Energy & Fuels 2003, 17, 225-239

225

Emissions of Batch Combustion of Waste Tire Chips: The Afterburner Effect Jefferson Caponero and Jorge A. S. Teno´rio* Polytechnic School, University of Sa˜ o Paulo, Sa˜ o Paulo, 05508-900 Brazil

Yiannis A. Levendis College of Engineering, Northeastern University, Boston, Massachusetts 02115

Joel B. Carlson United States Army Natick Research, Development and Engineering Center, Natick, Massachusetts 01760 Received June 11, 2002

A laboratory investigation was performed on the emissions from batch combustion of waste tire chips in fixed beds to identify techniques and conditions that minimize toxic emissions. Tire derived fuel (TDF), in the form of waste tire chips (1 cm), was burned in a two-stage combustor. Batches of tire chips were introduced to the primary furnace where gasification and oxidative pyrolysis took place. The gaseous effluent of this furnace was mixed with streams of additional air and, subsequently, it was channeled into the secondary furnace (afterburner) where further oxidation took place. The arrangement of two furnaces in series allows for independent temperature control; varying the temperature in the primary furnace influences the type and the flux of pyrolysates. The additional-air mixing section between the two furnaces allows for mostly heterogeneous and fuel-lean combustion in the afterburner. Results showed that both the operating temperature of the primary furnace, in the range of 500-1000 °C, and the existence of the afterburner had marked influences on the emissions of pollutants. Results showed that for this fuel use of combustion staging, with an additional-air mixing section, had a very beneficial effect. It drastically reduced the emissions of CO (by factors of 3-10), the particulates (by factors of 2-5) and the cumulative PAH (by factors of 2-3). Many health-hazardous PAH components were practically eliminated. Overall oxidizing conditions prevailed and the minimum oxygen mole fraction never fell below 2% in the effluent of either furnace. The operating primary furnace temperature (pyrolysis temperature) also proved to be important, with temperatures at the low side of the 500-1000 °C range producing fewer pollutants, upon treatment in the afterburner.

Introduction The disposal of scrap tires is a challenging environmental problem, especially for the industrialized countries. Approximately 264, 164, and 32 million tires are disposed of each year in the United States, Japan, and Brazil, respectively.1-4 In the United States, this amounts to approximately one used tire discarded per person annually, and the rate has been increasing lately. Nowadays, the major disposal methods have been dumping tires in landfills or in scrap tire stockpiles, either legal or illegal. However, scrap tires are a favorable place for proliferation of rodents and insects; thus, they pose a potential health hazard. A more serious problem is the fire hazard that scrap tires pose. * To whom correspondence should be addressed. Phone: +55 11 3091 5546. Fax: +55 11 3091 5243. E-mail: [email protected]. (1) Senneca, O.; Salatino, P.; Chirone, R. Fuel 1999, 78, 1575. (2) Jang, J. W.; et al. Resourc., Conserv. Recycl. 1998, 1-2, 1. (3) Cempre-Compromisso Empresarial Para Reciclagem. Pneus. On line 15/Jan/1999. Available from World Wide Web. URL: http:// www.cempre.org.br/ficha8.htm [27/Jan/1999, in Portuguese]. (4) Ferrer, G. Resourc., Conserv. Recycl. 1997, 4, 221.

It has been reported that there are several billion waste tires stockpiled or in landfills, just in the United States. The increased usage of automobiles in the industrial countries is exacerbating this problem. Improvements in the tire construction technology and, also in retreating of tires (in limited applications such as heavy-duty vehicles and airplanes) have increased tire life. Even if such improvements themselves should help minimize the tire disposal problem, the actual rate of tire disposal does not seen to reflect this reduction.5 The composition of tires varies depending on their use. Natural and synthetic rubber, carbon black, steel, aromatic oils, stabilizers, sulfur, and zinc oxide are the major components. Automobile tires are made normally with styrene-butadiene copolymers (SBR) or styrenebutadiene copolymer/polybutadiene mixtures (SBR/BR), while the sidewalls are generally made with addition of natural rubber (NR). Table 1 shows the typical tire compound composition.6 (5) Caponero, J.; Tenorio, J. A. S. In Proceedings of the 55th Congresso anual da ABM; Associac¸ a˜o Brasileira de Metalurgia e Materiais-ABM; Sa˜o Paulo, 2000 [in Portuguese].

10.1021/ef0201331 CCC: $25.00 © 2003 American Chemical Society Published on Web 12/18/2002

226

Energy & Fuels, Vol. 17, No. 1, 2003

Caponero et al.

Table 1. Typical Tire Compound Composition (wt %)6 SBR carbon black extender oil zinc oxide stearic acid sulfur accelerator total

62.1 31.0 1.9 1.9 1.2 1.1 0.7 99.9

Presently the major alternative to disposal in landfills is using waste tires as a fuel. The tire-derived fuel (TDF) has been used successfully in cement kilns, lime kilns, paper and pump mills, iron foundries, copper smelters, and waste-to-energy plants. Combustion The reason that waste tires are targeted for combustion is their high-energy content (calorific value), which is in the range of 29-39 MJ/kg; this is equal or higher than that of most typical coals.7 The combustion behavior of the coal has been extensively studied over the last few decades and still is. Recent literature has reported on similarities in the combustion behavior of pulverized coal and tire crumb.8-13 The tires’ content of major elemental species (carbon, minerals, sulfur and nitrogen) is similar to that of coal or coke.14 The tire volatile mass (∼60%) is approximately two times higher than that of most bituminous coals. The major differences are the moisture content, markedly low in tires, and zinc content, markedly high in tires.5,8 The nitrogen content of tires is a fraction of that of coal, while the sulfur content is generally comparable. Test-firing of whole tires or shredded tires, has been conducted in various types of boilers, with or without co-firing with coal, see for example reports by Epri,15 Pirnie,16 Clark et al.,17 Lemieux and co-workers,18,19 Kearney,20 and Harding21 reviewed recent commercial demonstrations of TDF co-firing with coal in various types of boilers. The boilers best suited for using TDF (6) Willians, P. T.; Besler, S. Fuel 1995, 74, 1277. (7) Blumenthal, M. H. Tires. In Recycling Handbook; Inlund, H. F., Ed.; McGraw-Hill: New York, 1993; Chapter 18, pp 18.1-18.64. (8) Atal, A.; Levendis, Y. A. Fuel 1995, 74, 1570. (9) Levendis, Y. A.; Atal, A.; Steciak, J. Combustion and Inorganic Emissions of Ground Waste Tires. In Proceedings of the 20th International Conference on Coal Utilization and Fuel Systems, March 1995, Clearwater, FL. (10) Atal, A.; Steciak, J.; Levendis, Y. A. NOx and SOx Emissions from Pulverized Coal and Waste Tire: The Role of Devolatilization and Char Combustion Phases. In Proceedings of the ASME Heat Transfer Division, HTD, Nov 1995, San Francisco, CA; Vol. 317-2. (11) Levendis, Y. A.; Atal, A.; Carlson, J.; Dunayevskiy, Y.; Vouros, P. Environ. Sci. Technol. 1996, 30, 2742. (12) Levendis, Y. A.; Atal, A.; Courtemanche, B.; Carlson, J. Combust. Sci. Technol. 1998, 131, 147. (13) Courtemanche, B.; Levendis, Y. A. Fuel 1998, 77, 183. (14) Teng, H.; et al. Ind., Eng. Chem. Res. 1995, 9, 3102. (15) EPRI GS-GS-7538. In Proceedings of the Conference on Waste Tires as a Utility Fuel, 1991; Prepared by Electric Power Research Institute: Palo Alto, CA, Sept, 1991. (16) Pirnie, M. Air Emissions Associated with the Combustion of Scrap Tires for Energy Recovery; Prepared for the Ohio Air Quality Development Authority: Columbus, May, 1991. (17) Clarc, C.; Meardon, K.; Russell, D. Burning Tires for Fuel and Tire Pyrolysis: Air Implications. EPA-450/3-91-024 (NTIS PB92-145358); U.S. Environmental Protection Agency: Washington, DC, Dec 1991. (18) Lemieux, P. M.; Ryan, J. V. Air Waste. 1993, 43, 1106. (19) Lemieux, P. M. Pilot-Scale Evaluation of the Potential for Emissions of Hazardous Air Pollutants from Combustion of Tire Derived Fuel. EPA-600-R-94-070; U.S. Environmental Protection Agency: Washington, DC, April 1994. (20) Kearney, A. T. Scrap Tire Use/Disposal Study; Final Report: Prepared for the Scrap Tire Management Council, Sept 1990.

