1. Introduction
Biomass utilization as a source of renewable energy has received great attention due to its CO
2 mitigation and abundant availability [
1]. The biomass of lignocellulosic origin consists mainly of cellulose (20–50%), hemicellulose (15–40%), and lignin (15–35%) in addition to water and some organic and inorganic compounds [
2]. Such biomass can be converted into bioenergy or biofuel using thermochemical processes i.e., gasification and pyrolysis [
3]. Through the pyrolysis process, biomass is converted into tar, char, and pyrolysis gas; tar being the main product of the process with a heating value of up to 20 MJ/kg which makes it an expected source of biofuel [
4]. Unfortunately, due to its high oxygen content, biomass pyrolysis tar (BPT) has many drawbacks when used as biofuel such as instability, high viscosity, high acidity, and lower calorific value [
5].
Thus, similar to the catalytic hydrotreatment process for petroleum oil refining can also be used for BPT upgrading which is known as the catalytic deoxygenation (DO) process [
6]. DO process upgrades tar oil by removing oxygen, decreasing coke generation, increasing the H/C of the produced biofuel, and producing chemicals of high added value [
7]. Literature sources report that BPT contains a high amount of oxygenated compounds such as ketones, phenols, furans, organic acids, ethers, and aldehydes [
8]. Most of the previous studies focused on a single model compound for BPT-DO process analysis, such as furfural [
9], 2-methylfuran [
10], phenol [
11], quaiacol [
12], acetic acid [
13], 2-furyl methyl ketone [
14], and acetone [
15].
Some researchers explained the interaction between oxygenated BPT compounds and proposed reaction routes using a mixture of model compounds due to the complexity of BPT. However, the studies analyze the DO process for a mixture of model compounds using noble and non-noble metal catalysts [
16,
17] or zeolite catalysts [
18,
19]. Among the most recent works, Yu et al. found that the highest conversion in the hydrogenation of a mixture of acetic acid and furfural was 57% for furfural with selectivity to esters and alcohols of 66.5% when using Al
2(SiO
3)
3 as a support with 5% Pd-loaded catalyst. Moreover, they reported that the presence of other oxygenated molecules (such as acetic acid) led to only a small interaction effect [
20]. Wang et al. investigated DO of a mixture of p-crysol and acetic acid in the aqueous phase using carbon support loaded with Ru as an active catalyst, which yielded high selectivity towards methylcyclohexane showing the preference of DO reaction of p-crysol in the presence of acetic acid [
21].
In addition, Chen et al. tested the DO of a mixture composed of phenol, n-butanal, and acetic acid using a mixture of HZSM-5 zeolite and a carbon supported Ru-active catalyst. The yield of butyl acetate and butanol was 98.5% from n-butanal compared to 93.8% for cyclohexyl acetate and cyclohyxanol from phenol [
22]. A mixture of formic acid and phenol DO have also been investigated by other researchers using MCM-41 loaded with different active elements (Ru, Pt, and Pd) and showing that the increase of phenol to formic acid ratio decreases the degree of phenol DO [
23]. Another study reported high esterification activity and 100% DO yield using a mixture of propanoic acid and quaiacol with hierarchical ZSM-5 loaded with Ni-active catalyst [
24].
Moreover, two phenol mixtures (phenol/anisole and M-crysole/anisole) using Pd-loaded carbon and Pt-loaded Al
2O
3 were studied and showed weak interaction of model compounds and adsorption competition on active sites [
25]. In yet another work, ZSM-5 zeolite loaded with Ni
2P active catalyst was used for DO of a mixture of acetic acid and quaiacol, where competition between acetic acid and quaiacol on the active sites led to the inhibition in the quaiacol DO [
26]. In a recent study, interactions between model compounds were investigated for DO of two different mixtures, the first one composed of phenol and acetic acid and the second one composed of water, furfural, 4-ethylguaiacolis, phenol, acetic acid, and acetone as a simulated bio-oil, using HZSM-5 loaded with Ni
2P active catalyst and proposing a reaction network for the process. The authors found that HZSM-5 activity increased when Ni
2P was loaded into the surface of the blank zeolite, thus increasing the conversion of acetic acid and decreasing the conversion of phenol with the increasing temperature due to the adsorption competition on the active sites. Moreover, higher temperatures led to increased production of aromatics and substituted phenols through the alkyl substitution of benzene and phenol, respectively [
13].
