Experimental Investigation of Reaction ... - ACS Publications

Article pubs.acs.org/EF

Experimental Investigation of Reaction Confinement Effects on Coke Yield in Coal Pyrolysis Jeffrey LeBlanc,† John Quanci,‡ and Marco J. Castaldi*,† †

Combustion and Catalysis Laboratory, Department of Chemical Engineering, The City College of New York, City University of New York, New York, New York 10030, United States ‡ Suncoke Energy Inc., 1011 Warrenville Road, Sixth Floor, Lisle, Illinois 60532, United States S Supporting Information *

ABSTRACT: The pyrolysis of agglomerating coal was analyzed by thermogravimetric analysis (TGA) coupled to microgas chromatography (μGC) to determine the effects of reactor confinement on solid product yield, tar evolution, and gas composition. The primary volatile products generated from pyrolysis are studied using two TGA crucibles with height to width aspect ratios of 0.11:1.0 and 2.0:1.0 for heating rates of 1, 3, and 10 K min−1. Mass balances were determined from measurements of the solid residual, gaseous flow rates, and tar products captured via glass impingers. The measurements resulted in mass balance closures greater than 99%. The higher aspect ratio confinement provided a zone where the residence time of volatile species was extended to 0.35 s from 0.04 s for the low aspect ratio confinement. The extended residence time was found to increase the solid yield by 0.6−5.7% for the low and mid ranged volatile material coals by secondary tar reactions which form coke. There was a trend of increasing solid yield with a decreasing heating rate because of the shorter residence time of the released volatile material at the higher heating rates. A decrease in tar production by 2.1−2.5% was observed in the higher aspect ratio confinement. The differences in product evolution between the two confinements were determined to be due to reactions occurring in the range of 783−848 K, which produce hydrogen, methane, and coke. In the confinement of higher aspect ratio, hydrogen production increased by 10% and 30% for low and mid VM coals, respectively, along with a 40% increase in methane production for both coals. The higher productions of methane, hydrogen, and solid residual between 820 and 1100 K for increased residence time of 0.27−0.35 s are due to demethylation and recombination reactions during the formation of char from recondensed tar products.

■

INTRODUCTION Global steel production is forecasted to grow at an average rate of 1% year−1 until 2020. In those years, iron blast furnaces will rely on metallurgical coke for operation and remain the primary route for iron making. In the last 20 years, the coke making industry has seen the establishment of the largest, most productive, and lowest emission coke-making plants in history.1 The development of horizontal heat recovery ovens offers the added benefit of heat recovery from the volatile products and low emissions, thus being more economically efficient and environmentally friendly. However, traditional vertical byproduct coke ovens have slightly higher solid product yields. It is expected that the secondary reactions of escaping volatile products at the solid−gas interfaces are responsible for the improved yield. Therefore, it is desirable to investigate the effects of confinement reflecting the different coke oven geometries on secondary reactions to increase the coke yield from horizontal heat recovery configurations. Coal science pertaining to coke making has been studied for 50 years and has unveiled a substantial understanding of the coal pyrolysis mechanism.2−29 It is accepted that reactions within the coal particle or bed are dominated by free-radical pathways.30 Yet, the mechanistic and empirical understanding of free radical mechanisms in coal pyrolysis is not sufficient to predict tar and gaseous product distributions accurately for the current environmental and heat recovery optimization requirements.3 Therefore, models that predict depolymerization and © 2016 American Chemical Society

devolatilization of coal on the basis of the coal’s molecular constituents have been built.3−10 Mechanistically free radicals are generated in two ways, from homolytic breaking of covalent bonds that form two radicals, shown in eq 1, and by β-scission forming smaller radicals and/or unsaturated bonds, shown in eq 2.30 ̇ 2 + CH ̇ 2CH3 R−CH 2CH 2CH3 ↔ R−CH

(1)

R−CH 2CH 2CH3 ↔ RCH 2 + ·C2H5

(2)

Between temperatures of 673 and 823 K, the coal reaches a semimelted state called the Metaplast. In the condensed phase of the Metaplast, hydrocarbons decompose through free-radical mechanisms producing fragments of coal molecules.30−32 Evaporation of volatile fragments formed in the Metaplast governs the rate of tar devolatilization and impacts solid yield.4 Char formation occurs simultaneously in the Metaplast state, which produces carbon dioxide, water, and methane as byproducts. The pool of H, CH3, and C2H5 radicals developed from free radical abstraction (removal of an atom or group from a molecule by a radical) and β-scission reactions in the metalplast reacts to form hydrogen and light hydrocarbons.6,11,33 The end of tar production is determined by Received: March 29, 2016 Revised: July 14, 2016 Published: July 15, 2016 6249

