Hazardous Air Pollutant Formation from Pyrolysis of Typical Chinese

ARTICLE pubs.acs.org/est

Hazardous Air Pollutant Formation from Pyrolysis of Typical Chinese Casting Materials Yujue Wang,†,‡,* Ying Zhang,† Lu Su,† Xiangyu Li,† Lei Duan,†,‡ Chengwen Wang,† and Tianyou Huang§ †

School of Environment, Tsinghua University, Beijing 100084, China State Key Joint Laboratory of Environmental Simulation and Pollution Control, Beijing 100084, China § Department of Mechanical Engineering, Tsinghua University, Beijing 100084, China ‡

bS Supporting Information ABSTRACT: Analytical pyrolysis was conducted to evaluate the major hazardous air pollutant (HAP) emissions from pyrolysis of bituminous coal and a furan binder, which are the two most commonly used casting materials for making green sand and furan nobake molds in Chinese foundries. These two materials were flash pyrolyzed in a Curiepoint pyrolyzer at 920 °C and slowly pyrolyzed in a thermogravimetric analyzer (TGA) from ambient temperature to 1000 °C with a heating rate of 30 °C/min. The emissions from Curie-point and TGA pyrolysis were analyzed with gas chromatography mass spectrometer/flame ionization detector. Thirteen HAP species were identified and quantified in the pyrolysis emissions of the two materials. The prominent HAP emissions were cresols, benzene, toluene, phenol, and naphthalene for the bituminous coal, whereas they were m,p,o-xylenes for the furan binder. Xylenesulfonic acid, the acidic catalyst in furan binder, was found to be the major source of xylene emissions. Thermogravimetry-mass spectrometer monitored the evolution of HAP emissions during TGA pyrolysis. For both of the casting materials, most of the emissions were released in the temperature range of 350 700 °C.

’ INTRODUCTION The metal casting industry represents an important manufacturing component that produces numerous important casting products for everyday life. With the development of the Chinese economy, the metal casting industry in China has been growing rapidly in recent years. The total production of casting metals in China has increased dramatically from about 14 million tons in 2000 to 34 million tons in 2008.1 As the largest producer of metal casting products in the world, China produces about one-third of the world’s total production. The rapid expansion of the Chinese metal casting industry has raised major environmental concerns associated with this industry. Particularly, hazardous air pollutant (HAP: defined herein as the 188 compounds that are categorized in Title III: Hazardous Air Pollutants of the 1990 Clean Air Act Amendments) emissions from the metal casting industry have attracted increasing attention. The HAP emissions from foundries are mainly generated from pyrolysis of organic casting materials (e.g., carbonaceous additives and organic binders that are used to make sand molds and cores) during the casting process.2 4 Many researchers have studied the HAP emissions of various casting materials that are used in U.S. foundries, such as bituminous coal, cellulose, phenolic urethane binder (PUB), and collagen-based binder.5 9 The results show different casting materials exhibit distinct HAP emission characteristics and levels, and U.S. foundry personnel have been using these results to develop and select less-emitting casting materials for casting production in order to diminish their r 2011 American Chemical Society

HAP emissions.10 12 For example, cellulose has been used to partially replace bituminous coal, with the aim of diminishing benzene, toluene, ethylbenzene, and xylene (BTEX) emission;11 meanwhile, biodiesel based non-naphthalene PUB has been developed and used to replace the conventional PUB for diminishing polycyclic aromatic hydrocarbon (PAH) emissions.12 In contrast, very few studies have evaluated the emission characteristics of casting materials that are used in Chinese foundries. Data on the HAP emission of the Chinese casting materials are scarce. Presumably, the HAP emission characteristics and levels of Chinese casting materials could be quite different from those of U.S. casting materials, because they are different in terms of origins, natures, compositions, etc. The need exists for Chinese foundries to appraise the HAP emission characteristics of their casting materials, so as to diminish their air emissions by developing and selecting “cleaner” casting materials for production. Thus, the main objective of this study was to evaluate the HAP emissions of the two most commonly used Chinese casting materials, bituminous coal and furan binder, via analytical pyrolysis emission tests. Specifically, the two casting materials were flash pyrolyzed in a Curie-point pyrolyzer and slowly pyrolyzed in a thermogravimetric analyzer (TGA). The major HAP emissions were then Received: January 26, 2011 Accepted: June 29, 2011 Revised: June 29, 2011 Published: June 29, 2011 6539

