A study on toxic organic emissions from batchcombustion of styrene

Charlotte Westblad a,b, Yiannis A. Levendis b,*, Henning Richter c,Jack B. Howard c, Joel Carlson d

a Link€ooping Institute of Technology, Link€ooping, Swedenb 334 Snell Engineering Centre, College of Engineering, Northeastern University, Boston, MA 02115, USA

c Massachusetts Institute of Technology, Cambridge, MA 02139, USAd US Army Natick, RD&E Center, Natick, MA 01760, USA

Received 13 August 2001; received in revised form 11 June 2002; accepted 17 June 2002

Abstract

Results from a laboratory-scale investigation on batch combustion of styrene are reported herein. Limited quantities

of waste styrene monomer are incinerated, however this monomer is, also, the primary pyrolyzate during combustion of

waste polystyrene, the second most abundant polymer produced worldwide. Thus, its combustion-generated emissions

are of importance to the operation of hazardous waste incinerators and municipal waste-to-energy powerplants. This

work focuses on emissions of polycyclic aromatic hydrocarbons (PAHs), particulates, as well as carbon monoxide. To

investigate methods for minimizing such emissions, batch combustion of the monomer was conducted in a two-stage

muffle furnace. An additional air mixing chamber was installed between the two stages. Small quantities of the liquid

monomer were inserted in the primary furnace which served as a gasifier/burner. The furnace temperature was in the

range of 300–1000 �C and diffusion flames were formed under most conditions. Upon mixing with additional air,

combustion of unburned gaseous fuel and primary reaction products continued in the secondary furnace (afterburner),

which was kept at a constant temperature of either 1000 or 800 �C. Using this technique, conditions that minimize

emissions were explored and theoretical investigations on the fate of pollutants in the secondary furnace were un-

dertaken. Results revealed that combustion of styrene, which is a highly volatile fuel, occurred with the formation of

flames that were often non-anchored, unsteady and unstable. Emissions of organic pollutants, soot and CO were more

intense than in the case of the polystyrene combustion, studied previously under identical conditions, due to the ad-

ditional depolymerization/pyrolysis steps therein. The emissions from the secondary furnace exceeded those of the

primary furnace, consistent with the fact that a very significant fraction of the fuel conversion occurred in the secondary

chamber. Clear trends in the emissions of PAHs and soot, products of incomplete combustion, with the temperature of

the primary furnace (gasifier) were observed. Emissions were drastically reduced with lowering the gasifier temperature.

While final cumulative emissions of PAHs and soot accounted for more than one third of the mass of the fuel at high

temperatures, their concentrations at the exit of the afterburner were negligible when the primary furnace was operated

at 300 �C under pyrolytic conditions. In the latter case air was added to the afterburner. Numerical modeling based on a

complex reaction network was used for the description of the primary furnace as well as of the afterburner. Kinetic

analysis showed acetylene and benzene to be key species in the growth of PAHs. Formation of PAHs in the afterburner,

found experimentally, was reproduced by the model using a plug-flow assumption.

� 2002 Elsevier Science Ltd. All rights reserved.

Chemosphere 49 (2002) 395–412

www.elsevier.com/locate/chemosphere

*Corresponding author. Tel.: +1-617-373-3806; fax: +1-617-

373-2921.

E-mail address: yal@coe.neu.edu (Y.A. Levendis).

0045-6535/02/$ - see front matter � 2002 Elsevier Science Ltd. All rights reserved.

PII: S0045-6535 (02 )00311-9

Keywords: Styrene; Polystyrene; Combustion; Emissions; PAH; Kinetics

1. Introduction

Approximately 15 million tons of total waste poly-

mers (plastics) are generated throughout Europe each

year, while in the United States the corresponding figure

is 20 million tons (Williams and Williams, 1999). To

dispose such non-biodegradable wastes, incineration is

an advantageous method since it reduces their weight

and volume by 90–95%. In addition, sterilization and

detoxification are ensured in the case of medical waste.

Due to their high energy content, which is comparable

to that of petroleum, plastics are valuable combustibles.

Plastics increase the combustibility of waste stream

mixtures and, thus, enhance the thermal output of fur-

naces (Coghlan, 1994). If the incinerator is a waste-

to-energy plant, then the released heat is recovered to

generate steam, which may be either used to produce

electricity or used directly for heating purposes. How-

ever, there is concern about the formation and emission

of toxic combustion byproducts. Although the toxic

emissions from municipal, hazardous waste as well as

medical incinerators represent a relatively small portion

of environmental emissions, they are more likely to have

a significant local effect due to their location in prox-

imity to urban areas (Barton and Landrum, 1991;

Gordon and Reinhardt, 1991; Green, 1992).

Polyethylene, polystyrene, polypropylene and poly-

vinyl chloride (PVC) are the highest production volume

waste plastics. Over the past several years, research

has been conducted in this laboratory to investigate

the combustion behavior and the associated hazard-

ous emissions from these rather complex waste fuels

(Wheatley et al., 1993; Panagiotou and Levendis, 1994,

1998; Panagiotou et al., 1994, 1996a,b; Courtemanche

and Levendis, 1998; Shemwell and Levendis, 2000). It

was illustrated therein that the aforementioned polymers

exhibited distinct combustion behaviors and, as a con-

sequence, their emissions also differed drastically. Poly-

styrene formed more polynuclear aromatic compounds

(PAHs) and soot than the other common waste poly-

mers when burned, presumably because of the presence

of aromatic rings. The emitted soot was shown to be

highly agglomerated, most was supermicron in size

(Shemwell and Levendis, 2000), and thus, most can be

captured with conventional filters. Other researchers

have also studied combustion of polystyrene, for in-

stance Elomaa and Saharinen (1991), Bernhard et al.

(1998), You and Chiang (1998), as reviewed by Wang

et al. (2001).

Wang et al. (2001) studied the emissions of PAHs and

soot from batch combustion of polystyrene (in the form

of styrofoam) in a two-stage combustor, the two stages

being separated by an additional air mixing section.

While diffusion-type flames occurred, the calculated

global (overall) fuel equivalence ratio was slightly fuel-

lean in the primary furnace. The temperature in the

primary furnace was varied between 500 and 1000 �C.The temperature in the secondary furnace was set at

1000 �C, and the gas residence time therein was kept

constant at 0.7 s. They found that low molecular mass

PAHs were uniformly reduced in the second furnace,

while concentrations of a few large PAHs increased

under certain conditions. They also showed that the

temperature of the primary furnace, where polystyrene

gasification took place, was of paramount importance

to the formation and subsequent emissions of organic

species and particulates. The least amount of pollutants

were obtained with staged combustion at a primary

furnace (gasifier) temperature of approximately 600 �C,followed by treatment in the secondary furnace (after-

burner) operated at 1000 �C.The present work supplements that of Wang et al.

(2001) but attention was turned to the combustion of

styrene, the monomer precursor of polystyrene. This

approach allows to investigate the combustion of sty-

rene monomers decoupled from the depolymerization/

decomposition processes that are associated with pyr-

olysis of polystyrene, prior to its oxidation. Polystyrene

is a complex fuel to describe the chemical kinetics of its

pyrolyzates, which include other compounds besides

styrene. To the contrary, it is more realistic to assess

quantitatively the formation/oxidation of byproducts of

the combustion of pure monomeric styrene, for instance

by means of kinetic modeling. It has been shown by

a number of workers that thermal decomposition of

polystyrene gives high yields of its monomer precursor.

