Fuel 193 (2017) 178–186

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Evaluating the potential of waste plastics as fuel in cement kilns using

bench-scale emissions analysis
Ebenezer A. Asamany ⇑, Mark D. Gibson, Michael J. Pegg
Department of Process Engineering and Applied Science, Dalhousie University, Halifax, Nova Scotia, Canada

a r t i c l e i n f o a b s t r a c t

Article history: In an effort to reduce emissions and lower fuel costs, cement plants have explored the use of waste mate-
Received 25 October 2016 rials to displace coal and petroleum coke. A potential candidate is non-recyclable waste plastics.
Received in revised form 5 December 2016 Knowledge of potential changes in cement kiln parameters with a change in the composition of operating
Accepted 19 December 2016
fuel prior to full-scale application is necessary for obtaining permits and for plant preparedness. Most
bench-scale performance studies have targeted the prediction of full-scale stack emissions with variable
degrees of success. In this work, a bench-scale tube furnace and a heated grid reactor (HGR) have been
Keywords:
used to compare combustion and emission changes when waste plastics displace an equivalent amount
Waste plastics
Combustion
of heat from the coal/coke blend. Trends in the results from bench-scale experiments on residual volatile
Alternative fuel organic compound (VOC) emissions matched full-scale observations on kiln thermal performance. Bench-
Coal scale particulate matter (PM) emissions indicated that no significant changes were likely in full-scale
Cement kiln stack PM emissions when using waste-derived fuel (WDF). This bench-scale fuel characterization
Emission approach has been used to study the potential impact of waste plastics on the performance of a local
cement kiln.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Approximately 2.8 million tonnes of plastic waste is disposed of
annually in Canada. Of this, an estimated 719,000 tonnes are from
residential sources and the rest are from non-residential sources
including the industrial, commercial and institutional sector which
includes grocery stores and goods manufacturing plants [1,2].
According to the Clean Foundation [3], 490 million plastic bags
are used in Atlantic Canada each year. At 5 g per bag, this repre-
sents a potential energy estimate of 101 PJ/year based on reported
calorific values for plastic film. Most of these plastics end up in
landfills since they are usually unrecyclable. Redirecting such
waste to a local cement kiln for co-processing presents a dual solu-
tion to both cost and environmental concerns.
Clinker is produced in cement kilns heated to about 1450 °C. In
rotary kilns, raw materials are introduced into the slightly elevated
end of large rotating drums inclined downward towards the burner
through which fuel is fed to supply heat for the reactions. The
blending of different fuels with coal in thermal processes such as
clinker production for economic, environmental and chemical ben-
efits has been discussed in literature [4–6]. Factors considered for
⇑ Corresponding author.
E-mail address: E.Asamany@dal.ca (E.A. Asamany).
alternative fuels for co-firing include moisture and ash content,
calorific value, volatile content, and reactivity. The alternative fuel
is also evaluated based on the potential for ambient air pollution,
slagging and fouling, and corrosion. Generally, the performance
of WDF depends upon their ability to influence the reactivity of a
coal-fuel blend, and to reduce the formation of potentially harmful
emission such as nitrogen oxides (NOx), sulphur dioxide (SO2), air-
borne PM and polychlorinated dibenzo-p-dioxins and dibenzofu-
rans (PCDD/Fs). These emissions and their environmental and
human health effects have also been discussed in the literature
[6–8]. Most cement plants undergo rigorous trials to ascertain
the suitability of a selected alternative fuel. These trials may inter-
fere with normal plant operations and need to be carried out over
long periods before definitive trends can be observed. This is due to
the complex changes that regularly occur in kiln operations. Know-
ing the composition of the potential fuel before industrial trials
does not adequately inform operators, regulators and communities
with sufficient data to manage potential changes for a given fuel.
The overall objective was to use a combination of previous stud-
ies, ultimate analysis, tube furnace emission tests and HGR com-
bustion tests to evaluate and correlate fuel properties and bench-
scale emission behavior to expected changes in full-scale cement
kiln parameters.

http://dx.doi.org/10.1016/j.fuel.2016.12.054
0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
 E.A. Asamany et al. / Fuel 193 (2017) 178–186
Nomenclature
C&D Construction and Demolition
Coal-coke
1:1 mixture of coal and petroleum coke by mass
EPS Expanded Polystyrene
ESP Electrostatic Precipitator
GCMS Gas Chromatography Mass Spectroscopy
HDPE High Density Polyethylene
HGR heated grid reactor
LDPE Low Density Polyethylene
LHW liquid hazardous waste
MRF Material Recovery Facility
MSW Municipal Solid Waste
2. Kiln emissions and thermal performance
Approximately 50–60% of the CO2 is related to calcination of
CaCO3, and the rest is as a result of fuel combustion [9,10]. SO2, car-
bon dioxide (CO2) and fuel NOx generated during combustion are
governed primarily by the sulphur, carbon and nitrogen content
of the fuel. Based on a simple stoichiometric mass balance, the
expected mass (me) of gaseous oxide (xO) with molecular mass
(MMxO) produced from the oxidation of element x (N, C, S or H)
with molecular mass (MMx) is given by:
 
