EPA-450/3-75-071
July 1975
AIR POLLUTANT EMISSIONS
FROM BURNING SUGAR CANE
AND PINEAPPLE RESIDUES
FROM HAWAII
U.S. ENVIROINMENTAL PROTECTION AGENCY
Of fire of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
-------
EPA-450/3-75-071
AIR POLLUTANT EMISSIONS
FROM BURNING SUGAR CANE
AND PINEAPPLE RESIDUES
FROM HAWAII
by
Ellis F. Darley and Shimshon L. Lerman
Statewide Air Pollution Research Center
University of California
Riverside, California
Grant No. R800711
EPA Project Officer: James H. Southerland
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, N. C. 27711
July 1975
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - as supplies permit - from the
Air Pollution Technical Information Center, Environmental Protection
Agency, Research Triangle Park, North Carolina 27711; or, for a fee,
from the National Technical Information Service, 5285 Port Royal Road,
Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by
Statewide Air Pollution Research Center, University of California, River-
side, California, in fulfillment of Grant No. R800711. The contents
of this report are reproduced herein as received from Statewide Air
Pollution Research Center. The opinions , findings, and conclusions
expressed are those of the author and not necessarily those of the Environ-
mental Protection Agency. Mention of company or product names is
not to be considered as an endorsement by the Environmental Protection
Agency.
Publication No. EPA-450/3-75-.071
-------
i iii
Abstract
Whole sugar cane, sugar cane leaf trash, and pineapple leaf trash from
Hawaii were burned in an instrumented burning tower to determine the emission
factors for particulate matter, carbon monoxide, and hydrocarbons. Analyses
of benzo(#)pyrene and the trace metals beryllium, cadmium, chromium, copper,
and nickel were made from a few whole cane fires. Particle size distribution
of particulate matter was determined in two cane leaf trash fires.
Emissions in terms of pounds per ton of fuel burned and pounds per acre
burned are given for sugar cane in the following summary table. Emissions from
pineapple trash are given only in terms of pounds per ton of fuel burned since
the fuel was not collected on an area plot basis. Yields of benzo(a)pyrene
are given on the basis of micrograms per gram of particulate matter and pounds
per acre. Yields from the metals are shown on the basis of picograms per cubic
-9
meter of air through the sampling probe and also as pounds x 10 per acre.
Yields of pollutants from sugar cane in terms of pounds per ton of fuel
burned agree quite well with the yields from a number of agronomic crops that
have been burned in the tower. The same is generally true of pineapple trash,
except that the CO yields is a little higher than from most other herbaceous
fuels of comparable moisture content.
There is little or no data with which to compare the yields of benzo
(a)pyrene. Some previous work done in the burning tower by EPA staff indicated
a yield of about 0.18 grams per ton of landscape refuse burned. Recalculating
the yield from the present study gives a value of about 0.22 grams per ton of
fuel burned.
About 90 percent of the particles from sugar cane leaf trash were less
than 0.5 yM in diameter.
-------
Emission Factor Summary table.
Part
CO
HC
Ibs/ Ibs/
ton acre
Sugar Cane
Whole cane
99% conf.
levela 6.0-8.4 92-133 60
Leaf trash
99% conf.
level9 4.1-6.5 33-47 47
Pineapple
Head fires
% moist. ,
dry wt.
basis
9.6 6.4
16.9 8.5
26.7 23.3
Back fires
% moist. ,
dry wt.
basis
9.3 6.4
16.5 7.7
24.4 9.1
B(a)P Ni Cr
vqt Ibs/ pg/ Ibs/. pg/
g acre m3 acre m3
Whole
cane 73.1 .009 1.92 4.86 .60 1
Leaf
trash 79.0 .003 1.02 1.94 .35
a Figures show range of true mean of popul
b -9
Lbs x 10 per acre for each of the five
c ug/g of part icul ate
Ibs/
ton
.1-81.
.7-71.
100. .1
105.4
129.9
107.4
112.5
116.7
Ibs/
acreb
.62
.61
ation
metal
Ibs/ Ibs/ Ibs/
acre ton acre
2 843-1383 4.7-16.0 82-222
2 392-562 2.3-14.4 27-112
4.6
5.9
12.3
3.7
6.0
7.2
Be Cd Cu
pg/ Ibs/ pg/ Ibs/ pg/ Ibs/
m3 acreb nr acre" mj acre0
.22 .60 1.96 5.48 6.73 17.53
.32 .70 2.03 4.65 6.48 13.91
using Student's 't1
s shown.
-------
Introduction
In mid-729 the Emission Factors Unit, National Source Inventory
Sectijn, EPA, Durham, North Carolina in response to solicitations for
assistance from EPA Region IX and Hawaii Officials, requested that the
Center's burning tower be used to determine the emissions from burning
sugar cane from Hawaii. The work would be supported through an amendment
to an existing EPA Research Grant R800711, "Air Pollution from Forest and
Agricultural Burning." This was appropriate because of the unique sampling
capabilities available at the University.
A meeting held in North Carolina to discuss feasibility was attended
by representatives from the Emission Factors Unit, EPA Region IX (San
Francisco), and the University. It was decided that such a study was feasible
but would require the close cooperation of the sugar growers and other agencies
in Hawaii. The decision was made to proceed to simulate the burning of sugar
cane to determine the emissions so decisions could be made regarding the
Hawaii State Implementation Plan.