dx.doi.org/10.1021/es200310p | Environ. Sci. Technol. 2011, 45, 6539–6544

Environmental Science & Technology

ARTICLE

Table 1. Proximate and Ultimate Analysis of the Bituminous Coal and Furan No-Bake Mold Samplea ultimate analysis (wt%)b

proximate analysis (wt%) sample

a

moisture

ash

V.M.

F.C.

C

H

N

S

bituminous coal

4.12

8.92

31.52

55.44

70.06

4.29

1.06

0.54

furan no-based mold sample

0.14

97.57

2.16

0.13

1.03

0.09

0.05

0.05

V.M.: volatile matter; F.C.: fixed carbon. b Dry basis.

identified and quantified via gas chromatography mass spectrometer/flame ionization detector (GC-MS/FID). Thermogravimetry-mass spectrometer (TG-MS) monitored the evolution profiles of HAP emissions during TGA pyrolysis. Based on the testing results, the pathways of HAP formation from pyrolysis of the two casting materials were then presented. The HAPs identified herein may in part represent what the foundry workers and nearby neighbors are exposed to when the bituminous coal and furan binder are employed in the casting production.

’ MATERIALS AND METHODS Sample Collection and Preparation. All the casting materials used in this study were collected from a full-scale production foundry that employed both green sand molds and furan no-bake molds for casting production. According to the foundry’s personnel, the bituminous coal (originated from Inner Mongolia province of China) and the furan binder were manufactured by two major Chinese casting material suppliers, respectively; and they are widely used in many Chinese foundries. The bituminous coal was sampled from the storage bin of the foundry. Coal particles with diameter less than 0.1 mm were used in the pyrolysis tests. The furan no-bake mold sample was prepared by mixing 97% silica sand, 2% furfuryl alcohol based resin, and 1% acidic catalyst, following the foundry’s operation procedures. According to the supplier, the resin contained 60 80% furfuryl alcohol, small amounts of urea and formaldehyde were used to improve the resin’s property; the acidic catalyst was a 65% xylenesulfonic acid solution. The silica sand was a high silica round grain sand that was sieved to pass a 0.149 mm sieve. The sieved sand was then thoroughly washed with deionized water and dried at 105 °C in the air overnight before it was used to make the furan no-bake mold samples. After the curing, the hardened furan no-bake mold sample was crushed with pestle and mortar into individual particles; the furan binder coated sand particles were then used in analytical pyrolysis emission tests. The ultimate and proximate analyses for the coal and furan no-bake mold samples are shown in Table 1. Curie-Point Pyrolysis Emission Tests. About 5 mg of bituminous coal or 30 mg furan binder coated sand particles were tightly wrapped in a ferromagnetic foil that was then placed in a small quartz tube (4.0 mm i.d.) in the Curie-point pyrolyzer (JHP-22, Japan Analytic Ind.). The foil rapidly heated to its specific Curie-point temperature of 920 °C in a helium atmosphere. The temperature rise took less than 0.2 s and was highly reproducible. This abrupt temperature shock simulated some key features of the intense heating conditions at the vicinity of a casting surface, where casting materials were rapidly heated to temperatures of several hundred degrees Celsius up to the pouring temperature. The pyrofoil was held at 920 °C for 5 s to achieve complete pyrolysis of the casting materials. The gases emitted via this abrupt heating were carried by high-purity helium through a heated (250 °C) transporting tube (0.5 m in length) to