n  Y x  MM xO
me ¼  Et =½kg ð1Þ
Ef  MM x
where n is the stoichiometric mole ratio of x to xO, Et is the energy
required for the process [kJ], Ef is the energy content of the given
fuel [kJ/kg], and Yx is the percent composition of x in the fuel on a
dry basis from the ultimate analysis. When heated, a solid fuel loses
moisture and then releases volatile components, which ignite and
form a flame. The extent of thermal degradation of produced VOC
such as benzene, xylene and toluene is dependent on temperature
and the retention time within this zone. At elevated temperatures
(>1000 °C) the mixture of organic vapor and air is oxidized to form
CO, CO2, NOx and H2O [11]. The more extensive the degradation, the
lesser the VOC content in the combustion products. Given the vari-
ables related to the emission of residual VOC, theoretical calcula-
tions based on stoichiometry fall short in predicting their
quantities in combustion products.
Similar to residual VOC, it is difficult to theoretically quantify
PM emissions even where the composition and stoichiometry of
combustion are well known. Changes in fuel such as using WDF
may influence PM emissions, especially PM with a median aerody-
namic diameter 6 10 lm (PM10) [12]. Total suspended particulates
are a function of the percentage of ash, or mineral matter, in the
fuel, coupled with the unburned carbon resulting from combustion
and firing method used [6]. Fuels such as waste plastics with low
ash content can therefore be blended with high-ash coals to reduce
suspended particulates. It is also expected that a higher volatile
content will reduce unburned carbon in the products of combus-
tion. Gases such as SO2, NOx, and VOC emitted from cement kilns
can transform in the atmosphere to form secondary fine airborne
particulate matter with a median aerodynamic diameter 6 2.5 lm
(PM2.5) and ozone (O3) [6,13–16]. Carrasco et al. [17] and CRA [18]
report a broad variation of expected changes in gaseous emissions
such as SO2, NOx and CO in their literature review on effects of Tire-
derived fuels on cement kiln emissions. Both groups confirm that
the variation in results show that emissions generally do not only
depend on the quantity of their parent elements in the fuel but also
on their presence in raw materials, and the efficiency of the com-
bustion process. This makes ascertaining the contribution of fuel
179
PCDD/F polychlorinated dibenzo(dioxin/furan)
PDF plastic-derived fuel
PM particulate matter
RDF refuse-derived fuel
SHW solid hazardous waste
TDT Thermal Desorption Tubes
VOC volatile organic compound
WDF waste-derived fuel
WMR Wire Mesh Reactor
combustion to gaseous and particulate emission of a cement kiln
stack by field measurements challenging. Many sources of similar
emissions are present throughout the cement making process. It
also makes the scale-up of bench-scale experiments based on
emissions unsuccessful in many cases. The work of Jeschar et al.
[7] for example showed that the challenge of replicating actual kiln
conditions on small-scale may sometimes lead to full-scale emis-
sions results which do not conform to theoretical expectations.
Despite these limitations, bench-scale experiments are useful for
the comparative analysis of materials under similar controlled con-
ditions. The degree to which full-scale predictions from bench-
scale observations are acceptable depend on (1) the reasonable
simulation of the actual kiln burning zone conditions by the smal-
ler scale test as can be obtained in a heated grid or entrained flow
reactor [19] and (2) the full-scale kiln parameters of interest.
Parameters such as flue gas flow rate and kiln gas temperature,
which are not easily masked by the presence of other materials
can be reasonably linked to properties determined from bench-
scale analysis.
Predictive methods based on flue gas flow rates have been used
successfully by Ariyaratne et al. [20] to study effects of fuel compo-
sition, moisture content, calorific value and state on actual kiln
performance. In their work the effect of substituting coal with
refuse-derived fuel (RDF), waste wood, solid hazardous waste
(SHW) and liquid hazardous waste (LHW) using models on overall
heat and mass balances of the cement kiln, were validated by full-
scale trials. The characteristics of the fuels as received, satisfacto-
rily explained trends in the results on temperature, flue gas flow
rates and their consequent effect on clinker production rates. For
example, kiln gas temperatures were 24 °C, 28 °C, 22 °C and 95 °C
lower than the temperature in the coal reference case for RDF,
waste wood, SHW and LHW respectively [20]. According to Ari-
yaratne et al. [20] this drop in temperature matched the fuel heat-
ing values obtained from ultimate analysis. The higher the heating
value and lower the ash content for the solid fuels, the higher the
kiln gas temperature. An increase in the required mass of fuel
increases gaseous combustion products and leads to an overall
increase in flue gases. This increase in the total flue gas leads to
a reduction in thermal performance which translates to a reduction
in kiln gas temperature. In a second paper by Ariyaratne et al. [21],
the kiln gas temperature was impacted by the air requirement for
fuel combustion, which can also be deduced from the ultimate
analysis of the fuel. Additionally, the secondary air requirement
of a cement kiln generally increases with an increase in volatiles
to ensure complete combustion. This can be countered by a
decrease in mass for a fuel with a high energy content. Fuel air
requirement is also affected by a combined effect of the presence
of O and the molar H/C ratios for the given fuel [21]. Generally,
180 E.A. Asamany et al. / Fuel 193 (2017) 178–186
the higher the molar H/C ratio of a fuel, the lesser the molar air
requirement.
A combination of these factors in comparison with conventional
fuels could translate into the suitability of a potential WDF.
3. Materials and methods
3.1. The cement kiln
The cement plant looking to utilize waste plastic materials as
alternative fuel uses a long dry process kiln equipped with an Elec-
trostatic Precipitator (ESP). It has, in recent years, been operating
one dry process rotary kiln and produces 200,000–300,000 tonnes
of clinker annually. Based on these production estimates, the
