Modeling Study of High Temperature Pyrolysis of Natural Gas

Subscriber access provided by UNIV OF WESTERN ONTARIO

Kinetics, Catalysis, and Reaction Engineering

A Modeling Study of High Temperature Pyrolysis of Natural Gas Soumya Gudiyella, Zachary J. Buras, Te-Chun Chu, Istvan Lengyel, Sreekanth Pannala, and William H. Green Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.8b00758 • Publication Date (Web): 11 May 2018 Downloaded from http://pubs.acs.org on May 22, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

A Modeling Study of High Temperature Pyrolysis of Natural Gas Soumya Gudiyellaa, Zachary J. Burasa, Te-Chun Chua, Istvan Lengyelb, Sreekanth Pannalab, William H. Greena* a

Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139 b

SABIC Technology Center, 1600 Industrial Boulevard, Sugar Land, TX 77478

*Corresponding author: [email protected] KEYWORDS: High temperature pyrolysis process, Natural gas chemical kinetic mechanism, Chemical reactor network, Reaction Mechanism Generator (RMG), Acetylene, Ethylene, Polycyclic aromatic hydrocarbon (PAH)

ABSTRACT: High temperature pyrolysis (HTP) is commercial process to convert methane to acetylene. The HTP process consists of two reaction zones, followed by a quenching zone. In this work, a pilot scale HTP process was modeled to assess the effect of the amount of fuel burned and the cracking gas composition on acetylene and polycyclic aromatic hydrocarbon (PAH) production. The HTP process is simulated using a chemical reactor network, which consists of a series of ideal reactors. The composition of cracking gas in the second reaction zone varied from methane to hexane. The propensity of the feed to form acetylene vs. PAH at a given 1 ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 60

process condition was determined using a detailed chemical kinetic mechanism. The chemical kinetic mechanism was developed using an automated mechanism generation software package, the Reaction Mechanism Generator (RMG). Compared to existing pyrolysis mechanisms that can only be used to model the cracking of a finite number of species, RMG can be used to model the cracking of any arbitrary species consisting of carbon, hydrogen and oxygen. The modeling results showed that the C2 yield is largely independent of either overall φ or cracking gas carbon number. In contrast, the lumped aromatic yield appears to have a positive correlation with both overall φ and cracking gas carbon number. Sensitivity and rate of production analyses were performed to identify the important pathways that lead to the formation of aromatics for various feed compositions.

1. INTRODUCTION Natural gas is a fossil fuel that is primarily composed of methane. Based on the origin of natural gas, it can also have varying amounts of heavier alkanes and minor amounts of species such as carbon dioxide, nitrogen, helium, hydrogen sulfide and water1. Natural gas is widely used around the world for a variety of applications pertaining to power generation, transportation, and residential usage and as a chemical industry feedstock2. Due to an increase in natural gas reserves around the world, it is expected that natural gas will become more important as a chemical feedstock3. As a result, it is of interest to convert natural gas to useful fuels and chemicals such as syngas and acetylene. Acetylene can be converted to ethylene through catalytic hydrogenation4. Ethylene is a building block for polymers and a number of chemical intermediates such as acetaldehyde, ethanol, and ethylene glycol1, which can be used to make everyday products. 2 ACS Paragon Plus Environment

Page 3 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Non-catalytic high-temperature pyrolysis (HTP) is commercially used to convert methane to acetylene4-5. In HTP, acetylene can be produced from methane/natural gas either through a onestep or a two-step process4. The well-known one-step process for acetylene production is the partial oxidation process (POX) developed by BASF4. In POX, methane and oxygen are premixed at an equivalence ratio greater than one in a mixing zone, ignited in a burner, combusted for milliseconds before quenching the products using water or heavy oil6. Other researchers7,8 also investigated one-step high temperature pyrolysis of fuels such as methane and naphtha in an electrically heated tube reactor. The yields of acetylene, ethylene and carbon were measured and a simplified reaction model was developed to describe the pyrolysis of fuels and formation of products. The two-step process developed by HOECHST consists of two reaction zones, followed by a quenching zone4, 9. In the first reaction zone, a stoichiometric mixture of feed and oxygen are combusted in the burner to provide the heat required for the pyrolysis reactions in the second reaction zone. A separate hydrocarbon feed (cracking gas) is introduced in the second reaction zone. The cracking gas mixes with the hot effluent from the first reaction zone and pyrolyzes to form acetylene. The two-step process simplifies the combustion process, since the burner is always running at essentially the same condition. The two-step process has also been reported to produce higher yields of acetylene when compared to the POX process due to more and finer controls for the flame composition, cracking gas ratio and reactor design geometry10. However, the major drawback in non-catalytic HTP is the tendency of this process to produce soot11. Therefore, in this work, the two-step HTP process is modeled for natural gas and various cracking gas compositions and the yield of acetylene, ethylene and soot-precursors (aromatic hydrocarbons up to 2-rings) are determined.

3 ACS Paragon Plus Environment


Page 4 of 60

One-step partial oxidation of methane has been extensively investigated, both experimentally and numerically in batch reactors3,12,

13

flow reactors5,14-17 and premixed flames18-21. However,

limited studies are present in literature for the 2-step process. High temperature pyrolysis of hydrocarbons was investigated by Ktalkherman et al.22-25, where hydrogen and air were combusted in the first reaction zone, followed by addition of propane22, LPG23 or naphtha24 or their mixtures25 in the second reaction zone, to produce ethylene. The yield of ethylene was significantly higher when compared to the conventional steam cracking22 and pyrolysis methods24-25. Recently, Zhang et al.10 numerically investigated the 2-step process using coke oven gas as primary feed and methane as a secondary feed; they optimized the design of a jet-incross flow reactor to maximize the acetylene production. Zhang et al.26 investigated the effect of addition of wet gas consisting of ethane and propane, on the methane partial oxidation process and observed an increase in acetylene and carbon monoxide production. Hence all these studies show that the 2-step process and the composition of the cracking gas influence the yield of acetylene. A chemical kinetic model which can accurately predict the yield of acetylene and other products from the HTP process would be very helpful in improving this process, and for understanding how the process depends on feed composition. Several chemical kinetic models for natural gas have been presented in the literature. GRI-Mech 3.027 is a commonly used chemical kinetic mechanism for combustion of natural gas3,10,14, and consists of hydrocarbons up to propane. Other commonly used natural gas oxidation mechanisms are the Glarborg15, Leeds28, Petersen29, Konnov30 and Curran31-32 mechanisms (to cite a few). The highest carbon number additive in the natural gas mechanisms in literature is n-pentane (C5H12)33-34. The natural gas and cracking gas compositions used in this study are shown in Table 1 and Table 2. Natural gas used in this study 4 ACS Paragon Plus Environment

Page 5 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


consists of heavy alkanes up to C6 and cracking gas compositions include C1-C6 hydrocarbons and also includes alcohols such as butanol and amyl alcohol. To the authors’ knowledge no reaction mechanism has been presented in literature which includes the reaction kinetics for all the species shown in Table 1 and Table 2. Therefore, the primary objective of this work is to develop a comprehensive mechanism for combustion and pyrolysis of natural gas mixtures including all these components. It is desired that the model accurately predict the yield of acetylene, ethylene, aromatics (up to naphthalene) and other major intermediates formed in the 2-step HTP process. The other major contribution of this study is development of a methodology for modeling the 2step HTP process. The HTP reactor in this study was modeled using a chemical reactor network and a detailed natural gas mechanism, which includes elementary kinetics for all the species listed in Table 1 and Table 2. The natural gas mechanism is developed using the automated mechanism generation software, Reaction Mechanism Generator (RMG).35 The results from the HTP reactor model were validated against pilot plant reactor data and the accuracy of the model in predicting the major species such as methane (CH4), ethylene (C2H4), acetylene (C2H2), hydrogen (H2), carbon monoxide (CO) and carbon dioxide (CO2) is verified. The effect of various cracking gas feed compositions on C2 (acetylene and ethylene) and aromatics yield is also investigated. Finally, the dominant pathways for the formation of major soot precursors such as benzene (C6H6), indene (C9H8) and naphthalene (C10H8) are discussed. Table 1. Natural gas fuel composition used in this study Species Methane (CH4) Ethane (C2H6) Nitrogen (N2)

