Selective catalytic deoxygenation of palm oil to produce green diesel over Ni catalysts supported on ZrO₂ and CeO₂–ZrO₂: Experimental and process simulation modelling studies

Abstract

The selective deoxygenation of palm oil to produce green diesel has been investigated over Ni catalysts supported on ZrO₂ (Ni/Zr) and CeO₂–ZrO₂ (Ni/CeZr) supports. The modification of the support with CeO₂ acted to improve the Ni dispersion and oxygen lability of the catalyst, while reducing the overall surface acidity. The Ni/CeZr catalyst exhibited higher triglyceride (TG) conversion and yield for the desirable C₁₅–C₁₈ hydrocarbons, as well as improved stability compared to the unmodified Ni/Zr catalyst, with TG conversion and C₁₅–C₁₈ yield remaining above 85% and 80% respectively during 20 h of continuous operation at 300 ^oC. The high C₁₇ yields also revealed the dominance of the deCO_x (decarbonylation/decarboxylation) pathway. A fully comprehensive process simulation model has been developed to validate the experimental findings in this study, and a very good validation with the experimental data has been demonstrated. The model was then further utilised to investigate the effects of temperature, H₂ partial pressure, H₂/oil feed ratio and LHSV. The model predicted that maximum triglyceride conversion was attainable at reaction conditions of 300 °C temperature, 30 bar H₂ partial pressure, H₂/oil of 1000 cm³/cm³ feed ratio and 1.2 h⁻¹ LHSV.

Introduction

The continuous increase in today's energy demands and the use of fossil fuels for energy production leads to a rapid increase in the earth's temperature, which should be restricted to avoid a runaway greenhouse effect and an irreversible damage in our planet's climate [1]. This has resulted in an increased interest in alternative sources of energy in recent years [2] amongst which, biomass, such as animal fats, waste cooking oil and microalgal oils, is one of the most promising alternatives to produce carbon-neutral fuels [[3], [4], [5]]. This type of biomass is almost exclusively composed of triglycerides (TGs), mainly tri-esters of glycerol with fatty acids. The transformation of TGs into fatty acid methyl esters (FAME) is known as transesterification and leads to the production of biodiesel [6,7]. However, biodiesel comprises of oxygen-containing fatty acid ester molecules, which have a noticeably different chemical identity than conventional petroleum diesel, while the process also leads to the accumulation of significant amounts of glycerol, which is the main by-product [6,8,9]. Therefore, many attempts have been undertaken recently to develop alternative upgrading routes, such as cracking/hydrocracking and the deoxygenation of natural TGs for green diesel production [6,10,11]. In contrast to biodiesel, green diesel consists of diesel boiling range hydrocarbons (C₁₅–C₁₈), which are free of aromatic compounds and, as a result, it is relatively identical to petroleum diesel, but has a narrower boiling point range [6,12]. It also has a higher heating value, a higher energy density and an exceptionally high cetane value (80–90 due to the exclusive paraffinic content) [12,13].

Noble metals (mainly Pd) supported on high surface area supports, as well as conventional sulphided CoMo, NiMo and NiW catalysts have been proven to be very active and selective for the “stand-alone” triglyceride selective deoxygenation process, in which natural TGs and the petroleum fractions are hydrotreated separately [6,14,15]. However, the noble metal catalysts have limited availability and elevated costs, while they may also be sensitive to contaminants that are contained in the feedstock, like oxygenated compounds [10,14]. In recent years, increased interest has been expressed in cheaper, transition-metal catalysts, especially Ni-based ones [14,16,17]. To this end, various catalyst formulations have been employed, including Ni supported on mineral palygorskite [18], Fe-promoted Ni with high loading that also maintains a high metal dispersion [19], Pd-promoted Ni/SiO₂–Al₂O₃ [20], Ni/MgO–Al₂O₃ with various Ni loadings [21] and Ni supported on ZSM-22 as a bifunctional catalyst [22].

