Emissions of air pollutants from household stoves: Honeycomb coal versus coal cake

Article abstract of DOI:10.1021/es049942k

Domestic coal combustion can emit various air pollutants. In the present study, we measured emissions of particulate matter (PM) and gaseous pollutants from burning a specially formulated honeycomb coal (H-coal) and a coal cake (C-coal). Flue gas samples for PM_2.5, PM coarse (PM _2.5-10), and TSP were collected isokinetically using a cascade impactor; PM mass concentrations were determined gravimetrically. Concentrations of SO₂, NO_x, and ionic Cr(VI) in PM were analyzed using spectrometric methods. Fluoride concentrations were measured using a specific ion electrode method. PM elemental components were analyzed using an X-ray fluorescence technique. Total (gas and particle phase) benzo[a]pyrene (BaP) concentration was determined using an HPLC/fluorescence method. Elemental and organic carbon contents of PM were analyzed using a thermal/optical reflectance technique. The compositional and structural differences between the H-coal and C-coal resulted in different emission characteristics. In generating 1 MJ of delivered energy, the H-coal resulted in a significant reduction in emissions of SO₂ (by 68%), NO_x (by 47%), and TSP (by 56%) as compared to the C-coal, whereas the emissions of PM_2.5 and total BaP from the H-coal combustion were 2-3-fold higher, indicating that improvements are needed to further reduce emissions of these pollutants in developing future honeycomb coals. Although the H-coal and the C-coal had similar emission factors for gas-phase fluoride, the H-coal had a particle-phase fluoride emission factor that was only half that of the C-coal. The H-coal had lower energy-based emissions of all the measured toxic elements in TSP but higher emissions of Cd and Ni in PM_2.5.

Full text of DOI:10.1021/es049942k

Environ. Sci. Technol. 2004, 38, 4612-4618
SO (by 68%), NO (by 47%), and TSP (by 56%) as compared
2
x
Emissions of Air Pollutants from
Household Stoves: Honeycomb Coal
versus Coal Cake
to the C-coal, whereas the emissions of PM2.5 and total
BaP from the H-coal combustion were 2-3-fold higher,
indicating that improvements are needed to further reduce
emissions of these pollutants in developing future
honeycomb coals. Although the H-coal and the C-coal
had similar emission factors for gas-phase fluoride, the
H-coal had a particle-phase fluoride emission factor that
was only half that of the C-coal. The H-coal had lower energy-
based emissions of all the measured toxic elements in
TSP but higher emissions of Cd and Ni in PM2.5.
S U G E
Am-As Corporation, Portland, Oregon
X U X U
Environmental Biomarker Shared Resource, The Cancer
Institute of New Jersey/Environmental and Occupational
Health Science Institute, University of Medicine and Dentistry
of New Jersey (UMDNJ), Piscataway, New Jersey 08854
Introduction
China has the largest coal production and consumption in
the world (1). It has been estimated that more than 400 million
people in China use coal for cooking and heating (2, 3). Due
to the limited storage of petroleum and natural gas and
relatively abundant coal storage (1000 billion ton) in China,
it can be predicted that this coal-based energy structure in
China may last for many years to come. Coal combustion,
unfortunately, produces a wide variety of air pollutants
including particulate matter (PM), carbon monoxide (CO),
carbon dioxide (CO2), sulfur dioxide (SO2), nitrogen oxides
(NOx), and a large suite of organic toxics (4-8). The
environmental and public health concerns over the coal
emissions have motivated the coal research community to
develop technologies that can reduce toxic emissions from
coal combustion. One relatively inexpensive method is to
burn coal in formulated form instead of burning raw coal,
because formulation of raw coals can enhance coal combus-
tion efficiency and reduce air pollution emission potentials
(10, 11). Formulated coals are usually made in briquette form.
Our previous research demonstrated that a formulated boiler
briquette coal, compared to its raw coal precursor, sub-
stantially improved combustion efficiency when used in a
traditional steam-generating boiler and reduced emissions
of major air pollutants including PM2.5, SO2, CO, CO2, and
some toxic metals in PM (10, 11).
J U D I T H C . C H O W A N D J O H N W A T S O N
Division of Atmospheric Sciences, Desert Research Institute,
2215 Raggio Parkway, Reno, Nevada 89512
Q I N G S H E N G
Research Institute of Physical & Chemical Engineering of
Nuclear Industry, 168 Jintang Road, Tianjin, 300180, China
W E I L I L I U
Joint Graduate Program in Exposure Measurement &
Assessment, University of Medicine and Dentistry of
New Jersey (UMDNJ) and Rutgers University,
170 Frelinghuysen Road, Piscataway, New Jersey 08854
Z H I P E N G B A I A N D T A N Z H U
College of Environmental Sciences and Engineering,
Nankai Univesity, 94 Weijin Road, Tianjin, 300071, China
J U N F E N G Z H A N G *
UMDNJ-School of Public Health, Environmental and
Occupational Health Sciences Institute,
170 Frelinghuysen Road, Piscataway, New Jersey 08854
Honeycomb coal briquettes are usually used in household
cooking/ heating stoves. It is estimated that ∼70 million ton
of coal is consumed for domestic usage per year in China,
of which 50 million ton is in the form of honeycomb coal.
Honeycomb coal briquettes are developed historically from
a standpoint of ease in use (transportation, handling, storage,
and burning). Reducing toxic emissions has not necessarily
been considered by numerous local “manufacturers” across
China. In attempts to improve combustion efficiency and to
reduce toxic emissions, scientists from the Institute of Coal
Chemistry of the Chinese Academy of Sciences have for-
mulated a new honeycomb coal (H-coal) designed for
household use. The objective of this study is to characterize
the emissions from combustion of the newly formulated
H-coal. For comparison, the emissions from combustion of
a coal cake (C-coal), which was made from the same raw
coal as was for the H-coal, were also characterized. We
determined the concentrations, emission rates, and emission
factors of PM, SO2, NOx, benzo[a]pyrene (BaP), fluoride (F-),
and toxic elements in PM for the H-coal and the C-coal. The
elemental source profile of the emitted PM was also
characterized for both coals.