cement plant consumes approximately 1.35 PJ of thermal energy
annually– the equivalent of 45,000 tonnes of coal. The plant uti-
lizes a blend of coal and petroleum coke (coal-coke) as its primary
fuel. The location of the plant near a limestone quarry provides
better incentive to substitute fuels than raw materials. It has over
the years obtained permits for the use of waste shingles and crude
glycerol as alternative fuels.
3.2. The plastic-based waste materials
The waste plastic samples used for combustion tests included:
(1) Construction and Demolition (C&D) plastics comprising a ran-
domized mix of rigid and flexible plastics resulting from C&D pro-
jects and recovered from a C&D landfill. (2) Plastic film which were
made of LDPE and LLDPE film collected from curbside collection
programs at a Material Recovery Facility (MRF). (3) Mixed plastics
type (4-7) excluding PVC were mostly plastic containers collected
from MRFs. (4) Polypropylene based carpet backing was collected
from carpet recyclers. (5) Textiles and fabric were used-clothing
from not-for-profit collectors. (6) Expanded Polystyrene (EPS) cups
and packaging material collected from disposal bins at a local hos-
pital and university campus. Coal and coke samples were sourced
from the stock pile at the local cement plant, dried and pulverized
to 80% < 200 mesh screen. They were then mixed to a 50–50 pro-
portion. Fig. 1 provides an example of plastic based waste (plastic
film and plastic containers) before and after size reduction for
combustion. Table 1 shows the ultimate analysis of the various
plastic materials studied in tube furnace experiments. Table 2
shows the ultimate analysis of coal-coke and waste shingles which
are fuels already been used in the cement kiln on full-scale basis.
3.3. Preliminary analysis
Fig. 2 is a chart showing results from stoichiometric calculations
on gaseous emissions. The calculations show the potential contri-
bution of the waste plastic samples to the reduction of CO2, SO2
and fuel NOx emissions compared to coal-coke on an equivalent
heat basis. Due to the complex nature and different formation
routes of fuel NOx, the fuel NOx calculated was based on an
assumption that for all the samples, including coal-coke, the same
proportion of fuel nitrogen was converted to fuel NOx. A general
trend of emission reduction is observed for the selected plastics.
In cases where emission increases were obtained, they were seen
to be a result of an integrated effect of increased fuel requirements
per energy supplied and higher amounts of the parent elements in
the fuel composition relative to coal-coke. Lesser quantities of C&D
plastics, EPS, containers, and plastic film are required compared to
coal-coke whereas the carpets and used clothing require approxi-
mately double the amount of coal-coke required to supply an
equivalent heat.
4. Experimental
4.1. Apparatus
4.1.1. Lindberg bench-scale electric tube furnace
A Lindberg electric tube furnace was used for combustion tests.
It is 190 cm long and has a 5 cm internal diameter. Temperatures
within the furnace reach a maximum set temperature of
1400 °C. The ends of the hot metal tube are cooled with water
flowing through copper tubes. As shown in Fig. 3, the sample
holder is fitted on a long probe fitted with a K-type thermocouple
which doubled as a handle for adjusting samples into place in the
furnace. Exhaust gas from the sample combustion is swept by air
into a metal gas sampling chamber, cooled with Argon, and then
into the secondary chamber to be collected by measuring devices
attached to sampling ports. The end of the tube furnace through
which the thermocouple probe runs is clamped shut with an open
connection to dry air supply.
4.1.2. Heated grid reactor
A heated grid reactor built in-house and used for combustion
tests is shown in Fig. 4. Other workers have referred to similar
reactors as Heated Wire Mesh (HWM) [22], Wire Mesh Reactor
(WMR) [23], heated grid reactor (HGR) [24] or heated foil reactor
[19]. The HGR was used to generate fast heating rates to simulate
actual kiln conditions to study PM emissions behavior of fuel
blends. The hardware comprises a reactor core, a stand, and the
sampling chamber. The core is made up of stainless steel blocks
placed 3 cm apart and mounted on laminated fiberglass platform
which are fastened to a circular stainless steel base with stands.
There is a 1 mm-diameter circular opening at the center of the
platform through which the thermocouple in positioned during
measurements. Bare wire K-type thermocouples of 0.13 mm diam-
eter were positioned through the 1 mm opening and centered to be
in contact with the mesh during experiments. A stainless steel
ANSI 304 wire cloth which could withstand temperature up to
1000 °C was used as sample holder and resistance heater. The
mesh was woven to an aperture of 0.2 mm.
For each experiment, the wire cloth mesh was cut to
15 mm  30 mm mesh was cut, wrapped around a sample and
was slotted in between the electrode and the stainless steel blocks
where it was screwed in place. The power supply is Delta Elektron-
ika S6-40 capable of delivering maximum voltage and current of 6
V and 40 A respectively. A computer feedback and temperature
control was run in LabVIEW. The LabVIEW Virtual Instrument
ran with a set-point profile ramp for PID controllers. Heating rates
of 103 K/s were reached using this set-up.
4.1.3. Air quality monitor
The Dylos DC1700 air quality monitor is a particle counter
which employs the use of a laser to count the number of PM per
cubic centimeter of flue gas. It counts PM with sizes between 0.5
and 2.5 lm, which are considered in this work as small particles
and also counts PM with sizes >2.5 lm, called large particles.
4.1.4. Thermal Desorption Tubes (TDT)
Stainless steel (90 mm length and 6.35 mm diameter) TDTs
were used to collect various VOCs. The TDT were packed with
Tenax TA, which is a polymer resin suitable for trapping non-
polar VOCs between n-C-7 to C-26 in air [25]. A Markes Interna-
tional Unity-2 Thermal Desorption unit, coupled to a Thermo Trace
1300 gas chromatograph that was coupled to a Thermo ELITE ISQ
EI mass spectrometer was used to identify and quantify the VOCs
sampled by the TDT.
 E.A. Asamany et al. / Fuel 193 (2017) 178–186 181