Hβ zeolite has a high content of silica, a highly acidic nature, and a unique channel structure of large pores which provide suitable conversion of palm oil into hydrocarbons [
27]. Adding rare earth metals to the structure of zeolite decreased the dealumination, resulting in increased rate of hydrogen transfer and zeolite activity per weight [
28]. Deoxygenation processes usually run in liquid phase under high pressure for maintaining existence of the liquid phase and better sorption of hydrogen into it. In this work, a different approach was selected. Liquid products are still in the form of vapors when leaving the pyrolysis reactor. That is why a deoxygenation in gaseous phase was selected and tested. The present work provides important data regarding the use of BPT to solve the problem of the accumulation of it from the pyrolysis process of wooden chips. To the best of our knowledge, no previous work has been conducted on real biomass pyrolysis tar using Hβ zeolite. Modified Hβ zeolites using salts of rare earth metals will affect the general framework of zeolites, it will form bridges between atoms of rare metals and oxygen atoms, which leads to the strengthening and fixing of these metals on zeolites, thus increasing the stability of the catalyst. In addition to increasing the acidic bronsted sites of the catalyst because the ionic diameter of the rare metals will increase and thus the efficiency of the simultaneous reactions that will be conducted in this research will increase.
Thus, the present work is focused on the investigation of the BPT GDO using Hβ zeolite loaded with different amounts of Ce, La, and Nd active catalysts. The rare earth metal loaded Hβ zeolite catalyst was prepared using insipient wetness impregnation method and characterized before and after regeneration using X-ray diffraction (XRD), X-ray florescence spectroscopy (XRF), Ammonia–Temperature programmed desorption (Ammonia-TPD), Pyridine-Fourier Transform Infra-Red Spectroscopy (Pyr-FTIR), Thermo-Gravimetric and Differential Thermo gravimetric analysis (TGA and DTG), and Brunauer–Emmett–Teller analysis (BET). The effects of reaction conditions (process time, temperature, rare earth metal loading type, and amount) on the biofuel yield are studied. BPT characterization showed a highly complicated composition containing a wide range of oxygenated compounds with the main components including water, furfural, 4-ethylguaiacol, phenol, methylethyl ketone, cyclohexanone, acetic acid, and acetone. The main interaction pathways for acetone and BPT-GDO are also proposed.
3. Materials and Methods
3.1. Materials
Hβ zeolite (Si/Al = 25, particle size < 20 μm), cerium(III) nitrate hexahydrate (99.9%), lanthanum(III) nitrate hexahydrate (99.9%), and neodymium(III) nitrate hexahydrate (99.9%) were purchased from Sigma-Aldrich (Darmstadt, Germany). Acetone (99.8%), methanol (99.9%), ethanol (99.9%), isopropanol (99.9%), furfural (99.9%), 4-ethylguaiacol (99.9%), phenol (99.9%), methylethyl ketone (99.9%), cyclohexanone (99.9%), and acetic acid (99.9%) were procured from Microchem (Pezinok, Slovakia). The chemicals used are analytical reagents. Ultimate analyses of BPT, liquid, and solid products were conducted using a Vario Macro Cube® (Elementar, Langenselbold, Germany) Elemental analyzer. A CHNS module with a combustion tube temperature of 1150 °C and a reduction tube temperature of 850 °C was used. Calorific values of BPT, liquid, and solid products were analyzed using an isoperibolic bomb calorimeter (Fire Testing Technology Limited, East Grinstead, UK). BPT raw material originated from our previous research and the physicochemical analysis of BPT, liquid, and solid products from tar are presented in
Table 6.
3.2. Synthesis of Modified Hβ Zeolite
Hβ zeolite was obtained by calcination of zeolite β with ammonium (Si/Al ratio = 25) for 5 h at 550 °C. Three rare earth elements (Ce, La, and Nd) were loaded in different wt% on the surface of Hβ zeolite by the incipient wetness impregnation method [
60]. A suitable amount of rare earth metal nitrate hex hydrate (as an example 0.3099 g for the preparation of 1% Ce loaded Hβ zeolite) was dissolved in 40 mL of distilled water to fill the pore volume and added to 10 g of Hβ zeolite dropwise until the Hβ zeolite powder became wet and all the 40 mL was added; the wet powder was left for 24 h for neutralization and water maturation. Then, a hot plate was used to remove water vapor from the prepared catalysts and was then kept in an air-dry oven at 110 °C for 16 h to vaporize the rest of the water. Finally, the catalysts were calcined in an air muffle furnace at 550 °C for 5 h. Six catalyst samples were prepared (1 wt% Ce/Hβ zeolite, 5 wt% Ce/Hβ zeolite, 1 wt% All) and calcined at 550 °C for 3 h before being tested in the GDO packed bed reactor.