Concentration (mol %) 91.612 5.038 1.604 5 ACS Paragon Plus Environment


Carbon Dioxide (CO2) Propane (C3H8) 1,3-Propadiene (C3H4) n-Butane (C4H10) Isopentane (C5H12) Hexane (C6H14) Cyclohexane (C6H12)

Page 6 of 60

1.389 0.266 0.036 0.027 0.012 0.0075 0.0075

Table 2. Cracking gas composition used in this study No.

Species

1 2

Natural gas (NG, composition shown in Table 1) NG+Ethane (C2H6)

3

NG+Propane (C3H8)

4

NG+Butane (n-C4H12)

5

NG+Butene (1-C4H10)

6

NG+Pentane(Isopentane(iC5H12,0.003) and n-Pentane (C5H12,0.997)) Hexane (n-C6H14) NG+Hexane(n-Hexane (C6H14, 0.84)/Cyclohexane (C6H12, 0.16)) NG+Hexene (1-C6H12) NG+Butanol (C4H10O) NG+Amyl alcohol (C5H12O)

7 8 9 10 11 12

NG+Mix(n-Pentane(C5H12,0.20)/Hexane(C6H14, 0.54)/Butanol(C4H10O,0.26))

Concentration (mol %) 100 80/20 54/46 0/100 87/13 64/36 6/94 89/11 70/30 25/75 59/41 68/32 86/14 100 76/24 77/23 72/28 77/23 87/13 78/22

2. PILOT PLANT DATA The pilot plant data of the high temperature pyrolysis reactor was obtained from SABIC. The pilot plant data consisted of 71 runs. Natural gas was used as the fuel and oxygen was used as the oxidizer in the burner. The reactor heat duty was varied by changing the flow rates of the 6 ACS Paragon Plus Environment

Page 7 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


fuel and oxidizer. The pressure in the burner section varied from 1.8-3.1 atm. The pressure in the cracking gas region varied from 1.2-1.5 atm. The equivalence ratio in the burner was close to 1, and the overall equivalence ratio (φ) in the reactor ranged from 2-3.5. The cracking gas composition in the secondary reaction zone was also varied (as shown in Table 2). The product composition at the exit of the reactor was measured. The measured products included hydrogen (H2), methane (CH4), ethylene (C2H4), acetylene (C2H2), carbon monoxide (CO) and carbon dioxide (CO2). The inlet and outlet temperatures of cooling water in the burner and cracking gas regions were also measured. The overall residence time in the reactor is ~10 ms.

3. MODEL DESCRIPTION The model for NG HTP can be broken down into three components: 1.) A detailed chemical mechanism generated by RMG, 2.) A network of ideal reactors in series, 3.) An approximation for heat loss along the length of the pilot plant reactor. The flow diagram for developing the chemical kinetic mechanism is shown in section 3.1, the reactor network is shown in section 3.2 and the estimation of heat transfer coefficients is discussed in section 3.2.1. Each component is discussed in detail below, with special emphasis on the chemical mechanism.

3.1. Mechanisms RMG is an open-source tool for automatically generating gas-phase reaction mechanisms using a flux-based criterion.35 RMG’s kinetics database consists of 45 “reaction families” that describe how different functional groups can react. These families are populated with kinetic data in the form of general “rate rules”, and specific “training reactions”. At the beginning of this work, RMG was not well-suited for predicting the chemistry of aromatic compounds because the rate rules and training reactions in the database were almost entirely for 7 ACS Paragon Plus Environment


Page 8 of 60

linear/branched hydrocarbons/oxygenates, which are oftentimes not extensible to cyclic or aromatic molecules. More fundamentally, the existing reaction families may not capture all of the ways in which aromatic molecules can react or form. Due to these two initial limitations on RMG’s aromatic prediction capability (lack of relevant kinetic data and non-exhaustive reaction families) it was necessary to rely on a “reaction library” of aromatic reactions at first. The chemical mechanism of Narayanaswamy et al.36 for high temperature combustion of gasoline surrogate mixtures was converted into an RMG reaction (and thermochemistry) library for this purpose. The Narayanaswamy mechanism was developed on top of a base mechanism from Blanquart et al.37, and the combined mechanism has been validated against an array of experimental data sets for the combustion of linear and branched alkanes, as well as substituted aromatics. Importantly, the Narayanaswamy mechanism includes PAH formation up to four aromatic rings, unlike other well-known combustion mechanisms such as GRI-Mech27 and USC II38. Although many of the kinetic parameters in this mechanism are global (i.e., non-elementary) reaction estimates, especially for reactions involving PAH, the sources of the estimates are well-documented and can usually be traced back to a quantum mechanical (QM) calculation of the potential energy surface (PES) consisting of elementary chemical reactions. The overall strategy for obtaining a reliable, detailed chemical mechanism for NG HTP including PAH formation (at least up to indene/naphthalene) was therefore the following: 1.) Create an RMG mechanism for NG HTP, referred to hereafter as Mech v1, using the Narayanaswamy reaction library to capture PAH chemistry. 2.) Evaluate the performance of Mech v1 by combining it with the reactor network and heat loss model of the pilot plant, and 8 ACS Paragon Plus Environment

Page 9 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


compare the predictions and pilot plant measurements. 3.) Use Mech v1 to identify the critical pathways to the following aromatics/aromatic precursors in Narayanaswamy reaction library through sensitivity analysis: cyclopentadiene (CPD), benzene, indene and naphthalene. 4.) Manually add the missing aromatic chemistry identified in step 3 to RMG’s kinetic and thermochemistry databases in the form of new reaction families, training reactions and Benson-style group additivity values (GAV’s) and corrections for thermochemistry.39 5.) Create a new RMG mechanism for NG HTP, referred to hereafter as Mech v2, without relying on the Narayanaswamy reaction library, such that all PAH chemistry is derived from RMG’s database. Mech v2 consists entirely of either elementary steps or chemically-activated wellskipping steps automatically computed by RMG.35 The predictions of Mech v2 can be validated against the pilot plant measurements. This work represents the first attempt to predict PAH formation with RMG using all elementary reactions, for a natural gas high temperature pyrolysis process. The main advantage of a chemical mechanism consisting of all elementary steps, as opposed to one containing nonelementary, empirically-fitted steps, is twofold. First, it is easier to extrapolate such a mechanism to other reaction conditions because the real temperature (and sometimes pressure) dependence of each elementary step is already accounted for. Second, improvement of the mechanism is straightforward. For example, if a sensitive reaction is identified that relies on a general rate rule, the modeler can easily improve upon this estimate by performing a QM calculation of the reactants and transition state (TS) involved in the elementary reaction, using transition state theory (TST) to predict the high-pressure limit kinetics, and adding the new kinetic parameters to the RMG database in the form of a training reaction. For highly sensitive reactions, experimental measurements of the kinetics may be necessary as the accuracy of QM 9 ACS Paragon Plus Environment


