102

Energy & Fuels 2004, 18, 102-115

Emissions of Batch Combustion of Waste Tire Chips: The Hot Flue-Gas Filtering 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 RD&E Center, Natick, Massachusetts 01760 Received February 25, 2003. Revised Manuscript Received October 7, 2003

A laboratory investigation was performed on the emissions from the batch combustion of waste tire chips in fixed beds. Techniques and conditions that minimize toxic emissions were identified. Tire-derived fuel (TDF), in the form of waste tire chips (1 cm in size), was burned in a two-stage combustor. Batches of tire chips were introduced to the primary furnace, where gasification and combustion occurred. The gaseous effluent of this furnace was mixed with streams of additional preheated air in a mixing venturi, and it was then passed through a silicon carbide (SiC) honeycomb wall-flow filter that had been placed inside this furnace. Subsequently, it was channeled into a secondary furnace (afterburner), where further oxidation occurred. The arrangement of the two furnaces in series allowed for independent temperature control; varying the temperature in the primary furnace influenced the type and the flux of pyrolysates. The hot-flue-gas filtering section, ahead of the exit of the primary furnace, allowed the retention and further oxidation of most of the generated particulates and, thus, prevented them from entering the afterburner. Results showed that the combination of the high-temperature ceramic filter with the afterburner treatment was successful in reducing the emissions from the combustion of waste tires. Depending on the temperature of the primary furnace, the final emissions of CO were reduced by factors of 2-6, NOx emissions were reduced by factors of 2-3, particulate emissions were reduced by 2 orders of magnitude (both PM2.5 and PM10), and most individual polycyclic aromatic hydrocarbon (PAH) species emissions were reduced by more than 1 order of magnitude, with the exception of naphthalene, whose reduction was less drastic. The overall combustion effectiveness was enhanced, as evidenced by higher CO2 yields.

Introduction The combustion process of solid fuels, such as waste tires, is often incomplete, and undesirable products of incomplete combustion (PICs) are formed. This can be attributed to the combined effect of local temperatures, inadequate mixing of the fuel purolyzates and air, and local oxygen-starved conditions around the fuel.1 The combustion of tires (single or in stockpiles) emits copious amounts of acrid soot; most people have vivid images of intense black smoke emanating from burning whole tires (single or in stockpiles). Hence, past work at this Northeastern University laboratory was focused on investigating whether the batch combustion of chunks of tires (simulating whole tires in a small scale) does, indeed, emit more pollutants than the continuous combustion of tire crumb injected in a furnace. The results showed that the batch combustion of tire chunks resulted in much higher emissions of PICs, such as * Author to whom correspondence should be addressed. E-mail: [email protected]. (1) Neeft, J. P. A.; Makkee, M.; Moulijn, J. A. Fuel Process. Technol. 1996, 47, 1.

polycyclic aromatic hydrocarbons (PAHs) and particulate matter. However, whereas the combustion mode influenced the amounts, it did not influence the nature of the species produced. Similar types of emissions were detected from the combustion of both tire chunks and tire crumbs at comparable temperatures and gas residence times in the furnaces.2-5 This is the second part of an ongoing study that addresses the emission of pollutants from the combustion of waste tire chips and identifies techniques for their minimization. The first part of the study6 investigated the effect of two combustion stages, separated by a section where flue gas was mixed with additional air. Experiments were performed in a laboratory apparatus that was composed of two laminar-flow hori(2) Levendis, Y. A.; Atal, A.; Carlson, J. B.; Dunayevskiy, Y.; Vouros, P. Environ. Sci. Technol. 1996, 30, 2742. (3) Levendis, Y. A.; Atal, A.; Courtemnche, B.; Carlson, J. B. Combust. Sci. Technol. 1998, 131, 147. (4) Atal, A.; Levendis, Y. A. Fuel 1995, 74, 1570. (5) Levendis, Y. A.; Atal, A.; Carlson, J. B. Combust. Sci. Technol. 1998, 134, 407. (6) Caponero, J.; Tenorio, J. A. S.; Levendis, Y. A.; Carlson, J. B. Energy Fuels 2003, 17, 225-239.

10.1021/ef030043b CCC: $27.50 © 2004 American Chemical Society Published on Web 11/22/2003

Batch Combustion Emissions of Waste Tire Chips

Energy & Fuels, Vol. 18, No. 1, 2004 103

Figure 1. Schematic of the bench-scale experimental apparatus, consisting of two furnaces in series, a mixing venturi, a ceramic filter, and emission sampling stages.

zontal muffle furnaces, simulating two independent combustion stages (see Figure 1). The primary furnace acted as a gasifier/combustor for the tire chips that were placed therein. The combustion effluent (gases laden with particulates) contained PICs. This effluent was blended with additional preheated air in a venturi mixing section, and it was then channeled into a secondary furnace, which acted as an afterburner. Because most pollutants from the combustion of tires emanate during the stage of the homogeneous combustion of devolatilizates, rather than from the heterogeneous char combustion phase,7 this entire investigation is focused on the combustion of the devolatilizates (volatiles are 70 wt % of this tire). The performance of the afterburner was influenced by the operating temperature of the primary furnace, which was varied in the range of 500-1000 °C. It was shown that, in the case of this tire-derived fuel (TDF), it was beneficial to use combustion staging, in conjunction with an additional-air mixing section.6 Emissions of CO were reduced by factors of 3-10, particulate emissions were reduced by factors of 2-5, and cumulative PAH emissions were reduced by factors of 2-3. Operating the primary furnace/gasifier at temperatures in the low side of the 500-1000 °C range resulted in low pollutant yields, upon treatment in the afterburner. That study has been continued in this work, in search of additional techniques that further reduce PICs. Herein, the aformentioned work was complemented by the installation of a high-temperature barrier filter ahead of the exit of the primary furnace (see Figure 1). This ceramic honeycomb filter was used to remove the particulates from the flue gas. Such particulates, being primarily organic in nature (soot, tars, oils, waxes), were retained inside the filter, where they were eventually gasified at the elevated temperatures therein. Under such circumstances, the afterburner was able to treat a practically particle-free effluent. The following two subsections discuss (i) the perils of exposure to soot, to illustrate why filtration is important, and (ii) how soot removal is accomplished. (7) Atal, A.; Steciak, J.; Levendis, Y. A. In Procedings of the ASME Heat Transfer Division; American Society of Mechanical Engineers (ASME) International: New York, 1995; Vol. HTD317-2.