Domestic coal combustion can emit various air pollutants.
In the present study, we measured emissions of particulate
matter (PM) and gaseous pollutants from burning a
specially formulated honeycomb coal (H-coal) and a coal
cake (C-coal). Flue gas samples for PM2.5, PM coarse
(PM2.5-10), and TSP were collected isokinetically using a
cascade impactor; PM mass concentrations were determined
gravimetrically. Concentrations of SO , NO , and ionic
2
x
Cr(VI) in PM were analyzed using spectrometric methods.
Fluoride concentrations were measured using a specific
ion electrode method. PM elemental components were
analyzed using an X-ray fluorescence technique. Total (gas
and particle phase) benzo[a]pyrene (BaP) concentration
was determined using an HPLC/fluorescence method.
Elemental and organic carbon contents of PM were analyzed
using a thermal/optical reflectance technique. The
compositional and structural differences between the
H-coal and C-coal resulted in different emission character-
istics. In generating 1 MJ of delivered energy, the
H-coal resulted in a significant reduction in emissions of
Materials and Methods
Tested Coals. The raw coal used to formulate H-coal and
C-coal was from a mine located in Yangquan, Shanxi
* Corresponding author phone: (732)445-0158; fax: (732)445-0116;
e-mail: jjzhang@eohsi.rutgers.edu.
9
4 6 1 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004
10.1021/es049942k CCC: $27.50
2004 Am erican Chem ical Society
Published on Web 08/04/2004

After being pulverized into coal powders, the raw coal is
mixed with hydrate lime (Ca(OH)2, 15%), a combustion-
promoting agent (0.5%), a waterproofing agent (0.03%), and
a binding agent (11.5%). The mixture is then pressed into a
cylinder 110 mm in diameter and 90 mm high, with 10 holes
from top to bottom (Figure 1). Each H-coal weighs about
1500 g and can burn for ∼60 min, depending on the air supply.
To fit the household stoves used in the study, the raw coal
was shaped into coal cake (C-coal) by mixing raw coal (80%)
with clay (20%) as binding material. The C-coal weighed 400
g and can burn for 20 min. The H-coal, the C-coal, and the
raw coal from which the H-coal and the C-coal were made
were characterized by proximate and ultimate analysis
performed by the Institute of Coal Chemistry (Table 1). For
comparison, the range of the composition of other Chinese
honeycomb coals is also summarized in Table 1.
Test Stove. The stove chosen for the study was one of
those most widely used to burn honeycomb coal briquettes
and coal cakes in China. In winter months, this type of stove
is usually used for both cooking and heating. The experi-
mental setup is shown schematically in Figure 2. Coals were
loaded from the top of the stove into the burning chamber,
and the ash fell down into the bottom chamber of the stove.
The bottom chamber of the stove was separated from the
burning chamber for ash storage. An ash door was used to
control the air supply flow and thus the heat output (power).
The ash door is usually closed when the stove is not in use
and is usually open to provide more oxygen when the stove
is used for cooking and heating. The H-coal had a substantially
improved thermal efficiency versus the C-coal (35% vs 23%,
see Table 2), as a result of the special H-coal formulation
(e.g., adding combustion-promoting agent, more uniform
combustion in the stove combustion chamber). Each stove
was equipped with a flue, as shown in Figure 2. The majority
of the coal-emitted gas was exhausted through the flue
connected to the stoves. Coal combustion in cook stove is
not a steady-state process. An integrated sampling approach
was used to collect air pollutant emissions during an entire
burn cycle (from fire start to fire extinction) as described in
the following sections. This approach provides emission data
that can represent a realistic burning situation (8).
FIGURE 1. Honeycomb coal.
TABLE 1. Proximate and Ultimate Analysis for the Coals
Testeda
H-coal C-coal R-coalb typical Chinese H-coalsc
Proximate Analysis
33.85 30.21
1.67 2.85
ash
m oisture
13.58
2.05
8.17
76.2
25.4-49.3
1.80-4.39
5.15-11.9
41.4-56.7
volatile m atter 12.28 11.27
fixed carbon
52.2
55.67
Ultimate Analysis
56.09 62.3
carbon
78.22
2.7
0.76
0.73
1.96
54.3-63.6
0.28-2.6
hydrogen
oxygen
nitrogen
sulfur
2.99
3.43
0.66
1.31
2.64
2.76
0.83
1.58
0.17-0.94
a
b
All contents are in percent (%) by weight on a dry fuel basis. The
c
raw coal from which the H-coal and the C-coal were m ade. Ref 9.
Province. The raw coal contained powders smaller than 13
mm (accounting for about 30 wt %) and coal pieces with a
diameter ranging from 13 to 25 mm. The H-coal formulation
is patented (Chinese Patent No. ZL 97 1 20946.4; International
Patent No. C10L5/ 04) and is briefly described as follows.
PM Sam pling. A sampling probe was placed at the center
of the flue of the stove, facing the direction of the chimney
gas flow. As shown in Figure 2, the height of sampling port
was 2 m above ground level. The sampling point is 0.9 m
FIGURE 2. Schematic of sampling platform for household stove fueled with honeycomb and briquettes. The stove was a cylinder 24 cm
in diameter and 50 cm in height. The burning chamber inside the stove was 30 cm in height and 12 cm in diameter.
9
VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 6 1 3

stored in a freezer (-20 °C) and analyzed for Cr(VI) within
2 weeks. The detection limit of this method was 1.0 µg/ g.