Cullis and Hirschler (1981) found that no appreciable

amounts of volatile products from polystyrene are

formed below 300 �C. Above this temperature the

products consist mainly of the styrene monomer (42–

45%) together with small quantities of the dimmer,

trimer and tetramer. Booty et al. (1999) studied poly-

styrene oxidation in a two-zone continuous-feed flow

reactor and detected styrene between 400 and 675 �C,using gas chromatography coupled to mass spectrome-

try (GC–MS). From 400 to 600 �C the yield of styrene

was around 40%, reaching its maximum yield of 46% at

550 �C. At higher temperatures a drastic decrease took

place, reaching a yield of �1% at 675 �C. Klusmeieret al. (1986) identified styrene to be the main product in

the combustion of polystyrene, with a maximum yield

reaching 64% at 600 �C. Czernik et al. (1990) studied fast

396 C. Westblad et al. / Chemosphere 49 (2002) 395–412

pyrolysis of polystyrene at three different temperatures

(532, 615 and 708 �C), and they also identified styrene tobe the main product. However, they found yields of

styrene much higher (about 75%) than those obtained in

the aforementioned studies, possibly because of the use

of nitrogen as the fluidizing gas, instead of air. The

discrepancy in the observed yields, at the same furnace

temperatures, could be attributed to oxidation of the

depolymerization products including styrene in the

presence of air.

Combustion of liquid styrene also deserves to be

studied on its own right since there are incinerators that

burn waste styrene, such as ENSCO’s El Dorado plant

which uses rotary kilns and afterburners to typically

process 25 000 kg/h of hazardous wastes (ENSCO,

2001). Styrene (C8H8) is prepared from benzene and

ethane (both from petroleum) through ethyl benzene.

The world production in 1996 was 14 700 000 tons of

which 2/3 was used to produce polystyrene (www.ke-

mi.se/kemamne/styren.htm, 2000). Styrene is an irritant,

a narcotic and a neuropathic agent (www.cdc.gov/niosh/

pe188/100-42.html, 2000). Thus in an incinerator, in

order to avoid exposure to humans, the complete de-

struction of waste styrene or waste polystyrene must be

ensured. The correlation between emissions and incin-

eration operating parameters, such as temperature, gas

residence time, addition of excess air, incineration

staging, etc., should be investigated to achieve combus-

tion conditions as efficient as possible, while minimizing

hazardous emissions. A better understanding of the

chemical and physical processes that are responsible

for hazardous emissions, in particular polycyclic aro-

matic hydrocarbons (PAHs) and particulates, may lead

to improved methods for combustion control, and, po-

tentially, methods to reduce formation of these com-

pounds.

During incineration of liquid or solid waste, diffusion

flames form. Due to the lack of oxygen, there is no oxi-

dative attack on the precursors formed during the fuel

decomposition in diffusion flames. Therefore, unlike in

premixed flames, there is a greater tendency to soot at

higher flame temperatures. The most dominant physical

characteristic that affects soot formation is the temper-

ature–time history that the decomposing fuel undergoes

(Glassman, 1988; Glassman et al., 1994). Incipient par-

ticle formation controls the total mass of soot generated

in any process. From previous experiences in this labo-

ratory, it has been observed that once soot is formed

from the combustion of various fuels, it is very resistant

to oxidation at typical post-flame furnace conditions.

Thus, it might be beneficial to burn such solid or liquid

fuels at low furnace temperatures to avoid formation of

soot. However, in those cases an afterburner would be

needed to destroy products of incomplete combustion

emitted from a low-temperature primary burner. Such

two-stage incineration was studied in this work. Initial

gasification/combustion of the waste fuel took place in a

primary furnace, followed by mixing of the gaseous

products with additional air and further combustion of

the resulting charge in a secondary furnace (after-

burner). Different operational conditions were tested,

mostly kept similar to those of Wang et al. (2001) to

allow for a direct comparison.

The experimental results of the present work were

compared with predictions of a detailed kinetic model,

which describes formation and destruction pathways of

chemical species, including most observed PAHs. The

modeling was performed assuming plug-flow conditions

for the secondary furnace. The experimentally deter-

mined chemical composition at the outlet of the primary

furnace was part of the input parameters of the model

calculation. Moreover, an attempt was made to also

model the primary furnace since, unlike the previous

case of the depolymerizing polystyrene, this time the

initial gas-phase fuel therein was known (styrene).

2. Experimental techniques and procedure

2.1. Experimental apparatus and conditions

Batch combustion experiments, involving fixed pools

of liquid styrene fuel, were conducted in a horizontal

electric muffle furnace (1 kW max.) fitted with a quartz

tube, 4 cm in diameter and 87 cm long. This primary

furnace was connected to a secondary muffle furnace

(the afterburner), see Fig. 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, as shown in

Fig. 1. Mixing of the air from the jets and the gaseous

effluent of the combustion occurred in the venturi. Wang

et al. (2001) calculated that the four radially placed jets

at the venturi, based on measured temperatures at the

venturi centerline and at the jet exits, penetrate the ef-

fluent to the centerline of the venturi so that good

mixing of the streams in the venturi is achieved. The

mixed charge was then sampled at the exit of the furnace

and, subsequently channeled to the secondary furnace.

The wall temperature of the primary furnace was varied

in the range of 300–1000 �C. In most experiments the

secondary furnace was kept at 1000 �C, a few experi-

ments were conducted at 800 �C. Upon reaching the

predetermined wall temperature, a porcelain boat loa-

ded with a pool of 0.5 g of liquid styrene was inserted

from the tube’s entrance and it was positioned in the

middle of the quartz tube. To insert the sample quickly

in the furnace, the porcelain boat was placed at the end

of the inner surface of a half tube (a quartz cylinder

longitudinally split along the centerline). The other end

C. Westblad et al. / Chemosphere 49 (2002) 395–412 397

of the tube was mounted at the entrance of the glass

fitting of the furnace. At the start of every experiment,

the fitting was opened and the half tube, with the sample

mounted at its tip, was quickly inserted in the furnace.

To minimize the amount of oxygen that could slip into

the quartz tube when loading the half tube with the fuel,

the inlet was temporarily plugged with a ground glass

fitting.

The combustion experiments were conducted at at-

mospheric pressure at fuel-rich global overall fuel

equivalence ratios. The air flow rate in the first furnace

was 4 lmin�1, and the gas residence time of the gases

between the sample and the venturi was a fraction of a

second. The flow rate of additional air supplied at the

venturi, through the four jets, was 2 lmin�1. The cal-

culated Reynold’s numbers, Re, in the venturi were in

the range of 425–590, depending on the measured tem-

perature therein, which varied with the aforementioned

temperature of the primary furnace. The gas tempera-

ture between the two furnaces dipped to 250–300 �C.The gas temperature in the secondary furnace was

constant for nearly its entire length, 25 �C below its wall

temperature, which was set at set at 1000 �C for most

runs; the residence time of the gas therein was 0.7 s and

Re ¼ 80. Sampling of the involved gases was conducted

at the exits of both furnaces. While luminous diffusion

flame durations, over the boat, were in the order of less

than half a minute, sampling lasted for about 3–4 min to

provide sufficient time for complete carbon oxidation.