DOI: 10.1021/acs.energyfuels.6b00699 Energy Fuels 2016, 30, 6249−6256

Article

Energy & Fuels the availability of hydrogen donors in the condensed phase.10 Resolidification of the Metaplast yields methane, hydrogen, and coke via condensation reactions between aromatic clusters. Volatile products released from the condensed phase are presumed to continue to react via free-radical mechanisms while diffusing through the particle bed and/or near hot solid− gas interfaces. The impact of the bed diffusion lengths (i.e., confinement) on tar yield has been speculated,10 yet never quantified or linked to a commensurate difference in solid and/ or gas product. The use of various confinements to elucidate secondary pyrolysis in the gas phase and solid−gas interface were first presented in 2013.13 To our knowledge, the technique has not been used to explain the rate and nature of secondary reactions of volatile species, which increase solid yield. Previous experiments investigating these reactions involved varying pressure in wire mesh heating experiment.14,23 In earlier works, researchers varied the gas flow through a packed bed.34,35 The conclusion was that increases in ambient gas pressure at constant volatile residence time had little effect on pyrolysis yields, but changes in volatile residence time at constant pressure were very important. Relevant classes of thermochemical secondary reactions for volatiles diffusing through a coal bed are shown below. The cleaving of alkyl functional groups from tar products (where R is an aromatic moiety) and the breaking of a C−C bridge between aromatic moieties are important steps in secondary reactions. ̇ 2 + R′−CH ̇ 2 R−CH 2CH 2−R′ ↔ R−CH

product distribution data of volatile species stagnated in the solid−gas interface for various residence times. In this study, three United States agglomerating coals used in commercial coke making are analyzed by TG−GC to better understand the relationship between the residence times of volatile species for heating rates relevant for coke production, 1−10 K min−1. The variation of the geometric aspect ratio of the pyrolysis confinement enables the investigation of solid product yield dependence with the residence time of volatile products immediately above the reacting solid. The investigation includes measuring the tar and gas products at different residence time to discern the secondary reactions. The residence time variation provides unique insight on the product formation and the gaseous and tar compositions.

■

Coals. The three agglomerating coals in this study, obtained by Suncoke Energy, are used in the commercial production of metallurgical coke. The proximate analyses and carbon percent of the coals are shown in Table 1. The principal difference between the

Table 1. Proximate Analysis on a Dry Basis and Ranks of Three Agglomerating Coals low VM mid VM high VM

(3)

(4)

Also, demethylation by abstraction may occur to form methane. R−CH3 + R′−H ↔ R−R′ + CH4

VM

ash

FC

C%

18.93 28.02 38.94

7.332 7.446 7.107

73.73 64.53 53.95

84.42 82.32 79.58

chosen coals is the composition of the volatile matter (VM). Therefore, the coals in this work are referred to as low, mid, and high VM coal. The coals vary widely in VM; however, they only range between 79.6% and 84.4% carbon. The samples were tested asreceived and not demineralized. The impact from minerals is negligible with regard to the investigation. The samples were crushed via a mortar and pestle and sieved to a uniform particle size of 44 ± 7 μm. The samples were analyzed as a 6.5 mm × 1 mm pellet formed using a press and dye system. TG−GC and GC/MS System. The experimental setup used for investigation was a Netzsch Luxx simultaneous thermal analyzer 409PC close coupled to an Inficon 3000 μGC, equipped with a thermal conductivity detector (TCD). This system is shown in Scheme 1. The coal sample size was 40 mg. A vertically oriented sweep

Reactions portrayed in eqs 1, 2, and 3 form aromatic radicals, which may proceed through bimolecular recombination. In this step, hydrogen may be formed by condensation or graphitization. R−H + R′−H ↔ R−R′ + H 2

MATERIALS AND METHODS

(5)

Equations 4 and 5 could be modes of forming polynuclear aromatic hydrocarbons (PAHs) or large graphitized coke species. During chemical growth, PAH may form dimers, trimers, and tetramers by collision without producing a gas species.36 PAH growth may cause char species to recondense from the gas phase into the solid product. These secondary reactions are recognized to be important for determining product distribution between solid, liquid, and gas species from coal pyrolysis. An investigation of secondary reactions of volatile species occurring at the solid−gas interface using various pyrolysis confinements will allow the products of these reactions to be measured. Furthermore, a scaling factor may be derived to assess the extent of secondary reactions of escaping volatile products at the solid−gas interfaces in commercial coke ovens leading to possible design modifications to increase the solid yield. Two different confinements were utilized to provide insight into the secondary reactions as a function of increased residence time in the solid−gas interface. In a confinement of smaller aspect ratio (0.11:1.0), rapid removal of volatile species is favored, leading to minimal secondary reactions. The volatile species in a confinement with a larger aspect ratio (2.0:1.0) were not immediately carried away by the inert sweep gas, and therefore experienced increased residence time at the solid−gas interface. The literature does not have TG−GC, GC−MS, or

Scheme 1. Schematic of TG−GC System Used for Pyrolysis Study

gas of research grade argon (Airgas AR R300) was maintained constant at 30 mL min−1. The effluent of the TGA was connected to a quarter-inch passivated (Silcotek SilcoNert2000) stainless steel tube heated to 548 K constituting an inert, heated sample transfer line and has been shown to have no impact on the sample.37 The heated transfer line routed the TGA effluent to a series of impingers (SKC Midget Impinger, 225-35-1), which served as a condensation train 6250