TABLE 2. Operating Conditions of the Test Stove Using H-Coal
and C-Coala
The separate set of PM2.5 and PM10 samples that were
collected on quart-fiber filter substrates were analyzed for
organic carbon (OC) and elemental carbon (EC) using the
thermal/ optical reflectance (TOR) carbon analyzer at DRI
(Reno, NV). The TOR method and quality control details were
described by Chow et al. (13).
parameter
unit
H-coal
C-coal
fuel burn rate
flue gas tem perature
flue gas hum idity
kg/h
°C
%
1.49 (0.07)
183 (7)
3.13 (0.5)
0.95 (0.05)
15.7 (1.5)
10.46
21.92 (0.12) 23.26 (0.23)
7.67 (0.61)
35
1.2 (0.09)
155 (4)
4.03 (0.4)
0.8 (0.05)
13.8 (1.4)
11.51
flue gas velocity
m /s
flue gas em ission rate
flue gas em ission factorb
flue gross heat value
delivered heat value
stove therm al efficiencyb
m 3/h
m 3/kg
MJ /kg
MJ /kg
%
Collection and Analysis of SO2 and NOx in Flue Gas. The
SO2 emission was collected and analyzed following the
Chinese State Environmental Protection Administration’s
(SEPA) standard method, which is similar to the U.S. EPA’s
standard method and described in Zhang et al. (11). Triplicate
samples of flue gas were collected using an impinger with
a solution of formaldehyde, potassium biphthalate, and Na2-
CDTA. SO2 was then determined colorimetrically at a
wavelength of 575 nm.
5.42 (0.22)
23
a
b
Standard deviations (n ) 3) in parentheses. Calculated average
value.
away from the stove. The flue gas was drawn through the
sampling probe under isokinetic conditions and diluted in
a previously described dilution system (10). During the
dilution process, the flue gas was cooled to ambient tem-
perature. PM was collected and segregated by particle size
using a PM2.5/ PM10/ TSP cascade impactor. The impactor was
operated at a flow rate of 36 L/ min to maintain the
aerodynamic cut-point for each stage. The particles were
collected on glass-fiber filters (Whatman 934AH, Clifton, NJ).
Another set of PM2.5 and PM10 samples were also collected
for elemental and organic carbon analysis based on the
sampling system and method reported by Ge et al. (10).
Three replicated measurements of the size-resolved
particles were acquired for both raw coal and coal cake. Field
and laboratory blanks were used to allow evaluation of
potential contamination of blank filters and samples during
preparation, weighing, storage, sampling, transportation, and
laboratory analyses. All the filter samples and blanks were
placed in clean Petri dishes individually during transportation
in coolers and storage in freezers.
The concentrations of NOx were measured using a method
adapted from the NIOSH standard method 6014 (14). Flue
gas was collected at a flow rate of 0.3 L/ min. An inline
chromate filter was used to oxidize nitric oxide (NO) to NO2.
The total nitrogen dioxide was then collected in an impinger
containing a triethanolamine solution. The NO2 concentra-
tion was determined by a spectrophotometer at a wavelength
of 540 nm after addition of hydrogen peroxide, sulfanilamide,
and N-(1-naphthyl)ethylenediamine dihydrochloride (NEDA)
solutions. The detection limit of the method was 2.0 mg/ m3.
A calibration curve was prepared daily with NaNO2 solution
ranging from 0.1 to 20 µg/ mL.
Collection and Analysis of Fluoride in Flue Gas. Fluoride
(F-) in flue gas was sampled and analyzed using a standard
method certified by SEPA (16). Three citric acid-impregnated
cellulose ester membrane filters followed by a potassium
dihydrogen phosphate-soaked cellulose ester membrane
filter (with pad) were used to collect F- in PM2.5, PM2.5-10
,
TSP-PM10, and gas-phase F-, respectively. The PM2.5/ PM10/
TSP separator was used to separate the PM by size. The
sample was drawn from the flue gas isokinetically at a flow
rate of 36 L/ min. Particulate fluorides were fused in sodium
hydroxide and then dissolved in 10 mL of water and total
ionic strength activity buffer (TISAB). The pH of the buffer
was ∼5, a level at which F- was the predominant fluorine-
containing species. The buffer contained cyclohexylene-
dinitrilotetraacetic acid, which formed stable complexes with
Fe(III) and Al(III), thus removing interferences by freeing
fluoride ions from complexes with these ions (14). Gaseous
fluorides were extracted with 10 mL of water and TISAB. The
extracts were then analyzed using ion-selective electrode
method. Sodium fluoride was used to prepare the calibration
curve. The method detection limit was 4 µg/ m3.
Analysis of PM Sam ples. Prior to and following sampling,
the filters were weighed using an electronic microbalance
with a sensitivity of 1 µg. A 210Po radioactive source was used
to remove static charge during filter weighing. Before
weighing, all filters were conditioned for at least 24 h in a
temperature- (20 ( 2.0 °C) and relative humidity-controlled
(40 ( 5%) environment at the Research Institute of Physical
& Chemical Engineering of Nuclear Industry (Tian Jin, China).
The mass concentrations of PM2.5 and PM10 were determined
using the net weight increases of the sample filters and the
sampling volumes.
After the gravimetric analysis, the filters were analyzed
for 40 elements using an X-ray fluorescence (XRF) spec-
trometer at Desert Research Institute (DRI) in Reno, NV (12).
Vacuum deposit thin-film standards were used as calibration
standards, and polymer films were used as quality control
standards. Blank filters from the same manufacturing lot as
the sample filters were analyzed along with the PM sample
filters. An average blank spectrum was constructed and used
for spectral background subtraction of the exposed filters.
