Modelling of Ammonia Decomposition Using Ru- and Co-Based Catalysts in a Packed Bed Reactor

The carbon-free hydrogen economy is hinging on viable and feasible ways to transport hydrogen on-site. Currently, ammonia seems to be the most promising hydrogen carrier and is being explored through multiple routes in the literature; one of which being catalytic ammonia decomposition. Ammonia decomposition (AMD) has come up as an alternative for on-site, high purity hydrogen generation over the conventional reforming methods. Ru-based catalysts are proven to be the most active metal for the decomposition of ammonia, although costly. Recent works have also demonstrated at par conversions from Ba promoted Co-based catalysts against the Ru-based ones, with the former also showing higher stability over long reaction times. In this work, we develop an isothermal kinetic model for these two catalysts and explore the behavior of the decomposition process in a catalytic packed bed reactor in different operation regimes, with an aim of optimizing the reactor to maximize CO_x-free, high pressure hydrogen.

Materials and methods

The experimental data used for comparison and validation for both Ru-based & Co-based catalyst arrives from our previously published works. Experiments were performed in a lab-scale packed bed reactor unit that operates between 250-550 °C and 1-40 bar; for a wide range of operating conditions covering space time, ammonia and hydrogen feed partial pressures. A power law rate expression considering the effect of the proximity to the equilibrium was selected to model the reaction rate, based on the wide-spread applicability of it to the AMD process. A 1-D model was developed based on the following assumptions: (1) steady-state operation, (2) plug flow along the axial direction of the reactor with no radial variation (3) isothermal operation and (4) negligible pressure drops along the axial direction of the reactor. The experimental data was used to find the kinetic parameters: a (ammonia reaction order), b (hydrogen reaction order), k₀ (pre-exponential factor) and Ea (activation energy). A parametric estimation was performed using fminsearch function of Matlab^®, which finds the minimum of unconstrained multivariable function using derivative-free method, setting the sum of the squared residuals between the experimental data and the calculated data as the objective function. Ammonia conversion has been used as the fitting variable. Various Matlab^® subroutines were used to solve the system of differential equations: ode45, ode23 and ode113; with the latter offering the most promising performance. The 95% confidence intervals for all the estimated parameters were found using the nlparci function in Matlab^®. Finally, a 3-D gPROMS model was used to further explore the effect of reactor geometry, operating conditions and catalyst geometry in the reaction performance; by using the fitted kinetic parameters.

Results and discussions

A summary of fitted parameters for both the catalysts is presented below:

The 95% confidence intervals for the parameters were found to be at least three orders of magnitude lower than the parameters values, indicating high confidence in the values found. The resulting rate equation is similar to the Temkin-Pyzhev model that is widely used in the associated literature; with reaction rate being a function of positive ammonia concentration order and negative hydrogen concentration order respectively. The simulated and measured ammonia conversions from packed bed reactors for both Ru- and Co-based catalysts are in excellent agreement with each other, highlighting the validity of the model. It is predicted that the noble-metal catalyst reaches equilibrium conversion around 425 °C; which is at least 50 °C lower than the novel, cheaper Co-alternative. We further implement these parameters in the 3-D gPROMS model and find the optimal operating conditions for a catalytic packed bed reactor for both the catalysts.

Conclusions

The production of high purity H₂ from ammonia decomposition in packed bed reactor, using Ru-based and Co-based catalysts was studied theoretically and experimentally. A reliable kinetic model was implemented and validated with experimental data. The excellent agreement between modeling and experiments for both the catalysts corroborates the goodness of the developed kinetic and flow model in this work. An adequate range of operating conditions was selected to obtain high performance for both of the catalysts. Relevant information about the effect of the reactor and pellet geometries was gather for a potential scale up of the reactor.

Acknowledgements

The authors gratefully acknowledge the financial support provided by Saudi Aramco, and the resources and facilities provided by the King Abdullah University of Science and Technology.