Health-Related Effects of Particulate Matter. The particulate matter emitted from the combustion of waste tires is basically soot and tars, which are byproducts of the incomplete combustion of organic fuels. The soot is composed of carbon and adsorbed organic compounds, such as PAHs; however, it may contain traces of inorganic elements, such as metals, oxides, salts, adsorbed liquids and gases, and nitrogen and sulfur composites.8 Living beings are exposed to soot as they breathe polluted air, as they consume contaminated food, and through contact with the skin. Cigarette smokers and workers of industries that either use soot or generate it as a byproduct, as well as workers of transport companies (such as drivers of diesel-powered vehicles), suffer greater risks from being exposed to high concentrations of soot.9,10 As a component of polluted air, soot is also abundant in cities with large amounts of atmospheric pollution.11 Soot, as a result of its physical nature and chemical composition, has been associated with the increased risk of lung, bladder, and skin cancers.12 Several regulatory and scientific agencies have recognized soot as a carcinogenic substance. The USEPA Carcinogen Assessment Group included soot in their list of potentials carcinogens, the National Toxicology Program classified soot to be “recognized as a human carcinogen”, and the International Agency for Research on Cancer (IARC) also has classified soot to be carcinogenic (Group 1).13 The first case of skin cancer related to soot was identified more than 200 years ago, when an increased risk of cancer was detected in chimney sweepers in England. Since that first report, several epidemiological (8) Health and Environmental Effects of Particulate Matter Fact Sheet. Environmental Protection Agency (US-EPA): Washington, DC, 1999. (http://www.epa.gov/ttn/oarpg/naaqsfin/pmhealth.htm) (9) Boffetta, P.; Jourenkova, N.; Gustavsson, P. Cancer, Causes Control 1997, 8, 444. (10) Stober, W.; Abel, U. Int. Arch. Occup. Environ. Health 1996, 68, (Supplement), S3-S61. (11) Soots, Tars, and Mineral Oils, Eighth Report on Carcinogens. National Toxicology Program: 1998. (http://ntp-server.niehs.nih.gov/ htdocs/8_RoC/KC/SootsTars&Min.html) (12) Mastrangelo, G.; Fadda, E.; Marzia, V. Environ. Health Perspect. 1996, 104, 1166. (13) Soots, Supplement 7; The International Agency for Research on Cancer (IARC): Lyon, France, 1987. (Available online at: http:// 193.51.164.11/htdocs/Monographs/Suppl7/Soots.html.)

104

Energy & Fuels, Vol. 18, No. 1, 2004

studies have demonstrated the correlation between soot and cancer risk. More recent work on chimney sweepers in Sweden and Denmark showed a significant increase in lung cancer; similar studies in Germany and the United Kingdom reached the same conclusions. One of these studies also reported on the increase of the risk of esophagus cancer, primary liver cancer, and leukemia among the same workers.14 In two independent studies, coke soot was applied in the skin of mice, and tumors were produced in both studies. In another animal study, the implantation of wood-generated soot under the skin produced located tumors in female mice. Extracts of oil from soot produced tumors in mice upon hypodermic applications or upon inhalation of great amounts of these extracts.14 Soot was shown to be mutagenic in several laboratories. Soot extracts were mutagenic to bacteria and cultures of human linfoblasts. Extracts of particulate matter emitted from wood combustion were shown to cause damage to the DNA of ovarian cells of female hamsters.13 Hot Flue-Gas Filtration. High-temperature filtration of the effluents from waste incineration can remove particulate matter such as soot. Suitable barrier filters, such as ceramic honeycomb or foam filters, may allow continuous operation if the retained particles are carbonaceous and can be oxidized therein. High residence times and catalytic effects significantly enhance the soot/tar oxidation efficiency in the filter. Extruded ceramic monolith substrates currently are widely used for automotive and stationary emission control reactors, such as three-way catalysts, selective catalytic reactors (SCRs) for the reduction of nitrogen oxides (NOx), and diesel particulate traps. Monoliths are increasingly under development and evaluation for many new reactor applications, e.g., in chemical and refining processes, catalytic combustion, ozone abatement, etc.15 Cell configurations and properties of monoliths are described in terms of geometric and hydraulic parameters.16,17 These properties can be defined in terms of cell spacing, wall thickness, and cell density, which is the number of cells per unit of cross-sectional area. In designing monolithic catalysts, a balance of geometric parameters such as cell density or wall thickness is necessary to meet the constraints of external processing requirements, such as space velocity, flow rates, and pressure drop. The use of diesel particulate filters (DPFs) constitutes the most effective method for removing soot particles from diesel engines. Filtration occurs when the particleladen exhaust gas is forced through the porous walls that separate adjacent channels of the filter honeycombs. Soot particles are trapped in the entrance channels and filtered exhaust passes outward from the exit channels (see Figure 2). In diesel engine applications, the filter is installed in the tailpipe of a vehicle, which is typically operated below the auto-ignition temperature of soot (ca. 600 °C). Thus, soot accumulates (14) Summary of Data Reported and Evaluation. Cap. 5-Soots; The International Agency for Research on Cancer (IARC): Lyon, France, 1985; Vol. 35, p 219. (Available online at: http://193.51.164.11/htdocs/ Monographs/Vol35/Soots.html.) (15) Williams, J. L. Catal. Today 2001, 69, 3. (16) Cash, T. F.; Williams, J. L.; Zink, U. H. Society of Automotive Engineers (SAE) Brazil, Paper No. 982927, 1998. (17) Day, J. P.; Socha, L. S. Society of Automotive Engineers (SAE), Paper No. 910371, 1991.

Caponero et al.

Figure 2. Schematic showing one channel of a silicon carbide (SiC) honeycomb wall-flow filter.

therein and regeneration of the filter is required for continuous engine operation at acceptable exhaust backpressures. Regeneration techniques include periodic thermal destruction or aerodynamic removal by external means. Cordierite, mullite, and silicon carbide (SiC) are materials that are commercially available for diesel exhaust filters. Development of new materials for DPF-type filters is underway.15 Over the last 15 years, extensive work has been conducted in this Northeastern University laboratory in the field of diesel engine exhaust after-treatment, using ceramic filters. The filtration characteristics for such ceramic monoliths were determined to be excellent, and the pressure drop is acceptably low for unimpeded engine operation. SiC honeycomb wall-flow filters can handle elevated temperature gases (if needed, as hot as 1500 °C) and corrosive gases. Particulate retention efficiencies have been recorded to be in the 97%-99% range, as shown in Figure 6 of the work by Larsen et al.18 Such filters are used in the study herein. Materials and Experimental Methods Tire chips with dimensions on the order of 1 cm were obtained from a local source and included organic fabrics, such as nylon belts (see Figure 3). No metallic belts were included. Some physical and chemical properties are shown in Table 1. The elemental analysis resulted in the mass fractions that are shown in Table 2. All of the tests conducted in this study involved the batch combustion of tire chips in fixed beds. Preweighted sample amounts of 0.8000 ( 0.0005 g, consisting of a few chips (three chips with a volume of ∼ 0.8 cm3 each, to keep the total surface area approximately constant in all samples), were placed in porcelain boats and inserted in the quartz tube of the primary furnace, which was 4 cm in diameter and 87 cm long. The tube was placed at the centerline of a horizontal, split-cell, electric furnace (1 kW maximum power) (see Figure 1). To insert the samples quickly into the furnace, the porcelain boats were placed at the end of the inner surface of a tube with a U-shaped cross section, i.e., a quartz cylinder longitudinally split along the centerline. The other end of this U-shaped tube was (18) Larsen, C.; Levendis, Y. A.; Shimato, K. Filtration Assessment and Thermal Effects on Aerodynamic Regeneration in Silicon Carbide and Cordierite Particulate Filters. Society of Automotive Engineers (SAE), Paper 1999-01-0466, 1999. (Also see Soc. Auto. Eng., [Spec. Publ.] SP 1999, SP-1414.)

