Hydrogen-Free Deoxygenation of Bio-Oil Model Compounds over Sulfur-Free Polymer Supported Catalysts

Abstract Hydrotreatment of bio-oil oxygen compounds allows the final product to be effectively used as a liquid transportation fuel from biomass. Deoxygenation is considered to be one of the most promising ways for bio-oil upgrading. In the current work, we describe a novel approach for the deoxygenation of bio-oil model compounds (anisole, guaiacol) using supercritical fluids as both the solvent and hydrogen-donors. We estimated the possibility of the use of complex solvent consisting of non-polar n-hexane with low critical points (Tc = 234.5 ºC, Pc = 3.02 MPa) and propanol-2 used as H-donor. The experiments were performed without catalysts and in the presence of noble and transition metals hydrothermally deposited on the polymeric matrix of hypercrosslinked polystyrene (HPS). The experiments showed that the presence of 20 vol. % of propanol-2 in n-hexane results in the highest (up to 99%) conversion of model compounds. When the process was carried out without a catalyst, phenols were found to be a major product yielding up to 95 %. The use of Pd- and Co-containing catalyst yielded 90 % of aromatic compounds (benzene and toluene) while in the presence of Ru and Ni cyclohexane and methylcyclohexane (up to 98 %) were the main products.


Introduction
The fast depletion of fossil fuel resources as well as the high value of green-house-gas emissions leads to the development of novel ways for the production of energy and chemicals [1]. Biomass is considered to be one of the most permissive sources for transportation fuels and chemicals as it contains a variety of compounds and is characterized by a high energy content [2]. In spite of numerous ways to process the biomass, they are limited by the variety and complexity of the biomass composition and structure. Fast pyrolysis is considered to be one of the most promising methods, allowing the lignocellulosic biomass to be effectively converted into the liquid called bio-oil [3][4][5][6]. However, the complex composition of bio-oil makes the fast pyrolysis product inappropriate for direct use as fuel [7]. As far as the bio-oil is presented by the phenolic compounds formed during the depolymerization of lignin, high oxygen content and high acidity lead to such disadvantages as pure storage stability, low heating value, high viscosity, and low lubricity [8][9][10][11]. Because of this, the bio-oil needs to be upgraded through a decrease in the oxygen compound concentration (e.g. through deoxygenation).
There are numerous works devoted to the hydrodeoxygenation of either bio-oil or the bio-oil model compounds. Almost all studies are carried out in a hydrogen atmosphere in the medium of C 10 -C 14 hydrocarbons as a solvent [15,21,[26][27][28][29][30][31][32][33][34][35]. This leads to the formation of phenols and cyclic alcohols as a major product. While using some catalysts (e.g. noble metals supported on the activated carbon or carbon nanotubes), aromatic and cyclic compounds can be produced [15,21,[36][37][38][39][40] Due to the high viscosity of the solvents used as well as the low hydrogen solubility, the process requires harsh conditions (temperature and pressure) to be used. Typically, the hydrodeoxygenation is performed at a temperature of 300-400 ºC and a pressure of 1.0-2.0 MPa.
In the current work, we describe the conversion of model bio-oil compounds (guaiacol and anisole) in a hydrogen-free atmosphere in the medium of supercritical hexane and in the complex (hexane:alcohol) supercritical solvent in the presence of polymer-supported catalysts synthesized by the hydrothermal deposition of noble and transition metals in the polymeric matrix of hypercrosslinked polystyrene.

Deoxygenation procedure
Deoxygenation experiments were performed in a six-cell reactor Parr Series 5000 Multiple Reactor System (Parr Instrument, USA) equipped with a magnetic stirrer. In a typical experiment, 1 g of the bio-oil model compound was dissolved in 30 mL of solvent. The mixture was put into the reactor cell and 0.05 g of the catalyst was added. The reactor was sealed and purged with nitrogen three times in order to remove air. Then the nitrogen pressure was set at 3.0 MPa, and the reactor was heated up to 270 °C. After the reaching of the reaction temperature, the pressure increased up to 6.8-8.5 MPa depending on the solvent composition. The solvent composition varied from 0 to 30 vol. % of propanol-2 in n-hexane. In order to exclude the equilibrium shift, the process was performed for varying times -from 10 to 90 min with a 10 min interval.
The liquid phase was analyzed by GCMS using a gas chromatograph GC-2010 and mass-spectrometer GCMS-QP2010S (SHIMADZU, Japan) equipped with a chromatographic column HP-1MS with 30 m length, 0.25 mm diameter and 0.25 µm film thickness. The column temperature program was set as follows: initial temperature 120 °C was maintained for 5 min then the column was heated up to 250 °C with the rate of 5 °C/min and maintained at 250 °C for 5 min. Helium (volumetric velocity of 20.8 cm 3 /s, the pressure of 253.5 kPa) was used as a gas-carrier. The injector temperature was 280 °C, ion source temperature was 260 °C; interface temperature -280 °C.
The quantitative estimation of the deoxygenation process was performed using the relative reaction rate, substrate conversion and product yield calculated according to the Equations 1-3. The calculations were performed on the basis of the reaction mass-balance (Eq. 4).