Broad information concerning field burning of sugar cane was solicited
as a first step. Replies were received from the Director of Agricultural
Extension Service, University of Hawaii, the Chairman, State Board of
Agriculture, and the Director, Office of Environmental Control, State of
Hawaii. These letters indicated an enthusiastic willingness to cooperate
wherever needed, and as a result UCR and EPA representatives met in Honolulu
with representatives of the University, Hawaii Departments of Health (DOH)
and Agriculture (DOA), Hawaiian Sugar Planter's Association (HSPA), and the
Pineapple Growers' Association of Hawaii (PGAH). The burning tower facilities
at Riverside were described and several technical challenges countered and
satisfactory understanding and agreements achieved. In the discussion
-------
that followed, technical questions concerning weight of cane, ratio of
cane to leaf and the amount of material burned, cutting sample plots,
packaging, shipping, and arrangement of cane on the burning table were
considered. Since the burning table is 8 feet in diameter, and the scale
on which it rests has a fuel loading limit of 125 pounds, it was suggested
that cane from a 5 foot square plot could probably be set on the table.
Of prime importance was the need to have the cane arrive in Riverside in
as fresh a condition as possible which meant shipping by air freight.
Staff of DOA stated that we would probably need an import permit to move
cane onto the mainland. This proved to be the case and a permit was sub-
sequently obtained from the USDA Quarantine Office in Hoboken, New Jersey.
The principal quarantine requirement was that the cane be fumigated with
methyl brotaide under a 25 inch vacuum. Riverside county (California)
officials required that all residual cane be autoclaved at the conclusion
of each experiment.
Following a visit to cane fields during which time burning was observed
and sampling methods were again discussed, it was agreed that preliminary
work would be done to establish field sampling methods, moisture loss and
other possible physical effects of the fumigation, and packaging procedures.
Officials of HSPA and DOA would arrange with the growers for cutting cane
samples and delivering them to the DOA laboratories. DOA would assume
responsibility for the fumigation, packaging and shipping.
While the principal emphasis was to be placed on determining pollution
emission factors from sugar cane, it was requested and agreed that pineapple
trash also be burned in the tower at Riverside. Since the plant materials
from pineapple is dead and fairly dry, freshness was not a factor and the
material could be sent to the mainland by ship.
-------
3
Experimental Methods
Collection and Shipping Plant Material
A. Sugar Cane
The principal object of burning cane is to get rid of much of the
leaf material so that the cane stalk itself is relatively clean for factory
processing. Thus the great bulk of material consumed in a fire is dead leaf
material on the ground and those dead leaves still attached to the bottom
and midportions of the cane. Some green leaves in the top may also burn.
Therefore, in order to duplicate the field conditions as nearly as possible
on the burning table, sectioned whole cane, attached leaves, and leaf trash
on the ground were sent to Riverside.
Cane to be sent to Riverside was cut from commercial fields on Oahu on
the morning that the field was to be burned. Four plots measuring 5' x 5'
were selected about 40 feet in from the edge of the field and corner stakes
placed 2-1/2 feet on either side of the center of the planting furrow and
for a length of 5' along the furrow. All of the cane contained within the
vertical block above two of the plots was cut. As much of this cane was
too long and crooked to ship or place on the burning table, it was cut into
approximately 5-6 foot sections and segregated into the categories of bottom
cane, mid cane and top cane. Some pieces were necessarily much shorter.
Each category from each plot was weighed. The bottom cane was stripped
of the few leaves it may have had and then discarded because its inclu-
sion generally would have exceeded the weight limit for the burning table.
The leaves from the bottom cane were included in the leaf trash. The
mid and top cane as well as the loose leaf trash (top trash) and the
leaf trash on the ground (bottom trash) from a given plot constituted
the plant material for one fire. At the same time, top and bottom leaf
-------
4
trash only were collected from comparably sized plots so that emissions from
burning leaves alone without the cane could be compared with burning whole
cane. Each category of material was placed in separate polypropylene woven
bags and coded alphabetically by collection date. For example, D-l and D-2
were the two cane plots for the fourth shipment and D-3 was the leaf trash
material. All material was taken to the DOA laboratories, fumigated as noted
earlier, and then well aerated to remove all methyl bromide. The bags of
leaves were sealed in cardboard cartons and the bagged cane was rolled in
sheets of corrugated cardboard and bound. The packages were then delivered
to an air freight terminal in Honolulu.
A uniform scheduling was adopted to avoid weekend transit and to provide
a minimum time lapse between collection and tower burning of material. Cutting
was done on Monday or Tuesday, or both days, to provide a maximum of two ship-
ments during any given week. As an example, cane cut early Monday morning
was delivered to the freight terminal by late evening, flown to San Francisco
that night, arriving about mid-morning on Tuesday. Packages were transferred
to another plane which arrived in Ontario about 20 miles from Riverside, by
noon or shortly thereafter. We brought them to Riverside, held them overnight,
and burned the cane Wednesday morning. Tuesday shipments were burned Thursday
morning.
A total of 10 cane shipments (20 plots) and 19 leaf trash samples were
received from October through mid-December, 1972, at which time the sugar
harvest was stopped because of rains and to permit repair work in the factories.
It was intended to resume additional shipments for further burning trials by
March, but it was later decided that the immediate needs were satisfied with
the data obtained from the first set of fires.
The remaining two plots in each field were cut after the fire. Weights
of cane and ashes were obtained by DOA staff and compared with the amount
sent to R'verside in order to determine the weight of material consumed in
-------
5
the fire.
B. Pineapple Trash
Pineapple trash consists of the stumps and leaves that remain after
the final harvest of fruit from a given planting. Prior to replanting, the
stumps are knocked over, allowed to dry and then field burned to reduce
residue. DOA and PGAH staff, working with the growers determined average
weights and depths of fuel on the ground prior to burning so that fuel densi-
ties could be reasonably duplicated in Riverside. Rather than collecting
plant material on a plot basis as was done with sugar cane, two large bulk
shipments were made, and from these, individual fires were conducted using
16-20 pounds of material arranged in a 4' x 4' bed.
