Ammonia Decomposition for Hydrogen production in Catalytic Microchannels with Slip/Jump Effects

Abstract

The rarefaction effects on the catalytic decomposition of NH₃ in ruthenium-coated planar microchannels is numerically simulated in the Knudsen number range 0.015-0.03. A collocated finite-volume method is used to solve the governing equations, A concentration jump model derived from the kinetic theory of gases is employed to account for the concentration discontinuity at the reactive walls. A detailed surface reaction mechanism for ammonia decomposition on ruthenium along with a multi-component species diffusion model are used to study the effects of concentration jump coupled with velocity slip and temperature jump on the walls. The velocity-slip, temperature-jump and concentration-jump boundary conditions have miscellaneous effects on flow, temperature and species concentration fields. The results suggest that the velocity-slip boundary condition only slightly influences the species distribution at the edge of the Knudsen layer as well as inside the channel, while the temperature-jump boundary condition affects the heat and mass transfer characteristics the most. The concentration-jump effect, on the other hand, can counter balance the temperature-jump effects in some cases.

Keywords: Ammonia decomposition; Heterogeneous reactions; Velocity slip; Temperature jump; Concentration jump; Microchannel.

Intruduction

In order to meet the increased power demand for micro-devices in almost every field of engineering, scaling down of conventional power supplies to micro-heat engines, micro fuel cells, micro-turbines and combustors has been proposed as an efficient, safe and reliable energy delivery method for Micro-Electro-Mechanical-Systems (MEMS). Due to high energy density of hydrocarbons compared to Lithium batteries and higher operational cycles, microburners have been studied in recent years as heat and energy sources for portable devices (Ahn et al. 2005, Fernandez-Pello 2002, Maruta 2011, Miesse and Masel 2004 and Yin et al. 2004). The energy is either utilized by thermoelectrics for electric power generation or through endothermic reactions such as fossil fuel steam reforming or ammonia decomposition for hydrogen production for fuel cells. The push towards reducing emission levels from hydrocarbon combustion has resulted in an interest in hydrogen production to power fuel cells. To avoid anode catalyst poisoning in Proton Exchange Membrane Fuel Cells (PEMFC), the hydrogen feed should be carbon monoxide free (less than 50 ppm). Conventional steam reforming and water gas-shift reactions could be employed to produce carbon-monoxide-free hydrogen from hydrocarbons in industrial scales. However, the process costs as well as transportation and storage costs make the on-site production of hydrogen an attractive option for hydrogen supply of fuel cells. Hydrogen production from a single step process such as ammonia decomposition is quite attractive especially in small-scale devices. Ammonia has been produced and stored in liquid form for a long time and issues about production, transportation, handling and storage are well established. Although even small traces (as low as 13 ppm) of ammonia can degrade PEM fuel cell performance, it is shown that the platinum catalyst of the fuel cell is not poisoned by ammonia; but rather the decrease in performance is because H⁺ ions are replaced by NH4⁺ within the fuel cell anode catalyst layer (Uribe et al. 2002). Also, higher purity of available commercial ammonia makes it a better candidate for hydrogen production compared to methanol (Choudhary et al. 2001). Therefore, the availability, relatively easy decomposition with no need for added oxygen or steam and narrow explosion limits make ammonia a good candidate as a hydrogen carrier especially for portable devices. Scaling down the conventional reactors for hydrogen production is one approach. However, due to thermal and radical quenching at the walls, gas phase reactions are suppressed in gaps smaller than 1 ~ 2 millimeters (Fernandez-Pello 2002). Catalytic-wall reactors could also be employed to enhance reactions. Commercial ammonia makes it a better candidate for hydrogen production compared to methanol (Choudhary et al. 2001). Therefore, the availability, relatively easy decomposition with no need for added oxygen or steam and narrow explosion limits make ammonia a good candidate as a hydrogen carrier especially for portable devices. Scaling down the conventional reactors for hydrogen production is one approach. However, due to thermal and radical quenching at the walls, gas phase reactions are suppressed in gaps smaller than 1 ~ 2 millimeters (Fernandez-Pello 2002). Catalytic-wall reactors could also be employed to enhance reactions.

