Ex situ catalytic biomass pyrolysis using mesoporous Ti-MCM-41

Biomass has attracted considerable attention as energy, economic, and environmental asset, as result of its abundance and range of properties. The use of mesoporous catalysts during fast pyrolysis has been a highly important route to improve efficiency as well adding value to biomass. The addition of titanium to molecular sieves increases the efficiency of the pyrolysis reaction by improving production and selectivity of products of interest. This study aims at analyzing the catalytic pyrolysis products of elephant grass using titanium catalysts prepared at different Si/Ti molar ratios, i.e., 25 and 50. The material was supported on MCM-41 for the catalytic pyrolysis of biomass. The biomass pyrolysis reactions were performed in a micropyrolyzer coupled to a GC/MS analyzer. The Ti-MCM-41 samples were characterized by XRD, BET-specific area, and UV-visible. The distribution of pyrolysis products depended on process parameters such as temperature and catalyst type. The highest yield for hydrocarbon production, such as styrene, benzene, methylbenzene, and naphthalene, was observed at 600 °C using Si/Ti equal to 50.


Introduction
In recent years, the pyrolysis of lignocellulosic biomass has attracted great interest in the global energy demand, due to the applicability of its solid, liquid, and gaseous products as fuels and chemicals (Bridgwater and Peacocke 2000;Bridgwater 1996;Demirbas 2009;Oladeji et al. 2015;Samolada et al. 2000). Among several biomass options, the elephant grass (Pennisetum purpureum Schum) is a perennial tropical grass species that has attracted attention as energy crop for its high productivity and low production cycle (Quesada et al. 2004;Silva and Rocha 2010). It reaches heights of about six meters in short periods of time (Silva et al. 2015) and requires small production areas. Finally, cutting after 180 days provides abundant accumulation of biomass with desirable characteristics for energy purposes (Santos et al. 2014).
Mesoporous MCM-41 can be used to improve the quality of pyrolysis products. Recently, several catalysts such as zeolites, MCM-41 (Braga et al. 2017), and SBA-15 (Adam et al. 2006) have been researched for integrated catalyst pyrolysis. Furthermore, the addition of metals with high redox potentials into MCM-41 produces isomorphic substitutions that increase the catalytic activity in oxidative processes involving organic molecules of commercial interest, such as styrene, benzene, methyl benzene, and naphthalene (Beltrán et al. 2003). Iliopoulou et al. (2007) tested two types of Al-MCM-41 mesoporous sieves as catalysts in the pyrolysis of biomass and reported a significant improvement in the quality of  Strezov and Evans (2008) investigated the thermal conversion of elephant grass to biogas, biooil, and charcoal under two heating rates, i.e., 10 and 50°C/min. Their work showed that the pyrolysis of elephant grass can produce combustible biogas compounds with enough calorific value to provide the internal heat of pyrolysis. Braga et al. (2017) analyzed the catalytic pyrolysis products of elephant grass using WO 3 catalysts supported on rice husk ash and RHA-MCM-41, increasing the production of aromatic compounds. The use of catalysts in biomass conversion processes is an alternative to the direct production of potential chemical compounds. The approach adds value to the process since it increases the yield and quality of the produced bio-oil, in addition to granting selectivity manipulation, responsible for improving the control over the final product characteristics. Moreover, the choice of catalyst has an important effect on the pyrolysis products. Studies on catalytic pyrolysis using molecular sieves modified by transition metals or metal oxides, such as Ti-MCM-41, are limited. Nevertheless, these catalysts are promising materials for the pyrolysis of biomass and the method has gained increasing interest as an important industrial process (Fontes et al. 2014). Therefore, the aim of this study is to evaluate the catalytic performance of Ti-MCM-41 in the pyrolysis of elephant grass using Ti-MCM-41 using Si/Ti ratios of 25 and 50.

Synthesis and characterization of catalytic materials
Ti-MCM-41 samples were obtained by hydrothermal synthesis (Fontes et al. 2014) and characterized by XRD using a Shimadzu XRD-6000 instrument operating at 30 kV, 30 mA, and λ = 0.15406 nm. The diffraction patterns were obtained in the angular range 1 ≤ 2θ ≤ 10°using 0.02 steps. Nitrogen adsorption/desorption isotherms were obtained at 77 K using Micromeritics ASAP 2020 equipment and analyzed using the BET and BJH models to calculate the surface area and porosity, respectively. UV-vis analyses were performed in a Shimadzu UV-2450 infrared spectrophotometer.