The collected samples were then submitted for hexavalent
chromium [Cr(VI)] using a visible absorption spectropho-
tometric method, which was modified from NIOSH standard
method 7600 (14). The filter was placed in a 50-mL beaker
containing 5 mL of 2% NaOH/ 3% Na2CO3. The beaker was
covered with a watch glass and heated to near the boiling
point for 30 min. During the extraction process, the headspace
above the solution was purged with nitrogen to avoid
oxidation of Cr(III) in the sample. The resolution was cooled
to room temperature and mixed with 1.9 mL of 6 N sulfuric
acid and 0.5 mL of 2.5 mg/ mL diphenylcarbazide in water/
acetone (1:1). The resulting solution was diluted to 25 mL
with distilled water and analyzed with a spectrophotometer
at a wavelength of 540 nm. Before analysis, all samples were
Collection and Analysis of BaP in Flue Gas. Total BaP,
including both gas- and particle-phase BaP was collected
and analyzed. The sampling probe nozzle was connected
with a stainless steel sampling tube where the majority of
gas-phase BaP was condensed into particulate phase. The
sampling tube was equipped with a water jacket through
which tap water flew to keep the temperature of the sampling
tube lower than 40 °C. This design was to prevent possible
reactions between PAHs in the gas phase and NOx at high
temperature (17). The particle-phase BaP was collected on
a glass-fiber filter using an air-sampling pump at a flow rate
of 12.5 L/ min. An impinger containing 50 mL of cyclohexane
was placed between the glass-fiber filter and the pump to
capture the remaining vapor-phase BaP. Prior to sampling,
all parts of the sampling train were cleaned and rinsed with
anhydrous ethanol twice. This method was validated and
comparable to BaP measured using the U.S. EPA proposed
Modified Method 5, which is typically used for semivolatile
organic material sampling (18).
9
4 6 1 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004

3
3
TABLE 3. Concentrations (mg/m ) and Size Distributions of
TABLE 4. Concentrations of Air Pollutants in Flue Gas (mg/m )
Particles Emitted from Burning of H-Coal and C-Coal
during Burning of H-Coal and C-Coal
H-coal
C-coal
H-coal
SD
C-coal
SD
diameter (µm) mean SD
N
% of TSP mean SD
N
% of TSP
compound
SO2
mean
431.5
6.7
0.5
N
mean
862.0
8.2
0.3
N
PM2.5
PM2.5-10
PM10
TSP-PM10
TSP
38.4 3.8
10.2 6.6
48.6 5.4
10.2 5.9
58.8 2.4
3
3
3
3
3
65.4
17.3
82.6
17.4
100.0
8.5 4.3
18.9 8.7
27.4 7.6
59.1 9.2
86.5 5.3
3
3
3
3
3
9.8
21.8
31.6
68.4
100.0
196.0
1.4
0.3
3
3
3
246.0
1.0
0.1
3
3
3
NOx
fluoride
(gas-phase)
BaP
1.0E-03 8.9E-04
7
1.8E-04 1.9E-04
7
(PM2.5/ PM10) for the C-coal and the H-coal, respectively. The
OC contents in PM were 11.9%/ 7.3% (PM2.5/ PM10) and 2.5%/
4.0% (PM2.5/ PM10) for the C-coal and the H-coal, respectively.
The lower carbon contents of the PM2.5 and PM10 emitted
from the H-coal combustion might result from the lower
carbon content of the H-coal (Table 1) and higher combustion
efficiency. EC abundances are 2-6 times higher for C-coal
as compared to H-coal, suggesting incomplete combustion
in C-coal.
PM Elem ental Profile and Toxic Trace Elem ents. Source
profile data are important for source apportioning analysis.
The PM concentrations of five toxic trace elements and two
toxic ions (F- and Cr(VI)) are reported in Table 5. All these
toxic elements are included in the U.S. EPA hazardous air
pollutants (HAPs) list (21). All the measured air toxics
concentrations were higher in both PM2.5 and PM10 but lower
in particles larger than 10 µm in the C-coal emission. Lead
(Pb) and F- were the most abundant elements with con-
centrations 1 or 2 orders of magnitude higher than the other
elements in particles. PM2.5 arsenic (As), F-, and Cr(VI) were
also abundant elements in the C-coal, with concentrations
exceeding 1000 µg/ g. However, in the H-coal PM2.5 and PM10,
only Pb and F- had concentrations exceeding 1000 µg/ g.
The concentrations of the other elements in PM2.5 and
PM10 are reported in Table 6. Of those elements, Se, As, Cr,
Co, Cu, and Al are characteristic elements for emissions from
coal combustion. Except for Cu, the other five elements were
present at higher concentrations in the PM2.5 and PM10
emitted from the C-coal combustion.
After sampling, the probe nozzle and the sampling tube
were rinsed with 20 mL of doubly distilled cyclohexane for
5 times to recover BaP residuals condensed onto the sampling
parts. The cyclohexane extracts were combined with 50 mL
of cyclohexane for collection of vapor-phase BaP and stored
in a glass bottle. The glass-fiber filter was wrapped with
aluminum foil and stored at -20 °C for up to 1 week prior
to extraction with 50 mL of cyclohexane for 16 h at 98 °C
using a Soxhlet apparatus (19, 20). The cyclohexane extracts
were combined, evaporated to near dryness, and reconsti-
tuted in 1 mL of acetonitrile (ACN). The extracts were analyzed
using an (HPLC) system (model 638-50; Hitachi Co. Ltd,
Tokyo, Japan) with a fluorescence detector (RF-510, Shi-
madzu, Kyoto, Japan). The HPLC column used was an
Supelcosil LC-PAH column (4.6 × 250 mm) (Supleco Inc.,
Bellefonte, PA) under controlled temperature at 30 °C. The
chromatographic separation was performed under isocratic
conditions (methanol:water ) 90:10). The mobile phase flow
rate was 1 mL min-1. The injection volume was 10 µL. The
fluorescence detector was set at an excitation wavelength of
360 nm and an emission wavelength of 440 nm. BaP
concentrations were quantified through a calibration curve
prepared using external standards.
Determ ination of Em ission Rates and Em ission Factors.
Emission rates (mg/ h) of the measured pollutants were
calculated by multiplying the flue gas concentrations (mg/
m3) and the flue gas emission rate (m3/ h). Mass-based
emission factors (mg/ kg of coal) were obtained by dividing
the emission rates (mg/ h) with the fuel burn rates (kg/ h).
The energy-based emission factors (mg/ MJ) were derived
from mass-based emission factors and delivered heat values
(MJ/ kg).