This was determined by continuously monitoring the

CO2 emissions.

2.2. Combustion emissions monitoring

PAH emissions, as well as NOx, CO, CO2 and par-

ticulates (soot) from the combustion of styrene (liquid)

and polystyrene were monitored at the exits of the two

furnaces. The PAHs were sampled by passing half of the

combustion effluent through each of the sampling stages

consisting of a Graseby sampling head with a filter stage

and a glass cartridge containing 25 mg XAD-4 adsor-

bents. The sampling stage was placed adjacent to the

furnace, see Fig. 1, to minimize losses. Prior to each

sampling stage the effluent of the furnaces was mixed

with a 2 l min�1 flow of nitrogen gas. This dilution ni-

trogen flow took place in the annulus of two concentric

tubes; the inner tube was perforated and therefore, en-

abled the mixing of the nitrogen with the furnace effluent

(Fig. 1). Hence, the effluent was cooled and inerted.

Subsequently, the particulate emissions were trapped on

the upper portion of the sampling stage on a 90 mm

diameter by 1 mm thick Whatman glass fiber filter with a

nominal pore size of 0.45 lm. Gas-phase aromatic hy-drocarbon emissions were adsorbed on the bed of Su-

pelco XAD-4 resin. The length of the XAD-4 bed was

more than twice its diameter for effective adsorbtion.

Thereafter the effluent passed through a mildly heated

Permapure dryer, where the moisture was removed, it

was then monitored for NOx using a Beckman 951A

chemiluminescent NO/NOx analyzer, for CO and CO2

with Horiba infrared analyzers and for O2 with a

Beckman paramagnetic analyzer. The output of the

analyzers was recorded using a Data Translation DT-

Fig. 1. Schematic of the experimental apparatus. A two-stage furnace was used for batch combustion of styrene. The two stages were

separated by a section where additional air was mixed with the effluent from the primary furnace.

398 C. Westblad et al. / Chemosphere 49 (2002) 395–412

322 data acquisition board in a microcomputer. The

signals from the analyzers were recorded for the dura-

tion of each experiment and subsequently were con-

verted to partial pressures.

2.3. Extraction and concentration of PAH emissions

Following the combustion experiments, the filters

and resins were removed and placed in separate glass

bottles with teflon-lined caps, and stored at 4 �C. Priorto extraction with methylene chloride, 50 lg of internal

standard solution containing equal amounts of naph-

thalene-d8, acenaphthene-d10, anthracene-d10, chrysene-d12 and perylene-d12 was applied to each of the glass

bottles containing the samples. 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, a combustion blank was used 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 qual-

ified based on their concentration therein. Combustion

experiments were triplicated to ensure the reproducibil-

ity of the combustion technique.

A Dionex ASE 200 Accelerated Solvent Extractor

was used for extracting the organic compounds from

both the XAD-4 resins and the cellulose filter papers.

The XAD-4 resins and filters were combined in 33 ml

extraction cells (one cell per sampling stage). The ex-

traction cells were allowed to initially equilibrate at 40

�C in the ASE 200 system for 1 min, and were then filled

with methylene chloride and allowed to thermally

equilibrate at 40 �C and the cell was pressurized to 33

bar 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 ex-

traction 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 25 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 were thor-

oughly cleaned and inspected to ensure that small resin

particles were not 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).

2.4. Extraction recovery

The extraction efficiency is the amount of standard

recovered from the sample divided by the amount of the

known compounds in the standard solution. The aver-

age recovery and statistical evaluation of the blanks,

samples and standards for all of the samples reported

are provided in Table 1. Results show that the recovery

is well over 50% for all of the compounds. The National

Functional Guidelines recommend a recovery of greater

than 50% for the internal standards. The percent dif-

ference (ratio of standard deviation to average recovery)

is low for all compounds but acenaphthene, demon-

strating the very good reproducibility of this extraction

technique for the analysis of XAD-4 resins and filter

papers. The efficiency with which XAD-4 captured va-

pors of styrene monomer in the current setup was as-

sessed by vaporizing a known amount of styrene at the

temperature of 200 �C. Approximately a 50% efficiency

was recorded; the experimentally obtained styrene yields

were adjusted accordingly.

2.5. Analysis by gas chromatography coupled to mass

spectrometry

The GC–MS system consisted of a Hewlett-Packard

(HP) Model 5890 GC equipped with a HP Model 5971

mass selective detector. The GC–MS conditions and

data reduction were described previously (Panagiotou

et al., 1996a,b). The instrument was tuned in accordance

Table 1

The recovery efficiency of the internal standards resulting from the extraction of XAD-4 resins and filter papers with the ASE 200

automated solvent extractor

Internal standard Average recovery (%) Standard deviation (%) Difference (%)

Naphthalene-d8 85 9 11

Acenaphthene-d10 100 22 21

Phenanthrene-d10 100 13 12

Chrysene-d12 99 11 12

Perylene-d12 100 8 8

C. Westblad et al. / Chemosphere 49 (2002) 395–412 399

with EPA semivolatile criteria prior to the GC–MS

analysis of each set of samples. The instrument passed

initial and continuing calibration criteria and no data

were qualified as a result of calibrations. Relative re-

sponse factors were calculated and all criteria were met.

Each of the target compounds as well as the tentatively

identified compounds were quantified using the appro-

priate deuterated internal standard. In a departure from

the EPA method, the GC–MS system was run in the full

scan mode and not in a single ion monitoring mode.

This was done to ensure the identification, quantifica-

tion 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 1 lg of component per gram of fuel

combusted.

3. Results and discussion

3.1. Experiments in the presence of air in both furnaces

The first series of experiments was conducted under

oxidative conditions in both furnaces by introducing air

at the primary furnace as well as at the venturi (air–air).

The temperature in the primary furnace was varied from

500 to 1000 �C, while the temperature in the secondary

furnace was fixed at 1000 �C. The time delay between

insertion of the fuel and ignition was between about a

second at higher temperatures and a few seconds at the

lower temperatures. The flame duration was in the order

of 30 s. At the lower temperatures (500–700 �C) the

flame typically burned over the sample for a few seconds

and then propagated to the exit of the primary furnace.

At temperatures above 700 �C the flame first spread in

both directions (towards the entrance and the exit) and

then sparkled before it stabilized and moved for the exit.

During batch combustion of styrene, the oxygen con-

centration at the exit of the two furnaces experienced

transient profiles. This was indicative of the derivative of

the fuel devolatilization and subsequent conversion in

the gaseous diffusion flame. In these combustion experi-

ments, the minimum oxygen partial pressure at the exit

of the primary furnace decreased from 17% to 14% of

one atmosphere, while that at the exit of the secondary

furnace increased from 9% to 14%, as the temperature

increased from 500 to 1000 �C. Since the minimum ox-

ygen partial pressures at both the entrance and the exit

of the secondary furnace were always above 9%, this

furnace was never starved from oxygen. A great deal of

additional fuel oxidation occurred in the afterburner,

especially when the primary furnace was at the lower

temperatures.