Characterization of elephant grass
The elephant grass (Pennisetum purpureum Schum) (EG) originated from Northeastern Brazil was comminuted using a knife mill. The particle size range of 0.074-0.104 mm was selected. The proximate analysis was based on ASTM E871-82 (2006) and E1755-01 (2007) to determine moisture and ash, respectively. Fixed carbon was calculated by difference. An ultimate analysis was also performed in a Series II CHNS/ O Perkin Elmer 2400 analyzer to determine carbon, hydrogen, nitrogen, and oxygen by difference. The high heating value (HHV) was determined by a calorimeter bomb according to ASTM E711-87 (2004) and the bulk density by ASTM E873-82 (2006).

Pyrolysis of biomass
Elemental grass pyrolysis was performed on a CDS 5200 HP-R pyrolyzer (CDS Analytical, Oxford, Pennsylvania) at a temperature of 600°C. The sample was placed in the  center of a small quartz tube, keeping in position, using small glass wool buffers at the ends of the tube, heated by a platinum filament around the tube. The condensed gases resulting from the thermal decomposition of elephant grass pyrolysis were washed through the N 2 flow with a flow of 50 mL min −1 and stored in a trap at 300°C for 4 min. The pyrolysis condensate gases stored in the trap were injected through the 1:50 ratio split injector, analyzed in a gas chromatograph coupled to the mass spectrometer (GC/MS-Shimadzu QP 2010) equipped with SHR5XLB DB-5 chromatography column (30 m× 0.25 mm), with the following temperature programming: initial temperature of the analysis was set at 40°C, remaining at this temperature for 4 m, then at 10°C min −1 to 280°C, remaining at this temperature for 14, 50 min. The catalytic pyrolysis of the elephant grass was carried out following the same procedure as the conventional pyrolysis, and the pyrolytic condensed gases were directed to the catalytic bed, containing a quantity of catalyst of approximately 50 mg, through the opening of a valve that allows the connection of this with the pyrolysis system. The pyrolysis of the elephant grass was carried out in the presence of Ti-MCM-41 mesoporous sieve, with Si/Ti ratios of 25, 50 in the catalytic bed heated to 400 and 600°C.

Results and discussion
Ti-MCM-41 characterization  (100), which is typical of the hexagonal structure of MCM-41 materials.
The MCM-41 material and the sample with the lowest Ti content exhibit, in addition to the sharp (100) peak, other peaks corresponding to (110) and (200) reflections, characteristic of a highly ordered MCM-41 structure. As the content of Ti increases, there is a decrease in the intensity of the first peak, along with an evident broadening of all peaks, which can be attributed to a reduction in the long-range order of the structure. On the other hand, all the samples show large surface areas, of about 1087-521 m 2 /g and total pore volumes of about 0.25-0.59 cm 3 /g, which are typical of mesoporous materials. However, when the Ti content in the material increases, the surface area decreases, which is clearly correlated to the decrease in the structural order observed. Additionally, no peaks at higher angles were observed, indicating the absence of any crystalline phases containing titanium.  The crystallographic parameters obtained from X-ray diffraction are shown in Table 1. The values calculated for the cell parameter (a 0 ) of the modified materials increased with respect to that of the initial sample due to the titanium in corporation into the silica matrix. It has been reported that the increase in titanium content increased a 0 of the mesoporous material (Galacho et al. 2007), probably due to the length of the Ti-O bond (1.80 Å), which is greater than that of the Si-O bond (1.61 Å) (Tuel 1999).
The pure siliceous material and the sample with the lowest Ti content exhibit the sharp (100) reflection peak and other peaks attributed to the (110) and (200) reflections, characteristic of a highly ordered MCM-41 structure. As the content of Ti increases, there is a decrease in the intensity of the first peak, and an evident broadening of all peaks, which can be attributed to a reduction in the long-range order of the structure. On the other hand, all the samples show large surface areas and total pore volumes about 0.65-0.85 cm 3 /g, which are typical of mesoporous materials. As the Ti content in the material increases, the surface area decreases, which is clearly related to the decrease in the structural order observed.
Textural properties obtained by BET are shown in Table 2, along with the average pore diameter and the thickness of the silica wall, obtained by BJH (Brunauer et al. 1938;Barrett et al. 1951). All samples show specific area, total pore volume, and wall thickness typical of mesoporous materials. However, when the Ti content in the material increases the surface area decreases. Figure 2 shows the adsorption-desorption isotherms of MCM-41 and Ti-MCM-41 materials and the respective pore size distributions obtained by the BJH method for samples MCM-41 and Ti-MCM-41 with Si/Ti ratios of 25 and 50, respectively. In all samples that exhibit type IV isotherms with a sharp inflection at relative pressure (P/P 0 ) which is characteristic of MCM-41 materials were obtained according to Brunauer et al. (1938). Discrete hysteresis has been observed in all cases, causing differences in the adsorption and desorption isotherms. Meanwhile, the isotherms of samples Ti-MCM-41 (25), synthesized with the highest Ti contents, show the presence of a pronounced hysteresis loop with a sharp decrease of the desorption branch at p/p 0 0.45-0.50. According to Eimer et al. (2008a, b), increasing the Ti content enhances the pore volume.
The UV-vis spectra of the calcined Ti-MCM-41 samples prepared with different Ti contents are shown in Fig. 3. The titanium-containing MCM-41 samples revealed a band in the 210-230-nm range attributed to distorted tetrahedral Ti in the mesoporous structure, where titanium is found as an isolated species. Titanium is surrounded by silicon, forming oxygen bridges (Ti-O-Si). As the titanium content   (Blasco et al. 1995). The spectra also indicate that there is no clear formation of TiO 2 particles. The absence of a band at 330 nm indicates the absence of anatase in the samples.

