Investigation dry reforming of methane over nickel using a one‐dimensional model

In the field of catalysis, dry reforming, that is, methane reforming with CO2$\rm {CO_2}$ , is in the focus due to growing environmental concerns about oil depletion and global warming with a desire to produce synthesis gas. However, this process can lead to the formation of carbon, which can cause catalyst deactivation, especially at industrial conditions. Nevertheless, the key to develop a more coke‐resistant catalyst is a better comprehension of the reforming process at a molecular level. Regardless of all the investigations available in literature, the detailed path for the conversion of methane to syngas and carbon remains a controversial issue. Another problem in setting up a reaction mechanism is the difficulty to define the thermodynamic data for intermediate surface species and this leads to the development of thermodynamic consistent surface reaction mechanisms in literature where the thermodynamic data are not used to calculate the rate coefficients of the reverse reactions. Rather the Arrhenius parameters for the forward as well as backward reactions are explicitly given in the reaction mechanism to establish thermodynamic equilibrium. In this investigation, a kinetically consistent detailed surface reaction mechanism is developed which consists of 26 reversible reactions with the help of a one‐dimensional model, LOGEcat. Our previous work constructs the basis of the present investigation. Further, a detailed sensitivity analysis of reversible reactions and reaction pathways is performed to understand the mechanism better. The mechanism is validated for dry reforming of methane over nickel catalyst, however, it can also be used for other processes, such as, steam reforming and partial oxidation. The mechanism is tested by comparing the simulation results with the literature experiments and simulations in a wide range of temperature. The new developed kinetically consistent surface reaction mechanism is able to accurately express the dry reforming of methane over the nickel catalyst for complete range of temperature and also provide a useful insight into the key rate determining steps.


F I G U R E 1
The schematic illustration of the 1D model, LOGEcat based on the single-channel 1D catalyst model, which is divided into a finite number of cells along the length of the cylinder/pipe (z-axis) and each cell is treated as a perfectly stirred reactor (PSR).
Nevertheless, the key to develop a more coke-resistant catalyst is a better understanding of reforming process at a molecular level.Despite all the investigations given in open literature (Aparicio [1] and the references within), the detailed path for the conversion of methane to syngas and carbon remains a controversial issue [22].Therefore, further investigations are needed to understand the reaction pathways followed in a reforming process using different approaches.
Another problem in setting up a detailed surface reaction mechanism discussed in literature is the difficulty to define the thermodynamic data for intermediate surface species.This leads to the development of thermodynamic consistent surface reaction mechanisms in literature where the thermodynamic data are not used to calculate the rate coefficients of the reverse reactions, rather the Arrhenius parameters for the forward as well as backward reactions are explicitly given in the surface reaction mechanism to establish thermodynamic equilibrium.
In this work, a kinetically consistent detailed surface reaction mechanism is developed consisting all reversible reactions using a one-dimensional model described in Rakhi et al. [23].The Arrhenius parameters are given only for the forward reactions to calculate the forward rate coefficient, whereas for the reverse rate calculations, thermochemistry of the surface bound species is used and the Arrhenius parameters are not needed.The mechanism is tested for dry reforming of methane over nickel catalyst by comparing the simulation results with the literature [24] in a broad temperature range.The thermodynamic model calculates the reactant and product species concentration for the dry reforming of methane over the nickel catalyst for complete range of temperature close to the literature experiments.

SIMULATION SET-UP
The one-dimensional model, LOGEcat [25], is a part of the LOGEsoft software suite for chemical reaction calculations and is used to perform the simulations discussed in this paper.The pressure gradient along the inhomogeneity of the mixture is neglected because the diameter of the catalytic channel is small and the external diffusion is modeled by a thin layer indicated by a separate pore gas zone close to the wall.
The model is based on the single-channel 1D catalyst model.The schematic illustration of the 1D modeling approach is shown in Figure 1.The single channel is divided into a finite number of cells and each cell is treated as a perfectly stirred TA B L E 1 Summary of the thermodynamic data for all the surface species used in the present simulations (at 300 K). S is the entropy and H is the enthalpy.reactor (PSR).The conservation equations are solved in each PSR for each time step in addition to the 1D Navier-Stokes equations for flow velocity solved by an operator splitting method over all cells.For more details, we refer the reader to our previous work [23,26].Dry reforming process is investigated by simulating a Ni-coated monolithic catalyst.The applicability of the thermodynamic model is tested by comparing the simulation results with the literature.The geometric data, catalyst parameters, and simulation conditions are taken from Delgado et al. [24].