a downstream online gas chromatograph (Agilent 7890A network GC system) that was equipped with a flame ionization detector (FID) and a mass spectrometer (5975C MSD). The GC column was a DB-624 capillary column (30 m  0.32 mm i.d.  1.8 μm film thickness, Agilent Technologies, Inc.). The temperature ramp started with an isothermal step of 4 min at 50 °C, followed by a gradient of 5 °C/min to 150 °C, and then a gradient of 10 °C/min up to a final temperature of 230 °C that was then held constant for 28 min. The ionization mode on the MS was set to electron impact at 70 eV and the mass range was from m/z 30 to 450 scanned at 1 s 1. For tentative identification of HAP species, all mass spectra were compared to the NIST mass spectrum library. Next, the MS identified major HAP species were quantified with the GC-FID by comparing the peaks’ retention times and areas with those of standard solutions that contained target HAPs with known concentrations. Four replicates of Curie-point pyrolysis-GC-FID emission tests were conducted for each of the casting materials. Table 2 lists the average concentrations and standard deviations. TGA Pyrolysis Emission Tests via Sorbent Tube Capture and GC-FID Analysis. About 15 mg of bituminous coal or 150 mg of furan binder coated sand particles were pyrolyzed in a thermogravimetric analyzer (TGA) (TGA/DSC 1, Mettler-Toledo Inc. Switzerland) with a heating rate of 30 °C/min from ambient temperature to 1000 °C under 40 mL/min N2. This TGA protocol simulated some key features of the ramped heating within the bulk of mold, where casting materials are progressively heated during the casting solidification and mold cooling as the casting heat dissipates from the casting surface outward through the mold. During the TGA heating, preconditioned tubes with Tenax TA were used to collect the emissions in the TGA effluent. The captured HAPs were then quantified via a thermal desorber (Turbo Matrix 650, PerkinElmer Instruments) that was coupled with a GC-FID (Clarus 600, PerkinElmer Instruments). During the desorption, the sorbent tube was heated to 280 °C for 10 min, while high purity nitrogen gas was passed through the tube at a flow rate of 40 mL/min to desorb the analytes and concentrate them into a cold trap, which was packed with Tenax TA and kept at 25 °C with a Peltier cooler. The trap was then heated to 301 °C for 5 min. The HAPs were desorbed from the trap and injected into the GC-FID for quantification. The GC column was an Elite-624 capillary column (30 m  0.32 mm i.d.  1.8 μm film thickness, PerkinElmer Instruments). The temperature ramping program of the GC oven was the same as described in the Curie-point pyrolysis emission analysis. Target HAP species were quantified by comparing the peaks’ retention times and areas with those of calibration standards. The standard solutions were freshly made by dilution from commercial solutions in methanol, which were then loaded into Tenax TA tubes via a calibration solution loading equipment (TP-2040, Tianpu Instrument Ltd., China). The calibration tubes were then analyzed via the same thermal 6540

dx.doi.org/10.1021/es200310p |Environ. Sci. Technol. 2011, 45, 6539–6544

Environmental Science & Technology

ARTICLE

Table 2. Hazardous Air Pollutant Emissions of the Bituminous Coal and the Furan Binder during Curie-Point Flash Pyrolysis and TGA Slow Pyrolysis (Four Replicates for both Flash Pyrolysis and TGA Pyrolysis)a curie-point pyrolysis bituminous coal emissions (mg/g coal or furan binder)

a

mean

STD

TGA slow pyrolysis

furan binder Mean

bituminous coal

STD

ND

Mean

STD

furan binder Mean

STD

1

hexane

0.123

0.003

0.117

0.012

2

benzene

0.853

0.067

0.315

0.012

1.029

0.042

ND 0.115

0.013

3 4

toluene ethylbenzene

0.817 0.077

0.038 0.003

0.389 0.181

0.010 0.015

1.077 0.080

0.048 0.006

0.241 0.134

0.019 0.014

5

m,p-xylene

0.411

0.023

1.455

0.061

0.409

0.027

1.197

0.118

6

o-xylene

0.117

0.006

0.392

0.017

0.147

0.009

0.146

0.008

7

styrene

0.068

0.006

8

phenol

0.755

0.021

0.087

0.014

0.791

0.030

9

o-cresol

0.432

0.007

0.023

0.004

0.225

0.009

NQ

10

p,m-cresol

1.355

0.013

0.009

0.001

0.386

0.034

NQ

11 12

naphthalene 2-methylnaphthalene

0.539 0.231

0.008 0.003

0.077 0.013

0.005 0.001

0.073 0.023

0.003 0.003

NQ NQ

13

1-methylnaphthalene

0.092

0.005

0.029

0.002

0.055

0.004

NQ

sum

5.87

0.117

2.97

0.059

4.412

0.276

ND

NQ

ND NQ

1.833

0.114

ND: not detected; NQ: Not quantified due to the negligible amounts.