Batch Combustion Emissions of Waste Tire Chips

Figure 3. Photograph of the waste de-wired tire chips that were burned in these experiments. Table 1. Some Physical and Chemical Properties of the Tire Chips Used property

value

fixed carbon content volatiles content ash content apparent density heating value residue specific area

24.9% 69.7% 4.0% 0.4 g/cm3 39 MJ/kg 74 m2/g

Table 2. Composition of Tire Chips Fractions Content (wt %) element

tire chips

burned volatile 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.

connected at the entrance glass fitting of the furnace quartz tube. At the start of each experiment, the fitting was opened and the U-shaped tube, with the sample mounted at its tip, was quickly inserted in the furnace, so that it was parallel and concentric to the 87-cm-long quartz furnace tube. All experiments were conducted in filtered dry air. The airflow rate in the first furnace was 4 L min-1 (at standard temperature and pressure, STP), 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 a diameter of 2 cm and a length of 38 cm. 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 and the gaseous effluent of the combustion occurred in the venturi.19 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.20 The tire sample temperature was not (19) Wang, J.; Levendis, Y. A.; Richter, H.; Howard, J. B.; Carlson, J. B. Environ. Sci. Technol. 2001, 35, 3541. (20) Wheatley, L.; Levendis, Y. A.; Vouros, P. Environ. Sci. Technol. 1993, 27, 2885.

Energy & Fuels, Vol. 18, No. 1, 2004 105 measured; such a temperature 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 neighborhood of the gas temperatures in the corresponding experiments. This is inferred based on previous studies in this laboratory where both single particles and groups of tire crumb particles were fluidized and burned and their combustion behavior was monitored cinematographically and pyrometrically.3,4 The flow rate of additional air supplied at the venturi, through the four jets, was 2 L min-1. 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, depending on the temperature therein, and the Reynolds number was in the range of 75-85. In the experiments conducted herein, the particulates generated in the primary furnace were prevented from entering the afterburner by a high-temperature ceramic filter. The filter used in this study was a SiC honeycomb wall-flow filter manufactured by Ibiden (see the flow illustration in Figure 2). The temperature at the entrance of this filter was the same as that of the gas in the primary furnace, while the temperature at its exit dropped by 150-200 °C. This filter was selfregenerating as the carbonaceous particulated burned therein. It should be mentioned that ash generated during the combustion of the tire chips remained in the sample bed, inside the porcelain boat. If, in a practical application, ash were to accumulate in the filter, then the filter would need to be subjected to aerodynamic regeneration, in the manner described in ref 18. Sampling was conducted at the exits of both furnaces simultaneously. The effluent of the primary furnace was divided in two halves. One half was sampled at the exit of the primary furnace, and the other half was channeled to the secondary furnace; it was then sampled at its exit, as shown in Figure 1. Both sampling stages and their operating conditions were identical. All combustion tests were repeated in triplicate, and if one outcome was found to be in disagreement, it was eliminated. Combustion occurred in diffusion flames (nonpremixed), and this naturally can partially explain the sample-to-sample variabilities. Typical sample-to-sample variabilities were presented and discussed previously for the combustion of solid polystyrene (see refs 19 and 21). Such variabilities did not affect the trends, the overall magnitudes, or the stated conclusions of this study. Typical durations of the luminous diffusion flame over the boat were on the order of 1-2 min. Flames seemed to be remarkably steady throughout each experiment. Emissions of major gaseous species, such as O2, NOx, SO2, CO, and CO2, from the combustion of tire chips were monitored at the exits of the two furnaces. Upon traversing the sampling stage, moisture was removed by passing the effluent through an ice bath and a Permapure dryer that was mildly heated to 50-60 °C. The effluent was then monitored for NOx (using a Beckman model 951A chemiluminescent NO/NOx analyzer), SO2 (by means of a Rosemount Analytical model 590 UV analyzer), O2 (using a Beckman paramagnetic analyzer), and CO and CO2 (with Horiba infrared analyzers). The output of the analyzers was recorded using a Data Translation model DT-322 data acquisition board in a microcomputer. The system used DT VPI to support data acquisition from within a Hewlett-Packard model 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 each experiment. Subsequently, they were converted to partial pressures and, eventually, to mass yields (milligrams per mass of sample burned or pyrolyzed). (21) Wang, J.; Levendis, Y. A.; Richter, H.; Howard, J. B.; Carlson. J. B. Environ. Sci. Technol. 2002, 36, 797.

106

Energy & Fuels, Vol. 18, No. 1, 2004

Polycyclic aromatic hydrocarbons (PAHs), as well as particulates (soot, tars, oils, and, possibly, traces of minerals), that were emitted from the combustion of tire chips were monitored at the exits of the two furnaces. Soot and PAHs were sampled by passing half of the combustion effluent through each of the sampling stages at the exits of the two furnaces. Each stage consisted of a Graseby sampling head, encompassing a Fisherbrand paper filter and a glass cartridge containing Supelco XAD-4 resin. The sampling stages were placed adjacent to the exits of the two furnaces to minimize losses. Prior to the sampling stages, the effluent of each furnace was mixed with a 2 L min-1 flow (STP) of room-temperature diluent inert gas (nitrogen) to quench chemical reactions. Thus, the effluent was cooled to ∼100 °C at the entrance of the sampling head. Subsequently, the particulate emissions were trapped on the 11-cm paper filter, which has a nominal particle retention of 1-5 µm. Gas-phase PAH emissions were adsorbed on the bed of the XAD-4 resin. The length of the XAD-4 bed was more than twice its diameter, for effective capture of these species. Following the experiments, the filters and resins were removed and placed in separate glass bottles with Teflon-lined caps. They were then stored at 4 °C. Prior to extraction with methylene chloride, a 25 µL internal standard containing 50 µg each of deuterated naphthalene-d8, acenaphthene-d10, anthracene-d10, chrysene-d12, and perylene-d12 was applied to each of the glass bottles containing the samples. Extraction of the organic compounds from both the XAD-4 resins and the cellulose filter papers was accomplished with a Dionex model ASE 200 accelerated solvent extractor at the U.S. Natick Army Labs. Analysis was conducted by gas chromatography coupled with mass spectrometry (GC-MS). The analytical studies were conducted in accordance with EPA Method 8270A, as specified in the analytical methodology SW846. The method was simplified to remove surrogate analysis procedures in the absence of potential matrix effects normally observed in environmentally obtained soil and water samples. Analytical data was reviewed based on the EPA's Volatile/Semivolatile Data Validation Functional Guidelines. Any analytical data that failed to meet those standards were rejected and the combustion experiments were repeated. The GC-MS system consisted of a Hewlett-Packard model HP 6890 gas chromatograph that was equipped with a Hewlett-Packard model HP 5973 massselective detector. The GC column was a Hewlett-Packard model Ultra-2, which was composed of 5% phenyl-methylsiloxane 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 was 0.7 mL/min, and the linear flow rate was 32 cm/s (methane) at 100°C. The split flow was 2.0 mL/min, and the split ratio was 2.7:1. The instrument was tuned in accordance with EPA semivolatile criteria prior to the GC-MS analysis of each set of samples. The instrument passed initial and continuing calibration criteria. Each of the target compounds, as well as the tentatively identified compounds, were quantified using the appropriate deuterated internal standard. In a departure from the EPA method, the GC-MS system was operated in the full scan mode and not in a single-ion monitoring 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 ∼1 µg of component per gram of fuel burned. Values that are <1 µg/g are technically nondetectable. Analysis of the combustion system blanks indicated the presence of xylenes, bis(2-ethylhexyl) phthalate, and siloxanes in quantities sufficient to reject all positive results for these compounds in the combustion extracts. The standard solutions containing 100 µg each of five characteristic deuterated standards were diluted to 10 mL and analyzed both as an instrument blank and to provide an indication of extraction efficiency for each of the internal standards. The National Functional Guidelines