Species
The reactor with a circular catalyst is used and it has radius 5 mm and a reaction zone of 27-mm long.A single layer of washcoat is used and a sensitivity analysis is done to find the surface area per catalyst length.Nitrogen dilution is used.The surface site density,  for Ni is 2.6 × 10 −5 mol/m 2 .
Using the settings mentioned above, the simulations are carried out at 4 slpm (standard liters per minute, T = 298.15K and p = 1.01325 bar) by varying the reaction temperature in a range, T = [400-1200] K for different reactor conditions.The simulation results for the dry reforming case considered are discussed in upcoming sections.

SURFACE REACTION MECHANISM
Different molecular paths indicating the overall reaction [24,27] represents the methane conversion into a mixture of hydrogen, carbon monoxide, and carbon dioxide.In Delgado et al. [24], a detailed surface reaction mechanism was developed using the overall/global reactions.Their mechanism can be used to model the various reforming processes, which covers all the ways from total oxidation to pyrolysis.The kinetic scheme consists of 52 reactions with 6 gas-phase and 14 surface species and this scheme serves as the base for our study.A thermodynamic model is developed in this work by extending our previously developed surface reaction mechanism [26].In Rakhi et al. [26], the mechanism comprises of 21 reactions and was applied to steam reforming process only.Here, new reaction scheme involving carboxyl species as intermediate along with carbon formation paths have been added to extend the mechanism presented in Rakhi et al. [26] in order to make the mechanism applicable to oxidative, dry, and steam reforming.
The new mechanism consists of 26 reversible reactions in total with seven gas-phase and 14 surface species.The set of 21 reactions has been directly taken from Rakhi et al. [26].Five more reversible reactions have been added here.These reactions are taken from Delgado et al. [24] and the kinetic data remain same as Delgado et al. [24].The thermodynamic data are taken from Rakhi et al. [26] and for new species, COOH(s) is taken from Liu et al. [28].
The kinetic data for all the backward reactions involved in the reaction mechanism are calculated with the help of thermodynamic data listed in Table 1 for all the species.The details about calculation of forward and backward rate constants The main reaction pathways for reforming processes, such as, steam reforming, dry reforming, and oxidative reforming of methane over nickel catalyst.

F I G U R E 3
The concentration of reactant CH 4 for dry reforming of methane over a nickel catalyst is shown as a function of temperature along with the reference data from Delgado et al. [24].The unfilled squares represent reference simulations, filled squares: reference experiments, solid lines: reference equilibrium calculations and dash lines: LOGEcat calculations with the thermodynamic model.are discussed in our previous work [23,26] for both the approaches, that is, thermodynamically as well as kinetically consistent surface reaction mechanism.Hence, we refer the reader to Refs.[23,26] for more details.
The main pathways for all the reforming processes of methane on the nickel catalyst for syngas production are shown in Figure 2 [24], which is further discussed in upcoming sections.

RESULTS
The mechanism with 26 reversible reactions discussed in the previous section is utilized to perform the simulations in order to check the predictability of the model for dry reforming of methane over a nickel catalyst.The concentration of reactant species, CH 4 and CO 2 computed with the one-dimensional model, LOGEcat at the reactor outlet as a function of temperature along with the reference simulations and experiments from Delgado et al. [24] is shown in Figures 3 and 4.