Fig. 1. A selection of received waste plastic materials (film and containers) before and after fuel preparation by size reduction.

Table 1
Analysis results reported by EXOVA labs for prospective WDF: size-reduced containers, carpet, clothing, plastic film, C&D plastics and expanded polystyrene.

Containers Carpet Textile Plastic film C&D plastics EPS

Chlorine (mg/kg) 3687 714 502 257 160 83.6
Fluorine (mg/kg) 15.7 54.0 <15.0 78.0 67.7 68.7
Carbon (% dry) 80.4 50.78 60.99 80.14 65.48 91.57
Hydrogen (% dry) 12.7 7.29 4.49 13.05 10.63 7.64
Oxygen (% dry) 1.91 17.65 32.87 3.46 1.88 <0.5
Nitrogen (% dry) <0.5 4.82 <0.5 <0.5 <0.5 <0.5
Sulphur (%) 0.046 0.088 0.031 0.025 0.1 0.0135
Moisture (%) 0.32 0.59 0.36 0.38 0.13 0.19
Ash (%) 5.16 17.72 7.44 3.22 7.66 0.16
Calorific (MJ/kg) 41.57 22.73 22.33 41.41 42.15 41.50

Table 2 perature was kept at 10 L/min to sweep combustion products to

Analysis results reported by EXOVA labs for coal-coke and waste shingles, which are the exhaust. The set-up shown in Fig. 3 was used. The exhaust
already used as fuel in the cement kiln.
gases were swept by the air into a primary distribution chamber
Component measured Waste shingles Coal-coke that also serves to cool the exhaust gases. Argon gas was intro-
Chlorine (mg/kg) 220 183 duced into the primary chamber at 4 L/min to keep the exhaust
Fluorine (mg/kg) 58.3 85.2 gas at temperatures of 50 °C and below. This was done to mitigate
Carbon (% dry) 46.22 69.67 secondary reactions and also to bring products to temperatures
Hydrogen (% dry) 5.08 3.18
suitable for the measuring devices. The exhaust gases then flow
Oxygen (% dry) 10.11 7.11
Nitrogen (% dry) 0.50 1.68 into a 6 L secondary plastic chamber where they are drawn into
Sulphur (%) 1.23 3.52 the particle counter at a rate of 1.7 L/min. The secondary chamber
Moisture (%) 7.70 0.37 was to ensure that concentrations were within the safe limit of the
Ash (%) 41.18 13.43 detection for the attached devices. The VOC sample collected at
Calorific value (MJ/kg) 16.46 29.63
10 ml/min into the TDT attached to the primary chamber. The
other ports on the primary and secondary chambers were shut
off when not in use.
4.2. Procedure In the second set of experiments, samples of coal-coke, waste
plastics, waste shingles and blends of coal-coke with waste plastics
In the first set of experiments a heat equivalent of 2 g coal-coke and shingles were blended to an equivalent energy of 10-mg of
for each sample as shown in Table 3, was weighed and combusted coal-coke. Waste shingles were size reduced to particle sizes of
in a tube furnace. The temperature of the furnace was held at a 1.5–3 mm and the waste plastic samples comprised of 3–5 mm
maximum temperature of 1100 °C for the combustion of the sam- chips and fluff. 30% of the thermal energy required was supplied
ples for a residence time of 3 min. Dry air supply at ambient tem-
182 E.A. Asamany et al. / Fuel 193 (2017) 178–186

Fig. 2. Theoretically expected changes in gaseous emissions from the combustion of plastic based waste materials based on reported elemental analysis. NOx values refer to
fuel NOx only.

A 10-s cycle which included heating and a hold time of 2 s, at

1000 °C was used for this set of experiments.
Finally, clinker production and emission data collected from the
full-scale trials using waste shingles are also presented. The data is
used to discuss full-scale implications of WDF use in the cement
kiln.

5. Results

5.1. Tube furnace tests

Fig. 3. Schematic cross section of tube furnace and attached sampling chambers
used for combustion tests.