Caponero et al.

Figure 4. Emission profiles from a single experiment on the batch combustion of tire chips at the exits of the primary furnace (stage 1) and the secondary furnace (stage 2), both operated at 1000 °C. recommend a recovery of >50% for the internal standards; therefore, samples with an extraction efficiency of <50% for any of the internal standards must be repeated. In this work, the extraction recovery was in the range of 85%-100%. The size of the tire-combustion-generated particulates was measured using an Andersen impactor (1 ACFM (actual cubic feet per minute) nonviable ambient particle sizing sampler), which simulates the human respiratory system. This is a multistage impactor, which is composed of a series of perforated disks (stages). In each stage, the size of the holes and their distribution results in seven particle-size cuts (starting at 9, 5.8, 4.7, 3.3, 2.1, 1.1, 0.7, and 0.4 µm), and a fibrous filter that captures particles smaller than 0.4 µm is present. This instrument was used instead of the aforementioned sampling head in separate experiments. It was retrofitted at the exit of either the primary or secondary furnace, in separate experiments.

Results and Discussion In every experiment, a batch of waste tire chips was introduced to the preheated primary furnace. The tire chips were heated and ignited. Two distinct combustion phases were observed: (i) burning of the volatile matter (devolatilizates), and (ii) burning of the char, which mostly consists of carbon black. Concurrently, burning of the soot that was trapped inside the ceramic filter channels also occurred. The combustion phases of tires were studied in detail in previous work.4,7 In this work, the effects of three parameters on the emissions from the volatile combustion phase of tire chips were monitored: (i) the primary furnace temperature, (ii) the hot flue-gas filtering effect, and (iii) the effect of the postfilter afterburner treatment of the effluent. The primary furnace acted as a gasifier/combustor, and its temperature was varied in the range of 500-1000 °C. The temperature of the afterburner 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 this tire-derived fuel (TDF) resulted in transient (time-dependent) emission profiles. Figure 4 shows typical emission profiles for the O2, CO2, and CO gases observed in these experiments. Based on the sample mass of the TDF volatiles that were burned, the overall duration of their combustion (on the order of 1 min), and the air flow rate in the

Batch Combustion Emissions of Waste Tire Chips

Figure 5. Partial pressures (given as a percentage of atmospheric pressure) and emission yields for O2 at the exits of both the primary and secondary furnaces. Primary furnace temperatures ranged from 500 to 1000 °C; the afterburner temperature was kept constant at 1000 °C. Legend is as follows: (9), primary furnace, without ceramic filter; (b), secondary furnace, without ceramic filter; (0), primary furnace, with ceramic filter; and (O), secondary furnace, with ceramic filter.

primary furnace, the calculated global (overall) equivalence ratios in the primary furnace were mildly fuelrich (φ ) 1-1.2). Upon ignition of the fuel, the amounts of the products of combustion increased, reached a maximum, and, thereafter, decreased as the flame dwindled to extinction. Emission yields were obtained by integrating each recorded profile and then by dividing that value by the mass of the fuel burned, which, in this case, is only the volatile mass of the tire chips (69.7 wt %; see Table 1). Oxygen Consumption. The profiles of oxygen partial pressure at the exits of both furnaces (see Figure 5a) indicate that, globally, neither furnace atmosphere was ever starved for oxygen during the combustion events, because the minimum O2 partial pressures never fell below 2%. This suggests that local equivalence ratios in the flame must have been much richer than the φ ) 1-1.2 range of the global equivalence ratios that were calculated previously. This observation is to be expected for diffusion flames. Figure 5, as well as subsequent figures, display superimposed results that have been obtained at the exits of both the first and second furnaces, either in the presence or absence of the filter (i.e., four different sets of data are shown). The introduction of the ceramic filter induced a higher consumption of oxygen, as monitored at the exit of the primary furnace (see Figure 5b). This is attributed to the burning of the soot and tars trapped within the filter. 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 (see Figure 5a and 5b); both values already include the additional room air that was introduced in the venturi. This is explained on the basis

Energy & Fuels, Vol. 18, No. 1, 2004 107

Figure 6. Partial pressures (given as a percentage of atmospheric pressure) and emission yields for CO at the exits of both the primary and secondary furnaces. Primary furnace temperatures were in the range of 500-1000 °C; the afterburner temperature was kept constant at 1000 °C. Legend is as follows: (9), primary furnace, without ceramic filter; (b), secondary furnace, without ceramic filter; (0), primary furnace, with ceramic filter; and (O), secondary furnace, with ceramic filter.

that further conversion of volatile pyrolysates and primary combustion products occurs in the secondary furnace, thereby consuming oxygen. The effect of the ceramic filter on the emissions from the secondary furnace (afterburner) may be seen in Figure 5. At low primary furnace temperatures (500 and 600 °C), where many oil and tar particulates were generated, the afterburner alone was fairly successful at oxidizing them. Hence, the amount of oxygen consumed was comparable in the absence or presence of the filter. At high primary furnace temperatures (700-1000 °C), the particulates were mostly soot and, in the absence of the filter, the afterburner was not capable at destroying all of it. Thus, oxygen consumption in the apparatus was not as high as that in the case where the filter was present. Major Gas-Phase Emissions. Maximum CO partial pressures at the exit of the primary furnace were typically <4%, whereas those at the exit of the secondary furnace were much lower (see Figure 6a). The CO emission yields were drastically increased when the ceramic filter was present, especially at higher temperatures in the primary furnace and, thus, in the filter itself. This is partly attributed to oxidation of soot trapped therein (see a subsequent section). Previous work, including that of Larsen et al.,18 Smith,22 Nagle and Strickland-Constable,23 Neeft et al.,24 and Wang et (22) Smith, I. W. The Combustion Rates of Coal Chars: A Review. In Nineteenth Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1982; pp 1045-1065. (23) Nagle, J.; Strickland-Constable, R. F. Oxidation of Carbon between 1000 and 2000°C. In Proceedings of the 5th Carbon Conference; Pergamon Press: Oxford, U.K., 1961; pp 1, 154.