F I G U R E 4
The concentration of reactant CO 2 for methane dry reforming over a nickel catalyst is depicted as a function of temperature along with the reference data from Delgado et al. [24].The unfilled squares represent reference simulations, filled squares: reference experiments, solid lines: reference equilibrium calculations and dash lines: LOGEcat calculations with the thermodynamic model.

F I G U R E 5
The concentration of product H 2 for methane DR over a nickel catalyst is shown as a function of temperature along with the reference data from Delgado et al. [24].The symbols have same meaning as explained in the previous figure.
The methane consumption starts at ≈600 K.The concentration slowly decreases with increasing temperature similar to the literature simulations and experiments.A full consumption of CH 4 happens at temperature ≈1000 K.Note that the reforming for experiments as well as our calculations is away from the equilibrium calculations.The qualitative behavior for the species computed with our model is similar to the reference results, however, quantitative differences are noted between the calculations from thermodynamic and kinetic model.
Carbon dioxide concentration computed with the LOGEcat also show similar behavior as shown by methane and the concentration decreases with increasing temperature.However, a linear variations is noted for CO 2 from 400 to 800 K and a full consumption happens at temperature ≈1000 K. Similar to the methane consumption, the consumption of this species also starts at 600 K for the reference results.However, for our calculations, the species is being used even at low temperatures indicating the difference in the profile qualitatively.The reactants consumption leads to the formation of H 2 , CO, and H 2 O and these species have also been validated against literature and are discussed next.
Figures 5 and 6 illustrate the concentration of the products, H 2 and CO.With the consumption of the reactants, CH 4 and CO 2 , a gradual increase in the formation of H 2 and CO is noted in the considered temperature range.This leads to the thermodynamic equilibrium for temperatures above 1000 K.

F I G U R E 6
The concentration of product CO for methane DR over a nickel catalyst is depicted as a function of temperature along with the reference data from Delgado et al. [24].The symbols are explained in the previous figure.

F I G U R E 7
The concentration of product H 2 O for methane DR over a nickel catalyst is shown as a function of temperature along with the reference data from Delgado et al. [24].The symbols are explained in Figure 3.
The computed water concentration shown in Figure 7 indicates increase in the concentration profile in temperature range 400-700 K.The species concentration reaches a maximum value at around 800 K and after that starts to decrease with increasing temperature.The qualitative behavior of this species matches with the reference data.However, overall the simulations over-predicted the species concentrations in case of dry reforming of methane.
This hints towards the indirect path for formation of H 2 and CO through H 2 O for the thermodynamic model.Whereas, a direct oxidation is followed for the kinetic model discussed in literature [24].
The reforming with the thermodynamic model is different from the kinetic model.The reforming is close to the equilibrium calculations computed from the kinetic model for the considered reforming conditions.The reforming calculations with the thermodynamic model is away from the thermodynamic equilibrium.Further, the reaction sensitivity analysis can to be performed to explain the deviations in the species concentration with the thermodynamic model when compared to the kinetic model.

CONCLUSION
The mechanisms available in literature needs specification of Arrhenius parameters for forward as well as reverse reactions to calculate the rate coefficients.The thermochemistry of the intermediate species is not used to calculate revere rate coefficient, rather the Arrhenius parameters are used for such calculations.In this study, a thermodynamic model is developed to investigate dry reforming of methane over nickel catalyst.The model needs the Arrhenius parameters only for the forward reactions.The reverse rates are calculated by using the thermochemistry of the surface bound species.The developed mechanism is tested against reference results and the model is able to predict the reactant distribution for the whole temperature range considered for the simulations for dry reforming of methane over nickel catalyst.

A C K N O W L E D G M E N T S
Financial support by the federal ministry of education and research (Bundesministerium für Bildung und Forschung, BMBF) under the Grant Number 03SF0693A of the collaborative research project "Energie-Innovationszentrum" is gratefully acknowledged.The authors also thank the Graduate Research School (GRS) of the BTU Cottbus-Senftenberg for the partial financial support.
Open access funding enabled and organized by Projekt DEAL.