by either the plastics or shingles. The weight of the samples tested
are shown in Table 4. The HGR was mounted in a sampling cham-
ber with ambient atmosphere and pressure. It was used in con-
junction with the particle counter to compare the particulate
emission of the single fuel to the blends after combustion in air.
VOC detection data on all samples from combustion in the tube
furnace was obtained from GC–MS performed on samples collected
with the TDTs. The quantity of the detected VOCs was summed up
to obtain a total VOC in mg/kg of fuel. As shown in Fig. 5, Carpets
recorded the largest quantity of total VOC in exhaust (8683 mg/kg
fuel) and EPS recording the least (4 mg/kg fuel). Figs. 5–7 are com-
bined plots of VOC detected with the calorific value, moisture con-
tent and ash content of the given fuel. The residual VOC detected
increased generally with decreasing calorific value, increased
moisture content and increased ash content in similar fashion as

Fig. 4. Heated grid reactor main block and reactor sample chamber assembly.
 E.A. Asamany et al. / Fuel 193 (2017) 178–186 183

Table 3
Weights of 2 g-coal-coke heat equivalent for all
containers, plastic film, clothing, carpet and C&D
plastics used in tube furnace experiments.

Sample tested in tube furnace Weight (g)

EPS 1.43
Plastic film 1.43
Containers 1.43
Carpet 2.61
Clothing 2.65
C&D Plastics 1.41

Table 4
Sample composition and weights for heated grid reactor PM emission tests.

Sample tested in HGR Weight for HGR (mg)

Coal-coke only 10.0
Mixed plastics only 7.0
Waste Asphalt shingles only 14.0 Fig. 7. Combined graph of ash content and the total VOC detected for plastic-based
Coal-coke + 30% energy from plastics 7.0 + 2.2 material samples tested in tube furnace experiment.
Coal-coke + 30% energy from shingles 7.0 + 4.5

5.2. Heated grid reactor tests

Figs. 8 and 9 show interval plots for the count of small and large
airborne PM in the gases collected during the combustion of the
fuel samples in the HGR. The number of small particulates (0.5–
2.5 lm) and large particulates (>2.5 lm) in the combustion prod-
ucts were continuously measured over a period of five minutes.
This period included a 10 s heating cycle applied the sample in
the mesh. Three tests (n = 3) were conducted for each sample.
The five-minute average for each run was used in the interval plot.
The plots show considerable overlap in the intervals for the coal-
coke and the blends meaning that, with 95% confidence, the PM
counts recorded for the samples do not differ significantly.

5.3. Full-scale tests

In the absence of burn permits for the waste plastic fuels,

Fig. 5. Combined graph of energy content and the total VOC detected for plastic-
based material samples tested in tube furnace experiment.
results from the trials on the local kiln using up to 10% thermal
input of waste asphalt shingles were conducted. Fig. 10 shows that
the introduction of waste shingles in the energy mix did not
adversely affect the thermal performance of the kiln. This is shown
by the improvement and maintenance of clinker production over
the period. The shingles were primarily delivered into the main
burner zone of the cement kiln via pipes separate from the main
burner.
Data obtained from National Pollutant Release Inventory (NPRI)
[26] and plotted in Fig. 11 shows a reduction in the reported PM
emissions from the cement facility. These reductions cannot be
solely linked to the use of shingles however; they do suggest that
waste shingle use did not lead to marked increases in PM under
full-scale kiln operating conditions.