The peak CO2 mole fractions at the exit of the pri-

mary furnace varied between 3% and 5%, depending on

the temperature, which are at the low end of results of

detailed balancing calculations for combustion of sty-

rene at equivalence ratios (/) between 1 and 2 (16.1%

and 4.3%, respectively). Indeed, the global (overall)

equivalence ratio in the primary furnace was calculated

to be in the neighborhood of / ¼ 1:67, based on the

total amount of oxygen provided during the duration of

the combustion events. However, the fact that the fur-

naces were never oxygen-starved, as indicated by the

minimum oxygen mole fractions mentioned above,

points to even higher local equivalence ratios in line with

the observed CO2 mole fractions. Combustion of poly-

styrene resulted in peak CO2 mole fractions between 7%

and 15% (Wang et al., 2001), which is consistent with the

lower calculated global fuel equivalence ratio (/ ¼ 0:8)therein. In the present work, peak CO varied between

0.7% and 1.5% (7000–15 000 ppmv), while the peak CO

mole fractions during the combustion of polystyrene

were around 2% (Wang et al., 2001). If a low emission

standard (e.g. 100 ppmv) were to be maintained in the

effluent of an incinerator, waste styrene would have to

be fed in very small quantities over a period of time. The

high CO emissions encountered in this work illustrate

what happens when large batches of certain wastes are

introduced to incinerators. Generation of such upset

conditions is often the case.

The overall yields of CO and CO2 emissions were

obtained by integrating the time-dependent profiles re-

corded during the batch combustion events, and results

are shown in Fig. 2. Examining the exit of the primary

furnace, the minimum yield of CO2 was obtained at 500

�C whereas for CO the minimum yield occurred at 700

�C. Maximum CO2 yields were obtained in the vicinity

of 800 �C, where CO was at a minimum. At higher

temperatures, the yield of CO2 dropped mildly and CO

increased. This particular drop of the CO2/CO ratio

would be thermodynamically consistent with the in-

crease in temperature. The emissions of CO2 and

CO from the secondary furnace were much higher

than those from the primary one, illustrating additional

oxidation, in complete agreement with the associated

lower oxygen concentrations. The trends at the exit of

the secondary furnace show a minimum yield of CO2 at

700 �C, followed by nearly constant yields at higher

primary furnace temperatures. The CO yield from the

secondary furnace dropped with temperature, and at

1000 �C the CO yield was below that of the primary

furnace.

The cumulative emission yields of all detected aro-

matics, particularly PAHs from mass range of 116 amu

(indene) up to 278 amu (picene) are also plotted in Fig.

2. More than 50 semivolatile PAH compounds were

detected by GC–MS. Plots of selected major individual

PAH component yields are shown in Fig. 3. Often dif-

ferent species with identical molecular mass and chemi-

cal formula, i.e., isomers could be identified in the

extracts. The number of possible isomers and therefore

400 C. Westblad et al. / Chemosphere 49 (2002) 395–412

the challenge for chemical analysis increased signifi-

cantly with the number of carbon atoms, i.e., with the

molecule size. The analysis was further complicated by

the presence of partially substituted PAHs in the

Fig. 2. Yields (in mg/g of fuel) of cumulative PAHs (sum of all detected PAHs), particulates, CO and CO2 as well as styrene monomer

at the exits of the primary and secondary furnace, as a function of the primary furnace temperature. Air was introduced at both the

primary furnace and the venturi, with flowrates of 4 and 2 lmin�1, respectively. The secondary furnace temperature was 1000 �C.

C. Westblad et al. / Chemosphere 49 (2002) 395–412 401

Fig. 3. Yields of 16 most prominent PAH compounds (in lg/g of fuel) at the exits of the primary and secondary furnace, as a functionof the primary furnace temperature (in �C). Air was introduced at both the primary furnace and the venturi, with flowrates of 4 and 2lmin�1, respectively. The secondary furnace temperature was 1000 �C.

402 C. Westblad et al. / Chemosphere 49 (2002) 395–412

combustion products. These PAHs were generally sub-

stituted by short saturated or unsaturated carbon chains,

including methyl, with a significant number of possible

isomeric positions, in particular in the case of larger

PAHs. For instance, different methylphenanthrene iso-

mers could be identified unambiguously. Two trends

for the cumulative PAH emissions can be identified from

Fig. 2(a) the PAH yields both from the primary and

from the secondary furnace increased with the primary

furnace temperature, and (b) the yields from the sec-

ondary furnace were much higher. The same two ob-

servations were also true for the particulate yields, also

shown in Fig. 2. However, these trends were quite op-

posite from those of the emissions of unburned styrene

monomer; which markedly decreased with the tem-

perature of the primary furnace and were, also, dras-

tically reduced (by 15 times at 500 �C) by the afterburnertreatment. These profiles reveal that as the tempe-

rature of the primary furnace increased, additional

amounts the styrene were burned therein in diffusion

flames and, thus more products of incomplete combus-

tion (condensed structures, such as PAHs and soot) were

formed. Any unburned styrene was effectively destroyed

in the afterburner and hence, additional condensed

species formed therein. Apparently, oxidation reactions

also took place in the afterburner and as a result addi-

tional CO2 was emitted therefrom. It should be men-

tioned that between 500 and 800 �C the captured

particulates consisted of a mixture of oil, tars and grey–

black particles. Above 800 �C the particulates were

brown–black in color, consisting exclusively of tars and

soot.

The profiles of 16 major PAH component yields are

shown in Fig. 3. These are indene, naphthalene, biphenyl,

acenaphthalene, fluorene, phenanthrene, anthracene,

fluoranthene, acephenanthrylene, pyrene, benzo[ghi]flu-

oranthene, cyclopenta[cd]pyrene, benzo[b]-flouranthene,

benzo[a]pyrene, perylene and benzo[ghi]perylene. Ex-

amination of the emissions of these individual PAH

components, indicates that for most yields were low at

the lowest furnace temperatures of this study, and

thereafter, they increased with rising furnace tempera-

tures, in agreement with the cumulative PAH trend

shown in Fig. 2.

The treatment in the afterburner increased the yields

of all individual compounds. Within the same sample

at the exit of the primary furnace the following trends

were observed, which are in agreement with previous

observations on other fuels, including as polystyrene

and medical examination latex gloves (Longwell, 1982;

Levendis et al., 2001; Wang et al., 2001):

(a) The yield of naphthalene was very high, amounting

to a few percent of the sample mass; naphthalene

yields were much higher than those of acenaphthyl-

ene.

(b) The yields of phenanthrene were generally higher

than those of the remaining PAHs, occasionally ex-

ceeding even those of naphthalene.

(c) The yields of anthracene, fluoranthene and pyrene,

were comparable to each other, and were all less

than naphthalene.

(d) The yields of cyclopenta[cd]pyrene were by factors

of 2–3 lower than those of pyrene.

(e) The yields of benzo[a]pyrene were lower than those

of cyclopenta[cd]pyrene.

(f) The yields of perylene were even smaller.

