Elsevier

Fuel Processing Technology

Volume 178, September 2018, Pages 197-205
Fuel Processing Technology

Research article
Pyrolysis simulations of Fugu coal by large-scale ReaxFF molecular dynamics

https://doi.org/10.1016/j.fuproc.2018.05.011Get rights and content

Highlights

  • Pyrolysis of Fugu coal model with 23,898 atoms was explored by ReaxFF MD method.

  • Very slow heat-up and a series of long time simulations of 2 ns were performed.

  • Simulations qualitatively predict evolution profiles of coal tar and gases.

  • High temp. leads to over-predicted coal tar yield &under-predicted gas yield.

Abstract

This paper presents pyrolysis simulations of a multi-component structure model of Fugu sub-bituminous coal containing 23,898 atoms with ReaxFF MD. The slow heat-up and long-time isothermal simulations were performed using the GPU-enabled ReaxFF MD code of GMD-Reax and reaction analysis code of VARxMD. The simulated pyrolyzate evolution trends and pyrolysis stages proposed based on lumped pyrolyzates agree fairly with literature. Major reaction pathways for early generation of CO2, H2O, CH4 and H2 in the activation and primary pyrolysis stages are found closely associated with carboxyl and methoxyl groups, indicating the significant role of oxygen-containing groups in the initialization of Fugu coal pyrolysis. Compared with reported experiments, the coal tar yield is over-predicted, while the gas yield is slightly under-estimated. To shorten the simulation time from 2 ns to 250 ps, an increase of 400 K in average on the basis of temperature range of 1200–2200 K is needed that will result in the over-prediction for the yields of gas and tar, as well as under-prediction for the non-volatile yield. This work suggests that ReaxFF MD simulation is useful for revealing the overall scenario and detailed reactions in coal pyrolysis but cautions should be taken for proper elevated temperatures.

Introduction

Coal pyrolysis is the initial reaction step in most coal conversion processes such as gasification, liquefaction and combustion [1,2]. Improved insight into coal pyrolysis, especially the fundamental pyrolytic reaction mechanism, will be useful in developing clean and efficient technologies for coal utilization. Coal pyrolysis is well accepted as a radical driven process that involves myriad coupled reaction pathways with vast free radical intermediates generated [3,4]. Most radicals are of high reactivity and have very short lifetime so it is hard to capture them experimentally [4,5], even with state-of-the-art experimental approaches. Besides, quantum mechanics (QM) that has been used to study the pyrolysis mechanism of simple model compounds of coal [[6], [7], [8]] can hardly be used to study the complex coal pyrolysis system, for it is computationally very expensive and requires prior knowledge of possible reaction pathways. The reactive molecular dynamics (ReaxFF MD) method that combines molecular dynamics with ReaxFF reactive force field provides a promising approach for simulating the complex chemistry and diverse reaction pathways involved in coal pyrolysis system [[9], [10], [11], [12]].

The ReaxFF force field is proposed based on bond order by van Duin and Goddard et al. [9] that can smoothly describe the dissociation and formation of chemical bonds. ReaxFF force field has accuracy close to DFT method with much lower computational expenses than DFT, so it is applicable to large reactive system of >1000 atoms. More remarkably, the feature of no requirement for predefined reaction pathways makes it an appropriate choice and promising for simulating the complex coal pyrolysis system.