108

Energy & Fuels, Vol. 18, No. 1, 2004

al.,19 has shown that the oxidation of soot is a relatively slow process at the gas temperatures encountered herein, i.e., <1000 °C. The primary oxidation product of soot is CO, and the CO/CO2 ratio increases as the carbon surface temperature increases (see Tognotti et al.25 and Levendis et al.26). Thus, as the soot burns inside the filter, which is inserted at the end of the primary furnace, CO is released and is detected at the adjacent sampling stage at the exit of the primary furnace. Also, consistent with the aforementioned expectations, the CO emissions experienced an increasing trend with temperature, with the exception of the highest temperature point, which, for unknown reasons, did not quite support this trend. The combustion effluent was treated in the secondary furnace; therefore, most of this generated CO was oxidized to CO2. As a result, the CO yields with and without the filter were comparable at the exit of the secondary furnace. Thus, the final CO emissions of this two-stage combustion process were not significantly affected by the filter, which was placed ahead of the exit of the primary furnace. This behavior was also observed in the combustion behavior of other fuels (see Wang et al.19). To further reduce the final CO emissions, the flowrate of additional air at the venturi may need to be increased. Conversion of the fuel to highly oxidized components signifies a rather complete combustion process. The increase of the CO2 emissions at the exits of both the primary and secondary furnaces (see Figure 7a and b) shows that the overall combustion process was aided by the presence of the ceramic filter. This was especially evident at the exit of the primary furnace. Emissions of SO2 and NOx during the volatile phase of the batch combustion of the tire chips were measured only at the exit of the primary furnace, because only a single a set of pertinent analyzers was available. In the absence of the filter, the SO2 yields were only slightly affected by the operating temperature of the primary furnace (see Figure 8a). Emissions seemed somewhat higher at lower temperatures, perhaps because the volatile fluxes under those conditions were lower and volatile flames lasted longer, and, thus, the overall availability of oxygen in the furnace was higher (see Figure 5). This could have resulted in higher yields of oxidized sulfur in the form of SO2. The SO2 emissions increased in the presence of the filter (see Figure 8a); a possible explanation is that the soot and tars, which may contain sulfur from the vulcanization process, are collected in the filter during combustion. As they burn therein, this sulfur is released and forms additional SO2 inside the primary furnace, especially at the lower temperatures when, again, there is more oxygen in the furnace. In the absence of the filter, NOx emissions increased as the primary furnace operating temperature increased (see Figure 8b). This is expected, because temperature promotes the formation of “fuel-NOx” and, especially, “thermal-NOx”. In the presence of the filter, NOx emis(24) Neeft, J. P. A.; Nijhuis, T. A.; Smakman, E.; Makkee, M.; Moulijn, J. A. Fuel 1997, 76, 1129. (25) Tognotti, L.; Longwell, J. P.; Sarofim, A. In Twenty-Fourth Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1988; p 1207. (26) Levendis, Y. A.; Atal, A.; Carlson. J. B. Environ. Sci. Technol. 1998, 32, 3767.

Caponero et al.

Figure 7. Partial pressures (given as a percentage of atmospheric pressure) and emission yields for CO2 at the exits of both the primary and secondary furnaces. Primary furnace temperatures were in the range of 500-1000 °C; the afterburner temperature was kept constant at 1000 °C. Legend is as follows: (9), primary furnace, without ceramic filter; (b), secondary furnace, without ceramic filter; (0), primary furnace, with ceramic filter; and (O), secondary furnace, with ceramic filter.

Figure 8. Emission yields for SO2 and NOx at the exit of the primary furnace, for primary furnace temperatures in the range of 500-1000 °C. Legend is as follows: (9) without ceramic filter and (0) with ceramic filter.

sions decreased and also experienced a surprising decreasing trend, relative to increasing furnace temperature. Likely explanations for these observations include the possibility that a “reburning” effect occurred inside the filter. Reburning refers to a technique where

Batch Combustion Emissions of Waste Tire Chips

Energy & Fuels, Vol. 18, No. 1, 2004 109

Figure 9. Emission yields for particulate at the exits of both the primary and secondary furnaces. Primary furnace temperatures were in the range of 500-1000 °C; the afterburner temperature was kept constant at 1000 °C. Legend is as follows: (9), primary furnace, without ceramic filter; (b), secondary furnace, without ceramic filter; (0), primary furnace, with ceramic filter; and (O), secondary furnace, with ceramic filter.

a portion of the fuel input to a furnace (typically 10%30%) is added in the post-combustion zone (reburn zone). Combustion of this additional fuel generates oxygen-lean conditions, which lead to reduction of the NOx that was formed in the main combustion zone; this eventually generates N2.27 In these experiments, products of incomplete combustion of the fuel, in the form of soot and tars, collected inside the filter and, as they burned therein, may have acted as a reburn fuel. This reduced the NOx that was formed in the flame. The higher the temperature of the reburn zone (i.e., the filter temperature, in this case), the more effective the reburning. This may explain the decreasing trend of the NOx emissions with increasing primary furnace temperature (see Figure 8b). Other possibilities include reactions of NO2 with soot inside the wall-flow monolith filter. However, most of the NOx emissions detected in this work were identified as NO, and very little NO2 was observed. Particulate Matter Emissions. The effect of the ceramic filter was highly beneficial in reducing particulate emissions (see Figure 9). The highest reductions were observed at high primary furnace temperatures (800-1000 °C). Under these conditions, larger amounts of particulate matter were generated, as attested by the emissions in the absence of the filter, and consequently, they were trapped in the ceramic filter. Also, a higher fraction of this particulate matter was soot (instead of oils and tars). Soot is effectively burned inside the filter. Tars may also burn therein; however, they may partially gasify in the filter and then be transported downstream. The afterburner reduced the particulate emissions further. The final emissions of particulates from this two-stage combustion system were, indeed, very low in the presence of the filter. Particulate yields were reduced by up to 2 orders of magnitude (see Figure 9). This is the major contribution of the high-temperature barrier filter. As mentioned previously, in this particular work, the filter was continuously self-regenerated, because, at the prevailing elevated temperatures, the carbonaceous particulates that were collected therein were gasified. Under the conditions of these tests, the filter did not create any measurable pressure drop. (27) Flagan, R. C.; Seinfeld, J. H. Fundamentals of Air Pollution Engineering; Prentice-Hall: Englewood Cliffs, NJ, 1988.

