545675 Preparation and Testing of Ru/Al2O3 Catalysts for a Compact-Scale Fixed Bed CO2 Methanation Reactor

Monday, June 3, 2019: 4:48 PM
Republic ABC (Grand Hyatt San Antonio)
Alessandro Porta1, Leonardo Falbo1, Carlo Giorgio Visconti1, Luca Lietti1, Claudia Bassano2 and Paolo Deiana2, (1)Department of Energy, Politecnico di Milano, Milan, Italy, (2)ENEA, Rome, Italy

INTRODUCTION

In order to reduce the CO2 emissions and comply with international agreements, a strong increase in the use of renewable energy and carbon capture technologies is foreseen. Renewables, especially solar and wind power, are intrinsically fluctuating sources and this leads to periods of energy overproduction, during which energy needs to be effectively stored so as not to be lost. Nowadays, batteries are still inadequate to provide long term-storage capacity, and a valid solution may be provided by the chemical storage through the so-called power to gas technology (PtG). This process uses the excess renewable power to fuel an electrolyzer that produces gaseous H2 (and pure oxygen) from the splitting of H2O. To facilitate its storage and transportation, which are still an issue, according to the PtG concept, H2 is then used to make CH4. Methane has the huge advantage to have an already-existing infrastructure for its transport and use [1]. Furthermore, methane production allows the recycling of waste CO2 (coming from capture, biogas rectification, etc.) through the Sabatier reaction:

CO2 + 4 H2 ↔ CH4 + 2 H2O              DH0R= -165 kJ/mol

This reaction is generally carried out over Ni-based catalysts, but from a process intensification standpoint, also the more expensive Ru-based catalysts are of interest. Indeed, it is known that Ru is the more active metal for the Sabatier reaction [2], and this allows to operate the Sabatier reaction at lower temperatures with respect to Ni-based catalysts. This is of crucial importance, since both CO2 conversion and CH4 selectivity of the highly exothermic Sabatier reaction are limited by thermodynamics at high temperatures. The use of Ru-based catalysts paves the way for operations at low temperatures, possibly enabling high per-pass CO2 conversions even at low pressures.

Aiming at the optimization of the CO2 methanation process using a compact scale fixed bed reactor and very active supported Ru-catalysts, it is necessary to gather information about the catalyst selectivity, stability and threshold size so to rule-out the onset of internal mass transfer limitations. For this purpose, we have prepared catalysts using γ-Al2O3 spherical supports having different diameters (100, 800 and 2300 μm, Sasol Puralox) and using two different Ru precursors (Ru(NO)(NO3)3 and RuCl3). The obtained catalysts, prepared via impregnation, have been tested (as such or after crushing and sieving) in a fully-automated rig operating 24/7 in a fixed bed reactor configuration, operating in integral conditions to represent at best the industrial process conditions.

RESULTS AND DISCUSSION

The impregnation of supports bigger than 100 μm with a metal loading of 0.5 wt% of Ru led to catalysts with an eggshell configuration, i.e. all the introduced active phase was located on the outer shell of the catalyst pellets. This result was observed using both metal precursors (Ru chloride and nitrosyl nitrate) and confirmed by SEM/EDX analysis. Using the chloride precursor, an active layer thickness as low 4% of the pellet radius was obtained, while using Ru(NO)(NO3)3 a thicker shell was obtained.

In order to tune the thickness of the impregnated layer, the impregnating solutions containing Ru(NO)(NO3)3 and RuCl3 have been acidified, and linear correlations have been observed between the active layers thickness and the molar amount of introduced acid, at constant concentration of the metal precursor in the impregnating solution. By strongly increasing the acidity of the solution, homogenously impregnated pellets with a metal loading of 0.5 wt% could be eventually obtained (Figure 1).


Figure 1. Effect of the acid/metal ratio on the active layer thickness for a 0.5 wt% Ru/Al2O3 catalyst of 2300 μm prepared from a) Ru(NO)(NO3)3 and b) RuCl3

These results have been explained considering the strong ruthenium precursor-support interaction, which results in the fast adsorption of the Ru ions during the impregnation of the outer layers of the alumina support.

At first, the catalytic performance of the two materials prepared using Ru(NO)(NO3)3 and RuCl3 on the 100 μm powder (and no acid addition) has been evaluated at P = 0 barg, T = 250-430°C, H2/CO2 = 3.9, GHSV = 5 L(STP)/h/gcat. The nitrate-based catalyst reached steady performance within the first hours of time on stream, while the chloride-based sample required more than 150 h to reach a stable performance, due to the slow elimination of Cl- ions from the catalyst surface. For this reason, even though the steady state performance of the two systems after the conditioning transient was comparable, the chloride containing samples have been excluded from the rest of this study. Notably, both samples show an almost complete methane selectivity in all the investigated temperature range, with CO being the most abundant byproduct (always below 2%). Small amounts of C2H6 were also detected below 350°C, due to the C-chain growth reactions.

