Improved generation of large-scale atomistic representations and pyrolysis/combustion simulations of Illinois coal and coal char using the Reaxff reactive force field
![open_access](/assets/open_access_icon-bc813276d7282c52345af89ac81c71bae160e2ab623e35c5c41385a25c92c3b1.png)
Open Access
- Author:
- Castro Marcano, Fidel
- Graduate Program:
- Energy and Mineral Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 21, 2012
- Committee Members:
- Jonathan P Mathews, Dissertation Advisor/Co-Advisor
Jonathan P Mathews, Committee Chair/Co-Chair
Ljubisa R Radovic, Committee Member
Caroline Elaine Clifford, Committee Member
Robert John Santoro, Committee Member
Adrianus C Van Duin, Committee Member - Keywords:
- coal structure
coal pyrolysis
char combustion
molecular modeling
ReaxFF
reactive molecular dynamics - Abstract:
- A highly automated molecular generation approach was implemented and coupled with reactive force field methods to create a new computational capability that enabled the investigation of structural transformations and chemical reactions involved in coal pyrolysis and char combustion. The present work demonstrated the applicability and utility of this new computational capability for examining at the molecular level the complex chemistry associated with coal pyrolysis and char oxidation and combustion. In this investigation, Illinois no. 6 Argonne Premium coal, the world’s most well-studied coal, was evaluated using atomistic representations of both the coal and the coal char created for this purpose. Orientation and stacking issues were also explored utilizing molecular representations of several Argonne Premium coals and an anthracite coal. An extensive review of the chemical and physical structural features of Illinois no. 6 coal was created covering aromatic and aliphatic components, functional groups and heteroatoms, molecular weight distribution, nature of the cross-linked network, porosity, surface area and density. Illinois no. 6 coal is vitrinite-rich (85%, dmmf) with a normalized elemental composition of C100H77.3O13.1N1.5S1.2 and a high organic sulfur content of 2.5% (dmmf). Nuclear magnetic resonance (NMR) analyses reported an aromaticity of 72% with 15 aromatic carbons (three to four fused aromatic rings) and 5 attachments per cluster, and an average cluster molecular weight of 316 Da. X-ray photoelectron spectroscopy (XPS) determined that organic oxygen, nitrogen, and sulfur forms are primarily ether and phenolic, pyrrolic and pyridinic, and aliphatic and thiophenic type structures, respectively. Small angle neutron scattering (SANS) analyses found that Illinois coal is mostly microporous and 129Xe NMR confirmed that the pore structure of Illinois coal consisted of two distinct regions with average pore diameter of 6 and 10Å. These analytical data from the literature were used to construct a large-scale coal molecular model based on an improved automated construction approach in an effort to move toward capturing the continuum structure over a large scale. The model contains 50,789 atoms within 728 diverse molecules and is the largest, most complex coal representation constructed to-date. The aromatic ring size distribution was based on multiple previously published high-resolution transmission electron microscope (HRTEM) lattice fringe micrographs and was duplicated with automated construction protocols (Fringe3D) in molecular modeling space. Additional structural data was obtained from the abundant literature assessing this Argonne Premium coal. Organic oxygen, nitrogen, and sulfur forms were incorporated primarily into the polyaromatic structures according to XPS and X-ray absorption near-edge structure spectroscopy (XANES) data. Aliphatic carbons were distributed among cross-links and pendant alkyl groups based on the combination of laser desorption ionization mass spectrometry (LDIMS), ruthenium ion catalyzed oxidation, elemental analysis, and NMR data to agree with literature data. Construction of coal molecules was performed using Perl scripts adapted and improved from earlier work in Materials Studio to eliminate personal bias and improve the accuracy and the scale of the structure generated. The Illinois coal model contained >50,000 atoms (C26860H20897O2502N412S330, atomic H/C=0.778 and O/C=0.093) in 728 cross-linked aromatic and hydroaromatic clusters exhibiting a broad and continuous molecular weight distribution ranging from 100 to 2850 Da with a sharp peak at ~350-400 Da, calculated values for Mn and Mw of 522 and 861 Da respectively, aromaticity of 75%, and a simulated helium density of 1.32 g/cm3 in agreement with experimental data for Illinois no. 6 coal. A theoretical pyridine extraction yield, determined by a group contribution approach, was in agreement with the experimental value. The extract and residue representations were generated from the large-scale Illinois coal model and showed consistency with NMR, elemental analysis and LDIMS trends. The distribution of heteroatomic classes and double bond equivalents of the extract was consistent with extract experimental data from electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) collected and provided by the National High Magnetic Field Laboratory. These data further constrain the molecular weight of extractable material and was consistent with limited pyridine extractability and model heteroatom classes. The ReaxFF reactive force field was used to perform pyrolysis simulations at 2000 K on the constructed large-scale molecular model for Illinois coal to examine structural modifications and reactions associated with coal pyrolysis. This high temperature enabled chemical reactions to occur within a practical simulation time. The ReaxFF simulation was performed until about 60% of the cross-links had been disrupted primarily through thermolysis. For this coal pyrolysis was mainly initiated by the release of hydroxyl groups, dehydrogenation of hydroaromatic structures, and by cleavage of heteroatom-containing cross-links. The main pyrolysis products were hydrogen, methyl, ethylene, acetylene, formaldehyde, ethynol, alkylphenols, alkylnaphthalenes and alkylnaphthols, in agreement