Page 10 of 60

calculations (±0.5 kcal/mol at best) is often not sufficient for modeling purposes. The new training reaction will then be used in subsequent RMG jobs not only to describe the kinetics of the sensitive reaction in question, but will also be applied to analogous reactions. In this manner, both the specific mechanism of interest to the modeler improves with time, as does the RMG database, which will be used to construct mechanisms for other purposes. The following subsections describe the results obtained with Mech v1 and the improvements made to the RMG database as a result, specifically in relation to aromatic chemistry. The predictions of Mech v2 are then presented in the Results and Discussions sections. It is the hope of the authors that the methodology presented here, which allowed us to go from limited, empirical knowledge of a type of chemistry (aromatic chemistry in this case) to a detailed, physics-based mechanism, can serve as an example to future mechanism developers. Because the process being modeled consists of two distinct steps, burning of fuel for heat and cracking of hydrocarbons to produce C2’s, both Mech v1 and v2 were also generated in a twostep process, as shown in Figure 1. The mechanisms were initiated with a “seed mechanism” consisting of H2/O2 chemistry40. The first RMG job was run using the conditions of the burner, and relying on several libraries of small molecule thermochemistry that are not well-described by group-additivity estimates41, as well as a reaction library based on the Foundational Fuel Chemistry Model Version 1.0 (FFCM-1) for combustion of small hydrocarbon fuels (such as methane)42. The RMG-predicted burner exit composition is then mixed with 12 different cracking gas mixtures, representative of the pilot plant runs, in different ratios (“Low” versus “High” heat duty). A total of 24 different RMG jobs are spawned, each for a different cracking gas mixture, to simulate the chemistry in the cracking section. The cracker simulations also use 10 ACS Paragon Plus Environment

Page 11 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the same small-molecule thermochemistry and reaction libraries as the burner simulation. In the case of Mech v1, the Narayanaswamy reaction library is also used. Once all 25 mechanism construction jobs have converged, the resulting mechanisms are automatically merged into a single large mechanism such that overlapping reactions are not duplicated. Finally, if requested, libraries of aromatic reactions (or any reactions) can be appended onto the mechanism. This can be useful if the modeler wants to guarantee that certain chemistry is included in the mechanism that might have been overlooked by RMG’s flux-based criterion. The final merged (and possibly appended) mechanisms can then be used to simulate both the burner and cracker sections, for a variety of cracking gas mixtures. More detailed versions of Figure 1 specific to Mech v1 and v2 are included in the SI (Figures S1 and S2). The input files for a representative burner and cracker mechanism are also provided in the SI. Seed Mechanism Thermo Libraries Simulate Burner Conditions Reaction Libraries Iterate over 12 Cracking Gas Mixtures Mix in NG Cracking Gas

Simulate High Heat Duty Cracker conditions

…

Simulate Low Heat Cracker Duty conditions

Mix in NG + Additive* Cracking Gas

Simulate High Heat Duty Cracker conditions

Simulate Low Heat Duty Cracker conditions

Thermo Libraries Reaction Libraries

Merge Mechanisms

Append aromatic libraries (optional)

Overall Mechanism

Figure 1: General flow diagram for development of RMG mechanisms for 2-stage NG HTP. The composition of *additives are listed in Table 2. Burner conditions: P = 2.5 atm, T = 800-2900 K, stoichiometric fuel-air ratio, residence time = 5 ms. Cracker conditions: P = 1.5 atm, T = 14002200 K, residence time = 5 ms.



Page 12 of 60

3.1.1. Mech v1 Mech v1 consists of 797 species and 12,836 reactions. The agreement between the measurements and Mech v1 predictions of the pilot plant runs was deemed satisfactory (Figure S3 in SI) and sensitivity analysis was performed, Table 3 (see section 4.3. for a detailed description of how sensitivity was quantified). Briefly, Table 3 represents the most sensitive reactions with respect to the composition of the six major species (C2H2, C2H4, CO, CO2, CH4 and H2) and of four representative aromatics/aromatic precursors (CPD, benzene, indene and naphthalene), across all 71 simulated pilot plant runs. Remarkably, only 19 reactions (out of more than 10,000 reactions total in the mechanism) distinguish themselves as most sensitive across a wide range of pilot plant conditions. Importantly, 13 out of the 19 highly sensitive reactions are non-elementary, i.e., they are chemically activated well-skipping reactions43-52, with estimated rates. In various ways described below, previous mechanism developers converted these complex, multi-step chemical processes into simple, single-step effective reactions that empirically match some experiments. For the small molecules, almost all of the sensitive reactions (#1-9) come from the FFCM-1 reaction library42, including the three non-elementary reactions: #1, 6 and 7. The authors of FFCM-1 obtained kinetic parameters for reactions #1, 6 and 7 from low-pressure limit calculations on each surface43-45. As demonstrated by pressure-dependent calculations on these surfaces, for combustion temperatures and most relevant pressures ( 20,000, satisfying the highly-turbulent criteria for Eq. (2). Note that because Re is proportional to linear flow velocity, U, which is in turn proportional to the pilot plant heat duty (inlet flow of fuel and oxidizer and burner section), it is expected that h will roughly scale as (!"#$ %$&).'. Estimated values of ℎ at different heat duties according to Eq. (2) were then tuned to match the measured pilot plant heat loss in the burner and cracker sections for 42 of the 71 runs (quantified by the temperature rise of the cooling water). The ‘h’ estimate is unique to the current modeling study. For each heat duty, it was necessary to fit a separate ℎ for the burner (PSR1 and PFR1) and cracker (PFR2 and PFR3) sections (shown in Table S1 and S2 in supplementary information). The tuned cracker ℎ was well within the uncertainty of the estimate given in eq.(2), whereas the tuned burner ℎ was typically ~3-4 times higher, possibly due to significant radiative heat loss at the burner temperature (~3000 K). In reality the heat transfer coefficient would change across the length of the reactor. Computational Fluid Dynamic simulations would have to be performed to obtain the temperature profile along the reactor and the heat transfer coefficients determined accordingly, which is out of the scope of the current paper. Hence, a simplified methodology was used by fixing the heat transfer coefficients in the burner and the cracking gas section but there is scope of improvement in prediction of the heat transfer coefficients. Figure 8 compares the measured and predicted burner and overall (burner + cracker) heat loss after tuning ℎ . As mentioned earlier, heat loss is shown to increase with heat duty. All subsequent predictions used this model for heat loss.



Burner

Page 30 of 60

Overall (Burner + Cracker)

Increasing Heat Duty Increasing Heat Duty

Figure 8: Parity plots of normalized heat loss along burner and overall length of pilot plant reactor.