Then, the impregnated pellets prepared using acid-free or acid-containing solutions of Ru(NO)(NO3)3 and 800 and 2300 μm Al2O3 spheres have been tested (after crushing and sieving) so to evaluate the effect of the acid introduction on the catalytic activity. No differences in CO2 conversion nor in CH4 selectivity were observed, indicating a negligible effect of the acid on the catalyst performance.

Subsequently, homogeneously impregnated pellets have been tested to verify the absence of internal diffusional limitations. No significant differences in the catalytic activity were observed when varying the dimension of the pellets, indicating the absence of intraporous limitations with a Ru loading of 0.5 wt% up to a characteristic length of diffusion (d) of 383 μm (Figure 2a)

In order to push the system to more stringent conditions, we increased the catalytic activity by increasing the Ru loading of one order of magnitude. The metal loading of 5 wt% was obtained with four successive impregnations and in this case no acid addition was required to obtain a uniform Ru distribution along the pellet radius.

At first the performance of the catalysts at 0.5 and 5 wt% loadings supported onto 100 μm powders has been compared. Interestingly, no differences in CO2 conversion nor in CH4 selectivity were observed working under kinetically controlled regime and at constant space velocity per gram of Ru. This indicates that the specific reactivity of the catalyst does not depend on the Ru loading. In line with these results, HR-TEM evidenced that the dispersion of the Ru0 phase does not vary by increasing the metal loading by one order of magnitude (dp,avg = 1 nm).

Subsequently, these experiments were repeated with for the 5 wt% catalysts supported on 800 and 2300 μm Al2O3 spheres (P = 0 barg, T = 250-430 °C, H2/CO2 = 3.9, GHSV = 50 L(STP)/h/gcat). In this case, both conversion and selectivity diverge from the performance of the powdered catalyst already in the case of the 800 μm pellets (d < 130 μm), suggesting that the catalyst performance is affected by internal diffusional limitations (Figure 2b). This evidence was confirmed by means of a kinetic study, which showed that the activation energy evaluated for 2300 μm thick catalysts was half the one evaluated for the powdered catalyst with the same Ru loading. Notably, the selectivity to methane drops when using big pellets (though still remaining above 95%) due to the specular increase in the CO formation, which might be due also to a non-isothermal profile along the catalyst pellet generated by the strongly exothermic Sabatier reaction under conditions dominated by transport limitations.

These results are relevant both in terms of mechanistic understanding of the Sabatier reaction and in view of the intensification of the process. A detailed kinetic study (able to describe not only the CO2 conversion but also the CH4 selectivity) based on these findings as well as on spectroscopic evidences is currently being developed.

 

Figure 2. CO2 conversion as a function of temperature for homogenously impregnated γ-Al2O3 spheres of 100, 800 and 2300 μm. a) 0.5 wt% Ru, b) 5 wt% Ru

CONCLUSION

We show that the strong ruthenium precursor-support interaction leads to the formation of eggshell catalysts at metal loadings as low as 0.5 wt%. However, homogeneously impregnated pellets can be obtained by increasing the acidity of the impregnating solution. The reactivity of pellets prepared using acid-free or acidified solutions is similar. Increasing the metal loading by one order of magnitude prevent the formation of eggshell pellets and does not result in an increase of the metallic particle size. As a result, the CO2 conversion rate per gram of active phase is unaffected by the increase metallic density on the surface. However, in presence of a catalysts with Ru-loadings as high as 5 wt%, intraporous mass/heat transfer limitations strongly affects both the CO2 conversion rate and CH4/CO selectivity. An accurate description of these phenomena is pivotal for the process intensification.

REFERENCES

[1]         C. Janke, M.S. Duyar, M. Hoskins, R.J. Farrauto, Catalytic and adsorption studies for the hydrogenation of CO2 to methane, Appl. Catal. B Environ. 152–153 (2014) 184–191. doi:10.1016/j.apcatb.2014.01.016.

[2]         P. Frontera, A. Macario, M. Ferraro, P. Antonucci, Supported Catalysts for CO2 Methanation: A Review, Catalysts. 7 (2017) 59. doi:10.3390/catal7020059.

 


Extended Abstract: File Not Uploaded
See more of this Session: CO2 Hydrogenation
See more of this Group/Topical: General Submissions