Discussion
Sulfur dioxide is a major air pollutant in China. Epidemio-
logical studies in China have demonstrated that exposure to
SO2 is significantly associated with increased morbidity and
population risks of cardiovascular and respiratory mortality
(22, 23). In 1998, the annual average SO2 concentrations in
ambient air were found to be higher than 120 µg/ m3 in about
50 Chinese cities, as a consequence of burning untreated
raw coal with high sulfur content (11). Therefore, the primary
objective of the development of the H-coal was to reduce the
SO2 emission. In the present study, 15% of Ca(OH)2 was added
to the H-coal to capture the emitted SO2 at high temperature
(950-1100 °C) during the coal burning process. The produced
CaSO4 is fixed in the combustion slag and ash (24). In addition,
due to the addition of the combustion-promoting agent,
waterproofing agent, binding agent, and sulfur fixation agent
(Ca(OH)2), the sulfur content of the H-coal (1.31%) was lower
than that of the C-coal (1.58%) (Table 1). The transformation
of raw coal to the H-coal resulted in a significant SO2 emission
reduction as expected. The emission rate and mass-based
emission factor of SO2 for the H-coal were only about half
of those for the untreated C-coal (Table 7). Furthermore, in
generating 1 MJ of delivered heat, the H-coal emitted less
than one-third of SO2 from the C-coal. Therefore, the use of
the H-coal would substantially reduce the SO2 emissions,
even after taking account of the difference in sulfur content
between H-coal and C-coal.
Results
Concentrations of Air Pollutants in Flue Gas. The particle
concentrations and size distributions are listed in Table 3.
Although TSP concentration resulting from combustion of
the H-coal briquette was about two-thirds that from the
C-coal, the PM2.5 concentration for the H-coal was 4-5 times
higher. Approximately 65% and 83% of TSP from the H-coal
were PM2.5 and PM10, respectively, compared to 10% and
32% for the C-coal. Both the H-coal (17%) and the C-coal
(22%) produced similar portions of PM coarse (PM2.5-10). The
concentrations of PM2.5 and PM10 collected using the cascade
impactor were comparable to those using single-stage
impactors. On the basis of the results from the single-stage
impactors, the PM2.5 concentrations were 37.6 ( 2.7 and 7.4
( 1.5 mg/ m3, respectively, and the PM10 concentrations were
51.5 ( 6.1 and 23.9 ( 9.1 mg/ m3, respectively, for the H-coal
and the C-coal.
The concentrations of SO2, NOx, F- (gas), and B(a)P are
shown in Table 4. While NOx emissions were similar, the SO2
concentration emitted from burning the H-coal was nearly
half of those from the C-coal. An order of magnitude higher
B(a)P concentration was observed during the H-coal burning
however.
OC and EC Contents of PM2.5 and PM10. The EC contents
in PM were 3.9%/ 0.66% (PM2.5/ PM10) and 0.47%/ 0.24%
Particulate air pollution is a serious environmental and
public health concern in developing countries. In China,
9
VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 6 1 5

TABLE 5. Toxic Elements in Particles Generated from Burning H-Coal and C-Coala
H-coal
C-coal
in PM2.5
mean
in PM10
mean
in TSP-PM10
in PM2.5
mean
in PM10
in TSP-PM10
mean SD
element
SD
SD
58
71
320
mean
SD
SD
mean
SD
As
Cd
Ni
224
174
107
79
590
142
750
90
3 170
2 550
557
4150
490
47
2.7
30
30
30
20
40
2 370
313
1 680
30
678
2 050
302
659
25 500
3 940
365
710
54
720
5
380
6
0.01
30
283
144
570
1 660
231
10
466
730
300
4 300
3 670
60
Pb
5 120
2 990
253
5 900
2 580
434
2 080
180
29 900
4 420
1 210
379
10 100
4 310
1 650
40
1 650
1 050
640
40
100
2
F-
Cr
Cr(VI)
400
10
118
106
226
20
84.6
8
a
Unit in µg/g of PM.
TABLE 6. Elemental Concentrations in PM and PM Resulting from Burning H-Coal and C-Coala
2.5
10
H-coal
C-coal
in PM2.5
in PM10
in PM2.5
in PM10
element
mean
SD
mean
SD
mean
SD
bdl
2 780
3 350
12 100
725
80 600
bdl
847
3 470
550
bdl
35
4 100
29
87
11 400
584
mean
SD
Na
Mg
Al
Si
P
S
Cl
K
Ca
Ti
V
Mn
Fe
Co
Cu
Zn
Ga
Se
Br
Rb
Sr
Yt
127 000
bdl
1 320
3 490
bdl
302 000
105 000
174 000
562
bdl
53 300
bdl
1 870
1 850
bdl
84 200
149 000
31 200
152
bdl
2
150 000
bdl
34 200
bdl
77
bdl
2 400
1 480
1 570
3 940
394
126 000
49 200
5 410
1 480
bdl
bdl
19
1 840
64
75
30 900
519
1 620
4 160
39
4 170
2 560
611
1 090
565
27 700
85 300
4 130
223
bdl
2 960
14 700
33 900
816
367 000
bdl
11 900
5 760
628
bdl
29
7 770
31
332
947
bdl
865
bdl
41 800
169 000
13 400
258
bdl
bdl
2
128 000
126 000
130 000
472
bdl
1
bdl
bdl
32
410
53
31
781
11
279
2 700
76
259
763
1 030
1
18
18
1
27
5
782
13
255
7 450
67
315
1 540
989
bdl
337
2
52
79
46 200
901
2 190
261
128
36 300
808
1 260
7 200
9
1 260
108
32
1 050
80
4 870
29
95
678
452
24
125
bdl
57
1 840
61
bdl
bdl
2
98
80
bdl
147
1
9
14
bdl
bdl
12
40
6
bdl
23
7
Zr
12
37
Mo
Nd
Ag
In
Sn
Sb
Ba
La
Au
Tl
76
49
bdl
59
214
122
498
582
47
13
16
20
70
bdl
65
58
172
218
178
66
6
22
58
65
36
44
450
54
457
272
62
bdl
bdl
20
74
51
62
112
76
34
384
4
bdl
bdl
96
471
384
714
3 510
363
1 190
2 360
431
106
bdl
85
233
642
761
876
504
2 070
3 360
415
bdl
190
bdl
130
2 060
730
593
1 840
196
1
330
bdl
115
1 470
64
587
1 650
340
88
96
bdl
99
31
Ur
50
a
Unit in µg/g of PM. bdl, below detection lim ition.