are cyclones, stockers and fluidized beds. If tire-derived fuel (TDF) were to be burned in existing pulverized coalfired utility boilers, it would have to be ground to tire crumb. Previous studies in this laboratory examined the combustion and emissions of pulverized tires, i.e., in the form of tire crumb8-13 and found that extensive grinding is not necessary since tire particles as big as 250 µm burned as fast as much smaller coal particles (≈75 µm, which is the typical size used in pulverized coal boilers). Tire particles also burned somewhat hotter but released 3-5 times less NOx emissions, 10% less CO2 and comparable SO2 (as the sulfur contents of the particular fuels burned were similar). However, they released more PAH’s than coal. Such emissions were reduced by adding air, in excess to that needed for coal.11 Another effective method for minimizing such emissions was found to be co-firing tire crumb with coal in pulverized fuel flames.12 In summary, pulverized-flame type combustion behavior of tire crumb in the laboratory appeared similar to that of pulverized coal and emissions were to some degree comparable, at least under some conditions.8,10-12 Nevertheless, as tire pulverization is currently both challenging and costly, it is economically advantageous to burn tires either as whole units or shredded. Combustion of whole or shredded tires, i.e., chips or chunks with dimensions of 1-3 cm, is currently of technological interest. Combustion of whole tires has found applications in cement kilns and dedicated waste to energy plants. The high temperature of the rotary cement kiln (2000 °C on gases and 1450 °C on load) creates a propitious environment to oxidize the tire steel. Tires provide the kiln with extra energy and substitute part of the iron ore used as a row material. The zinc oxide, always present, functions as a mineralizer in the clinker production lowering the clinkerization temperature.22 Problems are often encountered with the carbon conversion efficiency. Also, only limited data has been reported in the literature on the emissions from such plants, particularly the organic air toxics. Yet, most people have vivid images of copious amounts of black smoke emitted from burning whole tires, either single or in stockpiles. Thus, past work at this laboratory aimed at investigating whether batch combustion of chunks (chips) of tires (simulating whole tires in a small scale) emits more pollutants than the aforementioned combustion of tire crumb in pulverizedfuel flames. Indeed, the results showed that batch combustion of tire chips resulted in much higher emissions of PAH and particulate matter.23 Generally, combustion of fine fuel particles in pulverized fuel furnaces offers better mixing of the fuel and the oxidant than combustion of chunks of solid fuel. On the other hand, it does not appear to influence the nature of the species produced. Similar species were reported for both tire chip and tire crumb combustion at elevated temperatures and comparable residence times.8,11,12,23 (21) Harding, N. S. Cofiring Tire-Derived Fuel with Coal. In Proceedings of the 27th International Technical Conference on Coal Utilization and Fuel Systems, March 4-7, 2002, Clearwater, FL; pp 477-488. (22) Caponero, J. Comportamento da lama de fosfatizac¸ a˜o no processo de produc¸ a˜o do clı´nquer de cimento Portland. Master Thesis, Universidade of Sa˜o Paulo, Sa˜o Paulo, Brazil, 1999 [in Portuguese]. (23) Levendis, Y. A.; Atal, A.; Carlson, J. B. Combust. Sci. Technol. 1998, 134, 407.

Batch Combustion of Waste Tire Chips

Batch combustion of chunks tire chips has also been contrasted to batch combustion of tire crumb and batch combustion of pulverized coal24 and results may be summarized as follows: (i) Batch combustion of TDF (either crumb or chips) in a muffle furnace generated PAH emissions, which were orders of magnitude higher than those from burning streams of tire crumb in a drop-tube furnace, depending on the bulk equivalence ratio, φ. In contrast, combustion of pulverized coal in batch mode (muffle furnace) and in streams (drop-tube furnace) resulted in comparable amounts of PAH’s. (ii) Batch combustion of pulverized coal or tire crumb generated cumulative CO emissions, which were much higher than those from burning streams of pulverized coal or tire crumb in a drop-tube furnace, at comparable feed rates. (iii) Batch combustion of tire crumb in fixed beds resulted in cumulative PAH emission that were an order of magnitude higher than those from batch combustion of pulverized coal, at comparable mass loadings. (iv) All of the monitored cumulative PAH and CO emissions from both fuels in batch combustion, originated from the volatile flames. Such differences in the emissions of PAH and, eventually, particulate matter can be understood by examining the combustion of tire chips. It begins with the rapid heating of the fuel. When a certain temperature is reached on the particle surface, pyrolysis occurs along with thermal oxidation reactions. When chips are subjected to temperatures in the range of 350-600 °C in an oxygen-lean environment, as the surface temperature increases, decomposition or degradation takes place. The devolatilizates from tires are composed of a wide variety of hydrocarbons, called primary products. These products then undergo secondary reactions, such as, thermal and catalytic cracking, re-polymerization, cyclization of alkyl chains, recondensation oxidation and reduction. Upon ignition, a diffusion flame forms around the tire chip, where a gradient of oxygen partial pressure exists. Fuel pyrolysis generally begins at relatively low temperature as the fuel approaches the flame front. The most accepted mechanism for soot formation from aliphatic fuels is through the formation of acetylene and polyacetylenes at a relatively slow rate. Aromatic fuels may form soot by a similar process, but also through a direct route involving ring condensation or polymerization reactions that build on existing aromatic structures.25 The increase of the pyrolysis rate with temperature and/or with the amount of volatiles in the solid fuel leads to an increasing tendency to form PAH and soot. Particles in the order of 10-20 nm form, which then flocculate and fuse to aggregates of soot. Simplified mechanisms of the soot formation, during the secondary reactions, are presented by WU et al.26 and MASTRAL et al.27 Some PAH components, either in the gas phase or condensed on soot particles, are among the most (24) Levendis, Y. A.; Atal, A. Emissions from Burning Tire-Derived Fuel (TDF): Comparation of Batch Combustion of Tire Chips and Continuous of Tire Crumb. In Proceedings of the 23rd Coal Utilization & Fuel Systems International Conference, March 9-12, 1998, Clearwater, FL. (25) Graham, S. C.; Homer, J. B.; Rosenfeld, J. L. J. Proc. R. Soc. London 1975, 344A, 259-285. (26) Wu, S. Y.; Su, M. F.; Baeyens, J. Powder Technol. 1997, 93, 283. (27) Mastral, A. M.; Collen, M.; Murillo, R. Fuel 1996, 75, 1533.