Figure 1. GC-FID responses for the pyrolysis products of (A) the bituminous coal, and (B) the furan no-bake mold sample during Curie-point flash pyrolysis. Refer to Table 2 for the names of identified HAPs.

desorption and GC analysis protocol used for the TGA emission analysis. TGA Pyrolysis mass Spectrometer Analysis (TG-MS). The same TGA pyrolysis protocol described above was employed to heat the bituminous coal and furan binder coated sand particles. Nevertheless, in this TG-MS test, the gaseous effluent from the TGA flowed to a downstream mass spectroscopy (OmniStar GSD 301 O, Pfeiffer Vacuum Inc. Germany) for emission kinetics analysis. The MS was operated at 70 eV and monitored selected m/z values that were considered to be representatives of HAP species identified in the emission of Curie-point and TGA slow pyrolysis tests.

’ RESULTS Curie-Point and TGA Pyrolysis Emission Analysis. The major HAP species identified in the emission of the bituminous coal and furan binder during Curie-point pyrolysis are shown in Figure 1. Table 2 presents the quantification results of 13 HAPs in the emissions of Curie-point pyrolysis and TGA pyrolysis. HAPs with low abundance that were only intermittently detected in TGA emissions were not quantified herein due to their negligible amount. For both the coal and the furan binder sample, the HAP species generated in TGA pyrolysis and Curie-point 6541

dx.doi.org/10.1021/es200310p |Environ. Sci. Technol. 2011, 45, 6539–6544

Environmental Science & Technology

ARTICLE

Figure 2. Mass loss profiles and mass spectrometer responses to the emissions of the bituminous coal during TGA slow pyrolysis (30 °C/min) under a nitrogen atmosphere.

pyrolysis were the same, but their yields and fractional compositions were different. Specifically, for the bituminous coal, the major HAP emissions were cresols, benzene, toluene, phenol, and naphthalene in descending order of magnitudes. In contrast, the predominant HAP emissions of the furan binder were m,p,oxylenes, whereas benzene, toluene, etc. contributed merely fractions. As shown in Table 2, the sums of HAP emissions generated in Curie-point flash pyrolysis were 25% and 38% higher than those generated in TGA slow pyrolysis for the bituminous coal and the furan binder, respectively. The difference was mainly attributed to the distinct heating conditions of these two pyrolysis techniques that could profoundly influence the thermal decomposition reactions of the casting materials, and thus yield different emission levels. For example, many researchers have shown that flash pyrolysis causes more extensive thermal fragmentations of coal structures and suppresses the secondary recombination reactions of free radicals, therefore generating more volatiles than does slow pyrolysis.13,14 Additionally, the TGA emission tests involved sorbent capturing and thermal desorption steps during which sample loss and dilution might occur. These would also result in the lower emission levels in TGA emission tests. These HAP species identified herein may in part represent what the foundry workers and nearby neighbors are exposed to, when the bituminous coal and furan binder are used to make sand molds for the foundry’s production. Indeed, previous studies have shown that the HAP emissions from analytical pyrolysis of several casting materials used in U.S foundries were essentially the same as those identified in the emissions from the foundries.3,5 9,11,12 Additionally, although the emission values determined in this analytical pyrolysis study only reflected the relative emissions of the casting materials under the specific test conditions and should not be directly used as emission factors for estimating emissions from actual operating foundries, they may serve as baselines against which the relative emission levels of

Figure 3. Mass loss profiles and mass spectrometer responses to the emissions of the furan no-bake mold sample during TGA slow pyrolysis (30 °C/min) under a nitrogen atmosphere.