Article

Energy & Fuels operating at 273 K and 1.0 bar. Any uncondensed gaseous effluent that exited the train was analyzed by the μGC−TCD. Gas sampling from the TGA effluent stream was performed at a frequency of 3 min, providing very good temperature resolution of the evolved chemical species. The gaseous species that were identified and quantified were hydrogen, methane, ethylene, and ethane. The 14 m OV-1 column of the Inficon 3000 micro-GC is capable of measuring C3 and C4 order hydrocarbons, but these species either are not present in the gas product or are below detection limits of 100 ppm. The solid coal samples were subjected to a constant heat rate of 1, 3, and 10 K min−1 up to 1373 K in the TGA. Commercial coke-making processes operate at a temperature up to 1400 K.1 For each test, the temperature was held isothermally at 1373 K until the differential weight loss (DTG) was reached and remained at zero. Tar collected from the series of impingers was measured gravimetrically and subsequently dissolved in 20 mL of acetone to prepare a solution for chemical characterization. A 2 μL solution was injected into an Agilent 7890B gas chromatography instrument (GC), equipped with as an Agilent 5977A mass selective detector (MSD). An Agilent HP-5MS (30 m × 250 μm ID × 0.25 μm film thickness) column was used for separation of the tar compounds and directly connected to the MS. The GC oven was held at 303 K for 18 min, and then heated at 2 K min−1 ramp to 493 K with three 10 min isotherms at 373, 403, and 453 K and a 30 min isotherm at 493 K. The MS ion source temperature was maintained at 473 K. The Agilent Chemstation software was used to identify the chemical compounds throughout the chromatograms by matching the mass spectra of the sample with the National Institute of Standards and Technology (NIST) database. Confinements. Two different confinements, shown in Scheme 2, were utilized to observe how volatile species interact as a function of

heating rate, and confinement to obtain representative statistical values. A statistical t test using a 95% confidence interval was used to determine significant difference between the two conditions. The pvalue is used to show the significance level; a lower p-value means greater significance. The averages and p-values for gaseous products reported are based on nine repetitions at 3 K min−1 for each confinement. The reported p-value is the resultant value from a statistical t test between the pan and cup confinement. Mass Balance. Excellent work previously reported in the literature performed on coal pyrolysis reached a 95% mass closure on pyrolysis products in an encapsulated pyrolysis unit.14 Similar balances were obtained in an entrained flow reactor for more rapid heating conditions.15 However, in many other reports on coal pyrolysis, the tar and water yields were determined by difference (i.e., not measured), and more often the mass balance was not reported. In this work, the masses of solid, condensable (tar and water), and gaseous products were measured and close to 99.2% or better for three agglomerating coals. The procedure used is identical to that used in recently published work.38 The solid product was directly measured as the residual mass in the TGA. The volumetric rate of inert gas was measured via a flow meter (Alicat M-100). The volumetric rate of the gas produced during the pyrolysis reactions was calculated using the μGC measurements and argon as an internal standard. The gas sampling was performed downstream of the condenser at normal temperature and pressure. The gas densities at normal temperature and pressure were used to calculate the total mass. Tar, tar aerosols, and water were collected in the condensation unit and measured gravimetrically. The measurement of condensable yield fluctuated by 30−50% depending on coal with the high VM coal, resulting in the greatest variability. Reviews of similar processes for efficient on-stream tar capture demonstrated the use of condenser trains.39−43 A cooling rate of 150 K s−1 immediately downstream of the heated and insulated transfer line was determined by iteration and found to be critical for tar capture. The impinger design was such that the internal tube ended in a nozzle so that the tar products impact the chilled inner wall of the glass tube container. A solvent was not used for tar capture in this work. The use of solvents to capture tar is recommended to suppress polymerization and oxidation reactions among captured tar species; yet this was determined to be unnecessary in our system due to negligible reaction extent at 273 K.39 The main disadvantage of using a solvent is the step of solvent removal, which adds uncertainty to the measured yields and composition. The average product distributions for three agglomerating coals are given in Table 2. These product

Scheme 2. (Left) Short (Pan) Confinement Used in TGA Pyrolysis Experiments; and (Right) Tall (Cup) Confinement Used in TGA Pyrolysis Experiments

Table 2. Pyrolysis Product Distribution for Three Coals Heated at 3 K min−1

increased residence time above the reacting solid residual. These configurations provide insight into the secondary reactions during devolatilization. The confinement with a smaller aspect ratio of 0.11:1.0 (14.3 mm in diameter and 1.6 mm in height), designated as the “pan”, was designed to favor rapid removal of devolatilization products leading to minimal secondary reactions. The confinement with a larger aspect ratio of 2.0:1.0 (6.5 mm in diameter and 13 mm in height), designated as the “cup”, creates a zone above the sample where the volatile species evolving are not swept away by the sweep gas. The region above the sample creates the opportunity for secondary reactions to affect the product distribution. The Biot numbers of each confinement are between 0.04 and 0.11, indicating a uniform temperature distribution (more description in the Supporting Information). This means the solid−gas interface reactions monitored are occurring at the sample temperature. The space above the sample in the cup confinement was 10 mm, and the approximated residence time is 0.35 s (calculation shown in Supporting Information). It is important to consider that the reaction of volatile species does not only occur in the crucible while stagnated above the sample, but also in space outside of the crucible, and both reaction sets determine the amounts of each gas species observed by the GC. However, reactions in space outside the crucible have equal possibility for experiments with each crucible; that is, the main difference is residence time near solid−gas interfaces. The tests were repeated in triplicate for each coal,

low VM mid VM high VM

solid (%)

tar and water (%)

gas (%)