Loading Burning Table and Ignition
A. Sugar Cane
Several methods of loading the burning table were tried with the first
shipments of whole cane. A piece of hardware cloth bent in a corrugated fashion
to a height of about 2 inches and then cut to 5 feet square was placed on the
table. In a preliminary trial, plant materials were placed on the screen at
a uniform depth in as loose a pile as possible, intermixing leaves and cane
pieces; the cane was essentially horizontal. The pile was ignited with a
small propane torch near the bottom, along one 5-foot edge. The material did
not burn at all well and some leaves on the bottom of the pile were not consumed
even though the screen provided ventilation underneath.
A light weight aluminum rack 5 feet square was built. The corner posts
were 3 feet high and the side rails were permanently fastened at 12 inches from
the bottom. A vertically adjustable wire mesh frame was fitted to the corner
posts; the mesh measured 2x4 inches. With the wire mesh in place near the
top of the rack, cane pieces could be threaded vertically downward through
the mesh so their cut ends rest on the corrugated screen. For the next few fires,
-------
6
the fuel categories were placed on the table in the following order and their
cumulative weights noted: (1) bottom trash was placed loosely on the screen;
(2) mid-cane pieces were laid at an angle around the inside of the rack with
one end resting on the rack rail and the other on the bottom trash; (3) top
cane was threaded down through the wire mesh, mid-cane, and bottom trash, care
being taken not to compact the bottom trash; and (4) the top trash was worked
in loosely between the top cane pieces and on top of the mid-cane and bottom
trash. Since cumulative weights were noted, differences between successive
values gave the weights of individual fuel categories. The loaded fuel was
then ignited by applying the propane torch to the bottom trash along one 5-foot
edge.
The next three fires were conducted with the above loading method. They
burned very well and all succeeding cane fires were loaded in this manner.
In addition, after the fire, weights of top and mid cane were noted as these
pieces were removed from the table in order to give an idea of the weight
loss within these two categories. In some later experiments, each single
cane pieee, both top and mid, was individually tagged and the weight recorded
before and after the fire.
One important change should be noted that occurred about half way through
the program. In the earlier fires, when mid cane pieces were very short and
would not stand against the rail easily, attached leaves were removed and added
to the fire but the cane was discarded without recording the weight. Some top
cane was also discarded if too short to stand in the mesh frame. Later it was
decided that all cane shipped to Riverside should be put on the table. When
the emission data were analyzed, adjustments were made for the discarded cane
by methods discussed latter.
Where leaf trash alone was burned, head firing (with the wind) and back
firing (against the wind) was simulated by placing the fuel on a sloped tray
-------
7
and igniting from the bottom or top edges, respectively.
B. Pineapple Trash
No special loading techniques were needed with this fuel except to
use the head and back fire tray as noted above. Sufficient trash was shipped
to conduct 13 fires at a loading df 16-20 pounds. Since the fuel had dried
down while waiting to be burned, in some trials the trash was remoistened
with a fine spray and allowed to equilibrate in a large polyethylene bag.
Sampling Procedures
A. Particulate Matter
1) Total Particulate.
Total particulate matter was collected isokinetically on standard
Type A glass fiber filters held in a modified HIVOL sampler positioned just
just outside the burning tower as described by Darley et al. A pneumatically
controlled globe valve in the sample line continuously regulates the rate of
air flow through the filter so that approximately 1/776th of the total flow
out of the tower is sampled isokinetically. From the weight of particulate
matter on the filter, the yield based on weight of plant material burned in a
given fire can be calculated easily. From this value, the usual method for
all previous studies has been to express yields in terms of particulate per ton
of fuel burned. This was done in the present study. In addition, yield was
also expressed in terms of pounds of particulate per acre since all of the
combustible material on a measured plot was accounted for.
When the particulate sampling system was first installed, efficiency of
the single filter was checked by placing two filters back-to-back in the holder.
The second filter was always clean. Recently, it was suggested that perhaps some
Darley, E. F., S. Lerman, 6. E. Miller, Jr., and J. E. Thompson. Laboratory
testing for gaseous and particulate pollutants from forest and agricultural
fuels. Proc. Intern. Symp. "Air Quality and Smoke from Urban and Fdrest
Fires." Nat. Acad. Sci. In press.
-------
8
condensible materials were passing through the filter because the gases
passing through the filter were not yet at ambient temperature. Thus, late
in this study, a second HIVOL sampler was placed in the particulate sampling
line downstream from the first filter. It was found that additional particulate
was collected on the second filter. Since the second filter was not part of
the sampling system when the technical aspects of the program were agreed
upon in Honolulu, and further, since the two-filter system is still being
evaluated, only particulate collected by the single filter is reported in
this report.
Several filter papers were supplied to us by EPA. These filters had
been treated to permit analysis of benzo(a)pyrene as well as certain trace
elements. After collection of particulates from given fires, the filters
were returned to EPA for analysis.
2) Particulate sizing
Only a few samples were taken to determine particle size distribution.
More detailed work on sizing as well as particle morphology with the scanning
electron microscope was scheduled if the cane shipments had been resumed in
March, but, as noted above, it was subsequently decided to eliminate the
second set of shipments.
The instruments used for particle sizing were the Brink Cascade impactor*
loaned by EPA, and the Weathermeasure Cascade impactor, loaned by Agricultural
Engineering, University of California, Davis.
The Brink 5-stage impactor used a relatively low air flow rate of 3 liters
per minute and the sample cups of each stage are small but relatively heavy.
