Chemical pathways in the partial oxidation and steam reforming of acetic acid over a Rh-Al₂O₃ catalyst

Highlights

• We study the kinetics of HAc activation with O₂ and H₂O on Rh.

• Deep oxidation occurs at 225–300 °C and adsorbed HAc reacts with Rh-O* sites.

• At 300–450 °C, O₂ is displaced from Rh and an O-lean chemistry is observed.

• At intermediate temperatures dehydration favors C-accumulation.

• Above 500 °C, SR is highly active.

Abstract

The catalytic partial oxidation (CPO) and the steam reforming (SR) of acetic acid (HAc) over a 2 wt% Rh/α-Al₂O₃ were investigated in an isothermal annular reactor. The experiments were performed at GHSV values up to 2.0 × 10⁶ l(NTP)/kgcat/h by feeding highly diluted HAc/O₂ and HAc/H₂O mixtures; temperature was varied from 200 to 850 °C. In analogy with the CPO of hydrocarbons, also the CPO of HAc showed a low temperature regime characterized by the oxidation of the acid, while the production of synthesis gas occurred above 450 °C. Deep oxidation was observed below 300 °C. In between 300 and 450 °C, a progressive decline of O₂ conversion was observed (a very unique feature), accompanied by decrease of CO₂ production and formation of CO with traces of H₂. C-accumulation was observed starting from 400 °C. The data suggest that an increasing fraction of the surface sites was active in an oxygen-lean chemistry. This could be explained by assuming that competitive adsorption displaced oxygen in favor of acetic acid.

Dedicated SR tests confirmed that the production of CO and H₂ became significant only above 450–500 °C, while the process was hindered, likely by C-poisoning, at lower temperatures.

Raman measurements confirmed that at the intermediate temperatures (450 °C) the surface was enriched by unsaturated C-species. The TPO profiles after CPO test at 450 °C indicated a major peak at 350 °C, suggesting the presence of C-species with intermediate reactivity or intermediate proximity to the metal particles, in between the highly reactive CH_X fragments on metal sites and the polymeric C-species stored on the support.

Introduction

Nowadays, fossil fuels represent the major energy and carbon source; environmental and political issues, though, keep high the pressure for developing alternative, environmentally-friendly and economically viable energy conversion routes. One of the possible scenarios is represented by the so-called hydrogen-based economy, relying on low C-impact hydrogen production technologies and a massive penetration of fuel cells as highly efficient energy conversion systems [1], [2], [3].

Hydrogen can be produced from a variety of feedstocks [4], [5]. These include fossil resources, such as natural gas and coal, as well as renewable resources, such as biomass and water.

In the last 10–15 years, the catalytic conversion of biomass-derived oxygenates has been investigated for the development of sustainable routes to H₂. Ethanol has been primarily and widely studied as a model biofuel for steam reforming and partial oxidation processes [6], [7], [8], [9]. In parallel, the steam reforming of bio-oils or bio-oil components has also gained increased attention. Pioneering studies on the steam reforming of acetic acid, one major component of the volatiles formed during the gasification/fast pyrolysis of different biomasses, were published in the ‘90s by Wang et al. [10] and Marquevich et al. [11]. Systematic studies were then initiated by several research groups including Takanabe et al. [12], Rioche et al. [13] and Basagiannis and Verykios [14], Kechagiopoulos et al. [15]. In those and several successive studies acetic acid was suggested as a viable feedstock for H₂ production, but coking was identified as a major issue, strongly limiting the catalyst stability. Since then, the literature has been addressing the deepening of reaction mechanisms, effect of nature of the supports, strategies to regenerate the catalyst and there has been a steady increase of the number of papers published on this topic.

The reaction scheme of acetic acid steam reforming on supported metal catalysts has been recognized as highly complex, likely due to the multi-functionality of the molecule (containing both one highly hydrogenated C atom and one highly oxygenated C atom). Seshan and co-workers [16] have recently reviewed the reaction pathways, based on previous results from the group [17], [18], [19], [20], [21], [22], as well as the results reported by Lemonidou et al. [23]. The proposed scheme is also coherent with the identification of reaction routes by and Basagiannis and Verykios [14]. The overall steam reforming process includes both reactions leading to gas phase products (initiated by a carboxylation reaction, making available CH_X fragments that can undergo steam reforming to CO and H₂, but also gasification reactions of several surface hydrocarbon or oxygenate intermediates) and highly undesired reactions (such as dehydration or ketonization) leading to precursors for solid C-species such as adsorbed ketene, ethylene or acetone species, eventually leading to coke build-up through condensation reactions. A list of stoichiometries and heat of reactions has been reported in [14] and extended in [24].