4. RESULTS AND DISCUSSION All of the results were generated using Mech v2 and using the reactor network shown in Figure 3. 4.1. Parity plots for CO, CO2, H2, C2H2, C2H4, CH4 The (dry) exit product concentration of major species for 71 runs were plotted against the pilot plant data as parity plots in Figure 9. The dashed line represents ±50 % margin for the measured concentration. The color of the symbols denotes the varied cracking gas compositions. The predicted concentrations of all the major species are within ± 50 % of the pilot plant data, which is a good metric for model performance. The predicted dry mole fraction of CH4, H2, CO and CO2 are close to the parity line (y=x). In other words, the model accurately captures the combustion of the fuel (CH4) and formation of major combustion intermediates (CO and CO2). The predicted concentration of methane for most of the runs is below the parity line and within 50 %, except for the runs where NG is used as cracking gas. If desired, this small discrepancy can be resolved by adjusting the rate parameters of a handful of 30 ACS Paragon Plus Environment

Page 31 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


reactions sensitive to CO and CO2 formation/decay within their uncertainty limits (#1, 2, 3, 6 and 8 shown in Table 5). Since the predicted methane concentration was within a factor of 2 for most of the runs, adjustments to the model were deemed unnecessary. The predicted concentration of acetylene for most of the runs was above the parity line and that for ethylene was below the parity line. However, the combined yield of acetylene and ethylene was close to the parity line (shown in Figure 10). As acetylene will be converted to ethylene at a later stage in the HTP process, the excellent agreement of the predicted total C2 yield with the pilot plant data, confirms the accuracy of our model for natural gas HTP. More accurately describing the distribution of C2’s between ethylene and acetylene will require a more detailed experimental data set against which to compare RMG’s predictions, such as the CH4 POX experiments of Kohler et al.14 This will be one of the objectives of a forthcoming chemistry-oriented publication.



Page 32 of 60

Figure 9. Parity plots of dry HTP reactor exit composition measured at the pilot plant and predicted by model.


Page 33 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 10. Parity plot of dry C2 exit composition (acetylene + ethylene) measured at the pilot plant and predicted by model. 4.2. C2 yield and lumped aromatic mass fraction vs. carbon number Both the predicted C2 and lumped aromatic yields at the exit of the HTP reactor for all 71 pilot plants runs are plotted as functions of overall equivalence ratio, φ, and cracking gas carbon number in Figure 11 and Figure 12, respectively. C2 yield is defined as the fraction of carbon atoms in the cracking gas that are incorporated into C2’s (either C2H2 or C2H4) at the exit of the HTP reactor. The lumped aromatic yield is simply the summed mass fraction of all of the aromatic compounds at the exit of the reactor (mostly consisting of phenylacetylene, benzene, phenylbutadiyne, styrene, indene, 1-phenylpropyne and naphthalene, in decreasing order of importance). Overall φ is defined as the ratio of the stoichiometric amount of O2 that would be needed to convert all of the hydrocarbons (both in the fuel and cracking gas) into CO2 and H2O, and the actual amount of O2 fed into burner section.



Page 34 of 60

Figure 11. Predicted C2 yield for 71 pilot plant runs as a function of overall equivalence ratio, φ, and cracking gas carbon number.

Figure 12. Predicted lumped aromatic yield for 71 pilot plant runs as a function of overall equivalence ratio, φ, and cracking gas carbon number. 34 ACS Paragon Plus Environment

Page 35 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The predicted C2 yield in Figure 11 has already been validated by comparisons to the actual pilot plant measurements in Figure 10. Interestingly, the C2 yield is always around 40-50%, largely independent of either overall φ or cracking gas carbon number. Unfortunately, there are no accurate pilot plant measurements against which to validate the lumped aromatic yields predicted in Figure 12. Nonetheless, the trends exhibited by the predictions are intuitively sensible: the lumped aromatic yield appears to have a positive correlation with both overall ϕ and cracking gas carbon number. In other words, at more fuelrich pyrolysis conditions with longer-chain cracking gases the propensity for aromatic formation is enhanced, entirely consistent with expectation. Of course, there are other variables to consider that are not captured by Figure 11, such as the degree of unsaturation of the cracking gas. For our current purposes, however, the two variables selected by intuition, overall ϕ and cracking gas carbon number, are sufficient to explain why certain pilot plant runs are expected to yield more aromatics than others. As mentioned in section 3.1.2, Mech v2 was constrained to species with 10 or less carbon atoms (e.g., naphthalene), mostly due to RMG’s current inability to accurately estimate the thermochemistry of large, fused polycyclic species.72 In order to approximate the uncertainty introduced by ignoring higher-order PAH, the Narayanaswamy mechanism, which includes PAH up to four rings, was used to simulate 38 of the 71 pilot plant runs (the other 33 runs could not be simulated because they contain additives not included in the Narayanaswamy mechanism). Figure S17 in the SI shows that the predicted aromatic yield increases by ~2 × when using the Narayanaswamy mechanism compared to Mech v2. However, both mechanisms predict the same qualitative dependence of aromatic yield on overall ϕ and cracking gas carbon number. 35 ACS Paragon Plus Environment


Page 36 of 60

4.3. Sensitivity Analysis Results Given the wide range of conditions simulated in this work (71 pilot plant runs), it was necessary to develop a standardized approach to quantifying sensitivity for two reasons: 1.) to ensure that simulations can be directly compared with one another, and 2.) to avoid being overwhelmed by data. This standardized approach is described below before presenting sensitivity results. Using the model described in section 3.2, sensitivity analysis was conducted on the exit concentration of the six major species measured at the pilot plant (C2H2, C2H4, CO, CO2, CH4 and H2), as well as of four representative aromatic/aromatic precursor species (CPD, benzene, indene and naphthalene). For a PFR, CHEMKIN reports time-integrated sensitivities of species concentrations to reaction rate coefficients. Therefore, sensitivity coefficients at the exit of either PFR2 or PFR3 (Figure 3) were used in this analysis, depending on in which PFR the species in question was mostly produced (PFR2 for C2H2, C2H4, CO, CO2, CH4, H2 and benzene; PFR 3 for CPD, indene and naphthalene). This choice of sensitivity coefficient is equivalent to perturbing each rate coefficient in the mechanism one-by-one, solving the PFR equations for the perturbed mechanism, and measuring the impact on the exit composition. The sensitivity coefficients obtained in this manner were further normalized by the maximum concentration of the species of interest along the PFR length. Normalizing by maximum rather than local concentration is preferred in this case, because many of the species (such as C2H2, Figure 5) have near-zero concentrations near the PFR entrance, which can result in arbitrarily large sensitivity coefficients if local normalization is used. For each of the ten species, the normalized, integrated sensitivity coefficients were ranked, and the top five from each of the 71 simulations were plotted against each other (Figures S4-S13 in SI). The ten resulting sensitivity plots (one for each species) could each include up to 355 unique 36 ACS Paragon Plus Environment

Page 37 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


reactions (71 simulations × 5 top sensitive reactions), but fortunately there was a great deal of overlap between simulations, and the actual plots included a manageable number of reactions, as seen in the SI. Furthermore, it was clear from these plots that certain reactions distinguish themselves as being extremely sensitive across a wide range of simulated conditions. It is these extremely sensitive reactions that are reported in Table 3 above (Mech v1) and Table 5 below (Mech v2). As shown, there is significant overlap even amongst the extremely sensitive reactions with respect to the different species, such that there are only a handful of important reactions for NG HTP across all 71 simulations. Subsequently, it is this handful of reactions that are worth deeper, future investigation. 4.3.1. Sensitivity quantified for key species All of the sensitive reactions for small molecules (#1, 4, and 6-8 in Table 5) were previously identified using Mech v1 (Table 3), and their kinetics are all from FFCM-142. The only exception is #6, OH + C2H2, for which pressure-dependent effects could be important at the pilot plant conditions, therefore a library of pressure-dependent kinetics for the OH + C2H2 system was used, based on the calculations of Senosiain et al.44 The effects of sensitive reactions #1-4 and 68 are generally straightforward. For example, #1 is a well-skipping reaction that converts a C1 species (methyl radical) to a C2 species (ethyl radical) and a hydrogen. Therefore, CH4 has a net negative sensitivity to #1, while the C2’s (C2H2 and C2H4) and H2 have a positive sensitivity to #1. Another important small-molecule reaction is #7, OH + CO ↔ H + CO2, which mostly runs in the cracking section, leading to positive sensitivity for CO and negative sensitivities for H2 and CO2. Unlike Mech v1, the sensitive reactions in Mech v2 for aromatics/aromatic precursor (#20-28) are all either elementary (#20, 22, 25, 27 and 28) or well-skipping based on master-equation 37 ACS Paragon Plus Environment


