Elsevier

Chemical Engineering Journal

Volume 359, 1 March 2019, Pages 496-504
Chemical Engineering Journal

Pareto-optimal design and assessment of monolithic sponges as catalyst carriers for exothermic reactions

https://doi.org/10.1016/j.cej.2018.11.109Get rights and content

Highlights

  • Multi-objective optimization combined with multiscale reactor model.

  • Pareto-optimal sponge designs for catalysis predicted.

  • Quantification of tradeoffs regarding conflicting design goals.

  • Low pressure loss of solid sponges cannot be maintained at production scale.

Abstract

Monolithic sponges combine low pressure losses and excellent heat transport properties and are consequently considered as promising catalyst carriers for fixed-bed reactors. Insights on how to design porosity and window size of monolithic sponges to resolve conflicting relations between low pressure losses, high thermal conductivities, and high space-time-yields (STY), i.e., a high catalyst inventory, are still unknown, especially at pilot or production scales. This study quantifies the outlined tradeoffs and assesses the potential of monolithic sponges as catalyst carriers compared to conventional packed beds of pellets. A state-of-the-art heterogeneous reactor model was applied in combination with a genetic multi-objective optimization algorithm to predict Pareto-optimal sets of sponge designs (max. STY, min. Δp,ΔTmaxΔTtol). As example, the methanation of CO2 was chosen. The Pareto-optimal set of sponge designs shows that small windows are necessary to obtain high space-time-yields comparable to the ones of conventional packed beds. As a consequence, the expected low pressure loss cannot be achieved. Because of excellent heat transport properties, which are weakly dependent on the throughput, monolithic sponges however allow stable operation under varying gas loads. The results demonstrate that monolithic sponges will probably not replace packed pellet beds of pellets for the steady-state production of chemicals. Instead, they provide a competitive option for small-scale, decentralized production for example within chemical energy storage and CO2 utilization.

Introduction

Monolithic sponges, also called open-cell foams, are considered to be promising catalyst carriers in fixed-bed reactors for exo- and endothermic conversions [1], [2], [3], [4]. Since they provide a continuous solid network and high porosities, monolithic sponges combine excellent heat transport characteristics with low pressure losses.

First studies at the lab scale already demonstrated that monolithic sponges lead to lower hot-spot temperatures and comparable or higher yields than conventional packed beds of pellets or extrudates [5], [6], [7], [8], [9], [10]. Regarding the design of monolithic sponges in terms of material, open porosity (volume of accessible void space per unit sponge volume), and window diameter (see Fig. 1 a) Mülheims and Kraushaar-Czarnetzki [6], Gräf et al. [9], and Fischedick et al. [11], among others, emphasized that a high thermal conductivity of the solid is the main parameter to obtain excellent heat transport properties. More recently, Bracconi et al. [12] used a Voronoi tesselation to artificially create realistic sponge geometries (see also Wehinger et al. [13]) with precisely controlled strut diameter. Based on their finite-volume simulations, they concluded that next to a low porosity, thick and preferably uniform strut diameters are required to allow unhindered conduction in the struts and to fully exploit the thermal conductivity of the material. Further, Donsìet al.[14] and Mülheims and Kraushaar-Czarnetzki [6] speculated that small window diameters are necessary to provide large specific surface area for effective heat and mass transfer between the gas and the solid phase. Patcas et al. [15] compared monolithic sponges, honeycombs, and packed beds of pellets regarding mass transfer and pressure loss. They concluded that sponges provide the best tradeoff between high mass transfer rates and low pressure losses because of their open but tortuous flow path.

Recently, Montenegro Camacho et al. [16] presented the use of monolithic sponges in a pilot reactor for the autothermal production of hydrogen from biogas, indicating that monolithic sponges now move from the lab scale towards the pilot and production scale, and that an efficient coupling of the monolithic sponges to the reactor wall (see [17], [18], [19]) can also be realized at larger scales. In contrast to the lab scale, the pressure loss and the space-time yield become important figures of merit at the production scale as they determine the operational and capital costs. Regarding the design of monolithic sponges, this leads to conflicting design goals (see Fig. 1b). For low pressure losses sponges should exhibit high open porosities and large window diameters. This conflicts with the need for a large specific surface area, or equivalently a high catalyst inventory, and thus small window diameters. On the other hand, high porosities substantially reduce the effective thermal conductivity of monolithic sponges and thus might cause heat transport limitations and severe hot spots. For open porosities higher than 90 %, the specific surface area and thus the catalyst inventory also decreases with the porosity.