6. Discussion

6.1. Expectations based on bench-scale results on VOC emissions

The presence of VOCs in exhaust is influenced by the combined

Fig. 6. Combined graph of moisture content and the total VOC detected for plastic-
based material samples tested in tube furnace experiment.
kiln gas temperature observed by Ariyaratne et al. [20] in full-scale
cement kiln studies.
effect of the initial volatile matter content, the rate of devolatiliza-
tion, and the extent of combustion. These factors depend physically
on the temperatures of operation, quantity, size and structure of
fuel particles and residence time among others. It is difficult to
ascertain VOC emissions resulting from the fuel only, since a sub-
stantial amount of organic matter is also present in the kiln raw
materials. Furthermore, it is unlikely that at full-scale kiln condi-
tions, VOCs resulting from incomplete fuel combustion would exist
184 E.A. Asamany et al. / Fuel 193 (2017) 178–186

Fig. 8. Count of small particles present per cubic centimeter of combustion products of tested coal-coke fuel and blend samples in HGR.

Fig. 9. Count of large particles present per cubic centimeter of combustion products of tested coal-coke fuel and blend samples in HGR.

900
Shingles addition

800

700
tonnes of clinker/day

600

500

400

300

200

100

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
day of month

Fig. 10. Clinker production in a month with the introduction of waste shingles. Production levels were maintained at 900 tonnes of clinker/day.
 E.A. Asamany et al. / Fuel 193 (2017) 178–186
Fig. 11. Annual total particulate matter in tonnes, released from the stack of the
local cement plant between 2011 and 2015, showing reductions which remain
insignificantly changed in the last two years.
at the stack. The conditions in the bench-scale apparatus are there-
fore not an actual representation of the actual kiln combustion but
they provide a tool by which the extent of combustion can be com-
pared for the fuels under similar conditions, and in the absence of
competing sources. This extent of combustion can be deduced from
the residual VOC detected in the combustion products. Trends in
VOC emission therefore could provide insight into the behaviors
solely attributable to the fuel materials under set conditions. The
total VOCs detected in exhaust was less for the samples with
energy content above 40 MJ/kg than for samples with energy con-
tent below 30 MJ/kg. This could be explained by the fact that
higher energy density results in a reduction of the fuel mass
required which reduces the potential VOC emissions expected.
Increased moisture content generally resulted in more residual
VOC emission. The presence of moisture reduces the combustion
efficiency of the fuel by the loss of heat to vaporization. This also
means that within a given residence time, the period of moisture
loss is extended for the materials with high moisture content,
thereby reducing the time and energy available for the oxidation
of volatiles after they are released. The reduced combustion effi-
ciency and the loss in oxidation time, can explain the relation
between increased moisture content and increased residual VOC.
VOC emissions also increased with increasing ash content. Gener-
ally, the higher the ash content, the less the energy content of a
given mass. This reduces the energy available for thermal degrada-
tion. Among the plastics with energy content > 40 MJ/kg (EPS, plas-
tic film and containers), the physical properties of the fuel seemed
to affect the level of VOC emissions. EPS, is light and has an aerated
structure. It recorded very little residual VOC. The low structural
density allows for better diffusion of heat and oxygen into the par-
ticle. The structure also enhances the removal of volatile matter
from the core of the particles for oxidation. Volatiles are therefore
consumed more effectively within the given residence time. Simi-
lar explanations could be given for plastic films which were less
dense compared to the chips of the rigid containers. The presence
of inerts and fire-retardant additives used in the production of
clothing and carpet are primary reasons why they were less com-
bustible as received, and therefore generated more residual VOCs
and more ash.
6.2. Expectations based on bench-scale results on particulate matter
A review of the median counts from the plots generated for PM
data shows a mix of additive and synergistic effects from fuel
blending. Given the results obtained on a bench-scale for PM emis-
sions in the HGR, it can be said that, though the variability of fuel
185
related PM emissions may increase with the addition of WDFs, the
overall behavior of PM emission will not differ significantly from
that of coal-coke. The results also show that it will be difficult to
match bench-scale PM emissions with a characteristic fuel prop-
erty. It can however be seen that the more of a sample combusted,