Page 38 of 60

simulations on an underlying elementary PES (#21, 23, 24, 26). None of the sensitive reactions #20-28 are empirically fitted and their kinetics can all be traced back to reliable sources. #20 and #21 produce CPD via vinyl radical + 1,3-butadiene and both come from route F in Table 4 (C6H9 PES). #22 and #23 are part of the pathway from propargyl radical recombination to benzene (C6H6 PES, route C). As expected, #22 and #23 are sensitive not just for benzene, but also for larger aromatics that depend on benzene formation (indene and naphthalene). #24 produces propyne that is later converted to propargyl radical (which can later recombine to form benzene), whereas #25-26 effectively consume propargyl radical by recombination with methyl radical to form a C4 that is eventually converted to either vinylacetylene or diacetylene. Finally #27 converts the adduct of phenyl + propargyl radicals into a radical that can undergo ring-closure to form indene (route H), and similarly #28 “activates” a closed-shell product of phenyl radical + vinylacetylene into a radical than can form naphthalene (route K). From this brief analysis of the most sensitive aromatic-related reactions the general outline of aromatic formation is already clear: benzene is formed through propargyl radical recombination and indene/naphthalene are formed through phenyl radical reaction with propargyl radical/vinylacetylene. ROP analysis of the most sooting run (NG + butane cracking gas) in the following section will confirm this understanding and provide some additional detail.


Page 39 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47


Table 5: Summary of most sensitive reactions for major species using Mech v2. + and – indicate net positive or negative sensitivity # 1 2 3 6 4 8 7

Sensitive Reactions

C2H2

C2H4

CO

CO2

H2

+ + + -

+ +

+

+

+

+

+ +

+

+ + -

Sensitive for: CH4 CPD Benzene

Indene Naphthalene

+ + +

-

+

20

+

21

+

22

+

+

+

23

+

+

+

24 25

+ -

26

-


Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Page 40 of 60

+

27

C

28

H

+

+


Page 41 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4.4. Rate of production (ROP) analysis of fuel and major species The ROP analysis was conducted for the maximum sooting run i.e., when butane was added as the cracking gas. The decay of methane in the burner section is shown in Figure S14 in the supplementary information. The consumption pathways for butane and formation of major single ring and two ring aromatics are discussed in this section. 4.4.1. Reaction path analysis of butane The ROP analysis of butane is shown in Figure 13 and was performed after addition of cracking gas (PFR2, ~20% of butane decay). The majority of butane undergoes hydrogen abstraction reactions to form 2-butyl radical. About 6% 1-butyl radical was also formed, which primarily undergoes beta-scission to form ethene and ethyl radical and minor amounts of 1-butene through loss of hydrogen atom (not shown in Figure 13). Minor amounts of acetylene and ethylene are also produced from decay of butane. 2-Butyl radical predominantly undergoes a C-C β-scission reaction to produce propene and methyl radical. Minor amounts of 1-butene are also formed from 2-butyl radical through loss of hydrogen atom.



Page 42 of 60

Figure 13. ROP analysis of butane in the cracking gas section (PFR2, ~20% decay of 1-butane). Secondary pathways with less than 5 % flux were neglected. Major products and aromatic precursors are shown by blue boxes. Pathways that are present in both reaction path analysis and sensitivity analysis are shown by the blue arrows. The ROP analysis of propene, which is one of the major products of butane is shown in Figure 14. Propene undergoes hydrogen abstraction reactions to form allyl and 1-propen-2-yl radical. Allyl radical undergoes hydrogen abstraction reactions to form propargyl radical. Minor amounts of acetylene are also produced from the decay of allyl radical. 1-Propen-2-yl radical reacts through a chemically-activated C-C scission reaction to form acetylene and also undergoes hydrogen abstraction reactions to form allene and propyne. The ROP analysis of other minor products from n-butane decay is shown in Figure S15 of the supplementary information. 1-Butene primarily decays through hydrogen abstraction reactions to 42 ACS Paragon Plus Environment

Page 43 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


form but-3-en-1-yl radical. About 45 % of but-3-en-1-yl radical undergoes a C-C scission to form ethylene and vinyl radical and another 45 % of but-3-en-1-yl radical undergoes subsequent hydrogen abstraction reactions to form 1,3-butadiene and vinylacetylene. Other significant pathways for decay of 1-butene include formation of allyl radical through C-C scission reaction and but-3-en-2-yl (C4H7(264)) through hydrogen abstraction reactions. But-3-en-2-yl radical undergoes hydrogen abstraction reactions to form 1,3-butadiene (Figure S15 in supplementary information). In summary, consumption of butane in the cracking gas section produces important products such as acetylene and ethylene and aromatic precursors such as propargyl and vinylacetylene. These aromatic precursors aided the formation of aromatics such as benzene, indene and naphthalene.

Figure 14. ROP analysis of propene in the cracking gas section (PFR2, 17% decay of 1-butane). Pathways with less than 15 % flux were neglected. Major products and aromatic precursors are shown by blue boxes.

4.4.3. Reaction path analysis of major single ring aromatics 43 ACS Paragon Plus Environment


Page 44 of 60

The total exit mass fraction of aromatics was predicted to be about 4 % for the maximum sooting run, i.e., when n-butane was added as the cracking gas. The mass fractions of the major single ring and two-ring aromatics predicted by Mech v2 are shown in Figure 15. The major single ring aromatics formed in the model are phenylacetylene, benzene, phenylbutadiyne, styrene, methylphenylacetylene and phenylvinylacetylene. The major two-ring aromatics consisted of only indene and naphthalene.

Figure 15. Predicted mass fraction of aromatics at the exit of the HTP reactor for the maximum sooting run, when butane was used as the the cracking gas. These aromatics accounted for 93 % of the total aromatics mass fraction. The reaction path scheme for formation of benzene is shown in Figure 16. The ROP analysis of benzene was performed where maximum concentration of benzene was attained (PFR2). From the reaction scheme it is evident that benzene is primarily formed from propargyl recombination pathway49. Phenyl radical is produced from benzene through hydrogen abstraction reactions. As shown in Table 4, the propargyl recombination pathway was added to the RMG database (Route C) and as a result led to the formation of benzene through an elementary reaction network, with 44 ACS Paragon Plus Environment

Page 45 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the isomers C6H6(294) and C6H6(303) as major intermediates. The sensitivity analysis results are in agreement with the ROP analysis, wherein benzene showed positive sensitivity to generation of C6H6(294) and C6H6(303) species (#22 and #23 in Table 5).

Figure 16. ROP analysis of benzene performed in the cracking gas section (PFR2, maximum concentration of benzene). Pathways with less than 5 % flux were neglected. Major products and aromatic precursors are shown by blue boxes. Reactions present in both sensitivity and ROP analysis are shown by blue arrows. The other major single ring aromatics phenylacetylene, phenylbutadiyne and styrene were formed from addition of phenyl radical to acetylene, diacetylene and propene respectively and their reaction schemes are shown in Figure 17. The major route for formation of phenylacetylene is addition of phenyl radical to acetylene, followed by H elimination. Phenylbutadiyne was formed from addition of phenyl radical to diacetylene and subsequent H loss. Accounting for the high-barrier trans-cis isomerization of C10H7(389) (Figure 2) prevented naphthalene formation from phenyl radical + diacetylene route. Diacetylene is formed from butane through subsequent hydrogen abstraction reactions. Styrene is formed from addition of phenyl radical to propene, 45 ACS Paragon Plus Environment


Page 46 of 60

followed by subsequent loss of methyl radical. Therefore, phenyl radical, acetylene, diacetylene and propene play a dominant role in the formation of major single ring aromatics.