approximately one million premature deaths are thought to
be attributable to particulate air pollution every year (1). The
use of coal and other solid fuels for residential cooking and
heating is a primary source of the indoor and outdoor
particulate air pollution (15, 25-27). Since there are still high
TSP levels in many outdoor and indoor environments in
China (>200 µg/ m3), monitoring ambient TSP is still an active
regulation requirement in China (26, 28). In 1996, SEPA issued
a revision of the ambient air quality standard (GB 3095-
1996) that requires daily or weekly monitoring of PM10
concentrations in major urban cities. We included measure-
ment of PM10 and PM2.5 emissions in this study, although
SEPA has not included PM2.5 in the current air quality
standards. However, exposures to fine PM (PM2.5) have been
associated with various health outcomes including mortality,
respiratory morbidity, and cardiovascular morbidity (29).
The emission rate, mass-, and energy-based emission
factors of TSP for the H-coal were lower than the corre-
sponding values calculated for the C-coal. As shown in Table
7, the energy-based emission factor of TSP for H-coal and
C-coal were 80.1 and 183.6 mg/ MJ, respectively. Therefore,
in generating 1 MJ of delivered energy, using the H-coal in
place of the C-coal would lead to a 56.4% reduction in TSP
emission. A recent study reported that combustion of different
honeycomb coals in Chinese household stoves could produce
50-80 mg TSP in generating 1 MJ of delivered energy (8).
These values are comparable to that for the H-coal used in
the present study. The combustion of the H-coal in the
household stove generated comparable PM10 to the C-coal
in terms of generating the same amount of delivered energy
(Table 7). However, the H-coal PM2.5 emission factor (energy-
based) was nearly 3 times that of the C-coal (Table 7). In
contrast, combustion of the C-coal produced a greater
amount of coarse particles (PM2.5-10 and TSP-PM10) and thus
led to a higher TSP emission factor. This finding suggests
that reduction of PM2.5 be considered in future modification
of the H-coal and for developing other new types of
formulated coals.
9
4 6 1 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004

TABLE 7. Emission Rates and Emission Factors of Measured Pollutants for H-Coal and C-Coala
emission rate (mg/h)
H-coal C-coal
mean mean
603 117
emission factor (mass-based, mg/kg)
H-coal C-coal
mean mean
401 97.8
emission factor (energy-based, mg/MJ)
H-coal C-coal
mean mean
52.4 18.1
pollutant
PM2.5
PM2.5-10
PM10
TSP-PM10
TSP
SD
59.1
SD
58.7
SD
39.4
SD
49.0
SD
SD
5.1
9.0
159
762
160
922
6770
105
102
260
377
815
1190
11900
113.0
4.8
119
104
127
73.0
3390
14.1
1.1
106
507
106
614
4510
70.2
5.5
68.6
56.1
61.4
24.8
2050
14.4
3.2
217
314
680
995
9920
94.3
4.0
99.9
87.1
106
60.9
2830
11.7
1.0
13.8
66.2
13.9
80.1
588
9.1
8.9
7.3
8.0
3.2
267
1.9
0.4
40.0
58.1
125
183
1830
17.4
0.7
18.4
16.1
19.6
11.2
522
2.2
84.2
92.1
37.2
3080
21.7
4.9
SO2
NOx
fluoride (gas) 8.3
0.7
0.2
BaP
1.6E-02 1.4E-02 2.5E-03 2.6E-03 1.1E-02 9.3E-03 2.1E-03 2.2E-03 1.4E-03 1.2E-03 3.9E-04 4.0E-04
a
A determ inistic approach was used to determ ine average em ission rates and factors. The standard deviations in the table only account for
the m arginal uncertainty in concentrations listed Tables 3-5 since larger variability in pollutant concentrations was observed than in param eters
listed in Table 2. A Monte Carlo approach m ay be helpful for estim ating the propagation of uncertainties in all param eters provided distributions
of these param eters are well-identified in future studies.
half that emitted by the C-coal. However, fluoride in PM2.5
emitted per 1 MJ delivered energy by the H-coal was about
two times that emitted by the C-coal (Figure 3). At the present
time, relative importances of fluoride in gas phase versus
particulate phase and in coarse versus fine particles are
unknown. Recent studies have shown that, besides direct
inhalation exposure to coal-released airborne fluoride,
ingestion of fluoride contaminated food due to deposition
of airborne fluoride was also a significant exposure route for
people who use coals for heating and cooking (33, 34).
Therefore, fluoride emitted in gas phase, in fine particles,
and in coarse particles may have different implications as to
which exposure route(s) is important.
Incomplete combustion of coal can produce a large
number of polycyclic aromatic hydrocarbons (PAHs). High
airbone PAH indoor cocentrations resulting from household
coal combustion has been associated with elevated incidence
of lung cancer in some regions of China (35, 36). The fuel-
weight based emission factor of BaP was 11 µg/ kg for the
H-coal combusiton and 2 µg/ kg for the C-coal combustion.
These values are at least 2-3 orders of magnitude lower than
the BaP emission factor for wood burning in the Unted States
(2-3 mg/ kg) (37). The energy-based emission factor for the
H-coal was also approximately 3 times that of the C-coal. A
recent study (38) indicted that the PAH emissions were
positively correlated with the volatile content of a fuel,
possibly a consequence of incomplete combustion for a fuel
with high volatile content, such as wood (39). The H-coal
had a higher volatile content (12.3%) than the C-coal (11.3%).
This might be a reason for the higher BaP emission factor
for the H-coal. However, PAH formation is a complex process
that can be affected by many factors including the organic
content of a fuel, combustion temperature, oxygen supplying
rate during combustion, and the structures of stoves, etc.
FIGURE 3. Emissions of measured toxic elements, in three PM size
fractions, from burning the H-coal and C-coal. (H, H-coal; C, C-coal;
the scale for F- and Pb is 1 order of magnitude higher than shown).