Microstructured reactors benefit from high process intensification, a wide reaction range up to explosion limits, reactor safety, faster process development and distributed production which make them suitable for highly endothermic and exothermic chemical reactions. As the push for further miniaturization continues, modeling such systems becomes more and more complicated, since new physical phenomena should be taken into account. One of the complications in dealing with micro-scale devices is that the common continuum assumption can break down as the characteristic length scale of these devices approaches the mean molecular free path. In such a case, the number of inter-molecular collisions decreases and eventually there comes a stage in which the number of collisions between molecules are rare compared to the number of collisions with the surrounding walls, in which case each molecule acts independently to bring forth the gas properties (Kennard 1938). This makes the gas lose its intimate contact with solid bodies such that the gas “slips” over the surface, and in the case of heat or mass transfer, a temperature or concentration jump is observed between the surface and the adjacent gas layer. In the slip flow regime, the continuum equations can still be employed but proper velocity slip and, temperature and concentration jump boundary conditions should be specified.

The effects of velocity slip and temperature jump on flow and heat transfer characteristics of nonreacting flows have been extensively studied in microchannels (Niazmand et al. 2010, Morini 2004, Renksizbulut et al. 2006, van Rij et al. 2009 and Yu and Ameel 2001). However, non-equilibrium transport in reacting flows still remains to be studied in-depth. In the case of multi-species transport, another important effect analogous to temperature jump should be taken into account, i.e., the concentration jump. There is very limited work on the concentration jump and its effects on catalytic reactions and the available literature has mainly focused on the temperature jump and velocity slip effects. The investigation of the concentration jump was initially performed by Kramersand Kistemaker (1943) based on the work of Maxwell on velocity slip and temperature jump. Concentration jump not only affects the rate of reaction and local species concentration, but also velocity slip and temperature jump in both reacting and non-reacting systems. Many rate-limiting adsorption/desorption reactions are very sensitive to local temperatures and hence the proper modeling and computation of temperature along with the local species concentration is vital for an accurate prediction of the behavior of such systems. Therefore, all of these non-equilibrium effects should be considered simultaneously in the simulation of microreactors. This is even more pronounced in catalytic reactions since all the reactions take place on the wall. The effect of temperature jump on the performance of reactive systems was investigated and verified experimentally by Shankar and Glumac (2003) using low-pressure catalytic combustion systems. The concentration jump phenomenon has been detected in simulations of reacting gas mixtures by Bird (1994) and Papadopoulos and Rosner (1996). Xu and Ju(2005 and 2006) derived a concentration slip model and investigated the rarefaction effects on the rate of catalytic reactions in the numerical modeling of hydrogen and methane oxidation. In their work, they considered the combustion of premixed stoichiometric mixtures at very low pressures ranging from 100 Pa to 0.2atm. The velocity slip and temperature discontinuity at the wall were modeled using the conventional mixture averaged boundary conditions, Li et al.(2008 and 2009) compared the effects of different of methane/air and hydrogen/air mixtures. They reported negligible slip/jump effects at low velocities and large channels (d=1mm) and a considerable temperature discontinuity at the wall close to the flame region due to high radial gradients. More recently, Qazi Zaseet et al. (2012b) studied the non-equilibrium slip/jump effects in catalytic oxidation of lean hydrogen/air mixtures under different operating conditions. The presence of a temperature discontinuity at the wall was reported to be the main factor in defining the concentration jump at the edge of the Knudsen layer.

Small hydrogen generating devices have been developed and tested in recent years to meet the needs for on-site production of hydrogen for portale devices (Ganley 2004a, 2004b and Sorensen et al. 2005). Different catalyst and metal supports have been proposed and examined for hydrogen production. In these experiments the ammonia conversion rate has been measured for different geometries and operational conditions. The underlying detailed chemical kinetics of ammonia decomposition on different catalytic surfaces, however, is not well documented yet. Few studies have reported elementary reactions for ammonia decomposition and usually the global rate of reaction (one step) is presented for ammonia decomposition.