Figure 17. ROP analysis of phenylacetylene, phenylbutadiyne and styrene. ROP analysis of phenylacetylene and styrene were performed at exit of the HTP reactor (PFR3, maximum concentration of species). ROP analysis of phenylbutadiyne was performed at the beginning of PFR3 (where the species concentration was maximum). Pathways with less than 10 % flux were neglected. Major products are shown by blue boxes. 4.4.4. Reaction path analysis of major two-ring aromatics The HTP model predicted that indene and naphthalene were the major two ring aromatics formed at the exit of the reactor. The reaction path analysis for indene is shown in Figure 18 and was performed at the exit of the HTP reactor. About 78 % indene was formed from phenyl and 46 ACS Paragon Plus Environment

Page 47 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


propargyl radical recombination route (route I in Table 4). Minor amounts of indene (~ 15%) were formed from addition of benzyl radical to acetylene (route I). The benzyl radical was formed from the phenyl radical + propene route, which is the new aromatic-catalyzed pathway (route J in Table 4).

Figure 18. ROP analysis of indene performed at the exit of the HTP reactor. Pathways with less than 10 % flux were neglected. Major products and aromatic precursors are shown by blue boxes. Reactions present in both sensitivity and ROP analysis are shown by blue arrows.

A simplified reaction path scheme for formation of naphthalene is shown in Figure 19. The complete reaction path scheme is present in Figure S16 of supplementary information. The ROP 47 ACS Paragon Plus Environment


Page 48 of 60

analysis of naphthalene was performed at the exit of the reactor, where maximum amount of naphthalene was formed.

About 59 % of naphthalene is formed from phenyl (C6H5) and

vinylacetylene (C4H4) addition network53 and 14 % of naphthalene is formed from hydrogen abstraction acetylene addition pathway (HACA, phenyl + 2 acetylene - H)60. Both of these reaction networks were added as training reactions to RMG prior to generating the final mechanism (Routes G and K in Table 4).

Figure 19. ROP analysis of naphthalene performed at the exit of the HTP reactor. Only the major pathways are shown. Major products and aromatic precursors are shown by blue boxes. Reactions present in both sensitivity and ROP analysis are shown by blue arrows. 48 ACS Paragon Plus Environment

Page 49 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


6. CONCLUSION Natural gas high temperature pyrolysis was modeled using a detailed chemical kinetic mechanism and a network of ideal reactors. The objective of the model was to predict the C2 and aromatic yield from the natural gas HTP process, for varied cracking gas compositions. The comprehensive natural gas mechanism developed in this work consists of additives up to hexane and a few alcohol species such as butanol and amyl alcohol. The mechanism was generated using the Reaction Mechanism Generator (RMG) software. The capability of RMG for aromatic chemistry was enhanced by adding 15 elementary reaction networks for formation of aromatics such as benzene, cyclopentadiene, indene and naphthalene. All of these reaction networks, except one, were based on quantum mechanical calculations in literature. The model was validated against pilot plant reactor data and the major species predictions (CO, CO2, H2, C2H2, C2H4 and CH4) were within ±50% of the pilot plant measurements for most of the runs. Importantly, the predicted C2 yield was very close to the experimental data for all runs. The effect of various cracking gas compositions on the C2 and aromatics yield was also determined. The C2 yield seems to be independent of either overall φ or cracking gas carbon number. On the other hand, the lumped aromatic yield appears to have a positive correlation with both overall φ and cracking gas carbon number. Sensitivity analysis was performed for all of the major species and aromatics for all of the 71 runs. Reaction path analysis was performed for the maximum sooting case, in which butane was fed as cracking gas. The model predicted noticeable yield of aromatics such as phenylacetylene, 49 ACS Paragon Plus Environment


Page 50 of 60

benzene, phenylbutadiyne, styrene, methylphenylacetylene and phenylvinylacetylene, indene and naphthalene, in decreasing order of importance. Benzene was primarily formed through propargyl radical recombination. Phenyl radical played a role in formation of indene and naphthalene through recombination reaction with propargyl and addition reaction with vinylacetylene, respectively. Propargyl radical and vinylacetylene were formed from decay of cracking gas (butane). All of the aromatics were formed through an elementary network of reactions, most of which were added as training reactions in RMG. The formation of aromatics in the HTP process is dictated by how readily the aromatic precursors such as propargyl radical are generated from the cracking gas and how hydrocarbon-rich (overall equivalence ratio) the cracker composition is.

7. SUPPORTING INFORMATION Mechanism generation methodology schematics, parity plots with the Narayanaswamy Mechanism, sensitivity analysis plots with Mech v2, ROP analysis of methane, 1-butene and naphthalene are provided in the supporting information. Mech v2 is also provided in Chemkin format.

8. AUTHOR INFORMATION Corresponding author *William H. Green. Email: [email protected]

9. ACKNOWLEDGEMENT The authors gratefully acknowledge SABIC for funding this study, and for providing the pilot plant data used to validate the model. 50 ACS Paragon Plus Environment

Page 51 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


10. REFERENCES 1. Siirola, J. J.; The impact of shale gas in the chemical industry. AIChE J. 2014, 60, 810-819 2. Mohr, S. H.; Evans, G. M. Long term forecasting of natural gas production. Energy Policy. 2011, 39, 5550-5560. 3. Cao, S.; Wang, D.; Wang, T. Simulation of partial oxidation of natural gas with detailed chemistry: influence of addition of H2, C2H6 and C3H8. Chem. Eng. Sci. 2010, 65, 26082618. 4. Lin, Y.; Abdelghani, M. Process for the production of ethylene from natural gas with heat integration. U. S. Patent 8,080,697 B2, December 20, 2011. 5. Berger, R. J.; Marin, G. B. Investigation of gas-phase reactions and ignition delay occurring at conditions typical for partial oxidation of methane to synthesis gas. Ind. Eng. Chem. Res. 1999, 38, 2582-2592. 6. Howard, W. B. Production of acetylene. U. S. Patent 3,242,224, March 22, 1966. 7. Holmen, A.; Rokstad, O. A.; Solbakken, A. High temperature pyrolysis of hydrocarbons. 1. Methane to acetylene. Ind. Eng. Chem. Process Des. Dev. 1976, 15, 439-444. 8. Holmen, A.; Rokstad, O. A.; Solbakken, A. High temperature pyrolysis of hydrocarbons. 2. Naphtha to acetylene. Ind. Eng. Chem. Process. Dev. 1979, 18, 653-657. 9. Pechtold, N.; Wimmer, R., Wirtz, R. Process for cracking hydrocarbons. U. S. Patent 3,240,836, March 15 1966. 10. Zhang, Q.; Wang, J., Wang, T., Enhancing the acetylene yield from methane by decoupling oxidation and pyrolysis reactions: A comparison with partial oxidation process. Ind. Eng. Chem. Res. 2016, 55, 8383-8394.