Energy & Fuels, Vol. 17, No. 1, 2003 227

problematic emissions of the combustion process. They are considered hazardous due to possible interactions with biological nucleophiles, resulting in inhibition of their regular functions.28 The carbon accounts for nearly half of the mass of atmospheric fine and ultrafine particles and it is present in basically two forms: organic carbon, which includes hundreds of compounds, and elemental carbon.29 Therefore, a challenge in TDF combustion/pyrolysis is the development of a controlled combustion technique that minimizes the emissions of PAH and soot. The emissions of NOx in tire combustion were reported to be lower in tire than in coal, burned under identical conditions. Courtemanche and Levendis found emissions NOx from TDF to be 4 times lower than those from coal.13 This was mostly related to the nitrogen content of the fuels. In that work it was also observed that the emissions of SO2 and CO2 were proportional to the sulfur content and the carbon content of both fuels. In summary, previous studies in this laboratory showed that burning chunks of tires can be problematic because of the copious amounts of particulate matter and PAH that are released. Soot, once formed, is difficult to burn in typical post-flame furnace conditions. Thus, this work concentrated on identifying combustion techniques and furnace operating parameters that reduce the emissions from burning waste-tire chips. Chips were again burned in batches in a horizontal muffle furnace. However, this time two furnaces were used. The primary furnace acted as a gasifier/combustor. The gaseous effluent of this furnace was mixed with additional air, in a mixing section, and was channeled to a secondary furnace (afterburner) where additional oxidation took place, see Figure 1, under well-mixed and overall fuellean conditions. It was expected that with the arrangement of two furnaces in series, separated by the mixing section, combustion emissions would be reduced. In these experiments the fuel mass loading in the primary furnace/gasifier was fixed while the temperature was varied in the range of 500-1000 °C; the residence time of the effluent gases in the secondary furnace (afterburner) was under 1 s and the temperature therein was 1000 °C. At the exits of the furnaces CO, CO2, NOx, SO2, polycyclic aromatic hydrocarbon (PAH), and particulate emissions were monitored. Materials and Experimental Methods Tire chips with dimensions in the order of a centimeter were obtained from a local source and included organic fabrics, such as nylon belts, see Figure 2. No metallic belts were included. Physical and chemical properties are shown in Table 2. The elemental analysis resulted in the mass fractions shown in Table 3. All of the tests conducted in this study involved batch combustion of tire chips in fixed beds. Preweighted amounts of 0.8 g, consisting of a few (typically 3) chips, were placed in porcelain boats and were inserted in the quartz tube of the primary furnace, 4 cm in diameter and 87 cm long. The tube was placed at the centerline of a horizontal split-cell electric furnace (1 kW max.), see Figure 1. To insert the samples quickly in the furnace the porcelain boats were placed at the end of the inner surface of a half tube (a quartz cylinder longitudinally split along the centerline). The other end of the (28) Mastral, A.; et al. Fuel 1998, 77, 1516. (29) Violi, A.; D’Anna, A.; D’Alessio, A. Chem. Eng. Sci. 1999, 54, 3433.

228 Energy & Fuels, Vol. 17, No. 1, 2003

Caponero et al.

Figure 1. Schematic of the bench-scale experimental apparatus.

Figure 2. A photograph of the waste dewired tire chips that were burned in these experiments. Table 2. Some Physical and Chemical Properties of the Tire Chips Used fixed carbon

volatiles

ash

apparent density

heating value

residue specific area

24.9%

69.7%

4.0%

0.4 g/cm3

39 MJ/kg

74 m2/g

Table 3. Composition of Tire Chips Fractions, wt % element

tire chips

burned fractiona

carbon black residue

carbon hydrogen sulfur others

85.8 7.3 2.3 4.6

87.4 10.2 2.4 0.0

82.2 <0.5 2.2 15.1

a

Calculated.

half tube was connected at the entrance glass fitting of the furnace quartz tube. At the start of each experiment, the fitting was opened and the half tube, with the sample mounted at its tip, was quickly inserted in the furnace, so that the half tube inside the furnace was parallel to the 87-cm long quartz tube. All experiments were conducted in air. The airflow rate in the first furnace was 4 L/min, and the residence time of the gases between the sample location and the venturi was a fraction of a second (0.7 to 0.4 s, depending on the furnace temperature). This primary furnace was connected to a secondary muffle furnace (the afterburner), as shown in Figure 1. The dimensions of the secondary furnace were 2 cm in diameter and 38 cm long. The effluent of the first furnace passed through a venturi (8 mm in diameter) where it was mixed with four radially positioned perpendicular jets of preheated air, discharging at the periphery of the venturi, see Figure 1. Mixing of the air from the jets with the gaseous effluent of the combustion occurred in the Venturi. The mixed charge was then sampled at the exit of the primary furnace, and subsequently channeled to the afterburner.

The wall temperature of the primary furnace was varied in the range of 500-1000 °C, while that of the secondary furnace was kept constant at 1000 °C throughout these experiments. The gas temperature in the section between the furnaces dipped to 200-250 °C. The furnace temperatures were continuously monitored using thermocouples mounted outside the quartz furnace tube. The temperature profile of the gas inside the muffle furnace had been measured in prior work using a suction pyrometer.30 The tire sample temperature was not measured is not easy to obtain because of the formation of the heavily sooting diffusion flame around the sample, which obstructs pyrometric measurements. Tire surface temperatures during pyrolysis are presumably in the range of gas temperatures in these experiments. This is inferred based on previous studies in this laboratory where both single particles and groups of particles of tire crumb were fluidized and burned and their combustion behavior was monitored cinematographically and pyrometrically.8,12 The flow rate of additional air supplied at the venturi, through the four jets, was 2 L/min. The calculated Reynolds number in the venturi was 425. The residence time of the gases in the secondary furnace was calculated to be in the range of 0.6-0.8 s, according to the temperature therein, and the Reynolds number was in the range of 75-85. The laminar flow was selected in order to isolate effects of single parameters, such as: the combustion temperature, the residence time, the existence of the afterburner, etc., due to the associated repeatable and well-characterized environment. To estimate the penetration of the additional air jets at the venturi and assess the mixing with the effluent gas from the primary furnace, the approach outlined by Lefebvre31 was used. Accordingly, the maximum penetration length (Ymax) of multiple perpendicular round jets into a tubular duct (in this case the venturi) is given by

Ymax ) 1.25djxJ

(

)

m ˘g m ˘ g+m ˘j

(1)

where dj is the diameter of the jets, m ˘ is the mass flow rates of the gas in the circular ducts (g) and the air in the round jets (j), and J is the momentum flux ratio given by

J)

FjUj2 FgUg2

(2)

where F is the density of gases and U is the velocity of gases. (30) Wheatley, L.; Levendis, Y. A.; Vouros, P. Environ. Sci. Technol. 1993, 27, 2885. (31) Lefebvre, A. H. Gas Turbine Combustion, 2nd ed.; Taylor & Francis: New York, 1998.