The impactor is essentially designed to collect from a steady-state source
over a long period of time in order to obtain sufficient material for accurate
*Mention of commercial products here and elsewhere does not constitute endorse-
ment th' -eof by EPA
-------
weighing. Since these fires lasted only 5 minutes, it was difficult to obtain
adequate samples. In one test, burning only the leaves of sugar cane, we
attempted to maintain a stedy-state fire by hand-feeding the leaves onto the
fire bed at a given rate over a period of about 25 minutes. To improve the
ratio between weight of the collection surface and the weight of particles
collected, the bottom of the collection cup of each stage was lined with an
aluminum fiol disc held in place with a retention ring. A specially designed
cyclone filter preceded the impactor and a Nuclepore membrane filter (0.4yM)
followed the last stage to provide a sixth stage for particle sizing. The
sampling probe was inserted into the stack at about the same position as the
sampling orifice of the isokinetic system.
The Weathermeasure impactor fits a standard HIVOL sampler. This feature
make its use much more feasible for sampling in the burning tower because
the high flow rate of up to 50 cfm permits an adequate sample in a short time.
The impactor was attached to a portable HIVOL sampler and this combined unit
was fitted with a suitable sample probe of such length to assure that collec-
tion was made at ambient temperatures. The probe was inserted into the stack
near the sampling orifice of the isokinetic system.. Again, sampling was done
only from the burning cane leaf trash.
B. Gases and Water Vapor
Carbon monoxide and carbon dioxide were sampled continuously on
separate Beckman Model 15A infrared gas analyzers and concentrations recorded
on Ester!ine-Angus strip chart recorders. Full scale for each gas analyzer
was 4000 and 50,000 parts per million, respectively. Total hydrocarbons
were analyzed continuously with a Beckman Model 109 hydrogen flame detector
hydrocarbon analyzer with concentrations being recorded on a Honeywell
Electronic 17 strip chart recorded. The analyzer was calibrated at full
scale of 1000 ppm. Integrating these values into calculations with air flow
-------
10
air flow and temperature at the collection site permits expressing results
in terms of pounds of pollutants per ton of fuel burned. As was done with
the particulate matter5 yields were also calculated in terms of pounds per
acre.
No attempt was made to determine sulfur dioxide because there is no
evidence to date to indicate that this pollutant is significantly involved
in agricultural burning.
-------
11
Discussion of Results
A. Sugar Cane
1. Yields of particulates, carbon monoxide and hydrocarbons.
Whole cane: As noted earlier, when the burning table was loaded
with whole cane, some mid and top cane pieces from about the first half of
the shipments were discarded because they were too short to place in an erect
position. In later fires all cane was used. When emission data on particulates
were calculated it was found that there was correlation between particulate:
yield and weight loss of the tops. Thus, it was possible to adjust the data for
the tops that were discarded since the weight of those tops could be fairly well
established by taking the difference between the weight cut from each plot in
Hawaii and the weight actually put on the table. Because all the leaves from
any discarded mid-cane pieces had been placed on the table, and because later
experiments demonstrated that weight loss of mid-cane pieces themselves was
neligible, it was decided that the discarded mid-cane would not have contributed
significantly to the emissions. Therefore the yields of particulate, carbon
monoxide and hydrocarbons presented in Table I take into account the probable
emissions from the discarded top cane pieces.
Particulate yields averaged 7.2 pounds per ton of fuel burned (112 pounds
per acre) with a standard deviation of 1.6 pounds (32 pounds per acre). At
the 99 percent confidence level, the true mean of the population would be
expected to fall between 6.0 and 8.4 pounds per ton of fuel burned (between
92 and 132 pounds per acre). The yield of particulates in pounds per ton of
fuel burned is not excessive and falls within the range of many other herbaceous
types of fuel that have been burned in the tower. One might have expected a
higher particulate yield because of the presence of the moist, green tissue in
the tops. The tops did burn and lose a significant portion of their total
weight, the range being from 8 to about 23 percent. One reason the yields
-------
12
Table 1. Yield of Participate Matter, Carbon Monoxide, and Hydrocarbon in Pounds of
Pollutants per Ton of Fuel Burned and Pounds per Acre when Burning Whole
Sugar Cane from Hawaii.
Plot number
B-2
C-l
C-2
D-l
D-2
E-l
E-2
6-1
G-2
H-l
H-2
M-l
M-2
N-l
N-2
0-1
0-2
P-l
P-2
Pounds
cane per
plot
218
172
184
230
241
211
211
163
193
123
139
138
148
132
142
93
122
122
143
Mean
Std. dev.
Using Student's 't1 values,
true mean of population
will fall between:
at 95% confid. level
at 99%
confid. level
Particulate
Ibs/ Ibs/
ton acre
__a
--
4.6
6.1
10.4
5.2
10.1
6.7
6.3
8.7
6.8
7.0
8.0
7.3
7.1
6.5
7.1
7.2
1.6
95
143
185
135
128
104
--
91
113
112
86
156
76
89
75
94_
112
32
Pollutant emissions
Ibs/ C0 Ibs/ Ibs/ HC Ibs/
ton acre ton acre
63.8
86.0
77.8
49.4
81.9
106.6
64.6
73.5
47.3
36.8
70.3
76.0
76.0
58.3
104.3
71.4
70.0
66.4
61.6
70.6
17.3
Particulate
6.3-8.1
6.0-8.4
95-130
92-133
62.9-78
60.1-81
722
1573
1694
865
1819
1838
1536
832
610
497
949
993
1258
711
2044
740
878
769
821
1113
481
CO
.3 881-1345
.2 843-1383
2.4
4.6
4.5
1.8
4.6
6.5
3.8
4.1
1.9
3.4
14.6
27.7
13.8
--
23.6
17.8
19.6
15.0
16.8
10.4
8.3
HC_
6.2-14.5
4.7-16.0
27
84
98
32
102
112
90
46
24
40
197
362
228
--
468
184
245
174
224
152
121
92-210
82-222
aFirst three particulate samples intentionally not taken; H-l was not isokinetic.