Influence of the solvent composition
The influence of the solvent composition on the deoxygenation of the bio-oil model compound was studied without the catalyst. The solvent composition varied from 0 to 30 vol. % of propanol-2 in n-hexane. The results of the experiments are presented in Figures 1 and 2, and in Table  1. While analyzing the kinetic curves it is obvious that the curve slope is similar for all cases which can indicate the close reaction rates as well as the similar conversion mechanism. The calculation of the relative reaction rates at 50% of substrate conversion (according to Equation 2) showed that the values obtained seem to be close to each other. Thus, the solvent composition practically does not affect the rates of bio-oil model compound conversion.
In spite of the solvent, the composition does not practically influence the reaction rate, the product composition strongly depends on the propanol-2 concentration. The analysis of the reaction products showed the preferred formation of the phenolic compounds (phenol and pyrocatechol) indicating the ether bond hydrolysis during the reaction (Fig. 3, 4) [55]. Methanol was also observed among the reaction products confirming the proposed reaction route. A negligible amount of aromatics (mainly benzene) were also obtained. When the propanol-2 concentration was lower than

Influence of the catalysts
The use of the catalysts synthesized by hydrothermal deposition leads to a significant increase in the rate of bio-oil model compounds conversion (see Table 2). The highest conversion rate for both anisole and guaiacol was observed while using Ni-and Pd-containing catalysts. Coand Ru-based catalysts showed a lower deoxygenation rate due to the lower catalytic activity of the metals in the heteroatom removal [33,34]. The analysis of the reaction products showed that the presence of 5%-Pd-HPS and 5%-Co-HPS resulted in the formation of up to 90 % of aromatic compounds (benzene and toluene) indicating the behavior of deoxygenation and transmethylation reactions [55]. It should be noted that while using Pd-and Co-containing catalysts, phenol, and pyrocatechol were observed in the reaction mixture (Fig.  5, 6). These results correlate well with the literature data for palladium and cobalt catalysts [18,26,30,34,36,51]. Meanwhile, the use of 5%-Ru-HPS and 5%-Ni-HPS led to the formation of 99 % of cyclic compounds (cyclohexane and methylcyclohexane). In this case, transmethylation and hydrogenation were the main reaction routs [55].    Moreover, a trace amount of cyclohexanol (up to 0.5 %) was formed in the presence of these catalysts (Fig. 7, 8).
The formation of cycloalkanes can be explained by the high hydrogenating activity of Ru and Ni [26,28,29,32,34,49,50]. Cobalt did not exhibit hydrogenating activity probably due to the lower degree of the reduction of the active metal (see Fig. 1S) [56,57]. Based on the deoxygenation product analysis, the following schemes of anisole and guaiacol conversion in the presence of hydrothermally synthesized catalysts can be proposed (Figures 9 and 10). Thus, the synthesized catalysts can be effectively applied in the bio-oil upgrading in the medium of supercritical complex solvent resulting in over 85 % of oxygen removal for 70 minutes.

Conclusions
The conversion of model bio-oil compounds in a hydrogen-free atmosphere in the medium of supercritical and in a complex supercritical solvent was studied in the presence of polymer-supported catalysts synthesized by the hydrothermal deposition on the polymeric matrix of HPS. The study of the solvent composition's influence on the anisole and guaiacol deoxygenation showed that     the solvent consisting of 80 vol. % of n-hexane and 20 vol. % of propanol-2 was optimally providing up to 98 % of substrate conversion. The bio-oil model compound processing without a catalyst results in the formation of phenolic compounds, while the use of 5%-Pd-HPS or 5%-Co-HPS and 5%-Ru-HPS or 5%-Ni-HPS provides the formation of aromatic and cyclic compounds respectively. The synthesized catalysts result in over 85 % of oxygen removal from bio-oil for 70 minutes. O H 3 C OH OH CH 3 CH 3 Figure 9: Possible scheme of anisole conversion (half arrow -in the non-catalytic process; solid arrows -in the presence of 5%-Pd-HPS and 5%-Co-HPS; dash arrows -in the presence of 5%-Ru-HPS or 5%-Ni-HPS).