Batch Combustion of Waste Tire Chips Calculations using measured temperatures at the venturi centerline and the jet exits showed that the four radially placed jets penetrate the effluent to its centerline. This is indicative of good mixing of the effluent and air streams in the venturi. Sampling was conducted simultaneously at the exits of both furnaces; the typical duration of the luminous diffusion flame over the boat was in the order of 1-2 min. Diffusion flames forming over the TDF sample bed appeared to be remarkably steady throughout each experiment. Monitoring of Combustion Emissions. Polycyclic aromatic hydrocarbon (PAH) emissions, as well as particulates (mostly soot and, possibly, traces of minerals) from the combustion of tire chips were monitored at the exits of the two furnaces. The PAH’s were sampled by passing half of the effluent through each of the sampling stages consisting of a Graseby sampling head with a filter stage and a glass cartridge containing XAD-4 adsorbent. The sampling stages were placed adjacent to the furnace to minimize losses. Prior to each sampling stage the effluent of the furnaces was mixed with a 2 L/min flow of nitrogen to be cooled and inerted, so that chemical reactions would be quenched. This dilution nitrogen flow took place in the annulus of two concentric tubes; the inner tube was perforated and, therefore, enabled the mixing of the nitrogen with the furnace effluent (Figure 1). Subsequently, the particulate emissions were retained on the upper portion of the sampling stage on a 11 cm diameter Fisher Scientific cellulose filter paper with a nominal particle retention of 1-5 µm. Gas-phase aromatic hydrocarbon emissions were adsorbed on the bed of Supelco XAD-4 resin. The length of the XAD-4 bed was more than twice its diameter to enhance adsorption. Exiting each sampling stage the effluent passed through an ice bath and a mildly heated Premature dryer, where moisture was removed. The effluent was then monitored for O2 using a Beckman paramagnetic on-line analyzer and for CO and CO2 with Horiba and Beckman infrared analyzers. The first stage effluent was also monitored for NOx using a Beckman 951A chemiluminescent NO/NOx analyzer, for SO2 by means of a Rosemount Analytical 590 UV analyzer. The output of the analyzers was recorded using a Data Translation DT-322 data acquisition board in a microcomputer. The integrated acquisition system used DT VPI to support data acquisition from within a Hewlett-Packard HP VEE visual programming environment and automatically formed Microsoft Excel data files corresponding to the different experiments. The signals from the analyzers were recorded for the duration of the experiment and subsequently were converted to partial pressures and, eventually, to mass yields (mg/mass of sample burned). Sampling of Particulate Matter. The fraction of particulate matter inhaled and kept into the human respiratory system as well as its place of holding varies as a function of size, shape, density and all physical properties that influence the aerodynamic dimensions of a particle.32-34 The size of the TDF combustion generated particulates was measured using an Andersen impactor (1 ACFM nonviable ambient particle sizing samplers), an equipment that simulates the human respiratory system. This is a multi-stage impactor, composed of a series of perforated disks (stages). In each stage the size of the holes and their distribution results in the particle size, stages 1-7 have the following size cuts: 9, 5.8, 4.7, 3.3, 2.1, 1.1, 0.7, and 0.4 µm. The apparatus has also a bottom filter which is assumed to retain any remained particle.32 This instrument was used instead of the aforementioned sampling (32) Operating Manual for Andersen 1 ACFM Non-Viable Ambient Particle Sizing Samplers; Andersen Instruments Incorporated: 1985. (33) Health and Environmental Effects of Particulate Matter. Fact Sheet; U.S. Environmental Protection Agency: Washington, DC, 1999. Available online at http://www.epa.gov/ttn/oarpg/naaqsfin/pmhealth.htm. (34) Schwartz, J.; Dockery, D. W.; Neas, L. M. J. Air Waste Manage. Assoc. 1996, 46, 927.

Energy & Fuels, Vol. 17, No. 1, 2003 229 head in separate experiments. It was retrofitted at the exit of either the primary or the secondary furnace. Extraction and Concentration of PAH Emissions. Upon completion of each combustion run, the filters and resins were removed and placed in separate glass bottles with Teflonlined caps, and stored at 4 °C. Prior to extraction with methylene chloride, a 25 mL internal standard containing 50µg each of naphthalene-d8, acenaphthene-d10, anthracene-d10, chrysene-d12, and perylene-d12 was applied to each of the glass bottles containing the samples. The XAD-4 resin and cellulose filters, used to trap the PAH emissions, were oven dried at 50 °C prior to use. To ensure the purity of the XAD-4 resin and cellulose filters, blanks of XAD-4 resin and filter were also extracted and analyzed. In addition, combustion blanks were performed in which the furnace was operated in the presence of the XAD-4 and filter, but with no fuel present. Target compounds that appeared in any of the blanks were appropriately qualified based on their concentration in any of the blanks. Combustion experiments were at least duplicated to ensure the reproducibility of the combustion technique. To substantially reduce the amount of methylene chloride in the extraction process, a Dionex ASE 200 Accelerated Solvent Extractor was used in place of the Soxhlet extraction system methods reported previously.8,11,35,36 This revision in the extraction method necessitated the quality assurance evaluation of the automated solvent extraction system to ensure the extraction suitability, efficiency, and reproducibility of the proposed method. The Dionex ASE 200 was used for extracting the organic compounds from both the XAD-4 resins and the cellulose filter papers. The XAD-4 resins were transferred to 33 mL extraction cells while the filter papers were transferred to 11 mL extraction cells. The extraction cells were allowed to initially equilibrate at 40 °C in the ASE 200 system for 1 min, then they were filled with methylene chloride and allowed to thermally equilibrate at 40 °C; the cells were subsequently pressurized to 34 atm (500 psi) for a period of 15 min. Following the 15 min soak time the cells were each flushed with 80% of the cell volume with fresh methylene chloride and finally purged for 90 s with nitrogen. The methylene chloride extracts were collected in separate bottles for concentration. Two extraction cycles were used per cell. The total extraction time for the two-cycle process was about 25 min and about 45 mL of methylene chloride were used for the XAD-4 resins while about 20 mL were used for the extraction of the filter papers. The original bottles that stored the combustion resins and filter papers were rinsed twice with 1 mL of methylene chloride and added to the vials containing the methylene chloride extracts. No more than 30 mL of XAD-4 resin could be placed within a 33 mL extraction cell due to the expansion of this resin in methylene chloride. The cells had been thoroughly cleaned and inspected to ensure that small resin particles have not become trapped between the stainless steel cell body and the seals in the end caps of the cell. The samples were concentrated under vacuum to a final volume of 10 mL for analysis by gas chromatography coupled to mass spectrometry (GC-MS). Analysis by Gas Chromatography Coupled to Mass Spectrometry. The GC-MS system consisted of a HewlettPackard (HP) Model 5890 GC equipped with an HP Model 5971 mass selective detector. The GC-MS conditions and data reduction were described previously.8,11,35,36 The GC was equipped with an auto sampler and tray. The GC column was a HP Ultra-2, 5% phenyl-methyl-silicone with a length of 25 m, an inside diameter of 0.2 mm and a film thickness of 0.33 µm. The column head pressure was 480 mbar, the flow rate 0.7 mL/min, and the linear flow rate 32 cm/s (methane) at 100 °C. The split flow was 2.0 mL/min and the split ratio (35) Panagiotou, T.; et al. In Proceedings of the Twenty-Sixth Symposium (International) on Combustion; 1996; pp 2142-2430. (36) Atal, A.; Levendis, Y. A.; Carlson, J.; Dunayevskiy, Y.; Vouros, P. Combust. Flame 1997, 110, 462.

230

Energy & Fuels, Vol. 17, No. 1, 2003

was 2.7:1. The instrument was tuned in accordance with EPA semi volatile criteria prior to the GC-MS analysis of each set of samples. In a departure from the EPA method, the GC-MS system was run in the full scan mode and not in a single ionmonitoring mode. This was done to ensure the identification; quantification and reporting of tentatively identified compounds. The use of the full scan mode does not significantly modify the method except to raise the lower reporting limit to about 1µg of component per gram of fuel combusted. Values shown in the Tables within this paper that are less than 1 µg/g are technically nondetects. Each of the target compounds, as well as each of the tentatively identified compounds, was quantified using the appropriate deuterated internal standard. Standard solutions containing 50 µg each of the 5 aforementioned characteristic standards were diluted to 10 mL and analyzed as an instrument blank and to provide an indication of extraction efficiency for each of the internal standards. Samples with an extraction efficiency of less than 50% for any of the internal standards were repeated. To assess the reproducibility of the PAH analysis, triplicate analyses were performed and averages were taken. Experimental errors result from a combination of sampling, extraction, concentration and analysis techniques. The experimental procedure was kept consistent in all evaluations to ensure the validity of relative trends. Extraction Recovery. The extraction recovery was well above 50% for the internal standards in every sample analyzed. The National Functional Guidelines recommend a recovery of greater than 50% for the internal standards. Remarkably the percentage difference (ratio of standard deviation to average recovery) was very low demonstrating the excellent reproducibility of this extraction technique for the analysis of XAD-4 resins and filter papers. The recovery efficiency and reproducibility of the automated solvent extraction technique described herein was judged to be a suitable replacement for conventional Soxhlet techniques based on quality assurance and quality control evaluations. The application of this technique would have to be substantially modified for the extraction of samples in which complex matrixes are present. The observed lower recovery efficiencies for the more volatile of the internal standards most likely result from the vacuum concentration of the samples from 50 to 10 mL. Modification of several of the other experimental parameters in the methodology of sample generation and preparation may be employed in future experiments to eliminate the need for a final concentration step and improve the recovery of the more volatile of the internal standards.