many alternative casting materials could be evaluated. This may help foundries to select less-polluting casting materials for their production in order to diminish the HAP emissions (refer to the Supporting Information (SI) for more discussion on the application of analytical pyrolysis techniques in evaluating HAP emissions from casting materials). TGA Slow Pyrolysis mass Spectrometer Analysis. The TG-MS results are shown in Figures 2 and 3 for the bituminous coal and furan no-bake mold samples, respectively. For the bituminous coal sample, the minor mass loss at temperatures of 50 170 °C was mainly attributed to the weakly bonded moistures (no significant MS peaks of hydrocarbons were observed in this temperature range). Then at the temperatures of around 350 °C where the active thermal decomposition of coal occurred, the coal began to lose considerable mass and release HAP emissions. As shown in Figure 2, the mass loss rate curve (derivative thermogrametric, DTG) and the evolution curves of 6542

dx.doi.org/10.1021/es200310p |Environ. Sci. Technol. 2011, 45, 6539–6544

Environmental Science & Technology HAPs exhibited very similar trends; most of the mass loss and HAP emissions occurred in the temperature range of 350 700 °C. In contrast, the furan binder sample lost considerable mass before the temperature reached 300 °C, but it only released negligible HAP emissions in this temperature range. Most of the mass loss before 300 °C was attributed to the evolution of water (m/z = 18) and carbon oxides (CO2 m/z = 44) (see Figure 3). CO was not monitored herein because nitrogen was used as the carrier gas in this study and their ion fragments (m/z = 28) is thus indistinguishable. From temperatures above 300 °C, the release of HAPs species such as BTEX and monomers of methylfurans (m/z = 82 2-methylfuran, m/z = 96 2,5-dimethylfuran) increased markedly, indicating the breakdown of the binder’s cross-link structures. As shown in Figure 3, the predominant HAP emissions were released in the temperature range of 300 700 °C when the furan binder was slowly heating. The results presented herein show that both of these casting materials began to release the HAP emissions when they were progressively heated to above 300 350 °C. This suggested that foundries might be able to diminish their HAP emissions by shortening mold cooling time (when such operations are allowed by quality control requirements), so as to minimize the volume of sand mold that would be heated to above 300 350 °C during the mold cooling period (refer to SI for more discussion).

’ DISCUSSION Mechanisms of HAP Formation from Pyrolysis of Bituminous Coal and Furan Binder. It is well-known that coal structures

consist of aromatic and hydroaromatic clusters (1 5 condensed rings) linked by aliphatic or ether bridges.15,16 When the coal is progressively heated to temperatures above 350 400 °C, the active thermal decomposition of coal begins, and the coal structure experience massive thermal breakdown and generates extensive molecular fragments. The fragments are then stabilized by the abstraction of hydrogen from other fragments and/or random recombination. Small stabilized fragments such as single ring aromatics could diffuse though the pores of coal particles and reach the coal surface where they vaporize. In this way, benzene, toluene, phenols, and other HAPs are released as emissions. The evolution of volatiles during coal pyrolysis has been extensively studied by many researchers. For more information, readers are referred to the abundant literature on this subject.13 17 Regarding the furan binder’s structure, Choura et al.18 have thoroughly studied the acid-catalyzed polymerization of furfuryl alcohol resin and proposed that cured furan resin consists of highly conjugated sequences of oligomers of methylfuran (see SI Figure S3a). These sequences form branches and cross-links via electrophilic substitutions at the unsaturations (SI, Figure S3b) and/or an interchain Diels Alder reactions (SI Figure S3c), leading to the black cross-linked polymer.18 During polymerization, the acidic catalyst that was xylenesulfonic acid herein is incorporated into the polymer structure via condensation reactions (SI Figure S3d).19 Based on the above information, we have proposed a structure of the cured furan binder as presented in Figure 4. The structure is composed of branched and cross-linked heterocylclic rings and aromatic rings. Starting from the schematic structure, the paths of HAP formation during pyrolysis of furan binder have been described as follows. During the ramped heating of TGA pyrolysis, the cross-linked structure of the furan binder progressively decomposed and released volatiles. SO2 (m/z = 64) began

ARTICLE

Figure 4. Schematic of the thermal decomposition pathways of the furan binder.

