See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/12635834

Particulates Generated from Combustion of Polymers (Plastics)

Article  in  Journal of the Air & Waste Management Association (1995) · February 2000
DOI: 10.1080/10473289.2000.10463994 · Source: PubMed

CITATIONS READS

34 228

2 authors, including:

Yiannis Levendis
Northeastern University
219 PUBLICATIONS   3,519 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Combustion View project

Assessment of blends of Hydrocarbon/CO2 as alternative natural refrigerants View project

All content following this page was uploaded by Yiannis Levendis on 04 February 2016.

The user has requested enhancement of the downloaded file.

 Journal of the Air & Waste Management Association

ISSN: 1047-3289 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/uawm20

Particulates Generated from Combustion of

Polymers (Plastics)

Brooke E. Shemwell & Yiannis A. Levendis

To cite this article: Brooke E. Shemwell & Yiannis A. Levendis (2000) Particulates Generated
from Combustion of Polymers (Plastics), Journal of the Air & Waste Management Association,
50:1, 94-102, DOI: 10.1080/10473289.2000.10463994

To link to this article: http://dx.doi.org/10.1080/10473289.2000.10463994

Published online: 27 Dec 2011.

Submit your article to this journal

Article views: 209

View related articles

Citing articles: 16 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=uawm20

Download by: [46.105.152.68] Date: 29 January 2016, At: 20:27

 Shemwell andPAPER
TECHNICAL Levendis
Particulates Generated from Combustion of Polymers (Plastics)
Brooke E. Shemwell and Yiannis A. Levendis
Department of Mechanical, Industrial, and Manufacturing Engineering, Northeastern University,
Boston, Massachusetts
ABSTRACT
This is an experimental study on the characterization of
particulate (soot) emissions from burning polymers. Emis-
sions of polystyrene (PS), polyethylene (PE), polypropy-
lene (PP), polymethyl methacrylate (PMMA), and
polyvinyl chloride (PVC) plastics were studied. Combus-
tion took place in a laboratory-scale, electrically heated,
Downloaded by [46.105.152.68] at 20:27 29 January 2016
drop-tube furnace at temperatures of 1300 and 1500 K, in
air. The nominal bulk (global) equivalence ratio, φ, was
varied in the range of 0.5–1.5, and the gas residence time
in the nearly isothermal radiation zone of the furnace was
≈1 sec. The particulate emissions were size-classified at
the exit of the furnace, using a multi-stage inertial par-
ticle impactor. Results showed that both the yields and
the size distributions of the emitted soot were remark-
ably different for the five plastics burned. Soot yields in-
creased with an increasing bulk equivalence ratio.
Combustion of PS yielded the highest amounts of soot
(most highly agglomerated), several times more than the
rest of the polymers. More soot was emitted from PS at
1500 than at 1300 K. Substantial amounts of soot agglom-
erates were larger than 9 µm. At 1500 and 1300 K, 35 and
29% of the soot mass, respectively, was PM2 (2 µm or
smaller). Emissions from PE and PP were remarkably simi-
lar to each other. These polymers produced very low emis-
sions at φ ≤ 0.5, but emissions increased drastically with
IMPLICATIONS
Soot emissions (both the primary particles and the ag-
glomerates) from the combustion of PS, PE, PP, PMMA,
and PVC plastics were found to be distinctly different. To
minimize soot emissions, it is imperative to burn PE and
PP in strongly fuel-lean conditions, that is, with plenty of
excess air and effective mixing. This also aids the com-
bustion of the other polymers, but does not appear to be
as imperative. The aromatic structure of PS causes the
most soot emissions, but the soot is highly agglomerated
and should be easily captured with fabric filters. Capture
of the fine (mildly agglomerated) soot of PP, PE, and
PMMA, however, may prove more difficult. Generally, soot
emissions from PVC appeared to be less of a problem,
especially the PM2 soot emissions, which were found to
be comparatively low.
94 Journal of the Air & Waste Management Association
ISSN 1047-3289 J. Air & Waste Manage. Assoc. 50:94-102
Copyright 2000 Air & Waste Management Association
φ, and most of the soot was very fine (70–97% of the mass
was PM2, depending on φ).
Emissions from the combustion of PMMA were com-
paratively low and were the least influenced by the bulk
φ, and 79–95% of the emissions were PM2. Combustion
of PVC yielded the lowest amounts of soot; moreover, only
13–34% of the mass was PM2. On a comparative basis, at
1500 K, the following ranges of particulate yields were
PM2: 19–75 mg/g of PS, 8–36 mg/g of PE, 1.5–47 mg/g of
PP, 11–20 mg/g of PMMA, and 2–8 mg/g of PVC, depend-
ing on φ. These comparative results demonstrate that PS
produces the highest amounts of fine particulates, fol-
lowed by PP, PE, and PMMA, and then PVC. Burning these
materials with excess oxygen drastically reduces the par-
ticulate emissions of PE and PP, substantially reduces those
of PS, and mildly reduces those of PMMA and PVC.
INTRODUCTION
This laboratory investigation examined soot emissions
from polymer combustion. Burning of these polymers is
of interest to fire safety sciences and to waste-to-energy
combustion, since most plastics are disposable and end
up in the municipal waste stream. The polymers investi-
gated were polyethylene (PE), polystyrene (PS), polypro-
pylene (PP), polyvinyl chloride (PVC), and polymethyl
methacrylate (PMMA). The first four of these polymers
are currently discarded in the United States at estimated
annual rates of 10, 2.7, 1.1, and 0.5 million tons, respec-
tively.1 Most applications for these polymers are in dis-
posable goods, such as packaging for food, detergents,
motor oil, plastic cutlery, film for grocery and trash bags,
cosmetics, other consumer products, and so forth. Appli-
cations are also found in durable goods, such as cables
and pipes. PMMA, produced at an estimated 0.2 million
tons per year in the United States, 2 is mainly used as glaz-
ing acrylic sheet for various applications ranging from win-
dow panels for stadiums to airplane windshields. It is also
used in adhesives. These plastics do not rust, dissolve, or
biodegrade. After having been used only one time, they
can remain a problem for centuries. Therefore, they cre-
ate serious landfill concerns and serious environmental
problems. Combustion in waste-to-energy plants is a
Volume 50 January 2000

