Hydrodeoxygenation of guaiacol as model compound for pyrolysis oil on transition metal phosphide hydroprocessing catalysts

Abstract

The gas phase hydrodeoxygenation (HDO) of guaiacol, as a model compound for pyrolysis oil, was tested on a series of novel hydroprocessing catalysts – transition metal phosphides which included Ni₂P/SiO₂, Fe₂P/SiO₂, MoP/SiO₂, Co₂P/SiO₂ and WP/SiO₂. The turnover frequency based on active sites titrated by the chemisorption of CO followed the order: Ni₂P > Co₂P > Fe₂P, WP, MoP. The major products from hydrodeoxygenation of guaiacol for the most active phosphides were benzene and phenol, with a small amount of methoxybenzene formed. Kinetic studies revealed the formation of reaction intermediates such as catechol and cresol at short contact times. A commercial catalyst 5% Pd/Al₂O₃ was more active than the metal phosphides at lower contact time but produced only catechol. A commercial CoMoS/Al₂O₃ deactivated quickly and showed little activity for the HDO of guaiacol at these conditions. Thus, transition metal phosphides are promising materials for catalytic HDO of biofuels.

Introduction

Pyrolysis oil from thermal cracking of biomass is attracting attention as an alternative liquid fuel because of the depletion of petroleum deposits and the increasing environmental concern with the burning of nonrenewable resources [1]. However, oxygen removal is required to upgrade pyrolysis oil because its high oxygen content (20–40%) leads to undesirable properties of the oil such as low energy density, and thermal and chemical instability [2]. The subjects of hydrodeoxygenation [3] and pyrolysis oil treatment [4] have been reviewed. The average composition of pyrolysis oils is 50–65 wt% organic components, that include organic acids, aldehydes, ketones, furans, phenolic compounds, guaiacols, syringols and sugar based compounds, 15–30 wt% water and 20 wt% colloidal lignin fraction [5], [6]. The phenolics content, a major part of the lignin fraction may reach 30% of the organic component [6], [7], [8]. Guaiacol is one of the most abundant of the lignin-derived products in biooil, present at levels of approximately 0.18–0.51 wt% in switchgrass and alfalfa derived pyrolysis oils [9].

Generally, there are two methods for removal of oxygen. In the direct deoxygenation method, which is generally conducted at atmospheric pressure, C–O bonds are broken without the assistance of a reducing gas such as hydrogen [10], and in the hydrogenation route aromatic rings are hydrogenated before removal of oxygen [11]. The former has been reported on tungsten (IV) compounds [12] and acidic zeolites such as HZSM-5 [13]. The latter process is carried out at high pressure and temperature, and is related to hydrotreating of petroleum feedstocks for removal of sulfur and nitrogen. Thus, this hydrodeoxygenation (HDO) process can potentially allow use of the existing petroleum refining infrastructure for processing and transportation [14]. Conventional sulfide catalysts for petroleum hydroprocessing [15] and precious metal catalysts [16] have been studied for their reactivity in guaiacol and model oxygenate compound HDO. Oxygenated groups in pyrolysis oil such as ketones, aldehydes, and organic acids require lower temperatures for elimination of the reactive functionalities but guaiacol type molecules and other phenolic species require higher temperature [17]. Phenyl–oxygen bonds are cleaved at 500–650 K using hydropressing catalysts under hydrogen pressure, in which the oxygen is ultimately removed as water. However, typical hydrodesulfurization catalysts, such as NiMoS/Al₂O₃ and CoMoS/Al₂O₃ were found to quickly deactivate by coke deposition in model HDO reactions because of the acidity of the reactant [18]. Sulfides on neutral supports including carbon, silica and alumina modified by K for HDO reactions have been reported [17], [18], [19], [20]. Yet the effect of modification for alumina supported catalysts has not been optimized. It was found that guaiacol coking reactions were negligible with molybdenum sulfide supported on activated carbon, and the catalyst showed good stability and potential for catalytic hydrotreating system [21]. Noble metal Pt as an active component was added to a conventional CoMoS catalyst and showed no significant improvement [17]. According to this study, monometallic and bimetallic noble metal catalysts supported on zirconia have lower coke formation than a CoMoS/Al₂O₃ catalyst and in particular Rh-containing catalysts demonstrate potential in biofuel upgrading. Given the propensity of biofuels to thermally decompose to form coke and coke-like matter, catalysts with supports that are less active for coke formation or more hydrogenating catalysts permitting a rapid transformation of dioxygenated reactants (guaiacol, catechol) into less coke-forming products (phenol) would be highly desirable.

Hydrodeoxygenation of actual biooils has been studied with two stages because of the thermal instability of the oil [22]. In a first stage, a stabilization process was carried out at low temperature to eliminate reactive compositions like ketones. In a second stage, deoxygenation of the phenolic-type molecules was carried at higher temperatures. Using diluted model oxygenated compound solutions for hydrodeoxygenation study will give more precise chemical information and avoid thermal polymerization reactions [23]. In the present study, guaiacol (methoxyphenol) is chosen as model compound for hydrodeoxygenation because guaiacol and substituted guaiacols constitute a relatively high concentration of the lignin-derived fraction (up to 0.5 wt%) and these have a high tendency to coke. Guaiacols contains two different oxygenated functions (phenolic and methoxy groups), so are challenging molecules to completely deoxygenate. Transition metal phosphides supported on neutral silica are a promising class of new hydroprocessing catalysts [24], [25], and it was of interest to investigate them for guaiacol catalytic hydrodeoxygenation in comparison to commercial catalysts such as CoMoS/Al₂O₃ and 5% Pd/Al₂O₃.