Hydrocarbon sample line came disconnected while sampling N-l.
-------
15
yield varied from 1.4 to 21.8 pounds per ton of fuel burned. From J-l
on, samples were burned chronologically as listed and a similar variation
is roted. Therefore we can only conclude that the range noted above is
normal for sugar cane burned in the tower for the fuel arrangements used.
The values themselves are not too different from burning cereal grain
straw and are at or below the value of 14 pounds/ton obtained from burning
a number of herbaceous and woody agricultural fuels in California.
Leaf trash: The pollutant emission yields, with the expected true mean
values at 95 and 99 percent confidence levels, are given in Table 2. The
yields of each in terms of pounds of fuel burned are a little less than from
whole cane. As noted above, the variation in hydrocarbon yield from leaf
trash is about the same as for whole cane.
As noted earlier, several of the leaf samples were burned as head and
back fires. Pollutant yields from the back fires were slightly less than
from head fires but the difference was not significant. This result agrees
with data from cereal grain straws wherein there was little or no difference
between head and back firing when the fuels were dry. With straws however,
as the fuel moisture level increased, particulate pollution from head firing
increased at a much greater rate than that from back firing.
2. Yields of benzo^)pyrene and some trace metals.
As noted above, special filters had been received from EPA which,
after sampling, were returned to the EPA laboratories for analysis of
benzo(a)pyrene and the trace metals nickel, chromium, beryllium, cadmium
and copper. The results of the analyses of filters used on whole cane fires
are given in Table 3; those from trash fires are given in Table 4. Benzo(a)
pyrene yields are expressed in micrograms per gram of particulate and pounds
per acre. Amounts of metals are expressed as picograms per cubic meter of
_g
air going out of the burning tower and as pounds x 10 per acre.
-------
Table. 3. Yield of Benzo(a) Pyrene and Certain Trace Metals Contained in the Particulate Matter from Burning Whole Sugar
Cane from Hawaii
Plot
number
C-l
C-2
D-l
D-2
E-l
E-2
G-l
G-2
H-2
M-l
M-2
N-l
N-2
Rounds
cane
per
plot
172
184
230
241
211
211
163
193
139
138
148
132
142
Mean
Std. dev.
B(a)P,
ug/g
0 _Q
Trace metals, pg/m and Ibs x 10
and Ibs/acre
M|/
ga
135.4
124.6
35.1
87.4
22.3
27.3
21.4
22.7
181.6
--
--
--
--
73.1
61.1
Ibs/
acre
.019
.018
.003
.012
.004
.003
.003
.002
.015
--
.009
.007
m
2.67
1.51
0
0
.93
2.71
2.21
2.75
1.47
1.00
4.16
2.39
3.22
1.92
1.24
Ni
Ibs/
acre
8.38
4.26
0
0
2.60
7.54
5.29
6.38
3.48
3.06
8.97
5.80
7.45
4.86
2.95
Cr
pg/
ITH
tr
tr
0
0
2.21
2.12
1.28
tr
0
1.15
0
.98
0
.60
.85
Ibs/
acre
tr
tr
0
0
6.17
5.90
3.06
tr
0
3.52
0
2.38
0
1.62
2.35
m^
.56
.42
0
0
.34
.47
.29
0
0
0
.32
.29
.20
.22
.20
Be
Ibs/
acre
1.76
1.18
0
0
.95
1.31
.69
0
0
0
.69
.70
.46
.60
.59
Cd
pg/
2.89
3.69
5.23
.32
1.27
1.48
1.08
1.32
1.57
1.19
2.30
1.43
1.74
1.96
1.30
/ acre
Cu
Ibs/
acre
9.07
10.41
10.17
8.40
3.55
4.12
2.59
3.06
3.72
3.64
4.96
3.47
4.03
5.48
2.89
P§/
5.89
7.53
9.68
7.18
1§.99
9.35
6.92
4.05
2.90
11.16
3.85
4.43
3.60
6.73
2.89
Ibs/
acre
18.49
20.73
18.83
18.84
30.70 _.
26.01
16.57
99.40
6.87
34.13
8.30
10.74
8.33
17.53
8.81
particulate collected
-------
13
higher than from leaf trash alone (see Table 2) may be done due to the open,
loose arrangement of the top leaves. We know from work with cereal grain
straws, wherein the fuel bed is much more compacted, that air dry material
may yield about 5 pounds of particulate per ton of fuel burned. If the
moisture content of these fuels is raised to about 50 percent on a dry
weight basis, the particulate yield may be increased by a factor of 3 or 4.
A second reason why particulate yields were not much higher may be that with
whole cane, quite a bit of large flakes of charred leaf material ("Hawaiian
Snow") was easily visible; this material was not produced when burning leaf
trash alone.
The carbon monoxide yield averaged 70.6 pounds per ton of fuel burned
(1113 pounds per acre) with a standard deviation of 17.3 pounds (419 pounds
per acre). At the 99 percent confidence level the true mean of the population
would be expected to fall between 60.1 and 81.2 pounds (843-1383 pounds per
acre). In terms of pounds per ton of fuel burned, this again is only a
moderate amount of CO and is similar to the yield from dry cereal grain straw.
As an example of the extreme in yields of CO that can be obtained from other
herbaceous fuels that have some green matter in them, asparagus fern and
tumbleweed approach 200 and 300 pounds, respectively.