3.3. Catalyst Characterization
Adsorption–desorption phenomena of N2 in a Surfer equipment (Thermo ScientificTM, Waltham, MA, USA) were used to test the BET surface area and pore volume. In the beginning, the catalyst samples were degassed at 200 °C for 1 h and N2 adsorption and desorption were started. TGA and DTG were conducted using a simultaneous TG/DSC analyzer (Netzsch STA 409 PC Luxx, Selb, Germany) for the loaded Hβ zeolite catalysts by increasing the temperature in the range between 25 °C to 800 °C at a rate of 10 °C every one minute under the nitrogen flow of 30 mL/min. A D5000 diffractometer (Cu anode, and Kα = 1.5406 Å, Siemens / Bruker Munich, Germany) was used for the X-ray diffraction test of the loaded Hβ zeolite samples before and after their regeneration to check the phases and d-spacing values. The angle of the instrument was changed in the range of (2θ = 10–60°) with a scanning rate of 0.03° at an operation voltage of 40 kV and current of 30 mA.
In addition, X-ray fluorescence spectroscopy analysis using an Axiosm AX XRF (PANalytical BV, Netherlands) instrument was used for elemental analysis of the catalyst samples to detect the amount of Ce, La, and Nd in the loaded Hβ zeolite samples before and after their regeneration. The ammonia temperature programmed desorption (Ammonia-TPD) method was used to test the catalyst acidity using a CHEM-BET 3000 instrument (Quantachrome, Boynton Beach, FL, USA). Catalyst sample of 0.1 g was degassed at 450 °C for 1 h in nitrogen atmosphere followed by cooling to 0 °C and passing gas mixture of ammonia and nitrogen of 1 mol% for 1 h. In addition, the catalyst samples were heated to 100 °C to achieve equilibrium and the temperature was then increased at the rate of 10 °C/min until 900 °C were reached and the desorbed ammonia was detected by a thermal conductivity detector (TCD). Pyridine–Fourier Transform Infra-Red Spectroscopy (Pyr-FTIR) was used for the quantitative analysis of Bronsted and Lewis acid sites using a Bruker Tensor II Instrument (Bruker Optics GmbH, Ettlingen, Germany). The resulting spectra were in the range of 4000–400 cm−1 with a resolution of 4 cm−1. First, 10 mg of the sample was taken without treatment with pyridine (as Wo) and after treatment with pyridine as W1. The pyridine absorbed is W1–Wo. Then, divide the weight of pyridine by the weight of the sample used for acid site identification. Second, recollect the amount of pyridine adsorbed by heating the sample (W1) and analyze the amount of pyridine by using chromatography method. Finally, by comparing the intensity of the representative peaks of Lewis and Bronsted acid sites and multiplying them with the amount of pyridine absorbed, the concentration of the acid sites was calculated.
3.4. Biomass Pyrolysis Tar Purification
The biomass type was wooden chips (oak wood) with a diameter 1–1.5 mm pyrolyzed at 550 °C with residence time of approximately 15 min with feedstock moisture content of 5%. BPT used in the present work was extracted from biomass pyrolysis liquid fraction captured in isopropanol from a previous work of our research group (Volatile products were passed through water-cooled condenser where most of the tar condensed. To achieve higher purity of the gas, pyrolysis gas leaving the condenser passed through series of scrubbers filled with isopropanol. After each experiment, isopropanol was removed from tar by vacuum distillation at 50 mbar) [
4] and purified to be used as raw material in the GDO reaction. BPT purification steps are described here. BPT liquid mixture contained considerable amount of isopropanol and a sample of 800 g of this mixture was taken and filtered under vacuum to remove any solids. Isopropanol was removed from the mixture by atmospheric distillation to the final temperature of 82 °C. Mass of the distillation residue was around 260 g in every run and this fraction contained a considerable amount of water which was removed in the following successive steps. First, the distillation residue was mixed with double the amount of dichloromethane and the mixture was left to settle for 3 h forming two phases.
Most BPT was extracted into the dichloromethane layer; dichloromethane was then removed from BPT by atmospheric distillation in the temperature range of 40–50 °C. To minimize the water content in the distillation residue even further, solid sodium sulfate was gradually added into BPT after distillation until it stopped settling on the bottom of the flask and its crystals started to float in BPT. The produced slurry was then filtered under vacuum. Thus, purified BPT with minimum water content was produced. Final mass of the product (purified BPT) was around 60 g in each run (Real tar certainly contains significant amount of water. However, to keep the catalyst testing conditions consistent, solids and water were removed. Thus, pump fouling was prevented and only produced water was captured among liquid products. Waxes and highly viscous compounds were also removed during tar purification, which also led to easier handling of the tar feedstock and lower fouling of the reactor and tested catalysts).