viable option for disposing of plastics, along with other
components of the municipal waste stream.
Combustion not only minimizes the volume of the
waste stream but can also be used to produce useful elec-
tricity. Plastics are excellent fuels, since they have a high
energy content. The products of their combustion, how-
ever, may contain gaseous and particulate pollutants. Gas-
eous pollutants include oxides of nitrogen (NOx), CO,
and unburned hydrocarbons (such as polynuclear aro-
matic hydrocarbons [PAHs], aliphatics, and so forth); par-
ticulate pollutants are predominantly soot with
condensed PAHs. Particulates are recognized as hazard-
ous pollutants because they may be small enough to be
deposited deep into the lungs and cause respiratory and
cardiovascular ailments. Moreover, some of the con-
densed PAHs are known to be carcinogenic.3 Studies have
shown that not only the PAHs, but also the carbon core
Downloaded by [46.105.152.68] at 20:27 29 January 2016
of the particulates, may be carcinogenic.4
The Clean Air Act seeks to reduce all microscopic air-
borne particles with diameters of less than 10 µm (PM10);
however, the regulatory standard set to enforce the law
tends to ignore the smallest and most dangerous particles,
less than 2 µm (PM2). 3 Although many advanced particu-
late removal devices may be employed to curtail the re-
lease of particulates into the environment (fibrous bags,
electrostatic precipitators, and so forth), their soot removal
efficiency is typically low for PM2 particles.3,6 Therefore, it
is essential to minimize the formation and/or control the
emissions of soot during fuel combustion.
Limited work in combustion and emissions of plastics,
at conditions relevant to power generation boilers, has been
reported in the literature. Previous work in this laboratory
has examined the combustion characteristics of the afore-
mentioned plastics in pulverized form. The combustion of
single particles7-9 or of groups of particles10 were monitored,
and flame sizes, temperatures, duration, and interparticle
flame interactions were recorded. Each of the plastics in-
vestigated was found to exhibit distinct combustion char-
acteristics. The gas-phase emissions from the combustion of
streams of particles were also measured, including PAH10-13
and CO, CO2, nitric oxide (NO), nitrogen dioxide (NO2),
and sulfur dioxide (SO2) emissions.14 This study comple-
ments the above investigations by measuring the particu-
late emissions from the same fuels, burning them in the
same furnace and under the same combustion conditions.
The goal was to compare both the magnitude of such emis-
sions from the five commonly used plastics and the par-
ticle size distributions. The morphology of the soot
agglomerates and the agglomeration tendencies were also
examined. On the basis of this limited data, some sugges-
tions are made for minimizing particulate emissions. This
is important in order to minimize the release of the micro-
scopic airborne particles most dangerous to humans.
Volume 50 January 2000
Shemwell and Levendis
REVIEW OF LITERATURE
Previous research has shown that plastics actually increase
the combustion efficiency of the municipal waste mix-
ture, as well as the thermal output of boilers. Studies have
also shown that increasing the percent of plastics in the
municipal waste fed to boilers reduces the amount of
airborne pollution from unburned particles and com-
pounds.15 Lea16 also showed that combustion of waste
plastics maximizes energy recovery (32 GJ/ton) through
waste-to-energy (WTE) conversion and provides substan-
tial volume reduction. Other studies have shown that
even combustion of PVC can be safe, if appropriate pol-
lution control equipment is used.5,15,17,18
The most important physical parameter that affects
soot formation in diffusion flames is flame temperature.
Reducing the flame temperature in diffusion flames de-
creases the pyrolysis rate of gas-phase hydrocarbons and
therefore decreases the rate of soot precursor formation.3
Regardless of the fuel being burned, soot inception in
diffusion flames occurs around 1350 K (±35 K).19 The
quantity of generated soot increases with temperature,
reaches a maximum at 1900–2300 K, and then decreases
again.17,19 After soot is formed on the fuel side of the re-
action zone, it may burn in the outer oxidation zone of
a diffusion flame or it may not burn completely if its
temperature drops below 1300 K.19 In the latter case,
smoke (for example, an aerosol of soot) will be emitted
from the diffusion flame. According to Chung and
Tsang,20 two approaches may be used to reduce soot for-
mation in polymer flames:
(1) lowering the rate of polymer devolatilization by
decreasing the effective fuel/air ratio in the
flame; and
(2) affecting important chemical and physical soot
formation steps in the gas phase by introducing
metal additives (soot suppressors). Fuel structure
has a significant effect on the sooting tendency
of diffusion flames.
The chemical structure of a fuel plays a role in its ten-
dency to form particulates;19 therefore, fundamental knowl-
edge of a fuel’s pyrolysis chemistry will allow prediction of
its relative tendency to soot. In diffusion flames, the ten-
dency to soot among fuel types is in the following order:
aromatics alkynes > alkenes > alkanes.19 In order to prop-
erly compare the effect of a fuel structure on sooting ten-
dency, the flame temperature must be kept constant. The
relative tendency of one fuel to soot compared to another
may primarily arise from a difference in the initial rate of
formation of aromatic rings. Formation of the initial rings
controls the rate of incipient soot formation, which, in turn,
determines the soot-volume fraction.
A study by Wu et al.1 on the pyrolysis kinetics of
plastic mixtures of municipal solid waste (MSW)
Journal of the Air & Waste Management Association 95
 Shemwell and Levendis
concluded that PE, PP, and PS each burn in a one-stage
reaction (breaking of carbon-carbon bonds), while PVC
burns in two stages (breaking of carbon-carbon and
chlorine-carbon bonds). Combustion of PS involves
two major steps: first, the PS melts and devolatilizes
to produce light hydrocarbon species; second, such
species vaporize into the gas phase and undergo com-
bustion reactions with oxygen. 20 The rate of
devolatilization is determined by the rate of heat trans-
fer to the PS. The rate of combustion is mainly con-
trolled by flame temperature and available oxygen
concentration. Previous combustion studies7-9 on single
PS particles in drop tube furnaces showed that particles
burned with bright spherical flame envelopes through-
out their combustion. Flames were associated with
long, bright wakes. After extinction of the volatile flame,
a string of (agglomerated) soot remained at the location
Downloaded by [46.105.152.68] at 20:27 29 January 2016
of the luminous wake and did not burn readily.8,9 PVC
particles burned with bright spheroidal flames; however,
no wakes were observed.9 Upon extinction of the vola-
tile flames, combustion of char ( ≈ 9–10% of the initial
mass) took place. Little or no soot appeared to remain
after the extinction of the flames. Single-particle com-
bustion of PE and PP occurred with small and dim
flames. Maximum envelope flame temperatures in
single particle combustion experiments, at a T gas of
1400 K, were about 2000 K for PS, 2100 K for PVC, and
2400 K for PE.
When burned in groups, the particles of these poly-
mers exhibited distinct behaviors.10 PVC particles burned
with individual envelope flames, even at high particle
number densities. PS particles in groups burned with in-
tensified wake formation, and they formed common
flames only at high particle number densities. Such
flames were very sooty. PE and PP formed group flames
(puffs) readily, even at low bulk equivalence ratios, when-
ever local particle number densities exceeded a minimum
threshold. Such puffs appeared to contain a cloud of lu-
minous, finely distributed soot. Groups of PMMA par-
ticles generally formed common flames, which appeared
thicker and with less defined boundaries than those of
PE and PS, possibly because of the high oxygen content
of PMMA, which may support combustion in the inte-
rior of the flame. In combustion of particles in groups,
the recorded average flame temperatures were below
those measured for envelope flames surrounding single
particles; for example, when only one particle was in the
furnace. Furthermore, the recorded flame temperatures
decreased as the polymer particle number density in the
furnace increased.10
Emissions from the combustion of the aforementioned
polymers were studied in this laboratory as functions of
bulk equivalence ratio (particle number density) and
96 Journal of the Air & Waste Management Association
temperature. Results on the PAH emissions have been re-
ported elsewhere,11-13 and those on CO, CO2, NOx, and SO2
emissions are reported in Courtemanche and Levendis.14
This study complements the above by assessing the par-
ticulate emissions from the same polymers burned in the
same apparatus under similar conditions. The bulk equiva-
lence ratio was the main parameter, which was varied, ex-
tending in both the fuel-lean and fuel-rich domain. A
previous study20 showed that soot yield (mass soot/mass
fuel burned) can be greatly reduced by adding small
amounts of air into the pyrolysis zone of the flame (hence
creating fuel-lean conditions). The addition of air increased
combustion reaction rates and decreased soot precursor
formation rates. The soot yield in that study was found to
be significantly reduced as the air flow rate was increased.
FUELS AND THEIR COMPOSITION
Powders of PS were obtained from Duke Scientific. PE,
PMMA, and PVC were obtained from Aldrich. PP powder
was previously produced in the laboratory from cryogenic
grinding of larger pellets. The chemical compositions of
the plastics and the sizes of the particles used in these
experiments are given in Table 1.
EXPERIMENTAL APPARATUS AND PROCEDURE
Particle combustion took place in a sealed, electrically
heated, laminar flow, drop-tube furnace (see Figure 1).
Three grams of powder were placed in a test tube. The
test tube was advanced by a syringe pump, and particles
were fluidized in a stream of air aided by a vibration sys-
tem. The particles flowed into the furnace through a
water-cooled injector. The steady-flow cylindrical streams
of particles burned in a 20-cm long, 3.5-cm wide alu-
mina-tube isothermal zone. It is important that the flow
through the furnace is laminar, because the most funda-
mental knowledge about the condition of a fuel to soot
can be obtained only from laminar flame systems.19 The
Reynolds numbers in the furnace were 12 and 8.5 at gas
temperatures of 1300 and 1500 K, respectively.
Furnace wall temperatures were measured by type S
thermocouples.21 Gas temperatures were measured by a
suction thermometer (aspirated shielded thermocouple)
and were ≈ 50 °C lower than the wall temperature.21-23
The axial temperature profile in the radiation zone of the
furnace was found to be isothermal along the centerline,
except for approximately the length of the first centime-
ter. (See the region between the top and bottom insula-
tion layers of the furnace depicted in Figure 1). For a
complete axial profile of this furnace’s centerline tempera-
ture, see Atal et al. 23 (Figure 2).
At the furnace exit, the combustion effluent was passed
through a glass accelerator tube, mixed with dilution air,
and introduced to an Anderson eight-stage inertial particle
Volume 50 January 2000
 Shemwell and Levendis

