Silica and zirconia supported catalysts for the low-temperature ethanol steam reforming
Graphical abstract
Ni confirmed the best performing active phase to promote ethanol decomposition and reforming already at low reaction temperature (<500 °C). The stabilization of the active phase in very dispersed form allowed to reach stable catalyst performance with time-on-stream. SiO2, thanks to no Lewis acidity and sufficiently strong metal–support interaction, demonstrated an interesting support for Ni under the selected operating conditions.
Introduction
The steam reforming of ethanol raised interest for the production of H2 from a renewable source. Broad efforts have been devoted to develop active and stable catalytic systems for this application. Among non noble metal catalysts Ni and Co exhibited the most interesting performance, suggesting to optimize their formulation to achieve better results [1], [2], [3], [4], [5], [6], [7], [8], [9].
The steam reforming process is composed of many different possible reactions [9], of which we may summarize the most relevant as follows:CH3CH2OH + 3H2O → 2CO2 + 6H2which may be seen as the sum of the syngas production and the water gas shift (WGS) reaction:CH3CH2OH + H2O → 2CO + 4H2CO + H2O ↔ CO2 + H2
The latter is exothermal, in contrast to the former. Therefore, when operating under the typical steam reforming conditions the WGS reaction is often at equilibrium. Its degree of advancement may be deduced from the CO/CO2 ratio.
Ethanol may be dehydrogenated to acetaldehyde, which can be further reformed. However ethanol may also undergo dehydration to form ethylene that can then polymerize to form carbonaceous deposits over the catalyst. Depending on the operating conditions other parasitic reactions induce coke formation, such as CO disproportion (Bouduard reaction), active at moderate temperature, or the decomposition of hydrocarbons, active at high temperature.
Differently prepared Ni-, Co- and Cu-based catalysts supported over TiO2 have been recently tested for the steam reforming of ethanol. The results evidenced that the highest activity may be reached in general with Ni as active phase, but its tendency to form C filaments remains a key problem, besides possible coking due to support acidity [10], [11], [12], [13], [14]. It was also underlined that the interaction strength between the support and the active phase, tunable with the preparation procedure, determines the success of a formulation. For instance, the same 10 wt% Ni/TiO2 sample was completely inactive when the support was calcined at 500 °C, whereas it was conveniently active and stable when calcined at 800 °C [12], [13], [14].
In addition, Ni proved much more stable against coking when prepared in very dispersed form also for the SR of CH4 [15], [16], [17], [18]. The possibility to disperse (and stabilize in dispersed way) the metal depends on the preparation method, but also on the support. ZrO2 and SiO2 demonstrated interesting supports for Ni, provided the right preparation route is chosen [12], [13]. Therefore, in this work we investigated different Ni-, Co- and Cu-based catalysts supported over ZrO2 or SiO2. In particular a mesoporous SBA-15 support was chosen, trying at least the partial confinement of the active phase into the mesoporous framework. On the other hand, the mean pore size of SBA-15 is much larger than the kinetic diameter of ethanol, so that no significant mass transfer limitations are expected during the reaction.
All the samples were calcined at 500 °C in order to keep the active phase as dispersed as possible and we focused on low temperature activity testing in the temperature range 300–500 °C. Indeed, the possibility to operate at relatively low temperature, i.e. at 500 °C or below, is very attractive to limit the energy input to the reactor. Nevertheless, coking activity is particularly high in such a temperature range, especially with Ni-based catalysts.
In previous investigations [12], [13], [14] we managed the coking item by high temperature operation (500–750 °C), so to favour the gasification of the coke deposits in case formed. On the contrary, in the present work we focused on the optimization of metal dispersion and its interaction with the support to improve catalyst stability towards coking. This approach was successfully applied also with Pt-supported samples in this application [19]. Extensive characterization by means of various techniques of both the fresh and spent samples allowed to compare the main physical–chemical properties of the catalysts and to comment activity/stability data.
Section snippets
Support synthesis
The SBA-15 support was synthesized as previously reported [20], [21], in the presence of Pluronic 123 (P123, Aldrich) as structure directing agent and calcined at 500 °C.
Zr(OH)4 was prepared by a conventional precipitation method [21], [22] at a constant pH of 10.
Addition of the active phase
The active phase was added to each support by incipient wetness impregnation with an aqueous solution of the metallic precursor (Ni(NO3)2·6H2O, Sigma Aldrich, purity ≥98.5%; Co(NO3)2·6H2O Sigma–Aldrich, puriss. p.a. ACS reagent; Cu(NO3)2
Textural, structural and morphologic properties
The main physical chemical properties of the prepared catalysts are reported in Table 1.
SBA-15-supported samples showed very high surface area. In particular, they were characterized by a IV-type isotherm with a H1-type hysteresis. This is typical of this support, which is a mesoporous material with a high surface area (≈700 m2/g) and a sharp pores size distribution, with a maximum located at ca. 6 nm. Sample Z–Ni exhibited a IV-type isotherm containing a H3-type hysteresis, typical of materials
Conclusions
A set of catalysts for the ESR has been prepared with variable support (ZrO2 and a mesoporous SiO2) and active phase (Ni, Co, Cu). The present data confirm Ni-based ones as the most promising catalyst for this reaction at 500 °C. Indeed, full ethanol conversion was achieved, with satisfactory C balance, stable performance and limited formation of byproducts. Different reaction paths have been observed (i.e. different byproducts) depending on the active phase, but also on the support. In
Acknowledgements
The authors are indebted with Regione Lombardia and the Consortium for Material Science and Technology (INSTM) for financial support. The valuable help of Dr. Cesare Biffi and Matteo Compagnoni is gratefully acknowledged. The authors are indebted with Prof. Giuseppe Cruciani for XRD analyses. The work was partly supported by H2FC European Infrastructure Project (Integrating European Infrastructure to support science and development of Hydrogen and Fuel Cell Technologies towards European
References (60)
- et al.
Appl. Catal. A
(2011) - et al.
J. Power Sources
(2009) - et al.
Int. J. Hydrogen Energy
(2009) - et al.
Appl. Catal. A
(2009) - et al.
Int. J. Hydrogen Energy
(2009) - et al.
Catal. Today
(2013) - et al.
J. Catal.
(2004) - et al.
Chem. Eng. J.
(2010) - et al.
Int. J. Hydrogen Energy
(2009) - et al.
Appl. Catal. B: Environ.
(2012)