78.6 ± 2.3 77.4 ± 1.7 68.3 ± 1.5

11.8 ± 2.9 15.2 ± 3.8 25.5 ± 7.4

9.4 ± 1.6 7.3 ± 0.79 5.5 ± 0.68

distributions were compared to other works with high mass closure who reported cumulative tar yields between 20% and 32% for bituminous and subbituminous coals15−17 and cumulative condensable yields of 18.3−23.7% for lignite.14 Tar Evolution Profile. The collection of tar via the impingers gave an overall aggregate mass that could only be measured at the completion of each test. By measuring the distribution between solid, tar, and gas, the aggregate may be divided to reflect its time-dependent evolution. This was done using the gas and solid residual measurements to determine the amount of tar released at the corresponding temperatures, which can also be representative of time through heating rates utilized. rtar(t ) = rsolid(t ) − rgas(t )

(6)

where rsolid is the rate of solid consumption (measured by DTG), rgas is the rate of gas production (measured by micro-GC), and rtar is the rate of tar production (by difference), which are all known as a function of 6251


Article

Energy & Fuels temperature. The critical aspect is that the total amount of tar that is projected evolving over time matches the measured amount, thus fixing the total mass evolved during the test. The equation below illustrates how the total measured mass was reconciled with that obtained via the time-dependent calculation:

in the differential thermogravimetry (DTG) plot in Figure 1a,b represents the primary devolatilization. The maximum rate of mass loss increases with increasing VM as expected. The maximum rate of mass loss shifts to lower temperatures as VM increases, which also has been shown previously by experiment using TGA46,47 and is aligned with an increase in fluidity shown by the Gieseler test.19 Metaplast resolidification occurs between 820 and 1100 K and yields gaseous products (later shown). For all coals, the DTG asymptotically approaches zero during resolidification, signifying that coke is produced. Low and Mid VM Coals. A statistical t test using a 95% confidence interval in Figure 1a,b shows the residual mass difference between confinements for the low and mid VM coals is significant. The effect of confinement on solid yield was observed to be greater for the mid VM coal than the low VM coal. In Figure 1a,b, the separation of the TG curves in the different confinements occurs between 790 and 820 K for each coal. This corresponds closely to the solidification temperatures measured by the Gieselier test, which are 778 K for the low VM coal and 768 K for the mid VM coal. In Figure 1a,b, the magnitude of the DTG (rate) measurement between 820 and 1100 K is higher in the pan as compared to the cup. These effects on solid yield for low and mid VM coals are due to tar species being retained at the solid−gas interface due to the increase of the residence time of volatile products with each other and the solid interface. Figure 2a,b shows the calculated time-dependent production profile of tar and gas for low and mid VM coals at 3 K min−1 in each confinement. Tar production initiates at 500 K. Tar is the only product until 750 K and continues to evolve until 900 K. To precisely discriminate between gas production and condensable production is one of the main findings of this work. The total gas production in Figure 2a,b does include CO and CO2 evolution. Gas production begins at 750 K for both coals and confinements. Gas production is observed to be higher in the cup as compared to the pan for the low and mid VM coals. This finding is consistent with alkyl cleavage of volatile species that partition into lighter gases (thus higher production) and heavier molecules that recondense with a commensurate reduction in tar measured. This increased gas production between 790 and 820 K indicates secondary reactions during the increased residence time in the cup confinement. Table 3 shows the average product distributions in percent by mass for the low and mid VM coals heated at 3 K min−1 per confinement. There is a 2.1% and 2.3% decrease in tar production in the cup as compared to the pan for the low and mid VM coals, respectively. Results from GC−MS characterization (shown in the Supporting Information) show more speciation in the tar from the cup than the pan. The cup confinement produced tar with a lower molar H/C ratio (1.50) than in the pan (1.55). Furthermore, the cup yields more pyrene, indene, indanol, and naphthalene derivatives, indicating PAH growth at the solid−gas interface during the increased residence time. At increased residence time, PAH products at the solid−gas interface in the cup are likely integrated into the solid product during pyrolysis. Figure 3 shows the evolution profiles of methane, C2 hydrocarbons (ethane and ethylene, separate evolutions shown in the Supporting Information), and hydrogen during pyrolysis at 3 K min−1 for the mid VM coal in both confinements. The hydrogen production is higher in the cup as compared to in the pan in the temperature range of 750− 1140 K. Hydrogen gas is produced from recombination

tf

∑ rtar(t ) = masst total tar measured

■

0

(7)

RESULTS AND DISCUSSION TG and DTG of Coal. Figure 1a,b shows the TG profile of two coals heated in the pan and cup confinement at 3 K min−1.

Figure 1. (a) TG and DTG of low VM heated at 3 K min−1 in pan and cup confinement, and (b) TG and DTG of mid VM heated at 3 K min−1 in pan and cup confinement.