Table 1. Fuel chemical composition and properties. conducted at nominal equivalence ratios of
0.5, 0.7, and 1.0, at a gas temperature of 1300
Polymer PS PE PP PMMA PVC
Chemical Structure C8H8 CH2 C3H6 C5H8O2 C2H3Cl K. Because of its high chlorine content (see
Table 1), PVC at φ = 0.9 resulted in a high
Particle size (• m) 212–250 250–297 250–297 180–212 212–250 particle number density, comparable to that
Fix. Carbon (%) 1 0 0 0 9
Volatiles (%) 99 100 100 100 91 of the other polymers at the fuel-rich condi-
Ash (%) 0 0 0 0 1 tions (φ = 1.5). Furthermore, the laboratory
Carbon (%) 92 86 86 60 38 experimental setup did not allow for higher
Hydrogen (%) 8 14 14 8 5
Sulfur (%) 0.04 0 0 0 0
temperature combustion of PVC, as the injec-
Nitrogen (%) 0 0 0 0 0 tor was often plugged by molten PVC. The ac-
Oxygen (%) 0 0 0 32 0 tual bulk equivalence ratios, as measured by
Chlorine (%) 0 0 0 0 56
monitoring the flow rate of air and fuel pow-
Heating value (MJ/kg) 44.5 40.5 43.4 25.8 19.2
der, are recorded on the plots of Figure 2. The
flow rate of air in the furnace was 2.9 and 3.3
impactor (Figure 1). The impactor separated the particles lpm at 1500 and 1300 K, respectively, to provide a constant
by size, in eight ranges between 9 and 0.4 µm. Particles residence time in the isothermal cavity of 1 sec.
smaller than 0.4 µm were collected by a fiberglass filter.
Downloaded by [46.105.152.68] at 20:27 29 January 2016