to evolve from the furan binder sample at temperatures of about 150 °C (See Figure 3) because the R-SO2OH band was relatively weak. The furan rings in the polymer progressively decomposed over the ramping temperatures of 100 to 500 °C, and then disappeared at even higher temperatures.20 The ring-opening generated chain fragments that could be further decomposed: oxygen in the chain fragments was eliminated primarily as water and carbon oxides, and the carbon chains were broken apart to form shorter alkenes and alkynes (refer to refs 21 24 for more detailed descriptions of the decomposition mechanisms and pathways of furan-based materials). In the present study, the major decomposition products of the chain fragments were monitored via TG-MS. As shown in Figure 3, the evolution profiles of H2O (m/z = 18) and CO2 (m/z = 44) exhibited pronounced peaks before 300 °C, indicating that considerable oxygen was removed at relatively low temperatures. In contrast, the evolutions of unsaturated hydrocarbons, for example, C3H6, C4H6, and C5H8 (m/z = 41, 55, and 68), mainly occurred above temperatures of 300 °C because the carbon skeletons required more energy to break up. Also from temperatures of around 300 °C, the evolutions of xylenes (m/z = 106), toluene (m/z = 92), 2-methylfuran (m/z = 82), and 2,5-dimethylfuran (m/z = 96) exhibited remarkable increases (see Figure 3). This indicates that the aliphatic bridges connecting furan rings and aromatic rings broke up, leading to the release of methylated furans and benzenes. The methyl furans were thermally unstable at high temperatures and would be further decomposed to products similar to those of the chain fragment decomposition. During TGA slow pyrolysis, the furan binder sample released HAPs over a wide temperature range, mainly within 300 700 °C. In contrast, during Curie-point pyrolysis, the temperature ramped rapidly to 920 °C within about 0.2 s. The reactions of the furan ring-opening, carbon bridge scission, and further decompositions of chain fragments occurred almost instantaneously. The abrupt heating caused extensive breakdown of the furan binder structure, and resulted in the release of large amounts of various emissions in a short pyrolysis time (5 s herein). Based on the discussion, we perceived that the major source of HAP emissions, which were predominantly xylenes for the furan 6543

dx.doi.org/10.1021/es200310p |Environ. Sci. Technol. 2011, 45, 6539–6544

Environmental Science & Technology binder tested herein, was xylenesulfonic acid that was the acidic catalyst of furan binder. Indeed, previous studies on pyrolysis of furan-based materials, such as 2-methylfuran and 2,5-dimethylfuran, have shown that the major pyrolysis products were CO and C2 C5 unsaturated hydrocarbons,21,22 which were not HAPs. Although benzene rings might be formed from the chain fragments20 and/or consecutive radical reactions,22 as depicted in SI Figure S3a and b, aromatic products from the pyrolysis of these furan-based materials were usually very small.21,22 In comparison, Dungan et al.7 studied the pyrolysis of a furan binder that was catalyzed with phenolsulfonic acid. They found that phenol was the predominant HAP species of the emission, whereas BTEX compounds were negligible or undetected. This substantiated the inference that the acidic catalyst plays a vital role in determining the nature and levels of HAP emissions from the pyrolysis of furan binders. Therefore, from the perspective of minimizing HAP emissions, more environmentally friendly acidic catalysts should be (developed and) used to replace the conventional ones, that is, arylsulfonic acids, which are commonly employed in present foundry operations.

’ ASSOCIATED CONTENT

bS

Supporting Information. Details of some methods, results, and discussion. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: 86-10-62772914; e-mail: [email protected]..

’ ACKNOWLEDGMENT This research is supported by the NSFC project (51008175) and special fund of State Key Joint Laboratory of Environment Simulation and Pollution Control (09Y01ESPCT).