NOx are responsible for numerous environmnetal prob-
lems including photochemical smog, acid rain, water quality
deterioration, global warming, visibility imparment, etc. (30).
In addition, NOx can react with atmospheric organic chemi-
cals to produce various air toxics including nitroarenes and
nitrosamines. As shown in Table 7, the emission rate of NOx
for the H-coal was slightly lower than that for the C-coal. On
a fuel-mass basis, the H-coal produced an emission factor
about 26% smaller than the C-coal. However, in generating
the same amount of delivered energy, the H-coal produced
only about half the amount of NOx that the C-coal produced.
Toxic trace metals in PM can exacerbate health damage.
The emission factors of the selected trace metals on a basis
of delivered energy are plotted in Figure 3. For all particle
size ranges, the Pb emissions from the H-coal and the C-coal
appear to be at least 1 or 2 orders of magnitude higher than
the emissions from the other metals. As shown in Figure 3,
the emission factors of all the metals in TSP are higher for
the C-coal than for the H-coal. The reduction resulting from
using the H-coal in place of the C-coal are 76% for As, 41%
for Cd, 13% for Ni, 75% for Pb, and 15% for Cr. PM2.5 emission
factors for As, Pb, and Cr were lower but Cd and Ni were
higher for the H-coal. Figure 3 shows that the vast majority
of the Cd, Pb, and F- were associated with PM10.
In China, more than 30 million people have suffered from
dental and skeletal flurosis due to exposure to fluoride
compounds emitted from coal combustion (31, 32). Coal
combusion produces both gas- and particulate-phase fluo-
ride. By comparing the energy-based emission factors of gas-
phase and particulate fluoride for the tested coals (Table 7
and Figure 3), we found that 80% and 66% of fluoride emitted
from the H-coal and C-coal, respectively, were present in
the gas phase, with similar concentration levels in generating
1 MJ of energy (Table 7). Total particulate-phase fluoride
emitted per MJ of delivered energy by the H-coal was about
The emission factors of both PM2.5 As and PMcoarse As were
several times lower for the H-coal than for the C-coal. One
of the additives in the H-coal was hydrated lime (Ca(OH)2),
which was primarily designed to capture SO2 emissions.
9
VOL. 38, NO. 17, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 6 1 7

Particles; Landsberger, S., Creatchman, M., Eds.; Gordon and
Breach Science: Amsterdam, 1999; Vol. 1, pp 67-96.
(13) Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier,
C. A.; Purcell, R. G. Atmos. Environ. 1993, 27A (8), 1185-1201.
(14) NIOSH. Manual of Analytical Methods, 4th ed.; U.S. Department
of Health, Education, and Welfare: Washington, DC, 1994.
Recent studies have shown that lime and hydrated lime can
capture arsenic oxide from coal combustion and produce
Ca3As2O8 and Ca2As2O7, which have higher boiling points
than As4O6 and tend to be left in the combustion slag and
ash (40, 41). This may explain the orders of magnitude lower
emisssions of As from the H-coal. Due to the linkage between
arsenic compounds and harmful toxicological effects, reduc-
tion of arsenic emission from coal burning is of importance,
especially in some regions of China where arsenic poisoning
occurs because of the domstic use of the coal with high arsenic
content (as summarized in ref 2).
(15) Ge, S. Source apportionment of ambient PM2.5
samples in Beijing,
China. Ph.D. Thesis, Oregon Graduate Institute of Science and
Technology, Beaverton, OR, 1992.
(16) SEPA of China. Standardized Environmental Monitoring and
Analysis Methods; State Environmental Protection Agency of
China (SEPA): Beijing, China, 1992.
(17) Atkinson, R.; Arey, J. Environ. Health Perspect. 1994, 102 (S4),
117-126.
Chromium has three most common forms: metallic
chromium [Cr(0)], trivalent chromium [Cr(III)], and hexava-
lent chromium [Cr(VI)]. Metallic chromium does not occur
naturally. Trivalent chromium naturally occurs in rocks, soil,
plants, and animals, etc. When Cr(III)-containing materials
are heated (including coal burning process), Cr(III) is oxidized
to Cr(VI), which is the most toxic form and has been found
to cause a wide variety of adverse health effects including
carcinogenic effects in humans (42). For this reason, the
present study attempted to measure Cr(VI) in addition to
the measurements of Cr element in all forms. The energy-
based emission factor of PM2.5-bound Cr(VI) was slightly lower
for the H-coal than for the C-coal, while the energy-based
emission factor of TSP-bound Cr(VI) for the H-coal was only
half the value for the C-coal (Figure 3). By comparing the
concentrations of total Cr and Cr(VI) in PM2.5 bound (Table
5), we found that 47% and 31% of the total Cr was in the form
of Cr(VI) for the H-coal and C-coal emissions, respectively.
Finally, we want to state that the present study used only one
formulation of H-coal and C-coal from one region. Hence,
the generalization of the findings from the present study
should be extremely careful without a systematic evaluation
of representative samples of other types of honeycomb coal
and coal cakes made in other regions of China.
(18) U. S. EPA. Determination of particulate emission from stationary
Sources. Code of Federal Regulations, Title 40, Part 50. Appendix
A, Method 5, 1996.
(19) Farant, J. P.; Gariepy, M. Am. Ind. Hyg. Assoc. J. 1998, 59. 758-
765.
(20) Sanderson, E. G.; Farant, J. P. J. Air Waste Manage. Assoc. 2000,
50, 2085-2092.
(21) U.S. EPA. Study of Hazardous Air Pollutant Emissions from
Electric Utility Steam-Generating Units; EPA-453/ R-98-004A;
Government Printing Office: Washington, DC, 1998.
(22) Wang, X.; Ding, H.; Ryan, L.; Xu, X. Environ. Health Perspect.
1997, 105, 514-520.
(23) Venners, S. A.; Wang, B.; Xu, Z.; Schlatter, Y.; Wang, L.; Xu, X.
Environ. Health Perspect. 2003, 111, 562-567.