The errors shown are the standard deviations based on repeated tests as discussed in the Materials and Methods. The error arises primarily from the variation in the coal from sample to sample. For each coal, the solid product yield is greater in the cup confinement than in the pan confinement, demonstrating the effects of increased residence time. The low and mid VM coals have the same residual mass in the cup even though they have different VM content. This illustrates that confinement is a very important consideration in evaluating TG data because the system can have considerable mass transfer limitations, which is manifested in the product distribution. The TG profiles for each coal and condition resemble the well-known result for coal pyrolysis.13,16 The major peak observed between 643 and 900 K 6252


Article

Energy & Fuels

Figure 3. Hydrogen, methane, and C2 hydrocarbons (ethane and ethylene) evolution measured by gas chromatography from mid VM heated at 3 K min−1 in pan and cup confinement.

linking10 (eq 5). Therefore, the higher hydrogen and methane production during increased residence times can be linked to the increased PAH and solid yields in the cup. Table 4 shows the average total production of gaseous products in the pan and cup confinements for low and mid VM Table 4. Total Production of Gas Species (mg/mg of Initial Coal Mass) at 3 K min−1 in Each Confinement for the Low and Mid VM Coal and the p-Value between the Different Conditions low VM - pan hydrogen p-value methane p-value ethane p-value ethylene p-value

Figure 2. (a) Tar and gas evolution of low VM heated at 3 K min−1 in pan and cup confinement, and (b) tar and gas evolution of mid VM heated at 3 K min−1 in pan and cup confinement.

Table 3. Average Product Distributions in Percent by Mass of the Low and Mid VM Coals Studied in Two Confinements Heated at 3 K min−1 solid (%) low VM coal - pan low VM coal - cup mid VM coal - pan mid VM coal - cup

78.6 82.4 77.4 82.4

± ± ± ±

2.3 0.92 1.7 0.98

tar and water (%) 11.8 9.7 15.2 12.9

± ± ± ±

2.9 4.9 3.8 4.0

gas (%) 9.4 12.9 7.3 11.1

± ± ± ±

5.6 × 10−3 6.6 × 10−3 0.9 × 10−3 2.0 × 10−4

low VM - cup mid VM - pan mid VM - cup 7.3 1.2 9.4 0.10 0.6 2.9 2.0 0.84

× 10−3 × 10−2 × 10−3

4.7 × 10−3

× 10−3 × 10−2 × 10−4

1.2 × 10−3

6.9 × 10−3

2.0 × 10−4

5.2 0.55 9.6 0.26 0.7 9.0 2.0 0.78

× 10−3 × 10−3 × 10−3 × 10−4 × 10−4

coals. For the mid and low coals, hydrogen production is 10% and 30% higher in the cup as compared to the pan, respectively. The C2 hydrocarbon production is higher in the pan than the cup and occurs at a higher temperature (Figure 3). This is aligned with an overall reduction of H/C ratio of the volatiles generated. Because of the high coking propensity of C2 hydrocarbons,33 it is possible that C2 hydrocarbon involvement in coke-forming reactions above 760 K decreases their overall C2 production in the cup. The mechanism of how paraffin species are incorporated into PAH and coke species is discussed elsewhere.33,36 The greatest difference is the yield of ethane and methane between confinements for the low and mid VM coal. The pan confinement forms more ethane and less methane than the cup for the low and mid VM coals. The reduction of ethane production in the cup only accounts for 10% and 20% of the increased methane production for the low and mid VM coals, respectively. Therefore, approximately 80−90% of the increase in methane comes from the demethylation of the volatile species during coke formation. Ethylene production between 700 and 900 K appears unaffected by increased residence time for the low and mid VM coal. This indicates that ethylene is less impacted by the secondary solid producing

1.6 0.84 0.79 1.2

reactions (eq 4) during PAH growth near the solid−gas interface and integration of tar into the coke product.44 The production rate of methane is 68% higher from the cup than from the pan and occurs at a lower temperature. Throughout the tar evolution sequence, the concentration of hydrogen donor species in the Metaplast decreases; that is, the parent molecule loses saturated and cycloalkane functional groups and becomes more aromatic, which has been measured by NMR.10,38 This decreases the hydrogen abstraction reactions within the Metaplast and shifts the reaction preference toward demethylation during the combination of large aromatic clusters. It is apparent that the early onset of demethylation in the cup leads to a higher methane and solid yield than the pan. The two main methane forming reactions are thermal cracking of bridges and alkyl groups (eqs 1 and 3) and the demethylation associated with moderate temperature cross6253


Article

Energy & Fuels

production is only slightly higher in the pan than in the cup confinement. From GC/MS analysis (shown in the Supporting Information), the molar H/C ratio of the tar in the pan (1.55) is lower than in the cup (1.57), which is the opposite finding with the lower VM coals. The H/C of the tar product decreased from the pan to cup for the lower VM coals due to an increase in secondary reactions such as PAH growth or cracking. Table 6 shows the average total production of gaseous

reactions at the solid−gas interface. In summary, an increase in methane and hydrogen production and a decrease in ethane production are considered a resultant of the reactions, which form the increase in the solid product in the cup confinement from entrapped tar. These reactions are likely thermal cracking of tar species (eq 1, 2, or 3) followed by bimolecular recombination (eq 4) into solid products, which are responsible for the increase in solid yield in the cup. High VM Coal. Figure 4 shows the TG profile for the high VM coal heated in the pan and cup confinement at 3 K min−1.