Biomass characterization
The results of the ultimate and proximate analyses are shown in Table 3. The ultimate analysis showed that elephant grass consisted of moderately high carbon (41.54%) and oxygen (34.35%), characteristic of lignocellulosic material. These results were similar to those obtained by characterization of elephant grass and pine wood sawdust (Wang et al. 2014).
The fast pyrolysis results of EG are shown in the chromatogram of Fig. 4. The main peaks were identified and listed in Table 4. All compounds are oxygenates of different chemical classes, such as furans, phenols, ketones, alcohol, and carboxylic acid, derived from lignocellulosic biomass. Benzofuran has been identified as the main product (24.7%), in addition to phenol (8.42%), cyclohexanone (5.03%), toluene (4.62%), and levoglucosan (3.53%). These products may have been derived from the decomposition of cellulose as well as from hemicellulose (Stefanidis et al. 2014). According to Braga et al. (2014), the presence of metals in the ashes can cause secondary reactions of vapor cracking, converting levoglucosan (LG), which is a primary product of the decomposition of cellulose lower molecular weight compounds. Na + , K + , and Ca 2+ affect the formation of these compounds. In addition, LG cracking can produce furan and derivatives according to temperature and residence time (Lu et al. 2011).
The catalytic fast pyrolysis converted some oxygenate products into aromatic hydrocarbons, confirming the deoxygenation efficiency of the catalyst (Fig. 5). The influence of Ti was observed in the conversion of oxygenates into important aromatics such as benzene, styrene, and naphthalene. The Ti-MCM-41 (Si/Ti = 50) catalyst presented the highest deoxygenation activity, which could be related to the Ti-O-Si species, as confirmed by UV-vis results (Fig. 3) and the higher specific area when compared to Ti-MCM-41 with Si/Ti of 25 ( Table 2). The catalytic bed temperature also plays an important role in the deoxygenation reactions since at 600°C for all catalysts the aromatic content was increased. The production of styrene at Si/Ti equal to 50 at 600°C must be highlighted, because of its importance as an industrial unsaturated aromatic monomer.

Conclusions
The synthesized materials were characterized by high degree of hexagonal ordering, high specific area and pore size. UVvis analyses revealed that it was possible to identify the presence of titanium in the structure. High added-value chemicals were found among the pyrolysis products, such as styrene, benzene, methyl -benzene and naphthalene. Temperature and different Si/Ti molar ratios supported on MCM-41have important effects on product yields. The catalytic pyrolysis using Si/Ti = 50 showed the highest yield for hydrocarbons compounds. The results presented herein showed that elephant grass is a promising biomass source for fast pyrolysis.