the more PM emissions detected. These variations should be
expected on a large scale because cement stack emissions depend
on many factors beyond the fuel characteristics. This makes the
prediction of PM emissions with respect to fuel change quite chal-
lenging. The work of Jeschar et al. [7] encountered a similar diffi-
culty. This uncertainty in behavior is also consistent with large
scale PM emissions reported by workers such as Carrasco et al.
[17] and CRA [18]. If the full-scale PM emissions with respect to
fuel change behave similarly to bench-scale emissions, then a
proper functioning ESP for example, should handle the PM variabil-
ity resulting from fuel changes satisfactorily well.
6.3. Full-scale implications
In summary, the bench-scale experiments using the tube fur-
nace and HGR provided observations which when matched with
expectations drawn from ultimate-analysis, show varied extents
of predictability in full-scale practice. The trends in residual VOC
emission with respect to the energy, moisture and ash content,
match well with the observations made by Ariyaratne et al. [20]
and Ariyaratne et al. [21]. For the fuels they studied, increased fuel
requirement, high moisture content and low ash content generally
resulted in increased kiln flue gas, lower kiln gas temperatures and
reduced clinker output. This similarity in the trends of bench-scale
measured residual VOC and full-scale thermal performance sug-
gest that prior to full-scale trials, the tube furnace set-up can be
used as a means to comparatively estimate the effect that the fuel
composition might have on the kiln’s thermal performance. Actual
full-scale tests conducted on waste shingles at the local cement
plant showed that kiln thermal performance could be maintained
or improved by the use of WDF. This was seen in the maintenance
of clinker production rates over the period of firing waste shingles.
A similar behavior is expected for blends of coal-coke and waste
plastics because compared to most of the waste plastic samples,
waste shingles contain less energy, more ash and more moisture
per equivalent heat, and because of the effect that these character-
istics can have on bench-scale combustion efficiency and full-scale
kiln thermal performance, it is fair to expect that firing equivalent
amounts of the tested waste plastics would not reduce clinker pro-
duction as a result of lowered fuel combustion efficiency or low-
ered kiln thermal performance. The fuel handling approaches for
the waste plastic fuel stream may however be different due to dif-
ferences in density and particle size of the fuels as received. There
is need for further studies on the comparative behavior of single
plastic and waste shingle particles for a better prediction. How-
ever, since the key to obtaining actual trial permits lies in the out-
come of a comparison of the new fuel to available data on fuels that
are already in use, this expectation for waste plastics, though not
elaborately arrived at, forms an initial argument for acquiring
actual full-scale trial permits. Concerning PM emissions, both
bench-scale and full-scale observations made show that for the
most part, changes that are solely attributable to fuel change are
not significant when co-processing WDF and coal-coke.
7. Conclusions
Expectations in emission characteristics of waste plastics as
potential WDF in cement kilns have been studied using theoretical
calculation and bench-scale testing. This approach provides useful
insight into the potential impact of energy content and other phys-
186 E.A. Asamany et al. / Fuel 193 (2017) 178–186
ical and chemical properties of waste plastic materials on gaseous
emissions, particulate matter and residual VOC emissions. The
results of bench-scale experiments showed favorable correlation
with observations of full-scale behavior of kiln thermal perfor-
mance parameters seen in the literature. For particulate emissions,
the bench-scale results show no significant deviations if up to 30%
energy of conventional fuels was replaced with waste plastics or
waste shingles. It is however difficult to match the behavior to
any fuel property besides the amount of fuel combusted. The
results on the residual VOC emission as observed in bench-scale
tests matched well with the trends observed by other workers on
the effect of WDF on full-scale kiln gas temperature profiles and
clinker production rates. Waste plastics with higher energy value,
lower moisture content and lower ash content yield less VOCs in
bench-scale tests. This indicates high combustion efficiency under
bench-scale conditions. The bench-scale residual VOC approach
can thus be used as a screening method to select appropriate
WDF for actual kiln tests by relating the bench-scale combustion
efficiency to the potential impact on the full-scale kiln thermal per-
formance. Furthermore, the ease of comparison associated with the
described bench-scale tests, presents a good resource for operators