Table 6. Total Production of Gas Species (mg/mg of Initial Coal Mass) at 3 K min−1 in Each Confinement for the High VM Coal and the p-Value between the Different Conditions high VM - pan hydrogen p-value methane p-value ethane p-value ethylene p-value

The effect of confinement on solid yield for the high VM coal is statistically insignificant. This is in agreement with the results for highly volatile bituminous coals (VM of 36.5−41.9 wt %) studied in various confinements.45 The ineffectiveness of the confinement on the high VM is attributed to the increased release rates, resulting in a residence time too short for the secondary reactions to be observed. The rate constant for secondary tar reactions (for example, eqs 1−5) between 773 and 873 K is on the order of 103 s−1.29 The residence time in the cup for the mid and low VM coal is approximately 0.35 s, which is sufficient to observe impacts of secondary reactions. The secondary reactions occur to a minimal or insignificant extent for the high VM coal where the residence time is only 0.22 s. The minimum residence time of 0.27 s (cup confinement for low VM coal at 3 K min−1) is required to observe a significant extent of secondary volatile reactions within this temperature range. Therefore, the phenomena observed in the low and mid VM are not observed for the high VM coal, and this experimental configuration developed can discriminate classes of reactions that have close characteristic times. The insignificant effect of confinement, despite the increase in tar concentration, for high VM coals is hereby attributed to increased volatile release rates. Table 5 shows the average product distributions in percent by mass for the high VM coal heated at 3 K min−1 per confinement. The difference in solid yield between the confinements is insignificant. Gas Table 5. Average Product Distributions in Percent by Mass of the High VM Coal per Confinement Heated at 3 K min−1 solid (%)

tar and water (%)

gas (%)

68.3 ± 1.5 68.9 ± 1.3

25.5 ± 7.4 23.0 ± 7.9

5.5 ± 0.68 5.0 ± 0.98

1.3 × 10−2 0.9 × 10−3 2.0 × 10−4

high VM - cup 5.2 5.6 1.3 0.66 1.0 0.3277 3.0 1.4

× 10−3 × 10−5 × 10−2 × 10−3 × 10−4 × 10−3

products in the pan and cup confinements for high VM coal. The pan confinement results in 35% higher hydrogen production than the cup. The difference of hydrogen measured in tar only accounted for about one-third of the increase hydrogen in the pan. Methane yields from the pan and cup are the same. Heating Rate. Figure 5a shows the TG of the mid VM coal heated at 1 and 10 K min−1 in both confinements. The effect of confinement on the solid residual persists at heating rates of 1− 10 K min−1. Figure 5b shows the DTG of the mid VM coal heated at 1 and 10 K min−1 in both confinements. The residence time of volatile species in the cup nominally decreases by the same factor at which the heating rate increases. Therefore, for the cup confinement, the residence time ranges from 0.11 and 1.0 s for the 1 and 10 K min−1 cases, respectively. Table 7 shows the solid yield differences between the cup and pan confinements for the low and mid VM coals at 1, 3, and 10 K min−1. The difference in solid yield in the cup is inversely proportional to heating rate (i.e., it increases as the heating rate decreases). The extents of secondary reactions decrease due to increased transport rates. The secondary reactions incorporate tar products into the solid product via cracking reactions, which produce more gas. The DTG increases by a factor of 10 and shifts by 40 K to a higher temperature as the heating rate increases. This shift to a higher temperature with increasing heating rate agrees with accounts in the literature using TGA46,47 and other apparatuses such as wire-mesh reactors, electrical strip furnaces,23 and prediction in FLASHCHAIN.5 The shift to higher temperature is attributed to a delay in the onset of the Metaplast5 caused by a negative relationship of heating rate and low temperature cross-linking reactions.10 In the case of secondary tar reactions, a faster tar evolution rate would decrease the residence time of chemical species above the sample. The separation of the TG curves of each confinement shown in Figures 5a occurs between 790 and 820 K for the mid VM coal heated at 1 and 10 K min−1. This shows that at vastly different release rates (from 1, 3, and 10 K min−1) the temperature of separation due to reaction at the gas−solid interface in the cup confinement consistently occurs within a 30 K temperature span for the mid VM and low VM

Figure 4. TG and DTG of high VM heated at 3 K min−1 in pan and cup confinement.

high VM coal - pan high VM coal - cup

7.0 × 10−3

6254


Article

Energy & Fuels

occurring between volatile coal pyrolysis products near solid− gas interfaces. Large tar species at the solid−gas interface thermally decompose between 790 and 820 K. The effect of these reactions is minimal for high VM coal and increased heating rates.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.energyfuels.6b00699. Deconvoluted C2 hydrocarbon (ethane and ethylene) evolution at three heating rates for the mid VM coal, residence time calculations, Biot number calculations, and list of identified compounds from GC/MS analysis of tar for three coals in two confinements (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (212) 650-6679. Fax: (212) 650-6660. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge Suncoke Energy for the financial support and collaboration for this research. We acknowledge the Combustion and Catalysis Lab in the Chemical Engineering Department of City College of New York for their collaboration in the experiment of this research.

Figure 5. (a) TG of mid VM heated at 1 and 10 K min−1 in pan and cup confinement, and (b) DTG of mid VM heated at 1 and 10 K min−1 in pan and cup confinement.