3.5. Gas-Phase Deoxygenation
The GDO process was carried out in a tubular stainless-steel reactor (length = 470 mm, inner diameter = 17 mm) packed with inert packing up to half of the reactor length; middle hot zone of the reactor was packed with cylindrical pellets of unloaded and rare earth metal loaded Hβ zeolite catalyst (length = 8 mm, inner Diameter = 8 mm). Schematic diagram of the GDO process is shown in
Figure 10. For each experimental run, liquid acetone or BPT was fed into the reactor at the rate of 0.31 mL/min using a membrane dosing pump, while H
2 gas was fed from a pressure cylinder at the flow rate of 33 L/h into the reactor filled with 15 g of catalyst pellets, the feed was controlled by a control valve (hydrogen flow was selected according to capability of our lab flow meters. Then acetone feed was calculated according to the Ratio of H
2/Acetone nearly equal to 64/36). The stainless-steel reactor is fixed inside an electrical tube furnace with a controller for the temperature and time (temperature was controlled and measured by the heating furnace during the experiments. However, Ni-based thermal sensor was inserted into open heated-up reactor outside of experiments. Temperature of the hotspot, where catalyst was located differed from furnace’s preset temperature by max. of 5 °C. Temperature dropped towards top and bottom of the reactor. For example, the temperature of inert filling located under the catalyst was approximately 10 °C lower than furnace preset temperature). The outlet of the reactor is connected to a system of three scrubbers filled with 50 mL of methanol; glass beds are fixed in a cryostat system at −10 °C to ensure the absorption of product vapors. In addition, scrubbers’ outlet is connected through a three-way valve for GC-gas sampling to a bubble flowmeter and the flue gas vent. Before starting any experiments, the reactor is inertized using N
2 gas at the flow rate of 25 L/h for 15 min to ensure the removal of air from the system. After each run, the catalyst is regenerated in the reactor at 800 °C for 2 h under airflow of 30 L/h to ensure the removal of char from the catalyst active sites. Each experiment was performed for 3 h in the temperature range of 200–400 °C and a total of six GC-gas samples (one sample taken for each 30 min) were taken for each run using a 40 cm
3 syringe and injected into the GC gas analysis system. Hydrogen pressure was kept at 110 kPa (abs). The whole apparatus was considered as atmospheric. Slight overpressure of inlet hydrogen was selected to overcome pressure loss in the catalyst bed and in the scrubbers. Finally, the accumulated products with methanol solvent in the scrubbers were collected and weighted for final GC-liquid sampling while the solid products from BPT were collected from inside the reactor for later analysis using an elemental analyzer. Catalyst was regenerated after each experiment. However, in some cases, the reactor was cooled down after the experiment without catalyst regeneration so the extent of coke formation could be evaluated. After the coked catalyst was weighed it was inserted back into the reactor and regenerated. Coke formed on the surface of the catalyst was calculated by weighing fresh and used catalyst and the difference is the coke weight. Parameters affecting the GDO of BPT studied included rare earth metal loading, reaction contact time, and reaction temperature. Each experiment was repeated three times and the average value was used.
3.6. Biomass Pyrolysis Tar and Product Analysis
GC analysis for the gas product samples was performed using Agilent GC 7890A9 (Agilent Technologies, Santa Clara, CA, USA) equipped with two columns (J& W W113–4362 260 °C, 60 m × 320 μm × 50 μm connected to a flame ionization detector (FID) with the temperature set to 250 °C, and Agilent PLOTQ + MOLSIEVE 260 °C, 65 m × 530 μm × 50 μm connected to a thermal conductivity detector (TCD) and temperature set to 190 °C) and helium as the carrier gas. Liquid samples collected from the scrubbers were tested using Agilent GC 6890 N (Agilent Technologies, USA) equipped with a PEG column at 90 °C and connected to FID at 240 °C. However, GC analysis was carried out at 3 temperatures (60, 90, and 200 °C). Higher temperature of the GC column would cause several compounds to overlap, especially alcohols. The lowest GC temperature was applied for measuring alcohol aldehyde content, ketones content, middle temperature for measuring acetic acid, toluene, and xylene and the highest temperature for measuring furfural, phenol, and others. Provided calibration was used to calculate content of each selected compound. Weight fractions of the main liquid components identified by GC were calculated by dividing the peak area of the identified component by the total peak area of the total identified peaks including methanol (Calibration lines were created for all monitored compounds (ethanol, acetic acid, phenol) using pure compounds of known weight dissolved in methanol). In addition, BPT and the liquid products were characterized using gas chromatography-mass spectroscopy (GC-MS QP2010S, Shimadzu, Tokyo, Japan) employing column specifications: chromatographic column type HP-1MS, diameter = 0.25 mm, length = 30 m, thickness = 0.25 µm. The oven temperature increased from the initial value of 40 °C for 15 min at the rate of 10 °C/min to 250 °C which was maintained for 10 min. Moreover, elemental analysis was conducted for BPT and the liquid and solid products. Karl Fischer’s titration method was used to determine the water content in BPT and liquid products using a Metrohm 870 Titrino Plus instrument (Metrohm, Herisau, Switzerland). Total mass of the detected organic compounds was subtracted from the total mass of the liquid. To analyze the process results and interactions, reaction conversion (
X), product distribution (
), degree of deoxygenation (
DOD), yield(
), water content (
), and contact time were calculated according to the following equations. Reaction conversion (
X) of acetone or BPT was calculated according to Equation (2):
where
nf is the amount of feed injected acetone or BPT in moles, and
nlp is the amount of liquid product.