Contrary to the experience of the study of Wang

et al. (2001), the afterburner did not reduce the PAH

emissions from the primary furnace because, unlike

polystyrene, a great deal of the combustion of styrene

took place in the afterburner instead of in the primary

furnace. It was evident from experimental observations

that evaporation/devolatilization fluxes from heating up

styrene were much larger than depolymerization/devol-

atilization fluxes from heating up polystyrene. This is

also supported by the fact that liquid styrene has a va-

porization temperature of 145 �C, while polystyrene hasa melting point of 237 �C and a minimum decomposi-

tion temperature of 364 �C (Beyler, 1988). Upon ignition

in the primary furnace, the styrene flames were unsteady

and unstable; often they propagated downstream in the

primary furnace, were anchored in the venturi, and oc-

casionally continued even in the afterburner. To the

contrary, upon ignition of polystyrene, flames formed

over the sample bed and were anchored there as they

were continuously fed by a visible stream of pyrolyzates.

Furthermore, styrene combustion durations, as recorded

by the online analyzers, lasted for about 50 s, while

polystyrene combustion durations lasted for 100 s. As a

result of the higher devolatilization fluxes and there-

fore higher fuel concentrations in the primary furnace,

emissions from burning styrene were much higher that

those from burning polystyrene. PAH emissions were

higher by factors of 3–12 (!) from the primary and sec-

ondary furnaces, respectively. Particulate emissions were

comparable from the primary furnace but were higher

from the secondary furnace by 2–3 times in the case of

styrene combustion. CO2 emissions were lower from

styrene, while CO emissions were comparable. Finally,

in both the styrene and polystyrene cases, PAH and

particulates emissions increased with primary furnace

temperature in the range of 600–1000 �C. These resultsindicate that the same furnace setup was more over-

whelmed in the case of the more volatile fuel and the

secondary furnace behaved rather as an additional

main furnace than as an afterburner. To reduce the

devolatilization flux from styrene and render the situa-

tion more manageable it would be sensible to operate

the primary furnace, where the liquid fuel is vaporized,

at a sufficiently low temperature. Reducing the flame

C. Westblad et al. / Chemosphere 49 (2002) 395–412 403

temperature in diffusion flames leads to a decrease of the

fuel pyrolysis rate and, therefore, of the formation rate

of potential soot precursors (Chung and Lai, 1992). This

was indeed observed in these experiments. Operating

the primary furnace at 500 �C reduced the formation

of products of incomplete combustion therein and

enabled effective operation of the secondary furnace

(afterburner); the remaining styrene was nearly quanti-

tatively destroyed while the formation of PAHs was

curtailed and maximum conversion to CO2 achieved.

In view of the above arguments, to further reduce the

devolatilization flux in the primary furnace one would

either have to lower its temperature, even below 500 �C,or use smaller quantities of fuel, or use an inert atmo-

sphere to prevent flame formation, or confine the fuel

somewhat to impede the release of volatiles. Some of

these methods were investigated herein and results are

described in the following subsections.

3.2. Experiments under pyrolytic conditions in the primary

furnace

Experiments were performed at lower (than the

above) primary furnace temperatures, i.e., 300–500 �C,with the secondary furnace operating at 800 �C. Tofurther minimize the formation of soot by avoiding a

flame in the primary furnace the gas therein was pure

nitrogen. The fuel vapors were then mixed with pure

oxygen at the venturi. The proportions of the nitrogen

and oxygen flowrates, 4.7 and 1.3 lmin�1 respectively,

were chosen to provide the afterburner with a baseline

entrance gas composition equivalent to that of air (21%

O2 and 79% N2), as it was the case in the previous ex-

periments. Upon evaporation of the sample, a flame

occurred after the venturi and further combustion took

place in the secondary furnace. The flame was again very

unstable traveling the length of the tube between the

venturi and the secondary furnace, seemingly extin-

guishing and re-igniting. Resulting emissions are shown

in Figs. 4 and 5. Total styrene emissions from the pri-

mary furnace were high (0.4 g/g of sample, i.e., primary

furnace emissions times two) and were rather constant

with temperature; they were subsequently virtually de-

stroyed in the afterburner. The trends of the emissions of

the products of incomplete combustion with tempera-

ture are similar to those experienced in the air–air ex-

periments, described before. The PAH and particulate

emissions decreased even further as the temperature of

the primary furnace was lowered from 500 to 300 �C.The lowest emissions from the combined system oc-

curred at a primary furnace temperature of 300 �C (af-

terburner at 800 �C). At these conditions conversion of

styrene to CO2 was nearly complete, consistent with the

lowest remaining O2 partial pressures and highest CO2

yields at the exit of the afterburner recorded in this

study. Individual PAH emissions under these conditions

were indeed very low, see Fig. 5, and no soot was gen-

erated in the clear blue flames that formed in this case.

Such conditions may be, indeed, ideal for the operation

of hazardous incinerator furnaces if, however, one can

afford a somewhat slower fuel processing rate dictated

by the low temperature of the primary furnace/gasifier.

At 300 �C, the total pyrolysis/combustion process time

was lengthened by a factor of 2, with the entire event

lasting 120 s. However, this timeframe is not excessively

long if one considers that burning the same mass of

polystyrene required also nearly the same time. Based on

this, reductions of the evaporation/combustion time-

frames may not be an issue, especially when one con-

siders the environmental benefits.

3.3. Experiments with the fuel confined in capsules

A series of experiments explored the benefit of re-

ducing the vaporization rates of styrene on the pollutant

emissions, as discussed above. These experiments were

carried out by burning styrene enclosed in a gelatin

capsule. The sample mass of styrene was 0.9 g and the

capsule weighed 0.15 g. The primary furnace tempera-

ture was varied in the range of 500–1000 �C, and the

secondary furnace was kept at 1000 �C. Combustion wasconducted in air. Results showed that this technique had

only a small impact on reducing pollutants from burning

styrene; likely, because the fuel mass was larger than in

the previous experiments. The emissions were compa-

rable to the air–air experiments with some reduction of

emissions noticed at the exit of the primary furnace,

particularly at the lower temperatures. Hence, while

these results supported the argument that reducing the

evaporation rates of styrene lowers the emissions, they

did not show sufficient improvement to warrant further

discussion.

3.4. Emissions of NOx

Yields of NOx from the combustion of styrene were

only recorded at the exit of the primary furnace. Com-

bustion generated some NOx, but it was always found to

be at low levels (3–10 ppm).

3.5. Experimental uncertainty

In all cases a minimum of two combustion experi-

ments were conducted; if the results were not in good

agreement a third experiment was performed, for some

even a fourth. In most cases, the third experiment was

added regardless of the agreement of the two runs, in

order to probe the repeatability of the technique. The

repeatability in measurements for CO, particulates, four

404 C. Westblad et al. / Chemosphere 49 (2002) 395–412

Fig. 4. Yields (in mg/g) of cumulative PAHs (sum of all detected PAHs), particulates, CO and CO2 as well as styrene monomer

at the exits of the primary and secondary furnace, as a function of the primary furnace temperature. Nitrogen (4.7 lmin�1) was in-

troduced at the primary furnace and oxygen (1.3 lmin�1) was introduced at the venturi. The secondary furnace temperature was

800 �C.

C. Westblad et al. / Chemosphere 49 (2002) 395–412 405

Fig. 5. Yields of 16 most prominent PAH compounds (in lg/g of fuel) at the exits of the primary and secondary furnace, as a functionof the primary furnace temperature. Nitrogen (4.7 lmin�1) was introduced at the primary furnace and oxygen (1.3 lmin�1) was in-

troduced at the venturi. The secondary furnace temperature was 800 �C.