ARTICLE

foundry core binders by analytical pyrolysis. Environ. Sci. Technol. 2007, 41, 7922–7927. (10) Fox, J. R.; Adamovits, M.; Henry, C. Strategies for reducing foundry emissions. Am. Foundry Soc. Trans. 2002, 110, 1299–1309. (11) LaFay, V. S.; Crandell, G. The addition of cellulose to molding sand when reducing seacoal for emission reduction during pouring, cooling and shakeout. Am. Foundry Soc. Trans. 2007, 115, 805–811. (12) Eppley, M. C.; Laitar, R. A.; Pahr, E. R.; Roush, D. C.; Tse, R.; Zaretskiy, L. S. Improved phenolic urethane cold box foundry resin system. Am. Foundry Soc. Trans. 2005, 113, 505–510. (13) Serio, M. A.; Hamblen, D. G.; Markham, J. R.; Solomon, P. R. Kinetics of volatile product evolution in coal pyrolysis: experiment and theory. Energ. Fuel. 1987, 1, 138–152. (14) Wiktorson, L. P.; Wanzl, W. Kinetic parameters for coal pyrolysis at low and high heating rates—A comparison of data from different laboratory equipment. Fuel 2000, 79, 701–706. (15) Given, P. H. Chemicals from coal. New Sci. 1962, 17, 355–357. (16) Speight, J. G. the Chemistry and Technology of Coal; Marcel Dekker: New York, 1994. (17) Gavalas, G. R. Coal Pyrolysis; New York: Elsevier; 1982. (18) Choura, M; Belgacem, N. M.; Gandini, A. Acid-catalyzed polycondensation of furfuryl alcohol: mechanisms of chromophore formation and cross-linking. Macromolecules 1996, 29, 3839–3850. (19) Handbook of Adhesive Technology; Pizzi, A., Mittal, K. L., Eds.; Marcel Dekker: New York, 1994. (20) Shindo, A.; Izumino, K. Structural variation during pyrolysis of furfuryl alcohol and furfural-furfuryl alcohol resins. Carbon 1994, 32, 1233–1243. (21) Lifshitz, A.; Tamburu, C.; Shashua, R. Decomposition of 2-methylfuran. Experimental and modeling study. J. Phys. Chem. A 1997, 101, 1018–1029. (22) Lifshitz, A.; Tamburu, C.; Shashua, R. Thermal decomposition of 2,5-dimethylfuran. Experimental results and computer modeling. J. Phys. Chem. A 1998, 102, 10655–10670. (23) Fadden, M. J.; Hadad, C. M. Rearrangement pathways of arylperoxy radicals. 2. Five-membered heterocycles. J. Phys. Chem. A 2000, 104, 6324–6331. (24) Liu, R. F.; Zhou, X. F.; Zuo, T. M. The pyrolysis mechanism of furan revisited. Chem. Phys. Lett. 2000, 325, 457–464.

’ REFERENCES (1) Modern casting staff report. 43rd Census of World Foundry Production - 2008. Mod. Cast. 2009, 99, 17 21. (2) Allen, G. R.; Archibald, J. J.; Keenan, T. Hazardous air pollutants: a challenge to metal casting industry. Am. Foundry Soc. Trans. 1991, 99, 585–593. (3) Pre-Production Air Emission Test Report, Parts 1 3; Technikon Environmental Development Center: McClellan, CA, 2000. (4) Tiedje, N.; Crepaz, R.; Eggert, T.; Bey, N. Emission of organic compounds from mould and core binders used for casting iron, aluminum and bronze in sand moulds. J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng. 2010, 45, 1866–1876. (5) Wang, Y. J.; Huang, H.; Cannon, F. S.; Voigt, R. C.; Komarneni, S.; Furness, J. C. Evaluation of volatile hydrocarbon emission characteristics of carbonaceous additives in green sand foundries. Environ. Sci. Technol. 2007, 41, 2957–2963. (6) Dungan, R. S.; Reeves, J. B. Pyrolysis of carbonaceous foundry sand additives: seacoal and gilsonite. Thermochim. Acta 2007, 460, 60–66. (7) Dungan, R. S.; Reeves, J. B. Pyrolysis of foundry sand resins: a determination of organic products by mass spectrometry. J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng. 2005, 40, 1557–1567. (8) Wang, Y. J.; Cannon, F. S.; Salama, M.; Fonseca, D. A.; Giese, S. Characterization of pyrolysis products from a biodiesel phenolic urethane binder. Environ. Sci. Technol. 2009, 43 (5), 1559–1564. (9) Wang, Y. J.; Cannon, F. S.; Salama, M.; Goudzwaard, J.; Furness, J. C. Characterization of hydrocarbon emissions from green sand 6544

dx.doi.org/10.1021/es200310p |Environ. Sci. Technol. 2011, 45, 6539–6544