Each collection plate and the filter of the particle impactor
were weighed before and after the experiment to determine
the amount of soot collected at each stage.
PS, PE, PP, and PMMA were all burned at 1500 K at
three nominal bulk equivalence ratios: 0.5, 1.0, and 1.5.
Please note that the bulk (global) equivalence ratio is de-
fined as
φ = (mfuel/mair) actual / (mfuel mair)stoichiometric
The bulk equivalence ratio, φ, is equal to the inverse
of the excess air ratio (or stoichiometric ratio), λ. PS
was also burned at 1300 K to observe the effects of
temperature on the soot emissions. The PVC tests were
RESULTS AND DISCUSSION
Plots of the nominal particle size distribution of soot are
shown in Figure 2. In these graphs, the mass of particu-
lates collected on each stage per mass of fuel introduced
to the furnace is plotted against the nominal particle size
(aerodynamic diameter of soot agglomerates) range for
each stage. The total amounts of soot collected from the
combustion of each polymer on all stages for the three
equivalence ratios employed are listed in Table 2. In the
same table, the total amounts of the PM2 soot are also
listed, along with the percentage of PM2 soot. Soot depo-
sition at the flat surface of the entrance to the impactor
(see Figure 1) and on the glass tube, which connected the
furnace to the impactor, was negligibly small. The dilu-
tion air was also introduced in the glass tube. The only
notable exception was the case of PS, during combustion
of which some large (several mm long) strings of soot were
seen to deposit in the aforementioned areas. Such large
strings of soot were collected and were found to account
for 40–50 wt% of the total collected soot, depending on
the φ. The formation of such soot strings in the wake of
burning PS particles was illustrated in Figure 7 of
Panagiotou et al.8 The calculated stoichiometric air/fuel
ratios (mass of air/mass of polymer) are as follows: PS 13.1,
PE 14.7, PP 14.7, PMMA 8.2, and PVC 5.2.