(24) Zhuang, Y. H.; Shen, D. X.; Xiao, P. L. Sulfur fixation during coal
briquette combustion. In Processing and Utilization of High-
Sulfur Coals III; Markuszewski, R., Wheelock, T. D., Eds;
Elsevier: Amsterdam, 1990; p 653.
(25) Fang, Y. Q.; Wang, Y. Q.; Wang, X. D.; Zhang, J. China Environ.
Sci. 1993, 4, 16-21.
(26) Florig, H. K.; Sun, G.; Song, G. Chemosphere 2002, 49, 1163-
1174.
(27) Sun, G. Effectiveness, efficiency and governance: an integrated
study of China’s air pollution management. Ph.D. Thesis,
Department of Engineering and Public Policy, Carnegie Mellon
University, Pittsburgh, PA, 2001.
(28) Zhang, J. J.; Hu, W.; Wei, F.; Wu, G.; Korn, L. R.; Chapman, R.
S. Environ. Health Perspect. 2002, 10, 961-967.
(29) HEI. Airborne Particles and Health: HEI Epidemiological
Evidence; Health Effects Institute: Cambridge, MA, 2001.
(30) U.S. EPA. Health and Environmental Impacts of NOx (http:/ /
epa.gov/ air/ urbanair/ nox/ hlth.html), 2003.
Acknowledgments
We thank the National Planning Committee of China for
supporting this study. Assistance and advice from following
individuals are greatly appreciated: X. S. Li, Y. W. Song, and
others from Research Institute of Physical & Chemical
Engineering of Nuclear Industry, Tian Jin, China. J.Z. was
supported in part by NIEHS Center Grant P30 ES05022.
(31) Hou, P. S. Fluoride 1997, 30, 229-232.
(32) Li, J. X.; Cao, S. R. Fluoride 1994, 27, 125-128.
(33) Ando, M.; Tadano, M.; Asanuma, S.; Tamura, K.; Matsushima,
S.; Watanabe, T.; Kondo, T.; Sakurai, S.; Ji, R.; Liang, C.; Cao, S.
Environ. Health Perspect. 1998, 106, 239-244.
(34) Ando, M.; Tadano, M.; Yamamoto, S.; Tamura, K.; Asanuma, S.;
Watanabe, T.; Kondo, T.; Sakurai, S.; Ji, R.; Liang, C.; Chen, X.;
Hong, Z.; Cao, S. Sci. Total Environ. 2001, 271, 107-116.
(35) Mumford, J. L.; He, X. Z.; Chapman, R. S.; Cao, S. R.; Harris, D.
B.; Li, X. M.; Xian, Y. L.; Jiang, W. Z.; Xu, C. W.; Chuang, J. C.;
et al. Science 1987, 235, 217-220.
Literature Cited
(1) Florig, H. K. Environ. Sci. Technol. 1997, 31, A274-A279.
(2) Finkelman, R. B.; Belkin, H. E.; Zheng, B. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 3427-3431.
(3) State Statistical Bureau (SSB). China Energy Statistical Yearbook,
1991-1996; China Statistical Publishing House: Beijing, 1998.
(4) Gammage, R. B.; Wachter, E. A.; Wade, J.; Wilson, D. L.; Ahmad,
N.; Sibtain, F.; Raza, M. Z. Environ. Int. 1993, 19, 133-145,
(5) Lan, Q.; Chen, W.; Chen, H.; He, X. Z. Biomed. Environ. Sci.
1993, 6, 112-118.
(6) Smith, K. R.; Apet, M. G.; Ma, Y. Q.; Wongsekiarttirat, W.; Kulkarni,
Energy 1994, 19, 587-600.
(7) Zhang, J.; Smith, K. R. J. Exposure Anal. Environ. Epidemiol.
1996, 6, 147-161.
(8) Zhang, J. F.; Smith, K. R. Environ. Sci. Technol. 1999, 33 (14),
2311-2320.
(9) Zhang, J.; Smith, K. R.; Ma, Y.; Ye, S.; Jiang, F.; Qi, W.; Liu, P.;
Khalil, M. A. K.; Rasmussen, R. A.; Thorneloe, S. A. Atmos.
Environ. 2000, 34, 4537-4549.
(36) Mumford, J. L.; Li, X.; Hu, F.; Lu, X. B.; Chuang, J. C. Carcinogenesis
1995, 16, 3031-3036.
(37) U.S. EPA. Residential wood stoves. In External Combustion
Sources in Compilation of Air Pollutant Emission Factors-42,
Vol. 1: Stationary Point and Area Sources, 5th ed.; U.S. EPA,
Office of Air Quality Planning and Standards: Washington, DC,
1996; Chapter 1.10.
(38) Oanh, N. T.; Reutergårdh, L. B.; Dung, N. T. Environ. Sci. Technol.
1999, 33, 2703-2709.
(39) Smith, K. R. Biofuels, Air Pollution, and Health: A Global Review;
Plenum Press: New York, 1987.
(40) Jadhav, R. A.; Fan, L. S. Environ. Sci. Technol. 2001, 35, 794-
799.
(41) Mahuli, S.; Agnihotri, R.; Chauk, S.; GhoshDastidar, A.; Fan, L.
S. Environ. Sci. Technol. 1997, 31, 3226-3231.
(42) Pellerin, C.; Booker, S. M. Enivron. Health Perspect. 2000, 108,
A403-406.
(10) Ge, S.; Bai, Z.; Liu, W.; Zhu, T.; Wang, T.; Qing, S.; Zhang, J. J.
Air Waste Manage. Assoc. 2001, 51, 524-533.
(11) Zhang, J.; Ge, S.; Bai, Z. J. Air Waste Manage. Assoc. 2001, 51,
534-541.
(12) Watson, J. G.; Chow, J. C.; Frazier, C. A. X-ray fluorescence
analysis of ambient air samples. In Elemental Analysis of Airborne
Received for review January 11, 2004. Revised manuscript
received June 23, 2004. Accepted June 25, 2004.
ES049942K
9
4 6 1 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 17, 2004