406 C. Westblad et al. / Chemosphere 49 (2002) 395–412

major PAHs and styrene is shown in Fig. 6. In most

cases the agreement was satisfactory, varying from a

few percent up to a factor of 2 in the worst case.

Trends have been unambiguously supported by all re-

peats. Errors may arise from the combustion technique

itself, such as small variabilities in the manual sample

insertion, variability in the sample ignition and, hence,

the ensuing flame formation, etc. Small losses and

variabilities may subsequently affect the sampling,

transferring and storage of filters, adsorbents, etc. Fi-

nally, some uncertainties may be expected in the ana-

lytical techniques.

4. Kinetic analysis

The development of procedures allowing for the ef-

ficient thermal destruction of styrene and styrene-based

waste requires a deeper understanding of the chemical

processes responsible for the consumption and forma-

tion of products of incomplete combustion, in particular

Fig. 6. Yields of particulates, CO, naphthalene, phenanthrene, fluoranthene and styrene (in mg/g of fuel) of repeated experiments at

the exits of the primary and secondary furnaces, as a function of the primary furnace temperature. Tafterburner ¼ 1000 �C in all cases.

C. Westblad et al. / Chemosphere 49 (2002) 395–412 407

of PAHs. The above described experimental results at

various primary furnace temperatures show significant

differences in the PAH concentrations at the exit of the

afterburner while it was constantly operated at a tem-

perature of 1000 �C. Concentrations of PAHs monitoredat the exit of the primary furnace were consistently lower

than those at the exit of the afterburner (Figs. 2 and 3).

However, styrene concentrations dropped dramatically

due to the afterburner treatment. Therefore, a link be-

tween these two findings cannot be excluded, especially

at the higher temperatures of the first stage. The ex-

perimental observations at the exit of the afterburner are

likely to be related to concentrations of specific species

at the inlet of the afterburner. Kinetic modeling was

used to identity these compounds and to assess reaction

pathways explaining the resulting PAH growth.

A complex kinetic model describing the formation

and growth of PAHs up to coronene (C24H12) has been

developed and tested for low pressure (Richter et al.,

1999, 2000a) and atmospheric pressure (Richter et al.,

2000b) conditions, taking into account the pressure-

dependence of chemically activated reactions. The at-

mospheric pressure version of the model has been

applied recently to the description of the afterburner in

the case of two-stage combustion of polystyrene using

the same experimental setup as in the present work

(Wang et al., 2001, 2002). The data obtained in the

present work differ significantly from the previous

studies because of the pronounced changes in PAH

concentrations under certain conditions, while in the

case of polystyrene combustion post-treatment had only

a limited effect. The kinetic analysis of the oxidative

post-treatment of polystyrene combustion effluents al-

lowed for a more realistic assessment of PAH oxidation

in the model (Wang et al., 2002), taking into account the

experimentally observed slow degradation of once

formed PAHs.

4.1. Approach

The use of two combustion stages allows the de-

scription of the experimental setup by two separate se-

quential processes. As discussed above, the first stage

(the primary furnace) consists in a preheat zone followed

by reaction in a diffusion flame-type combustion. The

second stage, the afterburner, can be described––at least

in a first approximation––as a plug-flow reactor with the

outlet of the first stage after injection of additional air as

its input. In the present work two experimental condi-

tions were examined in more detail: (a) 1st stage: sty-

rene/air, 500 �C; 2nd stage (afterburner): 1000 �C; (b) 1ststage: styrene/air, 900 �C; 2nd stage (afterburner):

1000 �C.In the primary furnace, a complex sequence of

chemical reactions is connected to physical processes

such as the evaporation of the fuel and diffusion phe-

nomena. The detailed description of the interaction of

physical and chemical processes is not straightfor-

ward; therefore no quantitative modeling of the pri-

mary furnace is possible at the current time. However, in

comparison to polystyrene incineration, the degree of

complexity is reduced in the present case of the com-

bustion of monomeric styrene. No depolymerization

step is necessary and no fuel in solid form is present,

styrene being liquid under standard conditions. There-

fore, different from the previous work (Wang et al.,

2001, 2002), the impact of the operating conditions of

the primary furnace was explored based again on a plug-

flow assumption. The calculated overall equivalence

ratio of the styrene/air mixture, / ¼ 1:67, was used at

the inlet. The combustion of styrene is exothermic, thus

the flame temperature is expected to be significantly

higher than the furnace temperature which was kept

constant at 500 or 900 �C. Establishment of partial

equilibria in the reaction zone of diffusion flames, in

particular of COþOH¡CO2 þH, has been shown in

the past (Mitchell et al., 1980), therefore ratios between

integrated experimental CO2 and CO concentrations

should allow for a reliable assessment of the tempera-

ture at which combustion occurs. In order to identify

the combustion temperature, the primary furnace was

modeled at different temperatures using as input a sty-

rene/air mixture with an equivalence ratio of / ¼ 1:67.CO/CO2 ratios predicted for an estimated residence time

of 0.7 s were compared with experimental data at the

outlet of the primary furnace. All model calculations in

the present work were conducted with the Senkin code

of the Chemkin collection (Kee et al., 2000).

A CO/CO2 ratio of 0.32 was predicted at a flame

temperature of 1200 K (927 �C) while the model calcu-lation at 1400 K (1127 �C) led to a value of 1.21. These

data are in a good agreement with the experimental

findings of 0.38 and 1.21 at furnace temperatures of 500

and 900 �C, respectively. The comparison of the model

predictions with experimental data at the exit of the

primary furnace showed agreement within an order of

magnitude for most PAHs. Nevertheless, the underpre-

diction of styrene by more than four orders of magni-

tude, illustrates the effect of imperfect mixing and

therefore the limitations of the plug-flow assumption for

the primary furnace. Similar to the experimental obser-

vations at primary furnace temperatures of 500 and 900

�C, the increase of the flame temperature from 1200 to

1400 K showed no clear trend of the predicted PAH

concentrations and led overall to comparable results for

both temperature settings. However, the comparison

between model predictions at both temperatures for

species not monitored experimentally revealed signifi-

cant differences. This may explain why the subsequent

afterburner treatment depended on the temperature of the

primary furnace. Particularly striking was the increase

of the predicted mole fractions of acetylene, ethylene,

408 C. Westblad et al. / Chemosphere 49 (2002) 395–412

benzene, phenylacetylene, methane and molecular hy-

drogen by the factors 11.0, 3.2, 25.9, 7.0, 22.6 and 3.7,

respectively, when the temperature was changed from

1200 to 1400 K.

In the next step, the predicted mole fractions of the

above listed species, N2, H2O, experimental data for

major PAHs as well as integrated average mole fractions

of O2, CO and CO2 were used as input for the modeling

of the afterburner at 1000 �C. Dilution by addition of airat the venturi between both stages was taken into ac-

count by correcting the computed input mole frac-

tions accordingly. The comparison of model predictions

with experimental PAH data at the exit of the after-

burner gave encouraging results. For this purpose,

measured gas-phase concentrations were converted to

mole fractions using the total number of moles flowing

through the afterburner. Particularly, the overall trend

of a more pronounced PAH formation as the tempera-

ture of the primary furnace was increased from 500 to

900 �C could be reproduced. The predicted evolution of

selected PAHs is shown in Fig. 7 in comparison to ex-

perimental data at the inlet and the exit of the after-

burner.