The hydrocarbon yield averaged 10.4 pounds per ton of fuel burned (152
pounds per acre) but the standard deviation of 8.3 pounds is comparatively
large. Similarly the range of the expected true mean at the 99 percent
confidence level is large, being 4.7 to 16.0 (121 pounds per acre). The
fires were run chronologically in the order listed in Table 1. At first
glance it would appear that from sample H-2 on, there has been a decided
shift in the operation of the analyzing system. This conclusion however,
is not borne out by results of the leaf trash fires (see Table 2). Plots
D-3, F-3, 1-1, and 1-2 were burned on the same day and the hydrocarbon
-------
14
Table 2. Yield of Particulate Matter, Carbon Monoxide, and Hydrocarbon in Pounds of
Pollutant per Ton of Fuel Burned and Pounds per Acre when Burning Sugar
Cane Leaf Trash from Hawaii.
Pounds leaf
trash per Ibs/
Plot number plot ton
A-l
A-2
A-3
B-3
D-3
E-3
F-l
F-2
F-3
G-3
H-3
1-1
1-2
0-1
J-2
K-l
K-2
M-3
6.3
13.0
12.0
10.1
11.1
21.3
11.5
11.8
11.0
13.0
9.5
11.0
10.3
10.5
10.9
9.5
9.5
9.3
Mean
Std. dev.
Using Student's 't1 values,
true mean of population
will fall between:
at 951 confid. level
at 95
% confid. level
6.4
3.8
8.0
5.5
3.9
10.4
2.7
5.0
3.2
4.1
4.3
3.3
7.3
5.5
6.8
4.8
5.3
5.3
2.0
Particulate
Ibs/
acre
34
39
62
43
32
b
24
42
27
43
37
26
60
48
54
37
38
40
11
Particulate
4.3-6
4.1-6
.3 34-46
.5 33-47
Pollutant
Ibs/
ton
..a
64.7
75.1
75.9
36.4
96.4
49.2
41.1
50.1
53.6
53.8
57.3
64.5
46.2
64.7
62.7
59.4
15,3
emissions
CO
Ibs/
acre
__
--
507
621
--
330
798
420
426
361
458
422
469
567
362
496
459
477
126
Ibs/
ton
__
0.8
21.8
4.6
2.1
7.1
7.6
0.6
5.1
1.4
9.5
6.6
9.0
23.4
--
17.7
8.4
7.5
CO
50.7-68.2
47.7-71.2
401-552
392-562
4.1-12
2.3-14
HC
Ibs/
acre
__
6
180
19
59
65
6
37
12
74
54
79
183
130
70
61
HC_
.7 33-106
.4 27-112
First three gas samples intentionally not taken. Particulate from H-3 and hydrocarbon
from K-2 were invalid due to equipment maladjustment.
Load was ^ojbled for this fire, therefore, calculation on acre basis is not applicable.
-------
Table 4. Yield of Benzo(a)Pyrene and Certain Trace Metals Contained in the Participate Matter from Burning Sugar Cane
Leaf Trash from Hawaii.
Plot
number
B-3
D-3
E-3
F-3
G-3
1-1
1-2
M-3
Pounds
leaf
trash
per
plot
10.1
11.0
21.3
11.0
13.0
11.0
10.3
9.3
Mean
Std. Dev.
B(a)P
, ug/g
and Ibs/acre
vg/
ga
60.9
177.0
41.8
76.1
32.1
112.2
52.9
79.0
50.5
Ibs/
acre
.003
.006
b
.002
.001
.004
.001
.003
.002
Trace metals
Ni
pg/
m3
2.58
tr
1.66
0
1.03
.50
0
2.40
1.02
1.08
Ibs/
acre
5.63
tr
0
2.36
1.17
0
4.39
1.94
2.30
, pg/m
Cr
pg/
m3
tr
0
.96
tr
1.85
0
0
0
.35
.69
Ibs/
acre
tr
0
--
tr
4.24
0
0
0
.61
1.60
and Ibs x 10 /'acre
Be
pg/
m3
.59
.29
.31
.36
.62
.25
0
.16
.32
.21
Ibs/
acre
1.29
.67
.66
1.42
.59
0
.29
.70
.51
Cd
pg/
m3
3.55
1.67
1.03
1.56
1.75
1.51
2.89
2.27
2.03
.83
Ibs/
acre
7.75
3.88
2.88
4.02
3.53
6.38
4.15
4.65
1.75
Cu
pg/
m3
8.57
4.03
6.61
7.07
10.29
4.85
5.53
5.16
6.48
2.12
Ibs/
acre
19.10
9.37
12.91
23.61
10.72
12.20
9.44
13.91
5.41
ug/g of particulate collected
\oad was doubled for this fire, therefore, calculation on acre basis is not applicable.
-------
18
The means and standard deviations of B(a)P and of each metal from each
fuel type were about the same. There was, however, a great variation with-
in B(a)P, nickel, chromium, and beryllium. The three metals were completely
below detection in some samples and there is no identified explanation for
this. It is known that there is some condensation in the probe and this may
be a possible explanation; but no analyses have been made of the materials
adhering to the probe to determine quantities of metals retained.
There are no comparative data for the yield of metals from fires in the
tower so it is not possible to comment on the significance of the yield
figures. There are, however, comparative data for B(a)P. Several years ago
staff of the EPA used the tower to burn a variety of materials, one of which
was landscape refuse which included lawn clippings, leaves, and other plant
trash. From two such fires they obtained 0.31 and 0.13 grams per ton of fuel
burned. If the 7 trash fires, for which there are data on both particulate
and B(a)P yields, are combined, the yield of B(a)P is 0.18 grams per ton of
fuel burned. This is not too different from the mean of 0.22 grams produced
in the landscape refuse fires.