with experimental observation. During pyrolysis the molecular weight distributions shifted to lower values as expected due to thermal decomposition to form smaller fragments. The thermal degradation of sulfurated and oxygenated cross-links was more substantial than that of alkyl linkages, in accordance with their higher reactivity. Analysis of sulfur form distribution showed that aliphatic sulfur decomposed more rapidly while thiophenic sulfur was more thermally stable in agreement with experimental data. The extent of decomposition for heterocyclic 5-membered rings was: 57 % for pyrrolic, 47% for thiophenic, and 29% for furanic type structures. The ReaxFF simulation was repeated on a sulfur-free model to further analyze the role of organic sulfur forms in Illinois coal pyrolysis. ReaxFF results showed that the rate of generation of light gases and tars was higher in the presence of sulfur. Further analysis of ReaxFF simulations showed that aryl and alkyl C-S bonds are weaker than aryl and alkyl C-C bonds. Hence, cleavage of the C-S bonds resulted in more extensive fragmentation leading to larger quantities of aliphatic and aromatic structures that evolved as light gases and tars within Illinois coal model (sulfur containing) compared to sulfur-free model. Therefore, sulfur atoms enhanced the reaction kinetics during coal pyrolysis. A devolatilized Illinois no. 6 coal char atomistic representation was generated using published HRTEM lattice fringe images and Fringe3D in conjunction with Perl scripts, and coupled with the ReaxFF reactive force field. Fringe3D facilitates the char structure generation process by producing a distribution of aromatic structures based on HRTEM lattice fringe image analyses. Perl scripts were used for incorporating heteroatom and aliphatic components to aid elimination of investigator bias, and facilitate a more rapid construction process. The char structure was constrained by a combination of elemental and NMR literature data. Chemical and physical parameters were found to be consistent with the experimental data. The ReaxFF force field for hydrocarbon combustion was used to perform simulations to examine the structural transformations and chemical processes associated with char combustion. In this initial work, very high temperatures (3000-4000 K) were selected for ReaxFF simulation under stoichiometric, fuel lean and rich combustion conditions. These elevated temperatures were chosen to observe chemical reactions proceed to completion within a computationally practical simulation time. It is expected that with computational gains longer simulations at more reasonable combustion temperatures could be obtained. The char oxidation process was mainly initialized by either thermal degradation of char structure to form small fragments, that were subsequently oxidized, or by hydrogen abstraction reactions by oxygen molecules and O and OH radicals. A more rapid oxidation and combustion of the polyaromatic structures occurred at fuel lean (oxygen rich) conditions compared with fuel rich combustion. Char transitions included 6-membered ring conversion into 5- and 7-membered rings that further decomposed or reacted with mostly O and OH radicals. To further evaluate the applicability of HRTEM as the basis for structural model generation molecular slice models for several Argonne Premium coals (Beulah-Zap, Illinois no. 6, Upper Freeport, Pocahontas no. 3) were also constructed to predict the pair distribution function (PDF) and to evaluate the fine detail of the frequency spectra via examination of intermolecular and intramolecular contributions. Atomistic representations were generated directly from published HRTEM lattice fringe images using Fringe3D and constrained by elemental analysis and NMR data from the literature. The constructed coal models were partially geometry-optimized to achieve realistic bond lengths but without displacement of coal molecules enabling the distribution of fringe length, stacking, and orientations to be duplicated in 3D modeling space. The resulting coal slice models, devoid of cross-links, captured a distribution of turbostratic crystalline dimensions with an average cluster size of about 1 nm, an average interlayer spacing ranging between 0.37 and 0.39 nm, and an average stacking number of ~2-3 in accordance with HRTEM and XRD literature data for Argonne coals. Simulated PDF agreed with experimental data obtained from Argonne National Laboratory. Analysis of the simulated intermolecular PDF contribution showed stronger intensities with increasing coal rank in agreement with the growth in the stacking number and stack height observed from low- to high-rank coals. The simulated intramolecular PDF contribution showed shorter peak amplitudes for low-rank coals in comparison to high-rank coals in agreement with the slight increase in stack width with coal rank in the bituminous range with more extensive increases for anthracite. To further examine these contributions, lattice models composed of pyrene molecules were also constructed via Fringe 3D and manipulated to directly investigate the effect of aromatic orientation distributions and stacking on the simulated PDF. Peak intensities of simulated intermolecular PDFs at the average interlayer spacing increased with the degree of alignment, consistent with the slight increase in the stacking number observed from low- to high-rank coals with a more dramatic transition for anthracite. The automated construction protocol enabled generation of large-scale coal molecular models and aided in capturing a wider range of the complex distribution of coal structural features. Linking high-accuracy, large-scale coal model construction approaches improved here with a computational method for large systems with dissociation and formation of chemical bonds enabled obtaining a versatile simulation platform that provided the ability to evaluate the complex chemistry associated with coal pyrolysis and char combustion. The work performed here was aimed at demonstrating the capability of this approach for examining the complex chemistry involved in coal pyrolysis and char oxidation and combustion.