Both transition metals and noble metals have been proven to be active in the steam reforming of acetic acid. Hu and Lu [25] have tested Ni, Co, Fe and Cu supported on alumina and verified that Ni and Co were more active than Fe and Cu based catalysts. The Ni/Al₂O₃ was more stable than Co/Al₂O₃ but all samples showed a certain tendency to solid carbon formation. Matas Güell et al. [21] have tested Ni/ZrO₂ catalysts and observed rapid deactivation, but the performance of the Ni catalysts improved by using La/Zr and Ce/Zr modified supports [16], [21], [22]. Verykios and co-workers [26] analyzed Ni and noble metal catalysts; by comparing the performance with Ni based catalysts, they found that noble metals were less active but more stable in terms of carbon formation. They found that Rh is the most active noble metal when compared with Ru, Pt and Pd. Basagiannis and Verykios [26], [27] also showed that the acidity of the support plays an important role in the process. Al₂O₃ was recognized to catalyze cracking reactions with formation of carbonaceous species on the catalytic surface; however, the doping with basic oxides as MgO, CeO₂ and La₂O₃ showed to have a positive effect by reducing coke formation.

Takanabe and coworkers [12], [17], [18] studied the steam reforming of acetic acid over Pt/ZrO₂. They proposed that the process followed a bi-functional mechanism, being the metal phase responsible for acetic acid molecule activation, forming acetate or acyl intermediates that undergo CC cleavage to generate carbon dioxide, carbon monoxide and methyl groups. The CH₃- species would then react with adsorbed hydrogen to generate methane. The adsorbed hydrogen could also combine with each other and form molecular H₂. Interaction with the support would instead promote the ketonization reaction, that, followed by oligomerization of the acetone, would lead to coking of the support. The formation of carbonaceous species at the metal/support interface was reported as a critical step in the deactivation process. In recent works focusing on the design of stable steam reforming catalysts [21], [22], the same group has shown how the redox properties of the support and the surface oxygen mobility can contrast the formation of interfacial coke deposits, thus significantly increasing the catalyst stable performance.

Lemonidou and co-wokers [23], [28] have studied the reaction process using Rh supported over CeO₂-ZrO₂ and La₂O₃/CeO₂-ZrO₂ mixed oxides which greatly suppressed the coking activity. The reaction pathway was investigated by transient and isotopic experiments and entirely associated with the catalysis on Rh. It was proposed that the conversion of acetic acid on Rh particles proceeded via decarboxylation and steam reforming of the CH_X intermediates with effective production of CO, CO₂ and H₂. Instead, dehydration to ketene was assumed to proceed at the periphery of the Rh particles, leading to the formation of acetone.

The effect of O₂-cofeed has been studied to a much lesser extent. Lemonidou et al. [23] performed autothermal reforming experiments to evaluate the coking tendency of the Rh/La₂O₃/CeO₂-ZrO₂ catalyst under different conditions and found a beneficial effect from the oxygen-enriched system.

In this work, we address a kinetic study of the conversion of acetic acid over Rh-supported catalysts with O₂ and/or steam. Our main background is on synthesis gas production via partial oxidation and steam reforming of C₁C₈ hydrocarbons over Rh/Al₂O₃ catalysts [29], [30], [31], [32], [33], [34]. The CPO-system is a consecutive reaction process consisting of deep oxidation (consuming oxygen and part of the fuels with production of CO₂, H₂O) and steam reforming + WGS/RWGS (consuming the rest of the fuel, H₂O and CO₂, with production of synthesis gas). In an insulated reactor, wherein a balance between heat released by oxidation and heat consumed by reforming realizes, synthesis gas may be obtained auto-thermally at temperature ≥ 700 °C. The use of Rh allows to reach thermodynamic equilibrium within few milliseconds contact time, which prompted in our laboratory the development of short contact time reactors for both the kinetic investigation (the isothermal annular reactor [29]), and the process verification (the auto-thermal reactor with monolith catalyst [35]). In principle, also oxygenates can be reformed auto-thermally, that is without any external enthalpy input [36]. There is thus a general interest in studying the fuel-rich oxidation of acetic acid, as well as the steam reforming. In this study we adopted the same Rh/α-Al₂O₃ catalyst that has been tested in previous studies on the CPO of light and liquid hydrocarbons; on the one side Rh is suitable for the expected resistance to C-formation, on the other side this formulation allows to directly compare the new kinetic evidence with the more stablished behavior of hydrocarbons. It is in fact interesting to understand how the nature of the functional group present in the fuel affects the modes of activation of the molecule and the kinetic dependences.

We herein adopt the same methodology of investigation as developed in our previous studies, consisting of tests aimed at exploring the effects of operating conditions on the whole CPO process in order to identify the prevailing stoichiometries and kinetic dependences, and tests on selected reacting subsystems (e.g. steam reforming tests) in order to better characterize the temperature regions where these are active and their kinetics.