Figure 1. General schematic of the laminar-flow, drop-tube furnace
and the inertial impactor used for combustion of cylindrical streams of
polymer particles and for measuring their particulate emissions.
PS
For PS at 1500 K, large amounts of soot were collected at
the 9 and 1–5 µm stages (see Figure 2a, top row). This
trend was consistent at both gas temperatures tested, 1500
and 1300 K. The amount of PM2 soot ranged from 16 to
35% of the total collected soot mass. The PM2 fraction
was the largest at the lowest φ of 0.6, and the smallest at
the highest φ of 1.6. At 1300 K, PS produced somewhat less
soot than at 1500 K, but the particle size distribution pat-
tern was similar to that at the higher temperature, with

Volume 50 January 2000 Journal of the Air & Waste Management Association 97
 Shemwell and Levendis
Downloaded by [46.105.152.68] at 20:27 29 January 2016

Figure 2. Size distributions of particulate emissions from the combustion of (a) PS at 1500 K, (b) PS at 1300 K, (c) PE at 1500 K, (d) PP at 1500 K,
(e) PMMA at 1500 K, and (f) PVC at 1300 K. Particulate emissions were collected with an inertial impactor and assessed gravimetrically. Conditions
varied from fuel-lean to fuel-rich; the bulk (global) equivalence ratio for each case is included in the plots.

98 Journal of the Air & Waste Management Association Volume 50 January 2000
 Shemwell and Levendis

Table 2. Cumulative soot yields and PM soot yields.

2
PMMA
Polymer total soot (mg/g) soot < 2·m (mg/g) < 2·m/total soot PMMA yielded the least amount of soot overall, aside from
PS at 1500 K, φ =0.6 91.3 32.5 35%
PVC. It produced more soot than PE and PP at the lowest
PS at 1500 K, φ =0.9 130.3 46.4 36% φ cases, but produced much less soot than PE and PP at φ
PS at 1500 K, φ =1.6 221.8 74.5 34% ≥ 0.9. Of the total amount of soot produced by PMMA,
PS at 1300 K, φ =0.5 105.1 18.9 18% the vast majority (79–95 %) was less than 2 µm, at all φs.
PS at 1300 K, φ =1.1 119.1 18.7 16% Overall, the sooting tendency of PMMA was not drasti-
PS at 1300 K, φ =1.8 244.4 71.4 29% cally affected by the equivalence ratio, possibly because
PE at 1500 K, φ =0.5 8.6 8.4 97% of this polymer’s oxygen content. These measurements
PE at 1500 K, φ =1.0 37.1 26.3 71% were consistent with photographic observations, which
PE at 1500 K, φ =1.5 43.2 35.7 83% showed formation of common flames (at all φs examined)
PP at 1500 K, φ =0.5 2.5 1.5 61%
containing apparently fine soot.
PP at 1500 K, φ =1.0 24.9 23.0 92%
PP at 1500 K, φ =1.4 49.6 47.2 95%
PVC
PMMA at 1500 K, φ =0.6 12.3 11.2 90%
PMMA at 1500 K, φ =0.9 11.9 11.4 95%
Generally, PVC produced the least amount of PM2 soot
PMMA at 1500 K, φ =1.4 26.5 20.8 79% of all the polymers. Most of the soot produced by com-
Downloaded by [46.105.152.68] at 20:27 29 January 2016

PVC at 1500 K, φ =0.5 16.5 2.1 13% bustion of PVC had aerodynamic particle sizes larger than
PVC at 1500 K, φ =0.8 20.2 4.7 23% 2 µm, at most equivalence ratios (see Figure 2b, bottom
PVC at 1500 K, φ =0.9 23.0 7.8 34% row). The particulate yields generally increased with bulk
equivalence ratio. The fact that the particulate emissions
from combustion of PVC were low, in comparison to
other polymers, was also suggested by photographic ob-
peaks around 1–2 and 9 µm. Generally, at high φs, larger servations of burning clouds of PVC particles. Even at
amounts of soot agglomerates with bigger aerodynamic high particle loadings in the furnace (stoichiometric or
diameters were collected. This is consistent with the pho- even higher φ), PVC particles burned with individual en-
tographic observations of Panagiotou and Levendis, velope flames, resisting the formation of large common
which showed that the tendency for longer wake forma- flames, which locally create oxygen-deficient environ-
tion in particle flames was more pronounced at high PS ments and release soot.9 This behavior was explained on
particle number densities in the furnace. Flame wakes, the basis of the release of hydrochloric acid (HCl) from
upon extinction, result in highly agglomerated, string- the combustion of PVC, which may act as a fire retar-
like soot in the furnace.8 dant at the flame boundaries.