Figure 10. Particle size distribution of the particulate emission at the exits of (a) the primary and (b) the secondary furnaces, both operated at 1000 °C. Shaded portion represents processing with the ceramic filter in place and unshaded portion represents processing without the ceramic filter in place.

Figure 11. Measured particle size distribution fitted as a Gaussian distribution, for (a) the primary furnace and (b) the secondary furnace, both operated at 1000 °C. Legend is as follows: (9) without ceramic filter and (0) with ceramic filter.

The size distributions of the particulate matter at the exits of both furnaces, as measured by the cascade impactor, are shown in Figure 10. In the presence of the filter, the particulate yields were reduced by 98%99.8% throughout the range of particle diameters stud-

110

Energy & Fuels, Vol. 18, No. 1, 2004

Caponero et al.

Figure 12. Microstructure of the particulate matter from the combustion of tire chips collected on the impactor stages in the exhaust gases of the primary furnace: (a) stage 3, (b) stage 4, (c) stage 5, (d) stage 6, (e) stage 7, and (f) bottom filter.

ied. This is true even for the submicrometer-sized particles (i.e., the most prejudicial size), because they are capable of reaching the alveoli level in the human respiratory system. In the presence of the filter, the amount of particulate mater was so drastically reduced that 10 combustion runs, without disassembling the collector apparatus, had to be performed to obtain sufficient amounts on each stage for gravimetric and subsequent chromatographic analysis.28 The afterburner treatment alone reduced the submicrometer-sized soot emissions by only 11% (see Figure 11). The ceramic filter also affected the range of the particulate distribution. The median particle size decreased from 1.17 µm to 0.60 (28) Even after 10 runs, the samples collected in the first three stages of the impactor did not offer good samples for scanning electron microscopy (SEM) analysis. More runs damage the samples, by overloading them, at all of the other stages.

µm for the primary furnace and from 1.12 µm to 0.59 µm for the afterburner. Scanning electron microscopy (SEM) analysis of particulate matter that was collected on each stage showed some differences between the samples from the primary and secondary furnaces (see Figures 12 and 13). The particulate matter consists of aggregates of spherules. Those that were collected at the exit of the primary furnace had spherule sizes of 210 ( 36 nm (see Figure 12). This size range is compatible with the values found in previous work in the absence of the ceramic filter.2 Similar spherule sizes measured in either the presence or the absence of the filter may, indeed, be expected on the basis that mostly agglomerated particulates are retained by the ceramic filters. The small fraction of small agglomerates that escape the filter may still have the initial size of spherules. There is evidence of

Batch Combustion Emissions of Waste Tire Chips

Energy & Fuels, Vol. 18, No. 1, 2004 111

Figure 13. Microstructure of the particulate matter from the combustion of tire chips collected on the impactor stages in the exhaust gases of the secondary furnace: (a) stage 3, (b) stage 4, (c) stage 5, (d) stage 6, (e) stage 7, and (f) bottom filter.

deposition of tars on the lower stages of the impactor (see Figure 12). The spherule sizes in the particulate matter that has been collected at the exit of the afterburner are larger (390 ( 100 nm). This increase in unit size is, perhaps, associated with an increase in the amount of tars and oils that seemed to have condensed on the particulate matter during the completion of the aforementioned 10 necessary runs. The SEM analysis of the samples collected on the last three stages of the collector and on the bottom fibrous filter does not show soot spherules anymore; instead, some type of coating is evident. This suggests that the spherules are covered with a layer of tars that make them appear bigger under the SEM microscope. This argument is consistent with the typical energy-dispersive spectrometry (EDS) spectrum of the particulate matter, which shows that, in all samples, the same components were

present at similar amounts (see Figure 14). Some elements, such as silicon, aluminum, potassium, and oxygen, originated from the filter; gold was present from the sputtering process, and carbon and zinc came from the collected particulates. Polycyclic Aromatic Hydrocarbon Emissions Behavior. The cumulative emission yields of all detected semivolatile polycyclic aromatic hydrocarbons (PAHs), spanning the mass range from 116 amu (indene) to 278 amu (benzo[b]chrysene and isomers), are shown graphically in Figure 15; PAH emission data for the exhaust gases from the primary and secondary furnace are presented in Tables 3 and 4, respectively. More than 50 PAH compounds were detected by GCMS. Combined amounts of PAHs, both in the condensed phase (collected on cellulose filter paper) and in the gas phase (adsorbed on XAD-4 resin), are shown in Figure

112

Energy & Fuels, Vol. 18, No. 1, 2004

Caponero et al.

Figure 14. Typical EDS spectrum of particulate matter emitted from the combustion of tire chips and collected on the impactor stages. Table 3. Aromatic Components in the Exhaust Gases of the Primary Furnace, in the Presence of the Ceramic Filter, at Various Primary Furnace Temperatures Content (µg/g of fuel burned) compound

500°C

600 °C

700 °C

800 °C

900 °C

1000 °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

275.6 33.3 17.4 161.0 2538.0 32.4 171.7 126.1 17.7 176.4 62.8 405.6 22.9 116.5 19.4 10.9 24.4 289.0 66.9 12.1 15.9 17.5 24.5 9.2 13.4 111.6 60.8 110.5 23.3 20.2 7.5 23.0 4.9 38.3 48.2 21.2 15.6 19.0 12.9 13.8 14.9 29.9 4.6 1.0 3.1 3.8 15.7 10.5 0.5 0.8 2.8

290.7 19.4 20.8 255.3 2787.1 37.4 178.7 127.9 12.9 158.8 56.7 403.7 25.1 121.2 14.0 8.0 26.7 310.6 78.0 8.0 17.1 19.0 23.7 9.3 13.5 128.0 69.1 132.1 19.8 16.0 7.0 26.6 4.8 40.4 56.1 11.4 28.1 19.4 14.6 17.4 18.3 37.8 5.3 1.1 2.2 7.9 19.5 7.1 0.7 1.6 1.5

141.9 22.5 9.9 181.0 2680.2 45.5 195.4 123.0 8.0 124.8 50.0 350.4 23.7 87.5 10.3 5.9 29.4 326.3 78.1 5.5 16.5 17.5 17.0 7.1 10.9 126.4 56.7 135.7 11.6 8.5 5.8 24.8 3.3 29.4 47.7 21.8 14.2 13.3 12.5 14.8 17.2 33.3 3.9 1.5 2.0 8.4 18.4 12.6 1.8 1.4 1.9

107.1 26.6 18.7 65.9 2714.6 33.7 70.2 52.2 5.3 148.8 31.7 301.1 47.4 78.9 9.9 5.1 24.3 292.7 51.3 4.0 8.5 9.4 10.1 3.0 6.8 98.8 33.0 83.1 2.2 4.2 2.0 17.5 1.4 11.8 29.8 10.9 18.6 20.7 8.7 7.6 12.7 17.6 2.4 1.0 1.5 6.9 9.1 3.8 0.5 2.0 0.7

