z. Chao, Preparation of pyridine and 3-picoline from acrolein and ammonia with HF/MgZSM-5 catalyst, Catal. Commun., 2016, 80, 10–14.
17. F. Jin, Y. Tian and Y. Li, Effect of alkaline treatment on the catalytic performance of ZSM-5 catalyst in pyridine and picolines synthesis, Ind. Eng. Chem. Res., 2009, 48, 1873–1879.
18. F. Jin, G. Wu and Y. Li, Effect of sublimation treatment on the catalytic performance of MFI catalysts in pyridine and 3-picoline synthesis, Chem. Eng. Technol., 2011, 34, 1660–1666.
19. Y. Zhang, X. Yan, B. Niu and J. Zhao, A study on the conversion of glycerol to pyridine bases over Cu/HZSM-5 catalysts, Green Chem., 2016, 18, 3139–3151.
20. Y. Liu, H. Yang, F. Jin, Y. Zhang and Y. Li, Synthesis of pyridine and picolines over Co-modified HZSM-5 catalyst, Chem. Eng. J., 2008, 136, 282–287.
21. C. Hsu, A. S. T. Chiang, R. Selvin and R. W. Thompson, Rapid synthesis of MFI zeolite nanocrystals, J. Phys. Chem. B, 2005, 109, 18804–18814.
22. X. Mu, D. Wang, Y. Wang, M. Lin, S. Cheng and X. Shu, Nanosized molecular sieves as petroleum refining and petrochemical catalysts, Chin. J. Catal., 2013, 34, 69–79.
23. M. Choi, K. Na, J. Kim, Y. Sakamoto, O. Terasaki and R. Ryoo, Stable single-unit-cell nanosheets of zeolite MFI as active and long-lived catalysts, Nature, 2009, 461, 246–249.
24. M. Firoozi, M. Baghalha and M. Asadi, The effect of micro and nano particle sizes of H–ZSM-5 on the selectivity of MTP reaction, Catal. Commun., 2009, 10, 1582–1585.
25. A. Zheng, Z. Zhao, S. Chang, Z. Huang, H. Wu, X. Wang, F. He and H. Li, Effect of crystal size of ZSM-5 on the aromatic yield and selectivity from catalytic fast pyrolysis of biomass, J. Mol. Catal. A: Chem., 2014, 383–394, 23–30.
26. Y. Ni, A. Sun, X. Wu, G. Hai, J. Hu, T. Li and G. Li, The preparation of nano-sized H[Zn, Al]ZSM-5 zeolite and its application in the aromatization of methanol, Microporous Mesoporous Mater., 2011, 143, 435–442.
27. N. Viswanadham, S. K. Saxena, J. Kumar, P. Sreenivasulu and D. Nandan, Catalytic performance of nano crystalline H–ZSM-5 in ethanol to gasoline (ETG) reaction, Fuel, 2012, 95, 298–304.
28. A. A. Rownaghi, F. Rezaei and J. Hedlund, Selective formation of light olefin by n-hexane cracking over HZSM-5: influence of crystal size and acid sites of nano- and micrometer-sized crystals, Chem. Eng. J., 2012, 191, 528–533.
29. L. Xu, Y. Jiang, Q. Yao, Z. Han, Y. Zhang, Y. Fu, Q. Guo and G. W. Huber, Direct production of indoles via thermo-catalytic conversion of bio-derived furans with ammonia over zeolites, Green Chem., 2015, 17, 1281–1290.
30. Q. Yao, L. Xu, Z. Han and Y. Zhang, Production of indoles via thermo-catalytic conversion and ammonization of bio-derived furfural, Chem. Eng. J., 2015, 280, 74–81.
31. X. Zhu, S. Liu, Y. Song and L. Xu, Catalytic cracking of C4 alkenes to propene and ethene: influences of zeolites pore structures and Si/Al₂ ratios, Appl. Catal., A, 2005, 288, 134–142.
32. S. Liu, Z. Zhang, M. Jia, X. Gao and J. Yu, ZSM-5 zeolites with different SiO₂/Al₂O₃ ratios as fluid catalytic cracking catalyst additives for residue cracking, Chin. J. Catal., 2015, 36, 806–812.

---