ReaxFF MD has been widely applied to explore the reaction mechanisms of molecular systems with thousands of atoms or more recently. In particular, this method has been used to probe the reactions of coal macromolecules in various environments [10,11,[13], [14], [15], [16], [17]]. Salmon et al. [13] first employed ReaxFF MD method to simulate the thermal decomposition of Morewell brown coal using simple model compounds and one relatively large model of 2692 atoms. This work reproduced thermal decomposition processes of defunctionalization, depolymerization, and rearrangement of the residue structures observed in experiments and explained the generation pathways of some gaseous compounds. Castro-Marcano et al. [10] simulated the pyrolysis of a large-scale molecular model for Illinois No. 6 coal using ReaxFF MD method to investigate coal pyrolysis chemistry and the effect of organic sulfur content. The coal model contains 51,529 atoms that was the largest coal model ever simulated with ReaxFF MD when published. The results showed that the pyrolysis of the coal was initiated by release of hydroxyl groups and dehydrogenation of hydroaromatic structures followed by breakage of heteroatom-containing cross-links. This work further demonstrates that ReaxFF MD method combining with large-scale molecular models can provide a useful tool in probing the complex chemistry involved in coal pyrolysis. Zheng et al. [11,18] of the authors' group also performed large-scale ReaxFF MD simulations to explore the initial reaction mechanisms and product distributions in pyrolysis of Liulin bituminous coal. By simulating the second largest coal model of 28,351 atoms, the product evolution tendency observed in the simulations showed fair agreement with the experiments reported in literature. The coal pyrolysis was found to be initialized primarily by bond dissociation of alkyl-aryl ether bridges in coal structure [11]. Moreover, the pyrolysis process of Liulin coal can be divided into four major stages based on the cleavage of bridge bonds during the slow heat-up ReaxFF MD simulations [18]. In our recent paper [19], by comparing simulation results of three Liulin coal models with different scale (2338, 13,498 and 98,900 atoms), Zheng et al. concluded that a large-scale model is vital for investigating the overall behavior in coal pyrolysis by ReaxFF MD simulation.

This work focuses on the pyrolysis properties of Fugu sub-bituminous coal. A series of ReaxFF MD simulations were performed using a multi-component structure model of 23,898 atoms for Fugu coal. The simulation trajectories were then analyzed to explore the product profile evolution and underlying pyrolytic reaction mechanism. The simulating methodology, including the coal structure model, the simulation details and the trajectory analysis method will be described in Section 2. The major simulation results including the pyrolyzate profile evolution, the bond breaking reactions evolution and generation pathways of major gaseous compounds and corresponding discussions will be presented in Section 3. The last section will give a brief conclusion of this work.

Section snippets

Construction of a coal structure model

Fugu coal is a sub-bituminous coal from Shaanxi province of China. The coal model used in this work is a multi-component structure model constructed on the basis of the analytical data from proximate and ultimate analysis, solid state 13C nuclear magnetic resonance (NMR) and solvent extraction experiments of Fugu coal. The Fugu coal model is comprised of 75 macromolecular components and 29 low molecular weight compounds as listed in Table 1. The 2D structures of representative components M-II (C

Results and discussion

To facilitate comparison with experimental results, the simulated pyrolyzates are lumped into five categories as C100+, C41–C100, C14–C40, C5–C13, and C0–C4, following the similar lumping rule in previous work of the authors' group [15]. The C0–C4 compounds refer to gas products which include both organic gas molecules with 1–4 carbon atoms and inorganic molecules such as H2, CO2, and H2O. The lumps of C5–C13 and C14–C40 denote light tar and heavy tar. The C41–C100 compounds correspond to

Conclusion

Using the GPU-enabled ReaxFF MD code of GMD-Reax and reaction analysis code of VARxMD developed in the authors' group, ReaxFF MD simulations were performed on a multi-component structure model containing 23,898 atoms to study the pyrolysis properties of Fugu sub-bituminous coal. Such simulations allow for obtaining an overall scenario of Fugu coal pyrolysis behavior, the detailed reactions or radicals for producing gas products and addressing the effects of elevated simulation temperatures to

Author contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Notes

The authors declare no competing financial interest.

Acknowledgement

This work was supported by China's National Key Research and Development Plan [2016YFB0600302] and National Natural Science Foundation of China under Grant [91434105].

References (28)

  • S. Bhoi et al.

    Molecular dynamic simulation of spontaneous combustion and pyrolysis of brown coal using ReaxFF

    Fuel

    (2014)
  • J. Liu et al.

    Reaction analysis and visualization of ReaxFF molecular dynamics simulations

    J. Mol. Graph. Model.

    (2014)
  • W.J. He et al.

    Analysis of tars produced in pyrolysis of four coals under various conditions in a viewpoint of radicals

    Energy Fuel

    (2015)
  • A.C.T. van Duin et al.

    ReaxFF: a reactive force field for hydrocarbons

    J. Phys. Chem. A

    (2001)
  • Cited by (103)

    View all citing articles on Scopus
    View full text