Page 52 of 60

11. Al-Hamamre, Z.; Voz, S.; Trimis, D. Hydrogen production by thermal partial oxidation of hydrocarbon fuels in porous media based reformer. Int. J. Hydrogen Energy, 2009, 34, 827-832 12. Liu, Y.; Wang, T. Analysis of the autoignition process under the industrial partial oxidation conditions using detailed kinetic modeling. Ind. Eng. Chem. Res. 2011, 50, 6009-6016 13. Li, Z.; Dai, Z.; Guo, Q.; Liang, Q.; Wang, F.; Experimental and numerical study of MILD combustion in a bench-scale natural gas partial oxidation gasifier. Fuel, 2017, 193, 197-205 14. Kohler, M.; Oswald, P.; Xu, H.; Kathrotia, T.; Hasse, C.; Riedel, U. Speciation data for fuel-rich methane oxy-combustion and reforming under prototypical partial oxidation conditions. Chem. Eng. Sci. 2016, 139, 249-260 15. Skjoth-Rasmussen, M. S.; Glarborg, P.; Ostberg, M.; Johannessen, J. T.; Livbjerg, H.; Jensen, A. D.; Christensen, T. S. Formation of polycylic aromatic hydrocarbons and soot in fuel-rich oxidation of methane in a laminar flow reactor. Combust. Flame, 2004, 136, 91128 16. Glarborg, P.; Bentzen, L. L. B. Chemical effects of a high CO2 concentration in oxy-fuel combustion of methane. Energy Fuels. 2008, 22, 291-236 17. Li, C.; Burke, N.; Gerdes, K.; Patel, J. The undiluted, non-catalytic partial oxidation of methane in a flow tube reactor – An experimental study using indirect induction heating. Fuel, 2013, 109, 409-416 18. Li, Q.; Wang, T.; Liu, Y.; Wang, D. Experimental study and kinetics modeling of partial oxidation reactions in heavily sooting laminar premixed methane flames. Chem. Eng. J. 2012, 207-208, 235-244


Page 53 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


19. Watanabe, H.; Arai, F.; Okazaki, K. Role of CO2 in the CH4 oxidation and H2 formation during fuel-rich combustion in O2/CO2 environments. Combust. Flame. 2013, 160, 23752385 20. Turbiez, A.; El Bakali, A.; Pauwels, J. F.; Rida, A.; Meunier, P. Experimental study of a low pressure stoichiometric premixed methane, methane/ethane, methane/ethane/propane and synthetic natural gas flame. Fuel, 2004, 83, 933-941 21. Stelzner, B.; Weis, C.; Habisreuther, P.; Zarzalis, N.; Trimis, D. Super-adiabatic flame temperatures in premixed methane flames: A comparison between oxy-fuel and conventional air combustion. Fuel. 2017, 201, 148-155 22. Ktalktherman, M. G.; Namyatov, I. G.; Pyrolysis of hydrocarbons in a heat-carrier flow with fast mixing of the components. Combust. Explos. Shock Waves. 2008, 44, 529-534 23. Ktalktherman, M. G.; Namyatov, I.; Emel’kin V. High-temperature pyrolysis of liquefied petroleum gases in the fast-mixing reactor. Int. J. Chem. React. Eng. 2011, 9, 1-23 24. Ktalktherman, M. G.; Namyatov, I. G.; Emel’kim V. A. High-selective pyrolysis of naphtha in the fast-mixing reactor. Fuel Process. Technol. 2013, 106, 48-54 25. Ktalktherman, M. G.; Namyatov, I. G. Simulation of hydrocarbons pyrolysis in a fastmixing reactor. Chin. J. Chem. Eng. 2015, 23, 941-953 26. Zhang, Q.; Wang, J.; Wang, T. Effect of ethane and propane addition on acetylene production on the partial oxidation process of methane. Ind. Chem. Eng. Res. 2017, 56, 5174-5184 27. Smith, G. P.; Golden, D. M.; Frenklach, M.; Moriarty, N. W.; Eiteneer, B.; Goldenberg, C. M.; Bowman, T.; Hanson, R. K.; Song, S.; Gardiner, Jr. W. C.; Lissianski, V. V.; Qin. Z. http://www.me.berkeley.edu/gri_mech/ 53 ACS Paragon Plus Environment


Page 54 of 60

28. Hughes, K. J.; Turanyi, T.; Clague, A. R.; Pilling, M. J. Development and testing of a comprehensive chemical mechanism for the oxidation of methane. Int. J. Chem. Kinet. 2001, 33, 513-538 29. Petersen, E. L.; Davidson, D. F.; Hanson, R. K. Kinetics modeling of shock-induced ignition in low-dilution CH4/O2 mixtures at high pressures and intermediate temperatures. Combust. Flame, 1999, 117, 272-290 30. Konnov, A. A.; Zhu, J. N.; Bromly, J. H.; Zhang, D. Noncatalytic partial oxidation of methane into syngas over a wide temperature range. Combust. Sci. Tech., 2004, 176, 10931116 31. Healy, D.; Curran, H. J.; Simmie, J. M.; Kalitan, D. M.; Zinner, C. M.; Barrett, A. B.; Petersen, E. L.; Bourque, G. Methane/ethane/propane mixture oxidation at high pressures and high, intermediate and low temperatures. Combust. Flame, 2008, 155, 441-448 32. Healy, D.; Curran, H. J.; Dooley, S.; Simmie, J. M.; Kalitan, D. M.; Petersen, E. L.; Bourque, G. Methane/ethane/propane mixture oxidation at high pressures and high, intermediate and low temperatures. Combust. Flame, 2008, 155, 451-461 33. Healy, D.; Kalitan, D. M.; Aul, C. J.; Petersen, E. L.; Bourque, G.; Curran, H. J. Oxidation of C1-C5 alkane quinternary natural gas mixtures at high pressures. Energy Fuels. 2010, 24, 1521-1528 34. Bourque, G.; Healy, D.; Curran, H.; Zinner, C.; Kalitan, D.; Vries, J.; Aul, C.; Petersen, E. Ignition and flame speed kinetics of two natural gas blends with high levels of heavier hydrocarbons. J. Eng. Gas Turb. Power. 2010, 132, 021504-1-11


Page 55 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


35. Gao, C. W.; Allen, J. W.; Green, W. H.; West, R. H. Reaction Mechanism Generator: Automatic construction of chemical kinetic mechanisms. Comput. Phys. Commun. 2016, 203, 212-225. 36. Narayanaswamy, K.; Blanquart, G.; Pitsch, H. A consistent chemical mechanism for oxidation of substituted aromatic species. Combust. Flame 2010, 157, 1879-1898. 37. Blanquart, G.; Pepiot-Desjardins, P.; Pitsch, H. Chemical mechanism for high temperature combustion of engine relevant fuels with emphasis on soot precursors. Combust. Flame 2009, 156, 588-607. 38. Wang, H.; You, X.; Joshi, A. V.; Davis, S. G.; Laskin, A.; Egolfopoulos, F.; Law, C. K. USC Mech Version II. High-Temperature Combustion Reaction Model of H2/CO/C1-C4 Compounds. http://ignis.usc.edu/USC_Mech_II.htm, May 2007. 39. Benson, S. W., Thermochemical Kinetics. John Wiley and Sons: New York, 1968. 40. Burke, M. P.; Chaos, M.; Ju, Y.; Dryer, F. L.; Klippenstein, S. J. Comprehensive H2/O2 kinetic model for high-pressure combustion. Int. J. Chem. Kinet. 2012, 44, 444-474. 41. Goldsmith, C. F.; Magoon, G. R.; Green, W. H. Database of Small Molecule Thermochemistry for Combustion. J. Phys. Chem. A 2012, 116, 9033-9057. 42. Smith, G. P.; Tao, Y.; Wang, H. Foundational Fuel Chemistry Model Version 1.0 (FFCM1), http://nanoenergy.stanford.edu/ffcm1, 2016. 43. Stewart, P. H.; Larson, C. W.; Golden, D. M. Pressure and temperature dependence of reactions proceeding via a bound complex. 2. Application to 2CH3 → C2H5 + H. Combust. Flame 1989, 75, 25-31. 44. Senosiain, J. P.; Klippenstein, S. J.; Miller, J. A. The Reaction of Acetylene with Hydroxyl Radicals. J. Phys. Chem. A 2005, 109, 6045-6055. 55 ACS Paragon Plus Environment