Reitzmann et al. [3] conducted early simulations of the production of phthalic anhydride with monolithic sponges at the production scale. They concluded that higher space-time yields are possible with monolithic sponges because the high porosity allowed high throughputs without significantly lower yields.

Philippe et al. [20] simulated Fischer-Tropsch synthesis in monolithic sponges and packed beds at the production scale. They found that under identical operating conditions sponges provide better heat control than packed beds at low flow rates, mainly because of the higher porosity and thusless catalyst material per reactor volume. Further, they outline the possibility of tuning the porosity of monolithic sponges to balance high heat transport rates and high catalyst inventories. Nevertheless, both studies compared only selected sponge designs and, due to the lack of available correlations at that time, applied premature or oversimplified models for heat transport in solid sponges.

To date, design guidelines for solid sponges regarding the open porosity and the window diameter to balance high catalyst inventories, low pressure losses, and high thermal conductivities to facilitate scale-up to the production scale have not been available. Therefore, a state-of-the-art multiscale reactor model for monolithic sponges is used in this study to solve the outlined multi-objective optimization problem (Fig. 1 b). The resulting Pareto-optimal sponge designs are analyzed to quantify the tradeoffs in the design of monolithic sponges. Further, the performance of selected Pareto-optimal sponge designs is further analyzed for the methanation of CO2 and compared to the performance of conventional packed beds under identical conditions. The comparison allows to assess the potential of monolithic sponges as catalyst carrier in fixed-bed reactors.

Section snippets

Reactor models

To generate the Pareto-optimal sponge configurations, a 2-d heterogeneous multiscale model was applied. It has been described in detail and validated in previous studies [21], [22]. In brief, the model considered coated monolithic sponges in a single cooled cylindrical reactor tube to resemble the application of sponges in a multi-tubular fixed-bed reactor.

In contrast to pseudo-homogeneous reactor models, heterogeneous reactor models explicitly consider heat and mass transport in the solid

Convergence of the NSGAII algorithm

As the NSGAII algorithm belongs to the class of genetic algorithms, it requires to set the number of individuals that form the set of Pareto-optimal solutions and the number of generations (iterations) over which the population is evolved. Both parameters should be high enough to populate the complete Pareto-front and to avoid dominated solutions that are not Pareto-optimal. Fig. 2 shows the influence of both the number of individuals (Fig. 2a) and the number of generations (Fig. 2b) on the

Conclusions

Monolithic sponges as catalyst carriers in fixed-bed reactors were found to be a promising option for small-to-medium scale conversions, especially for exothermic reactions. Since heat conduction in the continuous solid phase is the dominant mechanism for removing the reaction heat, they can be operated at lower gas loads than conventional packed beds of pellets. In addition, the dominance of heat conduction in the solid phase allows safe operation of monolithic sponges over a wide range of

Acknowledgements

This work was supported by the German Research Foundation (DFG) within the Research Training Group GRK 1860 ’Micro-, meso-, and macroporous nonmetallic Materials: Fundamentals and Applications’ (MIMENIMA).

Declaration of interest

None.

References (39)

Cited by (21)

  • Recent advances in CO<inf>2</inf> hydrogenation to value-added products — Current challenges and future directions

    2021, Progress in Energy and Combustion Science
    Citation Excerpt :

    For instance, particles with radial distribution of the active material in the porous matrix have been considered frequently. Amongst these, ‘egg-shell’ particles were regularly preferred, since they had active layers close to the catalyst surface, which reduce the heat and mass transport resistances [217,218]. In addition, using such particles led to lower consumption of expensive material in the reactor without reducing the performance.

View all citing articles on Scopus
View full text