Results and Discussion Upon its introduction to the preheated furnace, the fuel bed heated and ignited. Two distinct combustion phases were observed; burning of the volatile matter followed by burning of the char, which mostly consists of carbon black. During the devolatilization period a diffusion flame formed over the fuel bed, and a laminar plume of smoke was visible flowing to the exit of the furnace. The flame appeared to be rather steady and lasted for a period of approximately one to two minutes. The rapid mass loss during devolatilization of the tire chips resulted in insufficient mixing of the fuel pyrolysates and the oxidant. This was followed by a lengthy char combustion phase (in the order of 10 min). The two combustion phases were studied in detail in previous work.8,12,23 It was shown that the contributions of the char combustion phase to the organic pollutant emissions were insignificant. Thus, this work concentrated on the emissions from just the volatile combustion phase

Caponero et al.

Figure 3. Emission profiles of a single experiment from batch combustion of volatile matter of tire chips at the exits of the primary (stage 1) and the secondary furnace (stage 2), both operated at 1000 °C.

of tire chips. The effects of two parameters on these emissions were monitored: (a) the primary furnace temperature and (b) the effect of the existence of a secondary furnace (afterburner). The primary furnace acted as a gasifier/combustor and its temperature was varied between 500 °C and 1000 °C. The afterburner’s temperature was kept constant at 1000 °C in all tests and the gas residence time therein was calculated to be 0.7 s. All other conditions were kept the same in these experiments. The reader should keep in mind that batch combustion of the fuel resulted in transient (timedependent) emission profiles. Figure 3 shows typical emission profiles for O2, CO2, and CO observed in these experiments. On the bases of the mass of the TDF volatiles burned, the overall duration of their combustion, and the air flow rate in the primary furnace, it was calculated that global (overall) equivalence ratios were mildly fuel rich (φ ) 1-1.2). Upon ignition of the fuel, the emission levels increased, they reached a maximum and, thereafter, decreased as the flame dwindled to extinction. Emission yields were obtained by integrating each profile and dividing by the mass of the fuel burned; in this case only the volatile component of the tire chips. Oxygen Consumption. Globally, neither furnace’s atmosphere was ever starved for oxygen during the combustion events, as the minimum O2 mole fractions in both furnaces never fell below 2%, see Figure 4a. This indicates that local equivalent ratios in the flame were much richer than the φ ) 1-1.2 range calculated above. The oxygen partial pressure at the exit of the secondary furnace was always lower than the corresponding value at the exit of the primary furnace, Figure 4; both values already including the additional air introduced in the venturi. This is because further conversion of volatile pyrolysates and primary combustion products takes place in the secondary furnace, thereby consuming oxygen. Inorganic Emissions. Maximum CO partial pressures at the exit of the primary furnace were typically below 3%, while those at the exit of the secondary furnace were much lower, see Figure 5a. The CO emission yields were drastically reduced by the afterburner, the minimum reduction being by a factor of 3 and the maximum by a factor of 250, see Figure 5b. The

Batch Combustion of Waste Tire Chips

Figure 4. 4. Partial pressures and emission yields for O2 at the exits of both the primary and the secondary furnace, for different primary furnace temperatures (500-1000 °C). The afterburner temperature was kept constant at 1000 °C.

optimum operating point, for low CO emissions, appeared to be a low temperature in the first furnace (500-600 °C) followed by treatment in the afterburner. Both the primary furnace temperature and the existence of the afterburner affected the CO2 emission yield. Maximum CO2 partial pressures were recorded to be in the ranges of 6-8% and 9-13% at the exits of the primary and the secondary furnace, respectively, see Figure 6. Again, this is because additional combustion is archived in the afterburner. Yields of SO2 and NOx emitted during the volatile phase of batch combustion of tire chips were only recorded at the exit of the primary furnace, see Figure 7. Both these emissions were found to be generally low. There was no pronounced trend in SO2 emissions, while the NOx emissions appeared to increase with primary furnace temperature. The latter trend is expected, since flame temperatures increase with furnace gas temperatures. Increasing flame temperatures enhances both the generation of fuel-NOx and, particularly, thermal NOx. A large fraction of the fuel nitrogen and sulfur remained in the char and the ash, see Table 3 and ref 9. A detailed investigation of the emissions of sulfur and nitrogen oxides from both the volatile and the char combustion phase of TDF and coal was conducted in the past, and details are given in ref 9.

Energy & Fuels, Vol. 17, No. 1, 2003 231

Figure 5. Partial pressures and emission yields for CO due to the combustion of the volatile matter of tire chips at the exits of both the primary and the secondary furnace, for different primary furnace temperatures (500-1000 °C). The afterburner temperature was kept constant at 1000 °C.

Particulate Matter Emissions. The particulate yields (soot, tars, and oils) at the exits of the two furnaces are shown in Figure 8. The particulate emissions from the primary furnace increased with the furnace temperature, since the furnace temperature influences the flame temperature itself. Higher diffusion flame temperatures (in a certain temperature region above ≈1070 °C) produce more soot.37 The effect of the afterburner was highly beneficial in reducing the particulate emissions, especially when the primary furnace was operated at the lower temperatures (500-600 °C), see Figure 8. This is because at the lower primary furnace (gasifier) temperatures the generated particulate matter was apparently composed of oils and tars, forming a light brown cloud which was readily oxidized in the afterburner. At the highest primary furnace temperatures the particulate matter was mostly black soot, which is very resistant to oxidation at the temperature of the afterburner (1000 °C) and, thus, it was only partially destroyed therein. The efficiency of the afterburner in oxidizing the particulate matter was calculated as follows:

ηpm )

∆m mi

(3)

232

Energy & Fuels, Vol. 17, No. 1, 2003

Caponero et al.

Figure 6. Partial pressures and emission yields for CO2 due to the combustion of the volatile matter of tire chips at the exits of both the primary and the secondary furnace, for different primary furnace temperatures (500-1000 °C). The afterburner temperature was kept constant at 1000 °C.

Figure 7. Emission yields for SO2 and NOx due to the combustion of the volatile matter of tire chips at the exits of both the primary and the secondary furnace, for different primary furnace temperatures (500-1000 °C). The afterburner temperature was kept constant at 1000 °C.

where ηpm is the efficiency in oxidizing the particulate matter; ∆m is the difference between the particulate mass collected at the exits of the primary and the secondary furnaces; and mi is the inlet of particulate mass in the afterburner which, since the effluent was split in two, was assumed to equal the amount of particulate mass collected on the paper filter at the exhaust of the primary furnace. Assuming that pore diffusion is negligible for the small size of the soot particles, the intrinsic reaction rate is approximated by

Rin )

ηpm tS

(4)

where t is the residence time in the second furnace, calculated as 0.7 s in the present work, and S is the surface area of the particulate matter per gram, cm2/g. While soot is generally considered a nonporous material, some researchers have reported the existence of porosity. There has been evidence that soot has internal surface area and that and this area increases with carbon oxidation by O2 in fuel-lean flames.38-41 Recent (37) Glassman, I. In Proceedings of The Combustion Institute; Pittsburgh, PA, 1988; pp 295-311.

Figure 8. Emission yields for particulate due to the combustion of the volatile matter of tire chips at the exits of both the primary and the secondary furnace, for different primary furnace temperatures (500-1000 °C). The afterburner temperature was kept constant at 1000 °C.

measurements revealed that upon oxidation the surface areas of two diesel soots (one previously collected in our (38) Neoh, K. G.; Howard, J. B.; Sarofim A. F. In Proceedings of The Combustion Institute; Pittsburgh, PA, 1984; pp 951-957.