■

Table 7. Difference in Solid Yield between Cup and Pan for the Low, Mid, and High VM Coal at 1, 3, and 10 K min−1 heating rate, K min−1

mid VM (%)

low VM (%)

high VM (%)

1 3 10

5.7 3.2 2.4

5.4 1.8 0.6

3.7 0.7 0.6

REFERENCES

(1) Diez, M. A.; Alvarez, R.; Barriocanal, C. Coal for metallurgical coke production: predictions of coke quality and future requirements for cokemaing. Int. J. Coal Geol. 2002, 50 (1), 389−412. (2) van Heek, K. H. Progress of coal science in the 20th century. Fuel 2000, 79 (1), 1−26. (3) Niksa, S. FLASHCHAIN theory for rapid coal devolatilization kinetics. 3. modeling the behavior of various coals. Energy Fuels 1991, 5, 673−683. (4) Niksa, S.; Kerstein, A. R. FLASHCHAIN theory for rapid coal devolatilization kinetics. 1. formulation. Energy Fuels 1991, 5, 647− 665. (5) Niksa, S. FLASHCHAIN theory for rapid coal devolatilization kinetics. 2. impact of operating conditions. Energy Fuels 1991, 5, 665− 673. (6) Niksa, S. FLASHCHAIN theory for rapid coal devolatilization kinetics. 4. predicting ultimate yields from ultimate analysis alone. Energy Fuels 1994, 8, 659−670. (7) Niksa, S. FLASHCHAIN theory for rapid coal devolatilization kinetics. 5. interpreting rate of devolattilization for various coal types and operating conditions. Energy Fuels 1994, 8, 671−679. (8) Niksa, S. FLASHCHAIN theory for rapid coal devolatilization kinetics. 6. predicting the evolution of fuel nitrogen from various coals. Energy Fuels 1995, 9, 467−478. (9) Niksa, S. FLASHCHAIN theory for rapid coal devolatilization kinetics. 7. predicting the release of oxygen species from various coals. Energy Fuels 1996, 10, 173−187. (10) Solomon, P. R.; Hamblen, D. G.; Carangelo, R. M.; Serio, M. A.; Deshpande, G. V. General model of coal devolatilization. Energy Fuels 1988, 2, 405−422. (11) Gavalas, G. R.; Cheong, P. H.; Jain, R. Model of coal pyrolysis. 1. qualitative development. Ind. Eng. Chem. Fundam. 1981, 20, 113− 122.

coal, respectively. This confirms the temperature range of secondary reactions in the cup confinement (eq 4) is between 790 and 820 K, which again is near the solidification temperature determined by the Gieselier test.

■

CONCLUSION Agglomerating coal pyrolysis was studied by TG−GC with online tar capture at appropriate heating rates for coke-making up to 1373 K. The mass balance between solid, condensable, and gas products was close to >99% for the three agglomerating coals. The use of confinement with varying aspect ratios demonstrated an increase in the solid production for low and mid VM coals heated at 1 and 3 K min−1. This exhibited the time dependence of secondary coke-forming reactions near the gas−solid interface. The residence time of volatile products above the sample was varied using confinement and heating rate from approximately 0.01 s (in the pan at 10 K min−1) to 1.0 s (in the cup at 1 K min−1). The increased residence time results in increased methane, hydrogen, and solid production and decreased tar and ethane production for the low and mid VM coals. This provided insight into the classes of reactions 6255