43.2 19.3 15.3 21.6 2516.4 35.6 24.7 16.6 2.2 84.5 8.2 132.5 36.2 20.1 2.9 1.2 18.9 153.7 20.7 1.6 2.6 2.9 3.8 0.7 2.0 73.0 19.7 58.3 1.9 1.4 0.7 12.4 0.7 7.5 16.7 4.9 13.4 14.3 5.5 4.3 8.5 10.5 1.5 0.7 0.7 5.3 6.1 1.3 0.6 0.7 0.3

11.9 4.4 7.1 2.8 2416.7 10.3 12.5 9.3 1.3 63.0 5.5 83.1 28.7 19.4 3.2 1.4 12.3 98.8 12.9 1.1 1.7 1.9 2.2 0.0 1.2 21.9 5.1 15.6 1.1 0.0 0.0 2.7 0.0 0.0 3.5 0.0 2.8 1.3 0.9 0.0 1.7 1.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

total (mg/g of fuel burned)

5.28

5.72

5.20

4.54

3.46

2.87

Batch Combustion Emissions of Waste Tire Chips

Energy & Fuels, Vol. 18, No. 1, 2004 113

Table 4. Aromatic Components in the Exhaust Gases of the Secondary Furnace, in the Presence of the Ceramic Filter, at Various Primary Furnace Temperatures Content (µg/g of fuel burned) compound

500°C

600 °C

700 °C

800 °C

900 °C

1000 °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

1.0 3.3 1.0 0.9 1727.2 1.0 2.4 1.5 0.3 25.0 0.4 11.0 0.8 4.9 0.5 0.2 0.0 30.6 1.9 0.2 0.1 0.2 0.5 0.1 0.1 17.0 4.7 10.7 1.0 1.1 0.1 6.9 0.0 2.0 17.2 15.8 0.1 32.8 5.5 4.8 10.2 8.1 4.0 1.4 0.3 0.8 11.1 5.4 4.0 5.7 2.6

0.1 3.0 0.4 1.2 1512.9 0.6 2.5 1.8 0.1 29.5 0.1 8.3 0.8 5.3 0.1 0.0 0.0 23.3 1.3 0.0 0.0 0.0 0.1 0.0 0.0 3.8 0.2 1.9 0.1 0.1 0.0 0.2 0.0 0.0 1.5 0.1 0.0 5.2 0.7 0.5 1.6 0.7 0.9 0.3 0.0 0.0 1.2 0.5 0.7 0.5 0.4

0.1 2.8 0.0 0.9 1668.6 0.1 3.3 2.1 0.1 32.0 0.7 11.9 1.1 6.3 0.1 0.0 0.0 21.8 1.7 0.0 0.0 0.0 0.2 0.0 0.0 4.2 0.5 2.2 0.1 0.1 0.0 0.2 0.0 0.0 1.3 0.3 0.0 1.4 0.2 0.1 0.5 0.3 2.4 0.1 0.0 0.0 0.8 0.0 0.2 0.2 0.1

0.1 2.0 1.2 10.5 1598.4 1.4 6.3 4.3 1.4 48.1 2.0 29.6 6.6 11.5 1.0 0.0 1.1 33.6 3.7 0.0 0.0 0.5 1.4 0.0 0.0 9.8 1.8 5.2 0.5 0.5 0.0 1.0 0.0 0.1 0.7 0.4 0.6 3.0 0.4 0.2 0.8 0.0 2.1 0.2 0.0 0.0 0.7 0.3 0.7 1.1 0.4

1.5 2.6 1.5 8.9 1811.2 2.0 3.7 2.8 0.1 24.5 0.1 16.2 5.3 6.7 0.1 0.0 0.0 15.3 1.4 0.0 0.0 0.1 1.0 0.0 0.0 6.5 1.1 3.3 0.1 0.1 0.0 0.8 0.0 0.1 0.8 0.4 0.5 2.2 0.2 0.1 0.5 0.5 2.1 0.1 0.0 0.0 0.7 0.1 0.2 0.5 0.0

0.0 1.9 0.7 0.8 1309.3 0.0 2.9 2.0 0.0 23.7 0.9 9.8 1.4 5.1 0.0 0.0 0.0 14.5 1.8 0.0 0.0 0.0 0.6 0.0 0.0 4.1 0.7 2.5 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 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

total (mg/g of fuel burned)

1.99

1.61

15. The ceramic filter reduced the emissions of PAHs from the primary furnace at almost all operating temperatures. It also resulted in a somewhat different trend of emissions with temperature, this time monotonically decreasing with increasing operating temperature. Conesa et al.29 studied the evolution of volatile and semivolatile compounds from tire pyrolysis and showed that the formation of PAHs is primarily related to the pyrolysis process that precedes combustion. In followup work, Fullana et al.30 showed that these compounds are partially eliminated during combustion in the presence of oxygen. Through variation of the bulk air-fuel ratio (λ), which is the inverse of the global equivalence ratio (λ ) 1/φ), they found that the PAH emissions decreased as λ increased (i.e., φ decreased). Results of that work at reported bulk air-fuel ratios in the range (29) Conesa, J. A.; Fullana, A.; Font, R. Energy Fuels 2000, 14, 409. (30) Fullana, A.; Font, R.; Conesa, J. A.; Blasco, P. Environ. Sci. Technol. 2000, 34, 2092.

1.77

1.80

1.93

1.38

Figure 15. Emission yields for cumulative PAHs at the exits of both the primary and secondary furnaces, for different primary furnace temperatures (500-1000 °C); the afterburner temperature was kept constant at 1000 °C. Legend is as follows: (9), primary furnace, without ceramic filter; (b), secondary furnace, without ceramic filter; (0), primary furnace, with ceramic filter; and (O), secondary furnace, with ceramic filter.

of λ ) 0.82-1.19 (φ ) 0.84-1.22), and at a temperature of 850 °C, may be compared to those obtained in this