Batch Combustion of Waste Tire Chips

Energy & Fuels, Vol. 17, No. 1, 2003 233

Table 4. Comparison of the Intrinsic Reaction Rate of the Oxidation of the Particulate Matter Yielded during the Combustion of Tire Chips with Other Empirical Data from the Literature at an Afterburner Gas Temperature of 1000 °C results N. as a as a and Neeft Wang porous nonporous S.C. Smith et al. et al. material material Rin (µg/cm2 s) 14.8 ηpm (%) 100a a

2.00 100a

0.321 0.163 100a 23

0.556

8.2 78

Total oxidation assumed in the calculations.

laboratory42 and another obtained from NIST) reached values in the range of 300-400 m2/g.43 Such measurements were performed with two independent methods, CO2 adsorption and small angle X-ray scattering (SAXS). However, no such measurements were performed on this particular soot from tires. Thus, two limiting cases of pore penetration were explored in this work, upon defining an effectiveness factor, η, based on the THIELE modulus, φ.44 If pores exist, a value of η ) 1 would indicate complete penetration of the pores by the oxidizing gas. If pores do not exist or they are not accessible then a value of η ) 0 would indicate no penetration. The THIELE modulus was calculated45 using Nagle and Strickland-Constable46 kinetics at Tgas ) 1000 °C. The pore diffusivity was calculated based on the rather reasonable assumption of nanometer-size pores. For the reported size of 200 nm spherules, see below, the Thiele modulus was very small (1 × 10-8). Then as (φ tanhφ) f φ2 the effectiveness factor η f 1. This indicates complete penetration of oxygen in the soot spherules, i.e., Regime I of combustion.47 This result was shown to be valid even when parameters such as the pore size, the porosity, the pore tortuosity, etc., were perturbed in the calculation. For the limiting case where oxygen penetrates the interior of the particle (η f 1) and using an estimate of S ) 200 m2/g for the internal surface area throughout carbon conversion in eq 4, the intrinsic reaction rate for the oxidation of the particulate matter was calculated to be 0.56 µg/cm2s. For the other limiting case, where oxidation only takes place at the exterior surface of the particle, S ≈ 14 m2/g, the intrinsic reaction rate for the oxidation of the particulate matter was calculated to be 8.2 µg/cm2s. A comparison of both of these rates with values obtained using the semiempirical expressions of Smith48 (for a variety of carbonaceous chars), Nagle and Strickland-Constable46 (for pyrolytic graphite), and Neeft et al.49 (for diesel soot) is shown in Table 4.The (39) Wicke, B. G.; Grady, K. A. Carbon 1981, 25, 791. (40) Du, Z. Kinetic Modelling of Carbon Oxidation. Ph.D. Thesis, Mechanical Engineering, MIT, 1990. (41) Bonnefoy, F.; et al. Carbon 1994, 7,1333. (42) Larsen, C.; Levendis, Y. A. SAE Publication No. 1999-01-0466 also SP-1414, 1999. (43) Kandas, A. W.; Senel G.; Levendis, Y. A.; Sarofim, A. F. Unpublished results. (44) Thiele, E. W. Ind. Eng. Chem.1939, 31, 916. (45) Levendis, Y. A.; Flagan, R. C.; Gavalas, G. R. Combust. Flame 1989, 76, 221. (46) Nagle, J.; Strickland-Constable, R. F. Oxidation of Carbon between 1000 and 2000 °C. In Proceedings of the 5th Carbon Conference; Pergamon Press: Oxford, 1961; Vol. 1, pp 154-164. (47) Smith, I. W.; Tyler, R. J. Fuel 1972, 51, 312. (48) Smith, I. W. In Proceedings of the The Combustion Institute: Pittsburgh, PA, 1982; pp 1045-1065. (49) Neeft, J. P. A.; et al. Fuel 1997, 76, 1129.

Figure 9. Particle size distribution of the particulate emission due to the combustion of the volatile matter of tire chips at the exits of (a) the primary and (b) the secondary furnaces, both operated at 1000 °C.

former calculated rate was rather consistent with the rate for diesel soot reported by Neeft et al.49 while the latter calculated rate was consistent with that obtained for rather nonporous carbons such pyrolytic graphite.46 The differences appear to be based on the existence and availability of the internal surface area under various combustion conditions. In both cases the soot oxidation efficiency in the afterburner was calculated to be 78%. Size Distribution of the Particulate Matter. The size distribution of the particulate matter, when both furnaces were operated at 1000 °C, showed that half of the amount yielded at the exit of the primary furnace (agglomerated soot) was submicron in aerodynamic size, which is capable of reaching the alveoli level in the human respiratory system. The afterburner reduced the submicron soot emissions by 11%, see Figure 9. The afterburner also affected the range of the distribution, as it became more closed, i.e., less small and big particles were present. A significant fraction, more than 54%, of the total soot exiting the afterburner was submicron. The afterburner was more effective in reducing the supermicron particles; more then 27% of their mass was destroyed therein. It should be mentioned again here that higher particulate conversions efficiencies were obtained when the primary furnace was operated at lower temperatures, see Figure 8, but size distributions were not monitored. The SEM analysis of particles collected in each stage show no readily detectable difference among the samples

234

Energy & Fuels, Vol. 17, No. 1, 2003

Caponero et al.

Figure 10. Microstructure of the particulate matter from combustion of volatile matter of tire chips collected in the impactor stages in the exhaust gases of the first furnace. (a-h) Impactor Stages 0-7; (i) Bottom Filter. Spherules: 210 ( 40 nm.

of the first and second furnaces (Figures 10 and 11). The spherules of the agglomerated soot show a similar

diameter size of 210 ( 40 nm. Comparing the TDF soot spherules with the particulate emissions from the

Batch Combustion of Waste Tire Chips

Energy & Fuels, Vol. 17, No. 1, 2003 235

Figure 11. Microstructure of the particulate matter from combustion of volatile matter of tire chips collected in the impactor stages in the exhaust gases of the second furnace. (a-h) Impactor Stages 0 to 7. (i) Bottom Filter. Spherules: 210 ( 40 nm.

combustion of five common plastics Shemwell and Levendis found variations in the spherule size. The soot

spherules from the combustion of polystyrene exhibited a large size variation, from 200 nm down to a fraction

236

Energy & Fuels, Vol. 17, No. 1, 2003

Caponero et al.

Figure 12. Emission yields for PAH’s due to the combustion of the volatile matter of tire chips at the exits of both the primary and the secondary furnace, for primary furnace temperatures in the range of 500-1000 °C. The afterburner temperature was kept constant at 1000 °C.

of 50 nm. Polyethylene and polypropylene soot was 100-200 nm and that of PVC and PMMA was down to 50-100 nm.50 The fact that the supermicron sized agglomerates appeared to be preferentially destroyed in the afterburner suggests division of the fused soot aggregates of soot by oxidation of the weaker links. Polycyclic Aromatic Hydrocarbons Emissions Behavior. The cumulative emission yields of all detected semivolatile polycyclic aromatic hydrocarbons (PAH) covering the mass range from 116 amu (indene) to 278 amu (benzo[b]chrysene and isomers) are shown in Figure 12. More than 50 PAH compounds were detected by GC-MS. For brevity, only the combined amounts of PAH, both in the condensed phase (cellulose filter paper) and in the gas phase (XAD-4) are plotted in Figure 12. Generally, only 2- or 3-ring PAHs were present in the gas phase, which means that they were not retained on the filter, condensed on particulate matter. Heavier multiring compounds were found in the condensed (solid) phase, i.e., with the collected particulate. After an initial increase, the overall trend of the cumulative PAH emissions was to decrease with the gas temperature of the primary furnace. It signifies either oxidation of PAH, or scavenging to soot, which exhibited an increasing trend with temperature, see Figure 12. Conesa et al.51 studied the evolution of volatile and semivolatile compounds from combustion of TDF, and showed that the formation of several PAH’s is related with the pyrolysis process that precedes combustion. In further work Fullana et al.52 showed that these compounds are partially eliminated in the combustion due to the presence of oxygen. Varying the bulk air ratio λ, that is, the inverse of the global equivalence ratio (λ ) 1/φ), they found that the emission of PAHs decreases with the increase of oxygen content. Their results at bulk air ratios similar to this study (0.82-1.19) and a temperature of 850 °C showed that the yield magnitudes (50) Shemwell, B. E.; Levendis, Y. A. J. Air Waste Manage. Assoc. 2000, 50, 94. (51) Conesa, J. A.; Fullana, A.; Font, R. Energy Fuels 2000, 14, 409. (52) Fullana, A.; et al. Environ. Sci. Technol. 2000, 34, 2092.