to obtain preliminary data to acquire fuel trial permits and to
improve heterogeneous WDF selection and blending.
Acknowledgements
The authors acknowledge the financial support from Lafarge
Cement, Brookfield, Nova Scotia, Canada, the Canadian Plastics
Industry Association, the Natural Sciences and Engineering
Research Council (NSERC) of Canada, MITACS Canada, and the
Resource Recovery Fund Board (RRFB) now DivertNS, Nova Scotia.
Assistance with some of the analytical work from Thomas Barnett,
David Bowen and Courtney Wilson is gratefully acknowledged as is
technical support from Ray Dube in building the heated grid
reactor.
References
[1] Kelleher-Environmental. Plastics waste denominator study for Canada.
Canadian Plastic Industry Association; 2012.
[2] Gray D. Earth day, sustainability and plastics. Canadian Plastic Industry
Association; 2014.
[3] Waste Reduction Facts, in, Clean Foundation, Nova Scotia, Canada; 2011.
[4] Mullinger P, Jenkins B. Industrial and process furnaces - principles. In: Design
and operation. Elsevier; 2008.
[5] Mokrzycki E, Uliasz- Bocheńczyk A. Alternative fuels for the cement industry.
Appl Energy 2003;74:95–100.
[6] Tillman DA, Duong DNB, Harding NS. Solid fuel blending - principles, practices,
and problems. Elsevier; 2012.
[7] Jeschar R, Jennes R, Kremer H, Kellerhoff T. Reducing NOx and CO emission by
burning plastics in the calcining cement rotary kiln of a plant ZKG
international, 1999;52:534–49.
[8] Tillman DA. Cofiring benefits for coal and biomass. Biomass Bioenergy
2000;19:363–4.
[9] Narang L. Techno-economical assessment of alternative fuels in cement kiln.
In: Department of biotechnology and environmental sciences. Patiala: Thapar
University; 2012.
[10] Themelis N, Castaldi M. Use of alternative fuels in cement production in; 2014.
[11] Choi BS, Yi J. Simulation and optimization on the regenerative thermal
oxidation of volatile organic compounds. Chem Eng J 2000;76:103–14.
[12] Gibson MD, Heal MR, Bache DH, Hursthouse AS, Beverland IJ, Craig SE, et al.
Using mass reconstruction along a four-site transect as a method to interpret
PM10 in west-central Scotland, United Kingdom. J Air Waste Manage Assoc
2009;59:1429–36.
[13] Gibson MD, Guernsey JR, Beauchamp S, Waugh D, Heal MR, Brook JR, et al.
Quantifying the spatial and temporal variation of ground-level ozone in the
rural Annapolis Valley, Nova Scotia, Canada using nitrite-impregnated passive
samplers. J Air Waste Manage Assoc 2009;59:310–20.
[14] Gibson MD, Pierce JR, Waugh D, Kuchta JS, Chisholm L, Duck TJ, et al.
Identifying the sources driving observed PM2.5 temporal variability over
Halifax, Nova Scotia, during BORTAS-B. Atmos Chem Phys 2013;13:7199–213.
[15] Gibson MD, Haelssig J, Pierce JR, Parrington M, Franklin JE, Hopper JT, et al. A
comparison of four receptor models used to quantify the boreal wildfire smoke
contribution to surface PM2.5 in Halifax, Nova Scotia during the BORTAS-B
experiment. Atmos Chem Phys 2015;15:815–27.
[16] Franklin JE, Drummond JR, Griffin D, Pierce JR, Waugh DL, Palmer PI, et al.
BORTAS field experiment. Atmos Chem Phys 2011;14(2014):8449–60.
[17] Carrasco F, Bredin N, Heitz M. Gaseous contaminant emissions as affected by
burning scrap tires in cement manufacturing. J Environ Qual
2002;31:1484–90.
[18] Conestoga-Rovers&Associates, Ontario, Canada. Report on air emission
assessment for proposed scrap tire co-combustion; 2007.
[19] Anastasakis K, Kitsiou I, de Jong W. Fast devolatilization characteristics of ‘low
cost’ biomass fuels, wood and reed. Potential feedstock for gasification. Fuel
Process Technol 2016;142:157–66.
[20] Ariyaratne WKH, Melaaen MC, Tokheim L-A. The effect of alternative fuel
combustion in the cement kiln main burner on production capacity and
improvement with oxygen enrichment. Int J Chem Mol Nucl Mater Metall Eng
2013;7:184–91.
[21] Ariyaratne HW, Melaaen MC, Tokheim LA, Manjula EVPJ. Kiln process impact
of alternative solid fuel combustion in the cement kiln main burner -
Mathematical modelling and full-scale experiment; 2014.
[22] Cai H-Y, Guell AJ, Chatzakis IN, Lim J-Y, Dugwell DR, Kandiyoti R. Combustion
reactivity and morphological change in coal chars: effect of pyrolysis
temperature, heating rate and pressure. Fuel 1996;75:15–24.
[23] Zeng C, Chen L, Liu G, Li W, Huang B, Zhu H, et al. Advances in the development
of wire mesh reactor for coal gasification studies. Rev Sci Ins 2008;79:084102.
[24] Prins MJ, Li ZS, Bastiaans RJM, van Oijen JA, Aldén M, de Goey LPH. Biomass
pyrolysis in a heated-grid reactor: visualization of carbon monoxide and
formaldehyde using laser-induced fluorescence. J Anal Appl Pyrol
2011;92:280–6.
[25] Dohoo C, Read Guernsey J, Gibson MD, VanLeeuwen J. Impact of biogas
digesters on cookhouse volatile organic compound exposure for rural Kenyan
farmwomen. J Expos Sci Environ Epidemiol 2015;25:167–74.
[26] NPRI. National Pollution Release Inventory, Canada from <https://www.ec.gc.
ca/inrp-npri/>.