Article

Energy & Fuels (12) van Krevelen, D. W. Coal--Typology, Physics, Chemistry, Constitution; Elsevier: Amsterdam, 1993. (13) Leblanc, J.; Marinov, N.; Quanci, J.; Castaldi, M. J. A kinetic investigation of coal conversion to metallurgical coke: influence of particle size, density and porosity. AIChE Annual Meeting, San Francisco, CA, Nov 5, 2013. (14) Suuberg, E. M.; Peters, W. A.; Howard, J. B. Product composition and kinetics of lignite pyrolysis. Ind. Eng. Chem. Ind. Eng. Chem. Process Des. Dev. 1978, 17, 37−46. (15) Chen, J. C.; Niksa, S. Coal devolatilization during rapid transient heating. 1. Primary devolatilization. Energy Fuels 1992, 6 (3), 254−264. (16) Bautista, J. R.; Russel, W. B.; Saville, D. A. Time-resolved pyrolysis product distributions of softening coals. Ind. Eng. Chem. Fundam. 1986, 25 (4), 536−544. (17) Oh, M. S.; Peters, W. A.; Howard, J. B. An experimental and modeling study of softening coal pyrolysis. AIChE J. 1989, 35, 775− 792. (18) Juntgen, H. Review of the kinetics of pyrolysis and hydropyrolysis in relation to the chemical constitution of coal. Fuel 1984, 63, 731−737. (19) van Krevelen, D. W. Coal--Typology, Physics, Chemistry, Constitution; Elsevier: Amsterdam, Oxford, NY, 1981. (20) Fletcher, T. H.; Kerstein, A. R.; Pugmire, R. J.; Grant, D. M. Chemical percolation model for devolatilization. 2. temperature and heating rate effects on product yields. Energy Fuels 1990, 4, 54−60. (21) Deshpande, G. V.; Solomon, P. R.; Serio, M. A. Crosslinking reactions in coal pyrolysis. Prepr. Pap.-Am. Chem. Soc. Div. Fuel Chem. 1988, 33, 310−321. (22) Solomon, P. R.; Serio, M. A.; Despande, G. V.; Kroo, E. Crosslinking reactions during coal conversion. Energy Fuels 1990, 4 (1), 42− 54. (23) Anthony, D. B.; Howard, J. B. Coal devolatilization and hydrogasification. AIChE J. 1976, 22, 625. (24) Solomon, P. R.; Serio, M. A.; Carangelo, R. M.; Bassilakis, R. Analysis of the Argonne premium coal samples by thermogravimetric fourier transform infrared spectroscopy. Energy Fuels 1990, 4, 319− 333. (25) Cramer, B.; Eckhard, F.; Gerling, P.; Krooss, B. M. Reaction kinetics of stable carbon isotopes in natural gas-insights from dry, open systems pyrolysis experiments. Energy Fuels 2001, 15, 517−532. (26) Pather, T. S.; Al-Masry, W. A. The influence of bed depth on secondary reactions during slow pyrolysis of coal. J. Anal. Appl. Pyrolysis 1996, 37, 83−94. (27) Tromp, P. J. J. Coal pyrolysis. Thesis, University of Amsterdam, 1987. (28) Lohman, T. W. Modelling and parameter estimation of reaction kinetics in coal pyrolysis. J. Inv. Ill-posed Problems 1998, 7, 463−480. (29) Serio, M. A.; Peters, W. A.; Howard, J. B. Kinetics of vaporphase secondary reactions of prompt coal pyrolysis tars. Ind. Eng. Chem. Res. 1987, 26, 1831−1838. (30) Poutsma, M. L. Free-radical thermolysis and hydrogenolysis of model hydrocarbons relevant to processing of coal. Energy Fuels 1990, 4, 113−131. (31) Sundaram, M. K.; Froment, G. F. Modeling of thermal cracking kinetics. 3. radical mechanisms for the pyrolysis of simple paraffins, olefins and their mixtures. Ind. Eng. Chem. Fundam. 1978, 17, 174− 182. (32) Willems, P. A.; Froment, G. F. Kinetic modeling of the thermal cracking of hydrocarbons 1. calculation of frequency factors. Ind. Eng. Chem. Res. 1988, 27, 1959−1966. (33) Reyniers, G. C.; Froment, G. F.; Kopinke, F. D.; Zimmermann, G. Coke formation in the thermal cracking of hydrocarbons. 4. modeling of coke formation in naphtha cracking. Ind. Eng. Chem. Res. 1994, 33, 2584−2590. (34) Dryden, I. G. C.; Sparham, G. A. Carbonization of Coals under Gas Pressure. B.C.U.R.A. Monthly Bulletin, 1963, 27. (35) Howard, J. B.; Peters, W. A.; Serio, M. A. Coal Devolatilization Information for Reactor Modeling Final Report. EPRI Project, 1981.

(36) Frenklach, M. Reaction mechanism of soot formation in flames. Phys. Chem. Chem. Phys. 2002, 4, 2028−2037. (37) Senkan, S.; Castaldi, M. J. Formation of polycyclic aromatic hydrocarbons (PAH) in methane combustion: comparative new results from premixed flames. Combust. Flame 1996, 107, 141−150. (38) Jones, K.; Ramakrishnan, G.; Uchimiya, M.; Orlov, A.; Castaldi, M. J.; LeBlanc, J.; Hiradate, S. Fate of higher-mass elements and surface functional groups during the pyrolysis of waste pecan shell. Energy Fuels 2015, 29, 8095−8101. (39) Neeft, J. A. P.; Knoef, H. A. M.; Zielke, U.; Sjöström, K.; Hasler, P.; Simell, P. A.; Dorrington, M. A.; Thomas, L.; Abatzoglou, N.; Deutch, S.; Greil, C.; Buffinga, G. J.; Brage, C.; Suomalainen, M. Guideline for sampling and analysis of tar and particles in biomass producer gases. Energy Project ERK6-CT1999-20002, Version 3.3. (40) Miljevic, B.; Modini, R. L.; Bottle, S. E.; Ristovski, Z. D. On the efficiency of impingers with fritted nozzle tip for collection of ultrafine particles. Atmos. Environ. 2009, 43 (6), 1372−1376. (41) Ringer, M.; Putsche, V.; Scahill, J. Large-Scale Pyrolysis Oil Production: A Technology Assessment and Economic Analysis; National Renewable Energy Laboratory, 2006. (42) Boyer, A. F. Assoc. Techn. de L’Indus du Gaz en France 69th Congr., 1952; p 653. (43) van Krevelen, D. W.; Chermin, H. A. G.; Dormans, H. N. M.; Huntjens, F. J. Brennstoff-Chem. 1958, 39, 77. (44) Stein, S. E.; Griffith, L. L.; Billmers, R.; Chen, R. H. Pyrocondensation of anthracene. J. Org. Chem. 1987, 52, 1582−1591. (45) Shi, L.; Liu, Q.; Guo, X.; He, W.; Liu, Z. Pyrolysis of coal in TGA: Extent of volatile condensation in crucible. Fuel Process. Technol. 2014, 121, 91−95. (46) Van Krevelen, D. W.; Huntjens, F. J.; Dormans, H. N. M. Chemical structure and properties of coal: XVI. Plastic behavior on heating. Fuel 1956, 35, 462−475. (47) Van Krevelen, D. W.; Dormans, H. N. M.; Huntjens, F. J. Fuel 1959, 38, 165−182.

6256