Product distribution (
) was calculated according to Equation (3):
where
ni is the amount of product
i in moles, and
ntp is the total amount of products in moles.
Deoxygenation degree (
DOD) was calculated according to Equation (4):
where
molp is the mass of oxygen in liquid products, and
mof is the mass of oxygen in the injected feed.
Yields of liquid, gas, and water products (
) were calculated according to Equation (5):
where
mpi is the mass of product
i, and
mf is the mass of the feed (acetone or BPT).
Water content (
) was calculated according to Equation (6):
where
mw,lp is the mass of water in the liquid product, and
mlp is the mass of liquid product.
Contact time (
) for the GDO reaction was calculated according to Equation (7):
where
Vs is the volume of the void fraction between catalyst particles,
Qfv is the flow rate (mL/s) of vapor feed (acetone or BPT), and
QH2 is the flow rate (mL/s) of hydrogen gas at normal pressure.
4. Conclusions
The addition of rare earth metals into the structure of Hβ zeolite decreased the surface area of the catalysts depending on the diameter of the rare earth metal ion added; a large diameter led to a more significant decrease in surface area and pore volume and an increase in the average pore volume. Although the surface area of modified catalysts decreased, the mesopore size increased significantly as the added rare earth metals cover the external surface area and block micropores. DTG graphs for all catalysts are similar and indicate their high stability at high temperatures; moreover, their structure did not change due to the addition of rare earth metals. The addition of rare earth metals into the structure of Hβ zeolite increases the ratio of Lewis/Bronsted acid sites, i.e., an increase in the Lewis acid sites, and it can be concluded that the GDO reaction is controlled by the Lewis acidity. The regenerated catalysts have the same MFI structure as the fresh ones, which means that the catalysts can be reused after coke removal without a significant change in their activity. The highest activity (DOD of 79.5% and conversion rate of 88.7%) was achieved for the 1 wt% Ce/Hβ zeolite catalyst and the main products were aromatic hydrocarbons and aldehydes in the ratio of 55.8% and 29.5%, respectively. Contact time and temperature had the most significant effect on acetone GDO. A considerable amount of CO and CO2 are produced at 400 °C due to the cracking of carbonyl and carboxyl groups. The composition of BPT was detected using GC-MS characterization and the main components were water: 0.71%, furfural: 5.85%, 4-ethylguaiacol: 2.14%, phenol: 13.63%, methylethyl ketone: 5.34%, cyclohexanone: 3.23%, isopropanol: 4.78%, ethanol: 3.67%, methanol: 3.13%, acetic acid: 41.06%, and acetone: 16.46%. The yields of water, liquid phase, and gas phase were 18.33%, 47.42%, and 34.25%, respectively. The Ce active material inhibited acetic acid self-ketonization at 400 °C. However, the conversion of smaller molecules decreased as the micropores were closed by the Ce active material. Alkyl-substituted phenols and aromatic hydrocarbons achieved the highest conversion of 37.34% and 35.56%, respectively. While acetic acid dehydration led to the production of 41.32% of acetaldehyde in the liquid phase, 26.44% of CO2, 28.52% of CO, 26.31% of CH4, 4.51% of C3H8, and 3.87% of C2H6 were the main gas phase products. Low-temperature GDO favored the esterification reaction. The main interaction pathways for the BPT-GDO with the 1 wt% Ce/Hβ zeolite catalyst at 400 °C and a process time of 3 h have also been proposed.