114

Energy & Fuels, Vol. 18, No. 1, 2004

investigation under overlapping conditions. This comparison shows that the emission yields are in agreement with those measured at the exit of the primary furnace of the present work, especially for styrene, benzaldehyde, fluorene, dibenzo-thiophene, phenanthrene, anhracene, pyrene, 11H-benzo[b]fluorene, and benzo[ghi]fluoranthene. In these experiments, naphthalene was the most predominant PAH produced, and it presented a singular behavior. The presence of the ceramic filter led to an initial decrease in the emissions of naphthalene at the exit of the primary furnace, at the lower temperature range that was investigated (500-700 °C). As the temperature of the primary furnace increased, the emissions of naphthalene also increased until, at the higher temperature range of 800-1000 °C, they exceeded the emissions in the absence of the filter. Other PAH species, such as acenaphthylene, fluorene, phenanthrene, anthracene, fluoranthene, acephenanthrylene, and pyrene, exhibited an expressive decrease in their emissions in the presence of the filter. Yet, other PAHs, such as anthracene, showed an increase in emission at low temperatures, followed by a decrease at higher temperatures, whereas the emission of benzothiophene mildly increased at all temperatures. Treatment of the effluent in the afterburner was very beneficial in reducing all PAH emissions under the conditions herein (see Figure 15 for cumulative yields, as well as Tables 3 and 4 for individual species). The combination of the high-temperature ceramic filter with the afterburner treatment produced results that were superior to the afterburner treatment alone in the absence of the filter (see ref 2). However, the emissions of cumulative PAHs, shown in Figure 15, were almost unaltered by the presence of the filter; this is misleading, because it is almost entirely due to the contribution of one component: naphthalene. Naphthalene accounted for ∼94% of the total mass of PAH at the exit of the afterburner (see Tables 3 and 4), and it was only mildly affected by the presence of the filter. All other compounds, perhaps with one or two exceptions, were drastically reduced at the exit of the afterburner when the high-temperature filter was installed. For instance, phenanthrene emission was reduced by factors of 2-9, cyclopenta[cd]pyrene emission was reduced by factors of 5-10, and fluoranthene emission was reduced by factors of 1.2-9, depending on the temperature of the primary furnace. At primary furnace temperatures of >600 °C the final emissions of most PAH species were almost eliminated by the filter/afterburner treatment (see Table 4). The most notable exception is naphthalene, which was still present in large amounts; lessnotable exceptions include biphenyl, acenaphthylene, fluorene, phenanthrene, fluoranthene, pyrene, and benzo[b]fluoranthene. The cumulative PAH amount that was found condensed on particulate matter collected on each stage of the cascade impactor accounted for <2% of the soot mass at every size cut. This is evident by comparing the data depicted in Figures 10 and 16. The detected cumulative PAH amounts exhibited a drastic reduction when the ceramic filter was installed (see Figure 16), as reflected in the discussion of the previous section (naphthalene, which was resisting the trend, is predominately found

Caponero et al.

Figure 16. Profiles of cumulative PAH condensed on particulates at the various stages of the impactor, at the exits of (a) the primary furnace ((9) without ceramic filter and (0) with ceramic filter) and (b) the secondary furnace ((b) without ceramic filter and (O) with ceramic filter), both operated at 1000 °C.

in the gaseous phase). This reduction seems to be more significant for the bigger particulates (supermicrometersized, >1 µm), with the maximum reduction observed for agglomerated particulates ∼1 µm in size, at the exits of both the primary furnace and the afterburner. Durlak et al.31 reported a similar finding. The fact that smaller particulates are associated with larger amounts of condensed PAHs may be explained on the basis of their larger surface area. PAHs readily adsorb on the surface of particulates. The increase in the efficiency of the oxidation of the PAHs with the filter may be expected because adsoption on the particulate would increase their residence time in the reaction zone. The particulates are “functioning as an ad/absorbent for the PAH”. For the adsorbed PAHs, the efficiency of destruction seems to be high as they pass through the afterburner (i.e., >99%; see Figure 16). Carbon Balance. To assess the validity of this experimental procedure, an assesment of the carbon balance was undertaken. The carbon content present in the reactants was calculated using the chemical analyses of the tire chips and the carbon residue in the sample bed. The volatile pyrolysates were the fuel during these experiments and were composed of 87.4% carbon (see Table 2). The analyses of the exhaust gases of both furnaces gave the amount of carbon present in the products of the combustion reactions. The sum of the carbon content of the CO, CO2, PAHs, and particulate matter led to an average carbon balance of typically 80%-90%; an increasing trend is observed as the temperature of the primary furnace is increased (see Figure 17). The carbon balance of the process using the (31) Durlak, S. K.; Biswas, P.; Shi, J. D.; Bernhard, M. J. Environ. Sci. Technol. 1998, 32, 2301.

Batch Combustion Emissions of Waste Tire Chips

Figure 17. Relationship between the total carbon content in the combustion products and total carbon content in the reactants, i.e., the tire devolatilizates. Primary furnace temperatures varied in the range of 500-1000 °C; the afterburner temperature was kept constant at 1000°C. Solid symbols represent data without filter, and open symbols represent data with filter.

ceramic filter most often showed a higher sum of the carbon content in the products, perhaps because of better combustion and less losses by particle deposition on the walls of the sampling train (because less particles were present therein). Likely sources of error are concentrations of light hydrocarbons, which were not taken into consideration, as well as losses by the thermophoretic deposition of particulates and condensation of semivolatile hydrocarbons on cooler surfaces at the exits of both furnaces, and possible losses during sample collection and preparation. Conclusions Laboratory experiments were conducted to investigate the emissions from the batch combustion of waste automobile tire chips under mildly fuel-rich conditions (global equivalence ratios of φ ) 1-1.2) in the furnace. This is the second part of a study that has been undertaken to identify techniques and conditions for the low-emission burning of tire-derived fuel (TDF). Hot flue-gas filtering, using a high-temperature ceramic honeycomb filter, was implemented herein. The filter was inserted ahead of the exit of a furnace, in which batches of tire chips were burned. The filtered effluent was then channeled into an afterburner. The combina-

Energy & Fuels, Vol. 18, No. 1, 2004 115

tion of hot flue-gas filtration and afterburning treatment seemed to be beneficial in reducing the emissions of products of incomplete combustion (PICs). This combination increased the oxygen consumption and increased the carbon dioxide (CO2) yield at the exit of the afterburner, which is indicative of overall better combustion effectiveness. This combination effectively reduced the carbon monoxide (CO), nitrogen oxide (NOx), individual polycyclic aromatic hydrocarbon (PAH) yields, and it dramatically reduced particulate emissions (soot and tars). For instance, the untreated (i.e., without filter) emissions of pollutants from the primary furnace, when operated at 600 °C, were reduced by the combined filter and afterburner treatment (i.e., compare data denoted by the solid square symbols to data denoted by the open circle symbols in Figures 5-9, 15, and 16) as follows: CO emissions were reduced by a factor of 5 (to a final 10 ppm); NOx emissions were reduced by 32%; cumulative PAH emissions were reduced by 80%; and most individual PAH emissions were reduced by 96%99% (naphthalene emission was reduced by 48%). Particulates were trapped and burned in the hot ceramic filter. Thus, the filter was continuously self-regenerated (cleaned) and a constant backpressure was maintained. Ash remained in the fuel bed. If ash had accumulated in the filter, then provisions for its periodic aerodynamic regeneration could have been made. The size distribution of the particulate matter showed a reduction of 98%-99.8% throughout the range of particle diameters studied. For the PAHs adsorbed on soot, the efficiency of destruction also seemed to be very high, reaching >99%. Overall, these laboratory experiments revealed a viable technique for the low-emission combustion of waste TDF. Acknowledgment. The authors would like to thank Mr. Eric Wisnaskas for assistance with the chromatographic analysis. The authors would also like to acknowledge the Sa˜o Paulo Research Foundation (FAPESP) for financing Dr. Caponero (through Grant No. 99/ 000375-5). Partial support was also provided by the United States National Science Foundation (US-NSF) (through Grant No. CTS-9908962). EF030043B