Alumina (Al₂O₃) and zirconia (ZrO₂) are popular support materials for many catalytic reactions, due to their high specific surface area (which enhances metal dispersion) and other properties such as thermal stability under the conditions of the reaction [[23]], [24], [25]]. In our previous work [12], we investigated the green diesel production of palm oil via deoxygenation over Ni catalysts that were supported on ZrO₂, SiO₂ and γ-Αl₂O₃, using a fixed bed continuous flow reactor. Ni/ZrO₂ was found to exhibit a higher low-temperature activity and improved stability compared to the Ni catalysts that were supported on Al₂O₃ and SiO₂. Lercher and co-workers have also published interesting works on fatty acid hydrodeoxygenation over Ni/ZrO₂ catalysts [[26], [27], [28]]. First, they showed that Ni nanoparticles and the adjacent ZrO₂ support can act synergistically to catalyze the conversion of microalgae oils into diesel-range alkanes [26]. Subsequently, they studied the mechanism of palmitic acid reduction to n-pentadecane and concluded that the rate determining step can also be catalyzed via Ni and ZrO₂, with the carboxylic acid group adsorbing at an oxygen vacancy of ZrO₂, followed by α-Η abstraction and O elimination to yield ketene, which is then hydrogenated and decarbonylated over metallic Ni [27]. Lastly, they showed that Ni/m-ZrO₂ is quite more active than Ni/t-ZrO₂ for the conversion of stearic acid to n-heptadecane, which was attributed to a higher active oxygen defect concentration on the former catalyst [28]. Regarding other works reporting on the performance of ZrO₂-supported catalysts, Ni et al. [29] reported on the ability to tune the electron density of metallic Ni in Ni/ZrO₂ by varying the synthesis procedure and they found that Ni supported on t-ZrO₂ with an approximate excess oxygen deficiency and electron density could catalyze the deCO_x pathway to selectively produce diesel-range C₁₇ hydrocarbons (n-heptadecane) via the hydrodeoxygenation of stearic acid.

More recently, doping of the ZrO₂ support with CeO₂ has also been proven to be a reliable pathway to create substitutional defects, thereby generating oxygen vacancies and modifying the defect chemistry and electron density of the resulting Ni-based catalyst [[30], [31], [32]]. For example, Yakovlev et al. [33] performed a thorough catalyst screening of Ni catalysts for the deoxygenation of aliphatic and aromatic oxy-organics. They found that bimetallic NiCu catalysts are more active than the monometallic Ni ones, whereas the utilization of CeO₂ and ZrO₂ can enhance the catalytic activity via the additional activation of oxy-compounds on the support surface.

Most of the studies investigating the catalytic deoxygenation of bio-oil have so far been performed solely on an experimental basis without using theoretical modelling. However, performing numerical studies using software is beneficial, as it provides an understanding of parameter optimisation to produce renewable fuels, such as hydrogen from methanol steam reforming and formic acid decomposition [[34], [35], [36]], as well as bio-oil [37,38] via catalytic deoxygenation. Hafeez et al. [39] investigated, for the first time, the palm oil catalytic selective deoxygenation for the production of green diesel over a Ni/ZrO₂ catalyst using computational fluid dynamics (CFD). The results showed that the increase in temperature caused an increase in bio-oil conversion, which reached a maximum value of over 95% at the reaction temperature of 300 °C. Temperatures higher than 300 °C were accompanied by a loss of conversion, which was attributed to catalyst deactivation. Furthermore, CFD modelling methodologies were employed to correlate the catalyst deactivation with the reaction temperature. Plazas-González et al. [40] developed a model for the hydrotreatment of palm oil components for green diesel production and the reactions were modelled employing Aspen Plus. In order to predict the reaction behaviour, an equilibrium reactor was used, while the reactor temperature, H₂/oil molar ratio and pressure were studied using their model. The results revealed that the optimal conditions are attained at an average pressure range of 30–60 bar, 20:1 H₂/oil molar ratio and temperature range of 300–400 °C.

The current work investigates the catalytic deoxygenation of palm oil using Ni supported on ZrO₂ and CeO₂–ZrO₂ catalysts in a fixed-bed continuous flow reactor. The calcined, reduced and spent catalysts are characterized via XRD, N₂-physisorption, NH₃-TPD, CO₂-TPD, H₂-TPR, TPO, XPS and TEM to evaluate their physicochemical characteristics and active surface state, as well as the effect of CeO₂ modification of the support to the catalytic properties. A process simulation model is developed to validate the experimental results and a good model validation is observed. This work therefore aims to improve the activity of Ni/ZrO₂ catalysts via the support modification with CeO₂ and apply a theoretical model that is able to predict the influence of operating parameters on the catalytic performance without the need to conduct extensive and time-consuming laboratory experiments.