Figure 13. Profiles of cumulative PAH condensed on particulates at the various stages of the impactor, at the exits of (a) the primary and (b) the secondary furnaces, both operated at 1000 °C.

are comparable with the emissions at the exit of the primary furnace of the present work, especially for: styrene, benzaldehyde, fluorene, dibenzothiophene, phenanthrene, anhracene, pyrene, 11H-benzo[b]fluorene, and benzo[ghi]fluoranthene. In that work the effect of the temperature on the emissions during the oxidative pyrolysis was also studied. Their results showed a similar behavior to the emissions of the primary furnace herein, but with a maximum yield at 850 °C, instead of the ∼650 °C herein, see Figure 12. The lower temperature herein may be related to the higher oxygen partial pressure and the competition between formation and oxidation reactions. In contrast to those at the exit of the primary furnace, the PAH emissions from the secondary furnace were much lower, by factors of 2-4, especially at the lower operating temperatures of the former furnace. Thus the average destruction efficiency of PAH in the afterburner ηPAH was 67%, i.e., two-thirds of the PAH were oxidized or converted to soot. Thus, the effect of the afterburner was beneficial in minimizing the final emissions of PAH from this apparatus. The individual PAH components are presented in Tables 5 and 6. The major components were naphthalene, phenanthrene, acenaphthylene, fluoranthene, pyrene, indene, biphenyl, methylnaphthalene, fluorene, acephenanthrylene, benz[a]anthracene, cyclopenta[cd]pyrene, anthracene, benzo[a]pyrene, acenaphthene, and benzo[b]fluoranthene. Within the same sample at the exit of the primary furnace, it was observed that the

Batch Combustion of Waste Tire Chips

Energy & Fuels, Vol. 17, No. 1, 2003 237

Table 5. Composition of the First Furnace Exhaust Gases compound (µg/g fuel burned)

500 °C

600 °C

styrene benzaldehyde phenol indene naphthalene benzothiophene 2-methylnaphthalene 1-methylnaphthalene biphenylene biphenyl acenaphthene acenaphthylene dibenzofuran fluorene 2-methylfluorene 1-methylfluorene dibenzothiophene phenanthrene anthracene 3-methylphenanthrene 2-methylphenanthrene 2-methylanthracene 4H-cyclopenta[def]phenanthrene 4-methylphenanthrene 1-methylphenanthrene fluoranthene acephenanthrylene pyrene benzo[a]fluorene 11H-benzo[b]fluorene 1-methylpyrene benzo[ghi]fluoranthene benzo[c]phenanthrene cyclopenta[cd]pyrene benz[a]anthracene triphenylene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]fluoranthene benzo[e]pyrene benzo[a]pyrene perylene dibenz[a,j]anthracene indeno[7,1,2,3-cdef]chrysene indeno[1,2,3-cd]pyrene benzo[ghi]perylene anthranthrene dibenz[a,h]anthracene benzo[b]chrysene picene

314.4 38.0 42.4 196.4 2800.7 14.4 187.7 124.7 20.0 147.7 56.8 407.4 20.5 121.4 21.0 17.9 11.6 450.7 36.2 11.4 19.2 20.8 17.0 13.2 8.1 144.7 81.1 138.1 11.2 13.0 14.6 27.4 6.0 60.5 47.7 18.9 21.9 27.5 7.0 20.8 18.4 39.5 8.0 0.0 0.0 12.2 15.0 4.3 1.2 0.5 0.0

394.2 21.6 30.9 276.7 2917.6 28.3 259.1 186.2 32.3 205.7 86.8 611.6 31.3 226.4 33.9 20.6 40.5 646.2 56.8 18.6 36.0 39.5 61.4 25.4 30.0 268.4 162.1 271.9 53.7 23.5 14.6 54.5 12.7 113.5 119.9 13.4 30.9 56.3 31.8 44.6 39.0 81.6 13.2 2.6 5.8 13.3 40.5 31.1 1.9 4.4 2.4

total (mg/g fuel burned)

5.86

7.83

yields of naphthalene were the highest, much higher than those of phenanthrene, which in turn was comparable with acenaphthylene. These were followed by fluoranthene and pyrene, and indene, which were comparable. The yields of biphenyl were still lower, by factors of 2-4 compared to those of phenanthrene, but comparable to the yields of methylnaphthalene, fluorene, acephenanthrylene, benz[a]anthracene, cyclopenta[cd]pyrene, anthracene, benzo[a]pyrene, and acenaphthene. Yields of benzo[b]fluoranthene were even less. Comparing Tables 5 and 6, the reduction in the PAH emissions due to the afterburner treatment is evident. Almost all individual PAH experienced the general trend shown in Figure 12, for the cumulative PAH emission. However, while naphthalene was reduced by less than a factor of 2 and accounted for the majority of emissions from the afterburner, the reduction of some individual components in the afterburner was indeed striking. For instance, health-hazardous compounds,

temperature of the first furnace 700 °C 800 °C 269.2 21.3 28.8 362.5 2668.1 36.7 235.5 159.2 26.7 188.9 81.4 624.6 27.5 200.7 24.7 17.5 32.9 511.2 64.7 14.4 31.8 35.5 45.8 22.1 21.9 264.0 84.0 275.1 36.5 24.0 12.5 51.4 11.6 102.4 103.6 9.2 41.3 49.4 23.2 41.2 37.7 79.9 14.0 2.7 7.0 16.0 35.0 13.6 3.7 3.9 3.5 7.10

199.9 16.9 15.7 99.3 2292.6 0.0 157.6 100.3 15.8 192.1 63.4 537.7 38.0 213.2 34.4 18.3 28.9 642.8 148.0 15.5 34.0 39.8 40.0 22.7 21.6 279.6 156.8 287.2 38.0 46.1 14.7 54.3 11.2 99.9 116.2 9.1 61.5 54.5 25.4 42.6 37.7 84.8 13.1 3.6 7.2 30.2 32.8 17.4 7.5 6.6 5.0 6.53

900 °C

1000 °C

95.1 15.9 36.0 8.8 2007.8 38.5 55.5 32.3 8.1 124.9 21.0 282.5 37.2 74.3 8.7 4.7 34.6 374.3 76.2 5.7 7.8 9.0 21.4 5.2 5.7 206.1 82.1 198.6 14.4 14.2 3.8 43.2 4.9 50.8 80.5 10.9 53.1 42.4 26.1 24.5 30.1 55.4 7.5 4.5 9.8 29.5 23.6 7.3 6.9 5.9 4.0

18.4 15.2 53.1 25.9 1722.0 5.1 10.5 8.7 6.0 78.2 6.0 146.1 37.7 37.2 6.1 4.1 26.0 245.7 16.6 3.1 1.7 2.1 9.9 1.5 1.7 137.5 36.4 94.9 3.7 2.0 0.4 22.5 2.4 16.2 25.1 9.0 1.4 29.2 9.3 8.2 16.1 20.6 5.4 0.7 1.3 3.5 11.9 6.0 0.8 0.4 0.1

4.40

2.95

such as cyclopenta[cd]pyrene and benzo[a]pyrene, disappeared from the exhaust of the afterburner under most conditions. Size Distribution of the Particulate Matter: PAH Analysis. The analysis of the sum of the PAH (total PAH), condensed on the particulate matter collected at each stage of the impactor, showed a similar behavior to emission of particulate mass. Condensed PAH accounted for 1-2% of the soot mass at every size cuts. At the exit of the primary furnace a bimodal size distribution was observed again, but at the exit of the afterburner the amount of condensed PAH on particles monotonically decreased with increasing particle size. The fact that smaller particulates were associated with larger amounts of condensed PAH may be explained on the grounds that that such particles have higher surface area; Durlak et al.53 have reported a similar finding. For the adsorbed PAH the efficiency of destruction appears to be high in the afterburner, i.e., an average

238

Energy & Fuels, Vol. 17, No. 1, 2003

Caponero et al.