Page 56 of 60

45. Joshi, A. V.; Wang, H. Master equation modeling of wide range temperature and pressure dependence of CO + OH → products. Int. J. Chem. Kinet. 2006, 38, 57-73. 46. Knyazev, V. D.; Slagle, I. R. Kinetics of the Reaction between Propargyl Radical and Acetylene. J. Phys. Chem. A 2002, 106, 5613-5617. 47. Roy, K. Kinetische Untersuchung zur Hochtemperaturpyrolyse und -oxidation von Cyclopentadien und Cyclopentadienyl mit Hilfe der Stosswellentechnik. Stuttgart University, Germany, 1999. 48. Sharma, S.; Green Jr, W. H. Computed Rate Coefficients and Product Yields for c-C5H5 + CH3 → Products. J. Phys. Chem. A 2009, 113, 8871-8882. 49. Miller, J. A.; Klippenstein, S. J. The Recombination of Propargyl Radicals and Other Reactions on a C6H6 Potential. J. Phys. Chem. A 2003, 107, 7783-7799. 50. Kislov, V. V.; Mebel, A. M. Ab Initio G3-type/Statistical Theory Study of the Formation of Indene in Combustion Flames. I. Pathways Involving Benzene and Phenyl Radical. J. Phys. Chem. A 2007, 111, 3922-3931. 51. Aguilera-Iparraguirre, J.; Klopper, W. Density Functional Theory Study of the Formation of Naphthalene and Phenanthrene from Reactions of Phenyl with Vinyl- and Phenylacetylene. J. Chem. Theory Comput. 2007, 3, 139-145. 52. Kislov, V. V.; Islamova, N. I.; Kolker, A. M.; Lin, S. H.; Mebel, A. M. Hydrogen Abstraction Acetylene Addition and Diels−Alder Mechanisms of PAH Formation: A Detailed Study Using First Principles Calculations. J. Chem. Theory Comput. 2005, 1, 908-924.


Page 57 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


53. Mebel, A. M.; Landera, A.; Kaiser, R. I. Formation Mechanisms of Naphthalene and Indene: From the Interstellar Medium to Combustion Flames. J. Phys. Chem. A 2017, 121, 901-926. 54. Park, J.; Nam, G. J.; Tokmakov, I. V.; Lin, M. C. Experimental and Theoretical Studies of the Phenyl Radical Reaction with Propene. J. Phys. Chem. A 2006, 110, 8729-8735. 55. Kislov, V. V.; Mebel, A. M.; Aguilera-Iparraguirre, J.; Green, W. H. Reaction of Phenyl Radical with Propylene as a Possible Source of Indene and Other Polycyclic Aromatic Hydrocarbons: An Ab Initio/RRKM-ME Study. J. Phys. Chem. A 2012, 116, 4176-4191. 56. Buras, Z. J.; Chu, T.-C.; Jamal, A.; Yee, N. W.; Middaugh, J. E.; Green, W. H. Phenyl Radical + Propene: a Prototypical Reaction Surface for Aromatic-Catalyzed 1,2-HydrogenMigration and Subsequent Resonance-Stabilized Radical Formation. Phys. Chem. Chem. Phys. 2018, DOI: 10.1039/C8CP01159A 57. Merchant, S. S. Molecules to Engines: Combustion Chemistry of Alcohols and their Application to Advanced Engines. MIT, Cambridge, 2015. 58. Buras, Z. J.; Dames, E. E.; Merchant, S. S.; Liu, G.; Elsamra, R. M. I.; Green, W. H. Kinetics and Products of Vinyl + 1,3-Butadiene, a Potential Route to Benzene. J. Phys. Chem. A 2015, 119, 7325-7338. 59. Wang, K.; Villano, S. M.; Dean, A. M. Reactions of Allylic Radicals that Impact Molecular Weight Growth Kinetics. Phys. Chem. Chem. Phys. 2015, 17, 6255-6273. 60. Mebel, A. M.; Georgievskii, Y.; Jasper, A. W.; Klippenstein, S. J. Temperature- and pressure-dependent rate coefficients for the HACA pathways from benzene to naphthalene. Proc. Combust. Inst. 2017, 36, 919-926.



Page 58 of 60

61. Mebel, A. M.; Georgievskii, Y.; Jasper, A. W.; Klippenstein, S. J. Pressure-dependent rate constants for PAH growth: formation of indene and its conversion to naphthalene. Faraday Discuss. 2016, 195, 637-670. 62. Matsugi, A.; Miyoshi, A. Computational study on the recombination reaction between benzyl and propargyl radicals. Int. J. Chem. Kinet. 2012, 44, 206-218. 63. Ismail, H.; Park, J.; Wong, B. M.; Green Jr, W. H.; Lin, M. C. A theoretical and experimental kinetic study of phenyl radical addition to butadiene. Proc. Combust. Inst. 2005, 30, 1049-1056. 64. Mebel, A. M.; Kislov, V. V.; Kaiser, R. I. Photoinduced Mechanism of Formation and Growth of Polycyclic Aromatic Hydrocarbons in Low-Temperature Environments via Successive Ethynyl Radical Additions. J. Am. Chem. Soc. 2008, 130, 13618-13629. 65. Shukla, B.; Koshi, M. A highly efficient growth mechanism of polycyclic aromatic hydrocarbons. Phys. Chem. Chem. Phys. 2010, 12, 2427-2437. 66. Faravelli, T.; Bua, L.; Frassoldati, A.; Antifora, A.; Tognotti, L.; Ranzi, E. A new procedure for predicting NOx emissions from furnaces. Comput. Chem. Eng. 2001, 25, 613-618. 67. Fichet, V.; Kanniche, M.; Plion, P.; Gicquel, O. A reactor network model for predicting NOx emissions in gas turbines. Fuel. 2010, 89, 2202-2210. 68. Monaghan, R. F. D.; Ghoneim A. F. A dynamic reduced order model for simulating entrained flow gasifiers. Part I. Model development and description. Fuel. 2012, 91, 61-80. 69. CHEMKIN 15151, Reaction Design, San Diego, 2015 70. Sieder, E. N.; Tate, G. E. Heat Transfer and Pressure Drop of Liquids in Tubes. Industrial & Engineering Chemistry 1936, 28, 1429-1435. 58 ACS Paragon Plus Environment

Page 59 of 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


71. Bird, R. B.; Stewart, W. E.; Lightfoot, E. N. Transport Phenomena. 2nd ed.; John Wiley & Sons, Inc.: New York, 2006. 72. Han, K.; Jamal, A.; Grambow, C. A.; Buras, Z. J.; Green, W. H. An Extended Group Additivity Method for Polycyclic Thermochemistry Estimation. Int. J. Chem. Kin. 2018, 50, 294-303.



Page 60 of 60

Table of Contents (TOC)/Abstract graphic


Modeling Study of High Temperature Pyrolysis of Natural Gas

Recommend Documents