4.2. Analysis of reaction pathways

In order to gain insight in major reaction pathways,

rates of production were determined for species of par-

ticular interest using the post-processor corresponding to

the Senkin code, included in the Chemkin software

package (Kee et al., 2000). This approach allows to assess

quantitatively formation or consumption of selected

species as well as the contributions of individual reac-

tions at different residence times. This analysis showed

that acetylene was a key species in the growth process

to larger and larger PAHs such as pyrene and benzo-

[ghi]perylene via the hydrogen-abstraction/acetylene-ad-

dition (HACA) mechanism. This finding is consistent

with the computed higher acetylene concentration at

the exit of the primary furnace operated at 900 �C. In-terestingly and in opposition to high-temperature com-

bustion, hydrogen-abstraction occurs mainly by reaction

Fig. 7. Comparison between model predictions and experimental mole fractions at the inlet and exit of the afterburner. (a) First stage:

500 �C, naphthalene; (b) first stage: 500 �C, benzo[ghi]perylene; (c) first stage: 900 �C, pyrene; (d) first stage: 900 �C, cyclo-penta[cd]pyrene.

C. Westblad et al. / Chemosphere 49 (2002) 395–412 409

with OH radical while reaction with hydrogen radicals

proceeds in the reverse direction. For instance, the re-

actions of acenaphthylene with H to 1-, 3-, 4- and 5-

acenaphthyl radicals and H2 are acenaphthylene-forming

pathways, under the conditions of this work. This ob-

servation is consistent with the temperature-dependence

of the corresponding equilibrium constants. The equi-

librium constant of the reaction with H is increasing with

temperature while that of the reaction with OH is

decreasing. Additional acetylene is formed via unimo-

lecular decay of C6H5CHCH, i.e., C6H5 þ C2H2 ¡

C6H5CHCH. C6H5CHCH was found to be the domi-

nant product of the first step of styrene degradation oc-

curring mainly by hydrogen abstraction from styrene

(C6H5CHCH2) by reaction with H and OH. However,

the predicted mole fractions of C6H5CHCH were ex-

tremely low (<10�10) and the reverse reaction with H2 did

play no role in this case. Naphthalene, the most abun-

dant PAH at the exit of the afterburner, was shown to be

formed by reaction of two cyclopentadienyl moities, i.e.,

C5H5 þ C5H5¡C10H8 þ 2H. Cyclopentadienyl is the

product of unimolecular decomposition of phenoxy

radicals (C6H5O) (Lin and Lin, 1986). Phenoxy radicals

are formed by benzene oxidation via reaction C6H6 þO¡C6H5OþH. Benzene is a product of the unimolec-

ular decomposition of styrene (C8H8¡C6H6 þ C2H2)

and of its reaction with hydrogen (C6H6 þ C2H3 ¡

C8H8 þH). In addition, benzene is formed from phenyl,

a product of the above-mentioned C6H5CHCH decay.

The benzene concentration was predicted to increase

significantly in the afterburner reaching mole fractions of

6:7� 10�4 and 1:3� 10�3 at a residence time of 0.8 s,

using input parameters corresponding to temperatures of

the first furnace of 500 and 900 �C, respectively. A similar

increase through the afterburner was predicted for acet-

ylene with computed mole fractions at the exit of the

afterburner of 1:8� 10�3 and 2:0� 10�3. The measure-

ment of both acetylene and benzene is of great interest

due to their importance for PAH growth and will be

undertaken in future work.

5. Conclusions

This work dealt with observations on the combustion

and emissions of styrene monomer. The fuel was burned

in batches of 0.5 g in a laboratory muffle furnace setup.

Two furnaces were placed in series separated by a mix-

ing section, where additional air was introduced. The

primary furnace acted as the liquid fuel gasifier/combu-

stor. Gaseous products of incomplete combustion were

then well-mixed with additional air at a venturi and were

channeled to a secondary furnace (afterburner) where

further gas-phase reactions occurred. The temperature

of these two atmospheric pressure furnaces was varied

independently. This work was performed for two rea-

sons: (a) to study the combustion of liquid styrene waste,

which is burned in batches in hazardous waste inciner-

ators (ENSCO, 2001); and (b) to help the understanding

of the gas-phase chemistry during combustion of wide-

spread polystyrene consumer wastes. The reasoning in

the latter case is that styrene monomer is the major

depolymerization product of polystyrene upon heating.

Combustion of styrene and polystyrene were compared

at the same conditions.

Results showed that styrene, being more volatile,

ignited instantaneously and burned in a fraction of the

time needed for polystyrene, approximately 50%. Unlike

in the case of polystyrene, styrene diffusion flames were

not always anchored over the sample pool but were

rather unstable and often traveled down and up the

furnace tube. At times they would disappear and then

re-ignite. To the contrary, ignition of polystyrene took

time and ensuing diffusion flames were very stable

and formed over the sample bed. As a result of such

disparate ignition and combustion behaviors, emis-

sions of products of incomplete combustion were much

higher from styrene than polystyrene, often by factors of

2–10.

Different to the previous study of polystyrene

combustion (in the form of shredded styrofoam cups,

see Wang et al., 2001, 2002), the afterburner did not

reduce the final emissions of PAHs over those of the

primary furnace. This was due to the fact that a great

deal of the combustion of styrene, unlike polystyrene,

took place in the afterburner instead of in the primary

furnace.

To minimize the emissions of PAHs and soot from

this apparatus, a clear trend appeared upon variation of

the primary furnace temperature, while the temperature

of the secondary furnace (the afterburner) was kept

sufficiently high, at 800 or 1000 �C. As the temperatureof the primary furnace was reduced, the pollutant yields

from the overall system decreased. Operating the pri-

mary furnace at the rather low temperature of 300 �Cthe emissions of PAHs and soot were remarkably low at

the exit of the afterburner; this, in conjunction with low

unburned fuel emission and high carbon dioxide con-

centrations, indicated good combustion.

Kinetic analysis showed the importance of benzene

and acetylene in the growth process of PAHs. Forma-

tion of PAHs in the afterburner at 500 and in particular

at 900 �C was reproduced using a complex kinetic

modeling and making a plug-flow assumption.

Acknowledgements

The authors acknowledge Mr. Eric Wisnaskas, Mr.

Jun Wang, Mr. Francesco Ayala Sanchez and Prof. Leif

Thuresson for assistance. This project was funded by

410 C. Westblad et al. / Chemosphere 49 (2002) 395–412

NSF with grant CTS-9908962, Dr. Farley Fisher pro-

gram director.

References

Barton, R.G., Landrum, V.J., 1991. Medical waste manage-

ment and disposal. U.S. Environmental Protection Agency,

Office of Solid Waste, pp. 108–112, 121–122.

Beyler, G., 1988. Thermal Decomposition of Polymers. Hand-

book of Fire Protection Association, Quincy, MA.

Bernhard, M.J., Biswas, P., Durlak, S.K., Shi, J., 1998.

Characterization of polycyclic aromatic hydrocarbon par-

ticulate and gaseous emissions from polystyrene combus-

tion. Environ. Sci. Technol. 32, 2301–2307.