Table 6. Composition of the Second Furnace Exhaust Gases compound (µg/g fuel burned)

500 °C

600 °C

styrene benzaldehyde phenol indene naphthalene benzothiophene 2-methylnaphthalene 1-methylnaphthalene biphenylene biphenyl acenaphthene acenaphthylene dibenzofuran fluorene 2-methylfluorene 1-methylfluorene dibenzothiophene phenanthrene anthracene 3-methylphenanthrene 2-methylphenanthrene 2-methylanthracene 4H-cyclopenta[def]phenanthrene 4-methylphenanthrene 1-methylphenanthrene fluoranthene acephenanthrylene pyrene benzo[a]fluorene 11H-benzo[b]fluorene 1-methylpyrene benzo[ghi]fluoranthene benzo[c]phenanthrene cyclopenta[cd]pyrene benz[a]anthracene triphenylene chrysene benzo[b]fluoranthene benzo[k]fluoranthene benzo[a]fluoranthene benzo[e]pyrene benzo[a]pyrene perylene dibenz[a,j]anthracene indeno[7,1,2,3-cdef]chrysene indeno[1,2,3-cd]pyrene benzo[ghi]perylene anthranthrene dibenz[a,h]anthracene benzo[b]chrysene picene

36.7 9.3 2.1 4.9 1584.0 0.6 8.2 7.0 6.0 66.2 11.8 98.8 8.5 69.1 28.5 16.9 17.9 226.9 39.3 5.5 8.4 9.4 15.5 6.4 7.2 72.2 39.7 68.1 13.3 12.8 4.6 13.4 3.5 24.3 29.2 4.0 20.5 13.8 8.9 10.6 11.1 18.8 7.2 1.4 3.2 9.1 7.8 3.0 1.1 1.1 1.6

18.8 4.7 3.0 5.1 1457.0 1.2 3.3 2.7 0.7 49.9 2.3 14.9 2.3 10.4 15.9 9.4 1.1 148.3 3.6 4.4 1.5 1.7 2.7 0.7 1.3 19.1 4.0 13.6 0.9 0.2 0.1 0.5 0.0 0.0 1.2 1.3 0.5 0.6 0.0 0.0 0.2 0.2 4.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

total (mg/g fuel burned)

2.70

1.81

efficiency over 90%, see Figure 13. Particulates exiting the afterburner appear to be mostly carbon with only small amounts of condensed organics. Carbon Balance. For the assessment of the suitability of this experimental procedure, the check of the carbon balance represents an important tool. The carbon content of the reactants was calculated using the chemical analyzes of the tire chips and carbon residue. The volatile pyrolysates are the fuel during these experiments and were composed of 87.4% carbon, see Table 3. The analyses of the exhaust of both furnaces gave the amount of carbon present in the products of the combustion reaction. The sum of the carbon content of the CO, CO2, PAH, and particulate matter led to a closure of the average carbon balance by more than 81% with (53) Durlak, S. K.; Biswas P.; Shi, J. D.; Bernhard, M. J. Environ. Sci. Technol. 1998, 32, 2301.

temperature of the first furnace 700 °C 800 °C 4.6 6.9 1.2 4.9 1403.1 1.0 2.7 1.9 2.3 31.8 0.7 15.7 3.4 7.9 2.4 1.1 0.2 51.4 2.7 0.9 0.3 0.3 1.1 0.3 0.2 14.8 2.4 8.2 0.6 0.3 0.1 1.3 0.0 0.5 2.2 0.8 1.9 1.4 0.2 0.2 2.1 0.3 3.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.59

4.0 5.5 2.0 1.9 1296.1 0.6 7.0 4.8 2.7 43.6 0.5 16.4 3.3 7.5 10.6 0.9 0.2 66.3 3.0 0.7 1.1 1.6 0.8 0.5 0.2 15.6 2.8 7.8 0.0 0.2 0.1 1.1 0.1 0.8 2.1 0.4 1.8 1.1 0.0 0.1 0.3 0.1 3.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

900 °C

1000 °C

2.9 1.1 0.0 2.0 1390.0 1.2 4.4 3.4 0.7 52.1 1.9 23.9 2.8 15.9 1.8 0.0 1.3 57.0 6.0 0.7 0.4 0.3 1.5 0.3 0.3 7.5 1.4 4.3 0.0 0.0 0.0 0.3 0.0 0.0 1.0 0.8 0.7 0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

6.6 5.4 1.3 4.1 1290.0 0.4 3.5 3.1 2.0 37.8 1.1 28.7 4.9 10.9 2.2 1.0 0.5 141.0 4.8 0.6 1.0 1.0 1.4 0.6 0.8 45.5 10.2 22.5 0.5 0.5 0.1 1.2 0.2 1.0 1.8 0.7 0.0 0.3 0.0 0.0 0.0 0.0 2.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

1.52

1.59

1.64

an increasing trend with the increasing primary furnace temperature, see Figure 14. Light hydrocarbons, which were not taken into account, and possible loses during sample collection and preparation are likely sources of error. Conclusions Batch combustion of tire derived fuel, in the form of dewired tire chips (order of 1 cm), was performed in a two-stage electrically heated laboratory furnace. The two stages were separated by an additional-air mixing section. The gas temperature of the primary furnace varied in the range of 500-1000 °C, while the temperature in the afterburner was kept at 1000 °C. Small batches of TDF were loaded in the primary furnace where they ignited and experienced transient combustion, with devolatilization products burning in diffusion flames. The overall equivalence ratio was mildly fuel-

Batch Combustion of Waste Tire Chips

Figure 14. Relation between the total carbon content in the products and total carbon content in the reactants, for different primary furnace temperatures (500-1000 °C). The afterburner temperature was kept constant at 1000 °C.

lean. The soot-laden effluent was mixed with preheated air and it was channeled to the secondary furnace (afterburner) for additional burnout. Emissions were monitored at the exits of both furnaces. The results illustrated that both the operating temperature of the primary furnace and the existence of the afterburner itself had pronounced effects on the emissions. Increasing the primary furnace temperature appeared to increase the CO, NOx, particulate and, to a lesser extent CO2, emissions. Emissions of PAH first increased and then decreased with increasing primary furnace temperature. The existence of the afterburner, operated at 1000 °C and 0.7 s gas residence time, drastically

Energy & Fuels, Vol. 17, No. 1, 2003 239

reduced the CO, particulate and PAH emissions, while CO2 emissions increased, indicative of additional oxidation therein. Particularly, many multiring and reported health-hazardous PAH components, such as benzo[a]pyrene, were effectively eliminated by the afterburner treatment of the combustion effluent. The afterburner significantly reduced the mass of both submicron and supermicron particulates, and it drastically reduced the amounts of condensed PAH on those particulates. It was concluded that, under the operating conditions of this study, the PAH components were more effectively destroyed in the afterburner (oxidized or converted to soot) than the particulates (mostly soot). Hence, conditions that minimize the particulate yield from the primary furnace may be preferred, as they also minimize the overall emissions from the afterburner. In view of this argument an operating gas temperature in the primary furnace between 600 and 700 °C, followed by mixing with additional air and an afterburner treatment at 1000 °C may be recommended. This also minimizes CO emissions. Acknowledgment. The authors thank Mr. Eric Wisnaskas for technical assistance with the chromatographic analysis. The authors also acknowledge Sa˜o Paulo research foundation, FAPESP, for financing Dr. Jefferson Caponero (No. 99/000375-5). Partial support was also provided by US-NSF with grant CTS-9908962; Dr. Farley Fisher, Program Director. EF0201331