-------
19
3. Particle size distribution
One sample of participate matter was taken with the Weatherrneasure
5-staoe cascade impactor using a normal leaf trash fire. A Nuclepore membrane
filter (0.4 ym pore diameter) was installed below the 5th stage to assure
complete collection. For this impactor, particle cut-off sizes for each stage
2
were determined by calculations based on the theory developed by Marple.
A correction was made for a 50 cfm air flow and a mass density of 0.9 g/cc
was selected as a reasonable approximation. Particle size distribution is
plotted in Figure 1.
10.0
I
1,0
20 40 60 80 95 99 99,8 99.99
CUMULATIVE MASS PERCENT LESS THAN Dp
'Figure 1. Cumulative particle size distribution of smoke from
leaf trash of sugar cane burned at 20 percent moisture
(dry wt. basis). Particles collected with the
Weathermeasure HIVOL cascade impactor.
Most of the particles were sub-micron and about 90 percent were less than 0.5 y
Because of the high volume sampling rate, the Weathermeasure collects a large
sample so that a measureable amount of larger particles may be collected. Even
so a little less than 2 percent of the particles were larger than 2 ym.
2
Marple, V. A., A Fundamental Study of Inertial Impactors. Ph.D. Thesis Mech,
Eng. Dept. University of Minnesota, Dec. 1970
-------
20
A second test wa3 made with the Brink 5-stage cascade impactor using
leaf trash in a hand-fed steady state fire. Again, a Nuclepore filter was
installed below the 5th stage. The Brink collected particles only on the
4th and 5th stages and on the back-up filter. The distribution of the
particle size based on these data indicated that 92 percent of the particles
were below the 0.5 ytn range and 61 percent had a diameter below 0.25 ym.
The 92 percent figure agrees with the Weather-measure. Tmp'actor data.
1. Yield of particulates, carbon monoxide and hydrocarbons.
As noted in the section on methods, pineapple trash was burned on
the sloped tray, either as a head or back fire. Additionally, moisture was
added to the trash to raise the fuel moisture content to levels of about 15
and 25 percent, on a dry weight basis. The reason for wetting the fuel was
that the trash, including stumps, had become very dry between the date it was
received and the dates of actual burning. Whereas the leaves were dry and
brittle upon delivery to Riverside, the stumps were quite wet. The overall
moisture level of a trash sample would have been higher than that of the
leaves alone. It was therefore decided to moisten some of the trash in order
to obtain an indication of the effect of moisture on pollutant emissions.
Pollutant emissions from the several fires, Duncan's multiple range test
for significance of mean difference for head/back fires and moisture levels
are presented in Table 5. Yield data from two groups of fires are shown in
the table as a matter of interest but were not included in the analyses.
In one group, the fuel was heavily dusted with red soil and was thus atypical
of the bulk of trash sent to Riverside. In the other, moisture added to the
fuel was higher than intended and the fuel bed was difficult to ingnite.
-------
21
Table 5. Yield of Particulate Matter, Carbon Monoxide, and Hydrocarbon in Pounds
of Pollutant per Ton of Fuel Burned when Burning Pineapple Trash from
Hawaii as Head/Back Fires at Three Fuel Moisture Levels.
Head Fires
Fuel
Emissions,
* Part.
Low
8.8
10.4
10. 81
Medium
16.7
18.4
15.7
H1gh~
25.0
28.3
26. 91
30. 62
30. 72
Factors -
Variables
6.3
6.4
7.0
8.7
8.1
8.8
24.8
21.7
10.5
45.0
28.5
- 1.
2.
3.
CO
108
102
78
121
91
87
131
127
99
166
126
Back Fires
Ibs/
.7
.1
.0
.8
.0
.4
.9
.8
.3
.0
.0
HC
4.3
4.9
5.7
6.6
5.7
5.4
11.7
12.8
8.4
26.4
17.2
Duncan1
Head versus back fires
Moisture level low,
Moisture x (head/back)
Particulate,
%
8.1
10.4
12.31
15.6
17.9
16.1
25.0
23.8
28. 31
(no
s Multiple Range
medium, high (lo
Part.
6.2
6.6
7.0
6.5
7.4
9.2
8.2
CO
113.
100.
78.
112.
103.
121.
no.
9.9 122.
9.0 146.
back fires)
Test:
, med,
hi)
9
9
2
9
1
5
7
7
0
HC
3.
4.
4.
4.
5.
8.
6.
7.
7.
2
2
9
6
3
2
6
8
2
carbon monoxide, and hydrocarbon, Ibs/ton
Particulate
CO
Confidence level
Means
Head
Back
H1
Med
Lo
Head-hi
Back-hi
Head-med
Back-med
Back-lo
Head-lo
12.
7.
16.
8.
6.
23.
9.
8.
7.
6.
6.
11
71
15
12
38
23
05
53
70
40
35
95%
?
b
a
b
c
a
b
be
be
be
c
99%.
a
b
a
b
b
a
b
b
b
b
b
Means
110.1
112.2
123.3
106.3
106.4
129.9
116.7
105.4
112.5
107.4
100.1
HC
Confidence
Means
7.343
5.700
9.725
5.967
4.150
12.25
7.20
5.90
6.03
3.70
4.60
95%
a
b .
a
b
c
a
b
be
be
c
be
level
99%
a
a
a
b
b
a
b
b
b
b
b
1
Fuel heavily dusted with red soil; data not included in analyses.
Fuel at a higher moisture than intended and data are not included in analyses.
Screen inadvertently omitted from first fire.
Different letters within a factor grouping indicates significant difference between
factors; sane K-ttOi' inci'Icates no difference.