PE SEM Photographs of Collected Soot

Under fuel-lean conditions, φ = 0.5, PE produced very little
soot; however, as much as 97% of the total soot was PM2.
The total amount of emitted soot increased with φ, and at
all φs, the vast majority of soot was found to be PM2 in size.
Some larger soot was generated as φ increased, but gener-
ally, its level remained comparatively low. The observations
that (1) the overall soot emissions were very low at strongly
fuel-lean conditions, and (2) at all equivalence ratios the
vast majority of the soot emissions were sub-µ, are consis-
tent with combustion observations. Such photographic ob-
servations10 revealed the formation of common particle
flames at medium to high φs (φ > 0.6). These common flames
appeared to contain well-dispersed fine soot.
PP
The trends in soot emissions from PP were similar to
those of PE. Again, sub-µ soot was predominant at all
φs, and total emissions were surprisingly low at the
most fuel-lean conditions herein, φ = 0.5.
The microscopic structure of the particulate emissions
from the combustion of each polymer was observed with
a scanning electron microscope. High-magnification pho-
tographs were taken of particles collected on the 0.65–
1.1-µm stage, for all materials burned at the highest φ case
(see Figure 3). Apparently, the soot is deposited on this
stage as agglomerates, with aerodynamic diameters in the
aforementioned range, since individual spherule diam-
eters are seen to be much smaller. At first glance, the mi-
crostructure of the soot seems to be the same for all the
materials burned. The soot particles appear to have an
acinform, or grape-like, structure—small spheres that grow
upon each other. More careful examination reveals the
following differences. The soot from PE and PP bear the
most similarities. The size distribution of their spherules
seems to be rather uniform, in the range of 100–200 nm
(0.1–0.2 µm). The appearance of the soot aggregates gives
the impression that their surfaces are smooth. The soot
spherules from the combustion of PS appear to exhibit a

Volume 50 January 2000 Journal of the Air & Waste Management Association 99
 Shemwell and Levendis
Downloaded by [46.105.152.68] at 20:27 29 January 2016

Figure 3. SEM photographs of particulate emissions from the combustion of (a) PS at 1500 K, (b) PS at 1300 K, (c) PE at 1500 K, (d) PP at 1500 K,
(e) PVC at 1300 K, and (f) PMMA at 1500 K. These particulates were collected at the 0.65- to 1.1-µm stage of the inertial impactor. With the
exception of PVC, combustion conditions were fuel-rich.

large size variation, from 200 nm down to a fraction of
that size (50 nm). The soot of PVC exhibited a rather uni-
form distribution in the range of smaller sizes (50–100
nm). The soot of PMMA exhibited a wide distribution,
but in small sizes (100 nm). Its aggregates give the im-
pression of rough surfaces.
Comparison with Emissions of PAHs
In this section, the emissions of soot are contrasted to
the emissions of PAHs recorded earlier, 13 from the
combustion of the same polymers under similar con-
ditions. This comparison is of technical interest, since
condensed PAHs are regarded, to a certain extent, as
precursors to soot. 24
There are many similarities in the PAH and soot
emission yields from the polymers. PS produced more
PAHs and more soot than the other polymers, under
all equivalence ratios tested. PE produced smaller
amounts of both soot and PAHs at fuel-lean and stoichio-
metric conditions than PS. At fuel-rich equivalence ratios

100 Journal of the Air & Waste Management Association Volume 50 January 2000

the total PAH yields from the combustion of PS and PE
were comparable, but the soot yields from PS were much
higher and the soot was much more agglomerated. Scru-
tiny of the individual PAH compound yields revealed that
PE combustion emitted preferentially much larger amounts
of lighter 2-ring PAHs (such as naphthalene and
acenaphthalene) than PS combustion.13 This is consistent
with an understanding of PAH formation from PS and PE.
In PS combustion, formation of PAHs may be attributed
to a direct (and hence more expedient) root of intermo-
lecular condensation reactions of its aromatic rings. The
PS pyrolyzates readily form multi-ring PAH compounds
even in single particle flames. Such PAHs are, eventually,
more likely to form soot, which is difficult to burn even
at very fuel-lean conditions.9
In PE combustion, PAH formation may be attributed
to less prompt intermolecular aliphatic fragmentation
Downloaded by [46.105.152.68] at 20:27 29 January 2016
reactions.13 Such intermolecular reactions are less likely to
happen at fuel-lean conditions, where the particle mass
loadings are low. If aromatic rings do form therein, they
are likely to react with the plentiful oxygen and burn. PAHs
are much more likely to form (and survive) at stoichiomet-
ric/fuel-rich conditions where a large concentration of 2-
ring PAH compounds is detected. But even at the fuel-rich
conditions, formation of highly condensed PAHs and then
soot from PE will be less likely than from PS. The fact that
the soot of PS is more agglomerated than that of PE is mostly
attributed to the physical structure of the diffusion flames
of the former and their strong tendency to form flame
wakes. In the case of PVC, both the PAH emissions and the
soot yields are more than an order of magnitude lower than
those of PS as well as those of PE (the latter is the case at
fuel-rich conditions). As argued earlier,13 the formation of
PAHs from PVC is likely to happen through an intramo-
lecular cyclization route. Most PVC particles burn with in-
tense individual envelope flames, which are likely to be
partially mixed with oxygen because of the observed igni-
tion delay caused by the release of HCl.9 In such flames,
the PAHs may find sufficient oxygen to burn, and soot for-
mation is thus limited.
SUMMARY AND CONCLUSIONS
Combustion experiments with the five most common mu-
nicipal waste plastics were conducted in the laboratory. Pow-
ders of PS, PE, PP, PMMA, and PVC plastics were burned in
a laminar-flow, drop-tube furnace. Gas temperatures were
in the range of 1300 and 1500 K, and residence times ≈ 1
sec. The bulk equivalence ratio, φ, (proportional to the par-
ticle number density) varied from 0.5 to1.5.
SEM observations revealed that the particulates col-
lected in the impactor stages were in the form of agglom-
erates, consisting of small soot spherules. Soots from PP
and PE consisted of rather uniform size spherules
Volume 50 January 2000
Shemwell and Levendis
100–200 nm; soots of PVC consisted of uniform spherules,
but were smaller (100 µm). The soots from PS and PMMA
consisted of wide distributions of spherules; the former
were larger in size than the latter.
Soot yields from combustion of PS were by far the larg-
est in all size ranges and at all φs. This indicates the forma-
tion of a wide variety of agglomerate chain fragments.
Emissions from combustion of PE and PP were strikingly
similar to each other in particulate size distributions (both
primary particles and agglomerates). Very low emissions
were observed at substantially fuel-lean conditions (φ = 0.5),
when PE and PP particles burned individually (with no
particle-to-particle interactions). However, as φ increased, a
strong tendency was observed for group combustion (puffs),
which resulted in dramatically higher soot emissions. The
vast majority of the soot was fine (small agglomerates). Simi-
larly, the soot of PMMA was also fine; however, its concen-
tration increased gradually with φ. Soot yields from
combustion of PVC were low and exhibited a wide size dis-
tribution, typically favoring the larger particulate (agglom-
erate) sizes. Thus, the soots (both the primary particles and
the agglomerates) from the five plastics burned were found
to have distinctive signatures.
To minimize soot emissions, it is imperative to burn
PE and PP in strongly fuel-lean conditions; for example,
with plenty of excess air and effective mixing. This also
aids the combustion of the other polymers but did not ap-
pear to be as imperative. The aromatic structure of PS causes
the most soot emissions, but the soot is highly agglomer-
ated and should be easily captured with fabric filters. Cap-
ture of the fine (mildly agglomerated) soot of PP, PE, and
PMMA, however, may prove more difficult. Generally, soot
emissions from PVC are less of a problem, especially the
PM2 soot emissions, which were found to be comparatively
very low.
The drop-tube furnace that was used herein was se-
lected to facilitate these rather fundamental studies, where
combustion observations—temperature profiles, burnout
time histories, flame diameters, soot volume fractions,
combustion rates, and so forth7-10— were related to emis-
sions11-14 and in the work herein. Commercial combus-
tion units come in many different designs and
configurations, and their flow field is usually turbulent.
Additional work would need to be done, as the opportu-
nity arises, to identify the exact relation of a particular
commercial combustor to the characteristics of the appa-
ratus in our laboratory. It is the authors’ opinion, how-
ever, that some of the relative emission trends identified
in this study may hold true for such systems.
ACKNOWLEDGMENTS
This research was partially supported by EPA grant
# R-819245-01-0 and NSF grant BES-9705703. The
Journal of the Air & Waste Management Association 101
 Shemwell and Levendis

authors would like to acknowledge technical help from 15. Coghlan, A. “Can burning plastics be good for the world? New Scien-
tist 1994, 143, 20.
Bonnie Courtemanche. 16. Lea, W.R. “Plastic incineration versus recycling: a comparison of en-
ergy and landfill cost savings,” J. Hazard. Mater. 1996, 47, 295-302.
17. “Combustion study shows no PVC link with toxicity,” Modern Plas-
REFERENCES tics 1987, 64, 12-13.
18. “Incinerator operator: vinyl is harmless,” Modern Plastics 1989, 66,
1. Wu, C.H.; Chang, C.Y.; Hor, J.L. “On the thermal treatment of plastic
19-20.
mixtures in MSW: pyrolysis kinetics,” Waste Manage. 1993, 13, 221-
19. Glassman, I. Soot Formation in Combustion Processes. In Proceedings
235.
of Twenty-Second Symposium (International) on Combustion; The Com-
2. Seymour, R.B. Engineering Polymer Sourcebook; McGraw-Hill: New York,
bustion Institute, Pittsburgh, PA, 1982; pp 295-311.
1990.
20. Chung, S.L.; Tsang, S.M. “Soot control during the combustion of
3. Chung, S.L.; Lai, N.L. “Suppression of soot by metal additives during
polystyrene,” J. Air & Waste Manage. Assoc. 1991, 41(6), 821-826.
the combustion of polystyrene,” J. Air & Waste Manage. Assoc. 1992,
21. Cumper, J.G.; Levendis, Y.A.; Metghalchi, M. Characterization of the
42, 1082-1088.
Environment in a Laminar Flow Furnace with Applications to the Com-
4. Mayer, A. Particle Size Distribution Downstream of Traps of Different
bustion of Coal and Coal-Water Slurries; Vol. 148; American Society of
Design; SAE Paper No. 950372; Society of Automotive Engineers:
Mechanical Engineers Publication, Heat Transfer Division, 1998, pp
Warrendale, PA, 1995.
89-96.
5. Eller, R. “Solid waste management in Japan,” Waste Age 1992, 23, 53-
22. Atal, A.; Steciak, J.; Levendis, Y.A. “Combustion and SO2-NOx emis-
64.
sions of bituminous coal particles treated with CMA,” Fuel 1995, 74
6. Flagan, R.C.; Seinfeld, J.H. Fundamentals of Air Pollution Engineering;
(4), 495-506.
Prentice-Hall: Englewood Cliffs, NJ, 1998.
23. Steciak, J.; Levendis, Y.A.; Wise, D.L.; Simons, G. “Dual SO2-NOx con-
7. Panagiotou, T.; Levendis, Y.A. “A study on the combustion character-
centration reduction by calcium salts of carboxylic acids,” J. Environ.
istics of PVC, poly(styrene), poly(ethylene) and poly(propylene) par-
Eng. (N.Y) 1995, 121 (8), 595-604.
ticles under high heating rates,” Combustion and Flame 1994a, 99(4),
24. Haynes, B.S. In Fossil Fuel Combustion — A Source Book; Bartok, W.A.;
53-74.
Sarofim, A.F., Eds.; John Wiley & Sons: New York, 1989.
8. Panagiotou, T.; Levendis, Y.A.; Delichiatsios, M. “Combustion behavior
of poly(styrene) particles of various degrees of crosslinking and sty-
rene monomer droplets,” Combust. Sci. & Tech. 1994b, 103(1-6), 63-
Downloaded by [46.105.152.68] at 20:27 29 January 2016

84.
9. Panagiotou, T.; Levendis, Y.A. “Combustion of pulverized PVC and
poly(ethylene),” Combust. Sci. & Tech. 1996a, 112, 117-140.
10. Panagiotou, T.; Levendis, Y.A. “Observations on the combustion of
polymers (plastics): from single particles to groups of particles,” Com-
bust. Sci. & Tech. 1998, 137, (1-6), 121-148.
11. Wheatley, L.; Levendis, Y.A.; Vouros, P. “An exploratory study on the
combustion and PAH emissions of selected municipal waste plastics,” About the Authors
Environ. Sci. & Tech. 1993, 27 (13), 2885-2895.
12. Panagiotou, T.; Levendis, Y.A.; Carlson, J.; Dunayefskiy, Y.; Vouros, P. Yiannis A. Levendis is a professor in the Department of Me-
“Semi-volatile organic emissions from burning poly(styrene), chanical, Industrial, and Manufacturing Engineering at North-
poly(ethylene) and PVC particles at high temperatures,” Combust. eastern University. He holds a B.S. and an M.S. in mechani-
Sci. & Tech. 1996b, 116-117 (1-6), 91-128.
13. Panagiotou, T.; Levendis, Y.A.; Carlson, J.; Vouros, P. The Effect of cal engineering from the University of Michigan and a Ph.D.
Bulk Equivalence Ratio on the PAH Emissions from the Combustion in environmental engineering from the California Institute of
of PVC, Poly(styrene) and Poly(ethylene). In Proceedings of Twenty- Technology. Brooke Shemwell is a graduate research assis-
Sixth Symposium (International) on Combustion; The Combustion In-
stitute: Pittsburgh, PA, 1996; pp 2421-2430. tant in the Department of Mechanical, Industrial, and Manu-
14. Courtemanche, B.; Levendis, Y.A. “A laboratory study on the NO, facturing Engineering at Northeastern University.
NO2, SO2, and CO emissions from the combustion of pulverized coal,
municipal waste plastics and tires,” Fuel 1997, 77(3), 183-196.

102 Journal of the Air & Waste Management Association Volume 50 January 2000

View publication stats