Booty, M.R., Bozzelli, J.W., Park, B.-I., 1999. Polymer

pyrolysis and oxidation studies in a continuous feed and

flow reactor: cellulose and polystyrene. Environ. Sci. Tech-

nol. 33, 2584–2592.

Chung, S.L., Lai, N.L., 1992. Suppression of soot by metal

additives during the combustion of polystyrene. J. Air

Waste Manage. Assoc. 42, 1082–1088.

Coghlan, A., 1994. Can burning plastics be good for the world?

New Sci. 143, 20.

Courtemanche, B., Levendis, Y.A., 1998. Control of the HCl

emissions from the combustion of PVC, by in-furnace

injection of calcium/magnesium based sorbents. Environ.

Eng. Sci. 15 (2), 123–135.

Cullis, C.F., Hirschler, M.M., 1981. The Combustion of

Organic Polymers. Oxford University Press, New York,

pp. 117–119, 135.

Czernik, S.R., Piskorz, J., Radlein, A.G., Scott, D.S., 1990.

Fast pyrolysis of plastic wastes. Energy Fuels 4, 407–411.

Elomaa, M., Saharinen, E.J., 1991. Polycyclic aromatic hydro-

carbons (PAH) in soot produced by combustion of poly-

styrene, polypropylene and wood. J. Appl. Poly. Sci. 42,

2819–2824.

ENSCO, Environmental Systems Company, 2001. Personal

communication.

Glassman, I., 1988. Soot formation in combustion processes.

Proc. Combust. Inst. 22, 295–311.

Glassman, I., Nishida, O., Sidebotham, G., 1994. Critical

temperatures of soot formation. Soot formation in com-

bustion. In: Bockhorn, H. (Ed.), Springer Series in Chemical

Physics, vol. 59. Springer, Heidelberg, pp. 316–324.

Gordon, J.G., Reinhardt, P.A., 1991. Infectious and Medical

Waste Management. Lewis Publishers, Inc., Chelsea, MI,

p. 94.

Green, A.E.S., 1992. Medical Waste Incineration and Pollution

Prevention. Van Nostrand Reinhold, New York, p. 37, 45.

Kee, R.J., Rupley, F.M., Miller, J.A., Coltrin, M.E., Grcar,

J.F., Meeks, E., Moffat, H.K., Lutz, A.E., Dixon-Lewis, G.,

Smooke, M.D., Warnatz, J., Evans, G.H., Larson, R.S.,

Mitchell, R.E., Petzold, L.R., Reynolds, W.C., Caracotsios,

M., Stewart, W.E., Glarborg, P., Wang, C., Adigun, O.,

2000. CHEMKIN collection, release 3.6, Reaction Design,

Inc., San Diego, CA.

Klusmeier, W., Ohrbach, K.H., Kettrup, A., 1986. Combustion

experiments of polystyrene by use of a modified Bayer-Ici-

Shell apparatus. Thermochim. Acta 103, 231–237.

Levendis, Y.A., Atal, A., Carlson, J.B., Esperanza, M., 2001.

PAH and soot emissions from burning components of

medical waste: examination/surgical gloves and cotton pads.

Chemosphere 42, 775–783.

Lin, C.-Y., Lin, M.C., 1986. Thermal decomposition of methyl

phenyl ether in shock waves: the kinetics of phenoxy radical

reactions. J. Phys. Chem. 90, 425–431.

Longwell, J.P., 1982. The formation of polycyclic aromatic

hydrocarbons by combustion. Proc. Combust. Inst. 19,

1339–1350.

Mitchell, R.E., Sarofim, A.F., Clomburg, L.A., 1980. Experi-

mental and numerical investigation of confined laminar

diffusion flames. Combust. Flame 37, 227–244.

Panagiotou, T., Levendis, Y.A., 1994. A study on the combus-

tion characteristics of PVC, polystyrene, polyethylene, and

polypropylene particles under high temperature. Combust.

Flame 99, 53–74.

Panagiotou, T., Levendis, Y.A., Delichatsios, M.D., 1994.

Combustion behavior of polystyrene particles of various

degrees of crosslinking and styrene monomer droplets.

Combust. Sci. Technol. 103 (1–6), 63–84.

Panagiotou, T., Levendis, Y.A., Carlson, J.B., Dunayevskiy,

Y.M., Vouros, P., 1996a. Aromatic hydrocarbon emissions

from burning poly(styrene), poly(ethylene) and PVC parti-

cles at high temperatures. Combust. Sci. Technol. 116–117

(1–6), 91–128.

Panagiotou, T., Levendis, Y.A., Carlson, J.B., Vouros, P.,

1996b. The effect of the bulk equivalence ratio on the PAH

emissions from the combustion of PVC, polystyrene, and

polyethylene. Proc. Combust. Inst. 26, 2421–2430.

Panagiotou, T., Levendis, Y.A., 1998. Observations on the

combustion of polymers plastics: from single particles to

groups of particles. Combust. Sci. Technol. 137, 91–128.

Richter, H., Grieco, W.J., Howard, J.B., 1999. Formation

mechanism of polycyclic aromatic hydrocarbons and fulle-

renes in premixed benzene flames. Combust. Flame 119, 1–

22.

Richter, H., Benish, T.G., Mazyar, O.A., Green, W.H.,

Howard, J.B., 2000a. Formation of polycyclic aromatic

hydrocarbons and their radicals in a nearly sooting

premixed benzene flame. Proc. Combust. Inst. 28, 118–

129.

Richter, H., Benish, T.G., Ayala, F., Howard, J.B., 2000b.

Kinetic modeling of the formation of polycyclic aromatic

hydrocarbons. Am. Chem. Soc., Div. Fuel Chem. Preprints

45, 273–277.

Shemwell, B.E., Levendis, Y.A., 2000. Particulates generated

from combustion of polymers (plastics). J. Air Waste

Manage. Assoc. 50, 94–102.

Wang, J., Levendis, Y.A., Richter, H., Howard, J.B., Carlson,

J., 2001. Polycyclic aromatic hydrocarbon and particulate

emissions from two-stage combustion of polystyrene: the

effect of the primary furnace temperature. Environ. Sci.

Technol. 35, 3541–3552.

Wang, J., Richter, H., Howard, J.B., Levendis, Y.A., Carlson,

J., 2002. Polycyclic aromatic hydrocarbon and particulate

emissions from two-stage combustion of polystyrene: the

effects of the secondary furnace (afterburner) temperature

and soot filtration. Environ. Sci. Technol. 36, 797–808.

Wheatley, L., Levendis, Y.A., Vouros, P., 1993. Exploratory

study on the combustion and PAH emissions of selected

C. Westblad et al. / Chemosphere 49 (2002) 395–412 411

municipal waste plastics. Environ. Sci. Technol. 27, 2885–

2895.

Williams, P.T., Williams, E.A., 1999. Interaction of plastics in

mixed-plastics pyrolysis. Energy Fuels 13, 188–196.

You, J.H., Chiang, P.C., 1998. Polycyclic aromatic hydrocar-

bons and mutagenicity of soot particulates in air emissions

from two-stage incineration of polystyrene. Proc. Combust.

Inst. 27, 1343–1349.

412 C. Westblad et al. / Chemosphere 49 (2002) 395–412