-------
22
The particu1 ate yield ranged from 6.2 pounds per ton of fuel burned with
a low-moisture back fire to 24.8 pounds with a high-moisture head fire. The
means of head and back fires were 12.1 and 7.7 pounds, respectively, and this
difference was highly significant. Increasing levels of moisture per se
resulted in an increased production of particulate and the differences were
significant at the 95 percent confidence level. At the 99 percent confidence
level, particulate yield from the high-moisture fuel was significantly higher
than that from either the medium- or low-moisture, but yields of the latter
two were not different from each other. In the two-way analysis of head/back
fires versus moisture level, high-moisture head fires gave significantly higher
particulate yield than all other combinations at the 99 percent level; the
other combinations were not significantly different from each other. At the
95 percent level, the low-moisture head fires produced less particulate than
high-moisture backfires.
It appears from these results that if the moisture level of leafy trash
is about 15 percent, the particulate yield should not exceed 9 pounds per ton
of material burned and may approach 6 pounds of the fuel moisture were reduced
to below 10 percent. These values are not too different from those obtained
from sugar cane. If it ever became necessary to burn under higher moisture condi-
tions, backfiring would have a definite benefit over head firing.
One of the two excessively high-moisture head fires demonstrated how high
the yield of particulate could be. In this fire, the aerating screen had been
left off of the burning table inadvertently. This might somewhat simulate
a field condition if the trash had not been fluffed up. Mechanically
stirring and fluffing up the trash is the normal practice. Particulate
yield was 45 po1 nds, which was almost double the mean of the high-moisture
-------
23
head fires. When the screen was used on the second fire, participate yield
decreased drastically.
Carbon monoxide yields varied from 87.4 to 131 pounds per ton fuel
burned. There was, however, no significant difference between head/back
firing, between the three moisture levels, or in the interaction of firing
direction and moisture. The overall mean of all fires was 111.1 pounds,
which is a little more than 50 percent greater than the mean of 70.6
pounds obtained from burning sugar cane.
The trend in the yield of hydrocarbon followed very closely that of
particulates, the least (3.2 pounds) being obtained from a low-moisture
back fire and the most (12.8) from a high-moisture head fire. The means
of head and back fires were 7.3 and 5.7 pounds, respectively, and the
difference was significant only at the 95 percent confidence level. At
the same confidence level, hydrocarbon increased with increasing moisture
and the difference was significant only at the 95 percent confidence level.
At the same confidence level, hydrocarbon increased with increasing moisture
and the differences were significant. At the 99 percent level, hydrocarbon
yield from the high-moisture fires was significantly different from medium
and low-moisture fuels, but the latter were not different from each other.
Again, as with the particulate yield, the two-way analysis of head/back
fires versus moisture showed that the 99 percent confidence level, only
the hydrocarbon yield from the high-moisture head fires was different from
all other combinations and none of the latter were different from each other.
At the 95 percent level, the yield of hydrocarbon from low-moisture back fires
was significantly less than that from high-moisture back fires.
If pineapple trash is burned at moisture levels of 15 percent or less,
the hydrocarbon should not exceed about 8 pounds but could approach 4 pounds
-------
24
if the moisture were reduced to 10 percent. At a high moisture level,
hydrocarbon yields would be reduced significantly if back firing was used.
The excessively high-moisture fire, wherein the aereating screen was
not placed under the fuel, again demonstrated how high the yield of hydro-
carbon might be. The 26.4 pounds produced from this fire was more than
twice that produced from the high-moisture fires.
No attempt is made to interpret the results obtained from those fires
using trash covered with red soil; the data are included for information.
Particulate yields were comparable to fuels without soil, except for the
high-moisture head fire wherein the yield was reduced by half. Red soil
on the fuel resulted in less CO except in the high-moisture back fire.
-------
25
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
, EPA-450/3-75-071
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Air Pollutant Emissions From Burning Sugar Cane and
Pineapple Residues From Hawaii
5. REPORT DATE
July 1975
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Ellis F. Darley and
Shimshon L. Lerman
8. PERFORMING ORGANIZATION REPORT NO.
9.
PERFORMING OAGANJZATION.NAME.AND ADQRESS
Statewide Air Pollution Research Center
University of California, Riverside
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
R800711
12. SPONSORING AGENCY NAME AND ADDRESS
U. S. Environmental Protection Agency
Office of Air Quality Planning and Standards, MDAD
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Whole sugar cane, sugar cane leaf trash, and pineapple leaf trash from Hawaii were
burned in an instrumented burning tower to determine the emission factors for partic-
ulate matter, carbon monoxide, and hydrocarbons. Analyses of benzo(a) pyrene and the
trace metals beryllium, cadmium, chromium, copper, and nickel were made from a few
whole cane fires. Particle size distribution of particulate matter was determined in
two cane leaf/trash fires.
>Emissions in terms of pounds per ton of fuel burned and pounds per acre burned are
given for sugar cane in the following summary table. Emissions from pineapple trash
are given only in terms of pounds per ton of fuel burned since the fuel was not col-
lected on an area plot basis. Yields of benzoCa)pyrene are given on the basis of
micrograms per gram of particulate matter and pounds per acre. Yields from the metals
are shown in the basis of picograms per cubic meter of air through the sampling probe
and also as pounds x 10"^ per acre.
Yields of pollutants from sugar cane in terms of pounds per ton of fuel burned
agree quite well with the yields from a number of agronomic crops that have been
burned in the tower. The same is generally true of pineapple trash, except that the
CO yields is a little higher than from most other herbaceous fuels of comparable
moisture content.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Benzo(a)pyrene
Hydrocarbon
Berylluim
Chromium
Nickel
Carbon Monoxide
Particulate matter
Cadmium
Copper
Emissions
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
30
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |