Transient puffs of trace organic emissions from a batch-fed waste propellant incinerator

Article abstract of DOI:10.1016/S0045-6535(00)00228-9

Emissions data have been obtained from a waste propellant incinerator. The incinerator is a dual fixed hearth, controlled air incinerator equipped with acid gas and particulate scrubbing. Puffing has been evident in this waste propellant incinerator by spikes in the CO concentration. Transient puffs of organics may travel down the combustion chambers and lead to stack emissions. The major conclusions from this study are that (1) transient puffs are formed due to the semi-batch feed nature of the combustion process (causing a local oxygen deficiency) and high water content of the desensitized propellant; (2) in batch-fed combustors, puffs can contribute to most of the organic emissions (which are relatively low) measured with US EPA sampling and analytical methods; (3) it is estimated that batch-fed combustion contributes up to 7-18 times more emissions than steady-state combustion will generate; (4) by applying dispersion analyses to determine the amount of oxygen deficiency in the flame zone, the combustion zone concentration of CO during batch-fed operation could be as high as 160,000 ppm, compared to a measured peak stack concentration of 1200 ppm CO; and (5) an organic sample is collected and averaged over at least a 2-h period that smooths out the transient peaks of organics emissions during batch-fed operation. For emissions that are associated with long-term potential health impacts, this is an appropriate sampling method. However, if a compound has a short-term potential health impact, it may be important to measure the time-resolved emissions of the compound.

Full text of DOI:10.1016/S0045-6535(00)00228-9

Chemosphere 42 (2001) 559±569
Transient pu€s of trace organic emissions from a batch-fed
waste propellant incinerator
*
John R. Hart
Department of Toxic Substances Control, California Environmental Protection Agency, 301 Capitol Mall, Sacramento, CA 95814, USA
Abstract
Emissions data have been obtained from a waste propellant incinerator. The incinerator is a dual ®xed hearth,
controlled air incinerator equipped with acid gas and particulate scrubbing. ``Pung'' has been evident in this waste
propellant incinerator by spikes in the CO concentration. Transient pu€s of organics may travel down the combustion
chambers and lead to stack emissions. The major conclusions from this study are that (1) transient pu€s are formed due
to the semi-batch feed nature of the combustion process (causing a local oxygen de®ciency) and high water content of
the desensitized propellant; (2) in batch-fed combustors, pu€s can contribute to most of the organic emissions (which
are relatively low) measured with US EPA sampling and analytical methods; (3) it is estimated that batch-fed com-
bustion contributes up to 7±18 times more emissions than steady-state combustion will generate; (4) by applying dis-
persion analyses to determine the amount of oxygen de®ciency in the ¯ame zone, the combustion zone concentration of
CO during batch-fed operation could be as high as 160,000 ppm, compared to a measured peak stack concentration of
1200 ppm CO; and (5) an organic sample is collected and averaged over at least a 2-h period that smooths out the
transient peaks of organics emissions during batch-fed operation. For emissions that are associated with long-term
potential health impacts, this is an appropriate sampling method. However, if a compound has a short-term potential
health impact, it may be important to measure the time-resolved emissions of the compound. Ó 2001 Elsevier Science
Ltd. All rights reserved.
Keywords: Waste combustion; Stack sampling; Solid propellants; Health risk assessment
1. Introduction
assessment and up to a million US dollars for an envi-
ronmental impact report.
There has been tremendous controversy regarding
the burning of hazardous waste. Special interest groups,
the public and political leaders have all publicly opposed
hazardous waste combustion, regardless of numerous
test results that show that it can be done safely. In
California, as in other parts of the world, millions of US
dollars have gone into environmental studies of each
proposed facility: more than a million US dollars for a
trial burn, up to a million US dollars for a permit ap-
plication, up to a million US dollars for a health risk
As the management of hazardous waste has become a
real economic burden to manufacturing industries, it has
become important to develop adequate and competitive
processes to burn organic hazardous waste streams. The
predominant method of disposing of waste propellant,
munitions, explosives and other energetic material is to
``open burn'' or ``open detonate'' the material. This is
done on pads, pans or on the earth. Uncontrolled
emissions are generated when propellants are open
burned, such as organics, metals, particulates and acid
gases. An alternative to the open burning was developed
at Aerojet, located in Sacramento, California to burn
propellant waste streams in an enclosed combustion
device. The research was overseen by the State of
* Tel.: +1-916-322-5265; fax: +1-916-323-3392.
E-mail address: jhart1@dtsc.ca.gov (J.R. Hart).
0045-6535/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 2 2 8 - 9

560
J.R. Hart / Chemosphere 42 (2001) 559±569
Fig. 1. Schematic diagram of AerojetÕs waste propelant incinerator.
California Environmental Protection Agency, Depart-
ment of Toxic Substances Control. The incinerator,
equipped with acid gas and particulate scrubbing, is part
of a two-unit system. The ®rst unit desensitizes propel-
lant to remove much of the oxidizer and to recover
soluble materials such as ammonium perchlorate and
alumina. This is done by ®rst cutting propellant from
rocket motors with a high-pressure water jet, followed
by placing chunks of the propellant into a maceration
unit that extracts the propellant from the binder me-
chanically (and also may include additional aqueous
extraction in the maceration unit). The second unit is the
incinerator which is the subject of this study. Fig. 1
presents a schematic of the Aerojet incinerator. Because
most of the oxidizer is removed from the propellant, air
is required to be fed to the incinerator.
which can include dissolved aqueous propellant waste
up to 379 l/h (100 gal/h). Operating temperature in the
primary chamber is ꢀ1283 K (1010°C). The temperature
of the gas to the scrubbers from the secondary com-
bustion chamber (SCC) is controlled by an additional
water quench. The gas temperature of the secondary
chamber is ꢀ1422 K (1149°C). The gas temperature is
cooled to ꢀ811 K (538°C) by two venturi scrubbers in
series.
The refractory-lined combustion chambers are
heated by auxiliary fuel. During this study, the fuel used
was natural gas. Each chamber is equipped with a
Pyronics gas burner. The combined total heat duty of
both chambers is 4.75 million kJ/h (4.5 million Btu/h).
Depending on the waste heating value, the maximum
waste propellant input rate to the incinerator is 170 kg/h
(374 lb/h). The air ¯ow and velocity in the primary
chamber is designed to minimize the entrainment of
particulates to the secondary chamber. In addition,
during the feed cycle, the auxiliary burner to the pri-
mary chamber is reduced to low ®re and the under®re
air is reduced to minimize particulate entrainment.
About 95% of the ash produced in the primary chamber
remains in the primary chamber. A camera is installed
to view the progress of the reaction of the waste pro-
pellant buckets.
2. Description of facility
The incinerator is a dual ®xed hearth, controlled air
incinerator originally manufactured by the MacBay. The
combustion chambers are side by side, with a shared
wall separating the chambers. A passage way is in the
wall at the end of the ®rst chamber for the combustion
gas to enter the second chamber. The volume of the
primary chamber is 6.6 m3 (230 ft3) and the secondary
chamber volume is 4.4 m3 (155 ft3). The cross-sectional
area of the primary chamber is about 1.32 m2 (14 ft2)
and the secondary chamber is ꢀ0.88 m2 (9.5 ft2), and the
overall lengths of the primary and secondary chambers
are about 5 m each (16 ft). Waste propellant is fed in 19-
or 23-l (5 or 6 gal) polypropylene buckets. Buckets are
fed through the loading door, then pushed through to
the ¯oor of the PCC by a ram feeder. Temperature can
be controlled in the primary chamber by a water quench,
Two venturi scrubbers, in series, comprise the air
pollution control equipment. Caustic is added to the
scrubber liquor. The ®rst stage di€erential pressure drop
is ꢀ7.1 kPa (29 in. H2O), and the second stage venturi
pressure drop is about 27.4 kPa (110 in. H2O). The
second stage venturi is of the collision-type design,
where the gas is split equally and then brought back
together in opposed ¯ow. An induced draft fan follows
the venturi scrubbers, and is capable of moving 71 m3/
min (2500 scfm). The stack contains ports for gas

J.R. Hart / Chemosphere 42 (2001) 559±569
sampling probes, and ports for continuous emissions 3.1. As-®red composition of propellants
561
analyzers. Emissions which are continuously monitored
and recorded are oxygen, carbon monoxide, total hy-
drocarbons, sulfur dioxide and nitric oxide.
Waste propellants are generated when the propellant
is cut out of a rocket motor and desensitized by a high
pressure water stream (called hydrolasing). The propel-
lant is broken into chunks by this process, and a part of
the propellant is dissolved into the water. Both parts of
the waste propellant, the aqueous and solid streams,
have been tested in the incinerator. Table 1 lists the
average composition of the desensitized propellants,
after they have been removed from the rocket motors.
The table lists only the primary constituents. Others not
listed are in trace quantities in the propellant. The
heating values of the desensitized propellants are also
listed, which range from 3000 to almost 19,000 kJ/kg
(1300±8170 Btu/lb).
The facility has received a Research, Development
and Demonstration permit from the US Environmen-
tal Protection Agency and the State of California
Environmental Protection Agency, Department of
Toxic Substances Control. Three sets of tests to mea-
sure detailed stack emissions have been performed on
the incinerator under the permits. Currently, Aerojet
does not plan to pursue further operation of the in-
cinerator.
3. Results
SICBM propellant is fed directly (stated as ``Direct''
in the table) and with lime (SICBM-Lime in the table).
Standard missile sustainer and deload propellant were
also tested. Note that water is the major constituent in
the desensitized propellants, which, in part, causes the
emission pu€s to occur.
The ®rst series of tests (Radian Corporation, 1992a)
were performed in October 1991 on Minuteman pro-
pellant containing ammonium perchlorate, aluminum
and polybutadiene. The second series of tests (Radian
Corporation, 1992b) were performed in October 1992
also on Minuteman propellant (data are averaged in the
tables). The third set of tests (GenCorp Aerojet, 1993)
were performed on waste propellant from Small Inter-
continental Ballistic Missiles (SICBM), Hawk Booster
and Sustainer missile propellants and Standard Missile
(denoted SM or Missile Sustainer and Deload in tables)
components.
3.2. Operating conditions of incinerator during tests
Table 2 contains the average operating conditions
reported during the tests. Except for Minuteman pro-
pellant tests, an average and standard deviation is pre-
sented for most of the parameters. Waste feed rates
Table 1
Composition of desensitized propellants ± primary constituents only
Compound
Function
Hawk
Booster
(%)
Hawk
Sustainer
(%)
SM
Sustainer
(%)
SM
Deload
(%)
SICBM-
Direct
(%)
SICBM-
Lime (%)
Minuteman
(%)
Water
Poly(1,2 butylene)
glycol
Maceration
Binder
30±50
10±20
30±50
10±20
30±50
30±50
65±85
55±70
30±50
15±25
Carboxyl term.
polybutadiene
C36 Aliphatic deme-
ryl di-isocyanate
Hexamethylene
di-isocyanate
Binder
5±10
1±3
2±6
Binder
2±5
2±4
Propylene glycol
Polyether polyol
Aluminum
Binder
Binder
Fuel
10±16
2±5
20±30
9
5±10
13.7
14.6
12±18
7.9
5±15
1±6
3±7
0±2
20±30
8±16
Ammonium
perchlorate
Oxidizer
4.1
Nitroguanidine
Polyethylene glycol
Cyclotetramethylene
tetranitramine
Calcium salts
Approx. heating
value (kJ/kg)
Oxidizer
Binder
0
21.9
5.4
1.84
1±6
3±8
<0.1
1%_5%
3±8
Treatment
4±19
3700
15,100
10,500
15,300
19,300
3000
18,800

562
J.R. Hart / Chemosphere 42 (2001) 559±569

J.R. Hart / Chemosphere 42 (2001) 559±569
563
varied for the tests from 50 kg/h (23 lb/h) for Hawk
Booster propellant to 155 kg/h (70 lb/h) for SICMB-
Direct propellant. The primary combustion chamber
the two test runs, except for the Minuteman tests. The
Minuteman emissions numbers represent an average of
six test runs, all at approximately the same operating
conditions.
(PCC) average temperature ranged from 1277
K
(1004°C) for the SICBM-Lime propellant tests to 1322
K (1049°C) for the Standard Missile Sustainer tests. The
SCC gas residence time ranged from 1.13 s for the
SICBM-Lime tests to 1.31 s for the Minuteman tests.
The SCC average temperature was a constant 1435 K
(1162°C).
Although laboratory blanks were not utilized during
the Minuteman tests, methylene chloride is a common
laboratory contaminant which could account for the
relatively high emission rate reported during these tests.
The relatively high benzene concentration during the
Minuteman tests could be due to its formation as a
product of incomplete combustion (PIC) from toluene.
3.3. Average continuous emissions monitoring results
3.6. Semi-volatile organic emissions
The average continuous emissions monitoring results
are presented in Table 3. Oxygen ranged from 11.1%
(dry) in the standard missile deload tests to 14.1% in the
Standard Missile Sustainer tests. Average carbon mon-
oxide (CO) ranged from 1 ppm in the Hawk Sustainer
tests to 31 ppm in the Hawk Booster tests. Average
emissions of total hydrocarbons were in the order of
1 ppm.
Average reported values of semi-volatile organic
emissions are listed in Table 5. Many of the analytes were
not detected, including some of the dioxin and furan
congeners. However, analytes detected were naphthalene,
acenapthene, ¯uorene, phenanthrene, anthracene, ¯uo-
ranthene, pyrene, chrysene, benzo(b)¯uoranthene, ben-
zo(g,h,i)perylene, benzo(e)pyrene, 2-methylnaphthalene
and dioxins and furans reported as international toxic
equivalents of 2,3,7,8-TCDD.
3.4. Sampling and analytical methods
During all of the tests reported in this paper, the
stack gas was extractively sampled and analyzed with
standard methods published by the US Environmental
Protection Agency. Particulates were sampled with a
Method 5 train (a heated ®lter followed by condensers).
Semivolatile organics (including dioxins) were sampled
by Method 23 (a modi®ed Method 5 train consisting of a
heated ®lter followed by condensers) and analyzed with
high resolution (capillary column) gas chromatography
(GC), followed by high resolution mass spectroscopy
(MS). Volatile organics were sampled by the Volatile
Organic Sampling Train (using two small sorbent traps),
and analyzed by GC/MS.
3.7. Destruction and removal eciencies
Destruction and removal eciency (DRE) is the US
combustion performance standard (which is 99.99% for
compounds other than dioxins and PCBs) for organics
emitted during hazardous waste combustion. It is a
measure of the most dicult to burn compound (prin-
cipal organic hazardous constituent, or POHC), both in
and out of the combustor, on a mass rate basis. These
compounds were fed (spiked) with the propellants dur-
ing the tests. Data are presented in Table 6 for DREs
that have been demonstrated on the incinerator, even
though the compounds selected will not be found in the
waste propellants in signi®cant quantities.
Tests included current quality assurance and quality
control measures adopted by the US Environmental
Protection Agency (US EPA, 1992, 1993, undated;
American Society for Quality Control, 1994).
4. Discussion of experimental observations
The following observations are relevant to the data.
4.1. Transient pu€s analysis
3.5. Volatile organic emissions
Volatile organic mass emission rate results are pre-
sented in Table 4. Many of the analytes were not de-
tected for the majority of tests (presented as ND ``not
detected'' and less than the limits of detectability). Some
compounds were in fact detected during the tests (please
refer to Table 4 for details): chloromethane, trichloro-
¯uoromethane, carbon disul®de, methylene chloride,
chloroform, carbon tetrachloride, benzene, toluene, tet-
rachloroethene, 1,1,1-trichloroethane, trichloroethene,
bromodichloromethane, m,p-xylene and o-xylene. Non-
detect values are averaged over the two test runs, and
the distinct emission values are reported separately for
The production of transient emissions has been
studied by Linak et al. (1987), on a pilot scale kiln
burning batch-fed packets of material. Cundy et al.
(1991), have studied the same transient pu€ phenome-
non in a full-scale rotary kiln. Under starved-air con-
ditions as in the PCC in the incinerator studied, organic
emissions can be high (on the order of 20,000 ppm of
total unburned hydrocarbons in the pilot-scale kiln
studied by Linak (1987)). A batch-fed drum of organics
often produces a local oxygen de®ciency in a combustion

564
J.R. Hart / Chemosphere 42 (2001) 559±569

J.R. Hart / Chemosphere 42 (2001) 559±569
565

566
J.R. Hart / Chemosphere 42 (2001) 559±569

J.R. Hart / Chemosphere 42 (2001) 559±569
567
Table 6
POHC emissions and DRE for 1,2,4-trichlorobenzene
where / is dimensionless time. For large dispersion
ꢁD=lL ˆ 0:2†, C/ is calculated as 0.7. Let us apply these
results to the generation of a relatively stable molecule
such as carbon monoxide. Time resolved data are
available for CO emissions. The highest peak concen-
tration found in the combustion exit (stack) gas stream
during any test was ꢀ1200 ppm with a decay to its
steady-state value in 7 or 8 min as seen on the CO stack
monitor during the Hawk Booster tests. In order for this
to occur under the model for small dispersion, a pu€ of
160,000 ppm would have been generated during 0.1 s
(chosen as the pu€ duration) within the combustion
chamber. For large dispersion, a pu€ of 110,000 ppm
carbon monoxide would have been generated during the
duration of 0.1 s. This is certainly possible in microscale,
low oxygen regions of the ¯ame zone found during
batch-fed operations.
Input rate
(kg/h)
Emission rate (kg/h)
DRE (%)
1,2,4-trichlorobenzene with SICBM direct propellant
4.46
4.09
4.18
0.658E)6
0.100E)6
0.581E)6
99.999985
99.999998
99.999986
chamber. Pung has been evident in this studyÕs waste
propellant incinerator by spikes in the CO and unburned
hydrocarbons concentrations. The two phenomena that
may result from this are:
1. Transient pu€s of organics may travel down the kiln
as a pulse of organic emissions, and be somewhat dis-
persed (``¯attened'') by the time the stack is reached.
2. Organic fragments from partial post-¯ame reactions
in the hot gases might result in reformation reactions
when the gases are quickly cooled at the ``cold'' end
(where a large amount of cool secondary air is intro-
duced) of the incinerator.
The ®rst set of tests on Minuteman propellant also
had CO spikes when buckets of propellant were fed to
the incinerator. The second set of Minuteman propellant
tests did not show the CO spikes because the under®re
air was adjusted and the ammonium perchlorate feed
rate was reduced by a factor of 3. A greater degree of air
staging was used in these tests (a greater percentage of
air was fed farther downstream in the primary chamber).
The CO could have still been produced in the ¯ame zone
where sub-stoichiometric conditions exist, and then
preserved (essentially ``frozen'') downstream when the
relatively cold secondary air was injected (and more
secondary air than primary air).
Equilibrium analyses have shown that no emissions
should be detected for a combustion system at equilib-
rium (Hart and Franco, 1989), a condition that does not
fully exist for full scale combustion systems. But, if a
pulse, or pu€, of emissions is generated from batch-fed
operations, it could travel down the incinerator as a plug
(Chang et al., 1987). The incinerator in this study ap-
proximates a plug-¯ow reactor because the Reynolds
number is ꢀ170,000, and the length-to-diameter ratio is
ꢀ5 for the combination of the primary and secondary
chambers (about 3 for the secondary chamber alone).
In addition to the spikes produced when buckets are
fed to the incinerator, CO spikes are produced when the
ash is raked by the ``ash ram''. The ash ram can be set to
operate periodically at the end of the feed cycle in order
to ensure that all of the propellant surface area is ex-
posed. The ash ram operates cocurrently to the gas ¯ow
in the PCC. The ash ram raked about once per four feed
cycles, and can produce as high or higher a CO and
hydrocarbon spike as feeding a bucket of propellant.
This spike can actually be higher than a feed spike be-
cause there may be more unburned propellants in the
unraked bottom ash pile than in a bucket of waste feed.
The raked bottom ash is dropped out of the PCC once
every 3 or 4 h. A CO and hydrocarbon spike occurs
(always at the same time) 3 or 4 min after an ash is raked
by the ram.
4.2. Dispersion model
A dispersion analysis can be done to determine the
amount of oxygen de®ciency in the ¯ame zone that
caused the emission pu€s. The dispersion of a pulsed
input function has been studied previously for a plug-
¯ow reactor by Levenspeil (1972). For dispersion with-
out chemical reaction (this will generally not occur for
most organics, but is useful in analyzing CO emission
peaks), the governing equation is
oC
ot
o2C
ox2
ˆ D
;
where D is an axial dispersion coecient, C the CO
concentration, t the time and x is the distance along the
axis. For small dispersion (the pulse or pu€ traveling
down the kiln intact, and d=ꢁlL† ˆ 0:002), the solution
takes the form
4.3. Dispersion with chemical reaction
For dispersion with chemical reaction (all other fac-
tors remain the same), the governing equation (Le-
venspeil, 1972) is
"
#
2
1
ꢁ1 /†
4d=lL
C/ ˆ
exp
1=2
ˆ 6:31;
1=2
ꢁ2ꢁpd† †=ꢁlL†
ꢁD=lL†ꢁo2C=x2† oC=ox ktCn ˆ 0:

568
J.R. Hart / Chemosphere 42 (2001) 559±569
For ®rst order reactions, and small deviations from plug
¯ow, the solution reduces to
2
C=C0 ˆ exp‰ kt ‡ ꢁkt† ŠꢁD=lL†;
where k is a ®rst order reaction rate constant, x the
distance and n is the reaction order. For values of
1
D=lL ˆ 0:002, k ˆ 0:91 s (from Glassman, 1977) and
t ˆ 2 s, C=C0 ˆ 7:66  10 4. Therefore, a transient pu€
of 593 g/h (1.3 lb/h) of benzene could be generated and
frozen to create a ®nal emission of 454 mg/h (0.001 lb/h)
of benzene at the exit during the Minuteman propellant
test, with dispersion and chemical reaction. One can see
that large transients are not required in order to produce
observed levels in the exhaust.
Fig. 2. Representation of a typical transient pu€ of CO emis-
sions.
It is important to note that sampling to quantify
trace organics occurs over at least a 2-h period. There-
fore, an average sample is collected and reported that
smooths out the transient peaks of organics emissions.
For emissions that are associated with long-term po-
tential health impacts, this is an appropriate sampling
method. However, if a compound has a short-term po-
tential health impact, it may be important to measure
the time-resolved emissions of the compound.
emissions generated from the batch-fed operation is a
signi®cant combustion system design issue.
4.5. Amount of waste feed rate is governed by heat release
and transient pu€s
A practical consideration for the design of a waste
propellant incinerator is that only a small of amount
waste propellant can be fed in the buckets, depending on
the heat of combustion of the propellant (and the
maximum allowable temperature of the combustion
chamber). The size of the incinerator combustion
chamber governs the amount of waste feed possible
because of the pu€s generated during the batch-fed op-
eration. An obvious solution is to modify the feed pro-
cedures to a steady-state operation, as long as safety
hurdles could be overcome. If the propellant could be
blown into the incinerator through a burner (either dry
or as a slurry), and/or the amount of water in the feed
could be reduced, the transient pu€ and heat release
problems would be minimized.
4.4. Estimation of emissions generated by transient pu€s
compared to steady-state
As some transient data are available, it is possible to
estimate the percentage of emissions generated by batch
feeding buckets of the waste propellant compared to a
continuous feed, steady-state operation. This can be
done by observing the frequency of buckets fed and
comparing the CO emission peaks against the steady-
state peaks.
Depending on the propellant and the percentage of
propellant in the desensitized propellant, the maximum
frequency of feed to the incinerator is about 10 buckets
each hour. This means that the number of CO peaks due
to the feeding of buckets (or pulses of reactants) is about
10 each hour (not including the rake of the ash ram).
During the Hawk Booster tests, ®ve substantial CO and
hydrocarbon spikes were observed (actually due to the
ash rake, not due to the feed of buckets) and recorded.
Strip chart traces indicate that the maximum peak
height is ꢀ1200 ppm CO and 375 ppm total hydrocar-
bons (by FID), and that the peak subsides to a steady-
state value after 7 or 8 min (see Fig. 2). Therefore, one
can integrate under the curve to estimate the emissions
generated by batch feed compared to steady-state op-
eration. It is estimated that there were 18 times more
emissions resulting from the batch feed compared to
steady-state, continuous feed operation based on the CO
spikes observed. Hydrocarbon spikes indicate that seven
times more emissions were generated by the batch op-
eration than the steady-state value. The increased
5. Conclusions
The conclusions from this study are:
·
Transient pu€s are formed due to the semi-batch feed
nature of the combustion process (causing a local ox-
ygen de®ciency) and high water content of the desen-
sitized propellant.
·
·
Batch-fed combustion in an enclosed device can be
safely done on waste propellants, but at a fairly low
feed rate (up to 155 kg/h in this study).
Volatile organics detected during the tests were:
chloromethane, trichloro¯uoromethane, carbon di-
sul®de, methylene chloride, chloroform, carbon tet-
rachloride, benzene, toluene, tetrachloroethene,
1,1,1-trichloroethane, trichloroethene, bromodichlo-
romethane, m,p-xylene and o-xylene.

J.R. Hart / Chemosphere 42 (2001) 559±569
569
Chang, D.P.Y., Mournighan, R.E., Hu€man, G.L., 1987.
Thermodynamic analysis of post-¯ame reactions applied
to waste combustion. In: Proceedings of the 13th Annual
Research Symposium on Land Disposal, Remedial Action,
Incineration and Treatment of Hazardous Waste, US
Environmental Protection Agency, Cincinnati, Ohio, 6±8
May, pp. 234±245.
·
·
Semi-volatile organic emissions detected were
naphthalene, acenapthene, ¯uorene, phenanthrene,
anthracene, ¯uoranthene, pyrene, chrysene, ben-
zo(b)¯uoranthene, benzo(g,h,i)perylene, benzo(e)py-
rene, 2-methylnaphthalene and dioxins and furans
reported as international toxic equivalents of 2,3,7,8-
TCDD.
Cundy, V.A., Sterling, A.M., Lester, T.W., Jakway, A.L.,
Leger, C.B., Lu, C., Montestruc, A.N., Conway, R.B., 1991.
Incineration of Xylene/Sorbent packs. A study of conditions
at the exit of a full-scale industrial incinerator. Environ. Sci.
Technol. 25 (2), 223±232.
In batch-fed combustors, pu€s can contribute to
most of the organic emissions (which are relatively
low) measured with US EPA sampling and analytical
methods.
GenCorp Aerojet, 1993. Trial Burn Report for the Propellant
Thermal Processor: Desensitized Class 1.3 and Class 1.1
Propellants. vol. I, September.
·
·
It is estimated that batch-fed combustion contributes
up to 7±18 times more emissions than steady-state
combustion will generate.
Glassman, I., 1977. Combustion. Academic Press, New York,
p. 20.
By applying dispersion analyses to determine the
amount of oxygen de®ciency in the ¯ame zone, the
combustion zone concentration of CO during
batch-fed operation could be as high as 160,000
ppm, compared to a measured peak stack concentra-
tion of 1200 ppm CO. When chemical reaction is con-
sidered in the dispersion analyses, the combustion
zone production of benzene could be as much as
593 g/h (1.3 lb/h) to create a stack emission of 454
mg/h (0.001 lb/h) of benzene during the Minuteman
propellant tests.
Hart, J.R., Franco, G., 1989. The use of combustion modeling
in permitting of hazardous waste incinerators. In: Proceed-
ings of the Incineration of Industrial Waste Third National
Symposium, San Diego, CA, 3 March.
Levenspiel, O., 1972. Chemical Reaction Engineering, second
ed. Wiley, New York, pp. 272±278.
Linak, W.P., McSorley, J.A., Wendt, J.O.L., Dunn, J.E., 1987.
On the occurrence of transient pu€s in a rotary kiln
incinerator simulator. I. Prototype solid plastic wastes.
J. Air Pollut. Control Assoc. 37 (1), 54±65.
Linak, W.P., McSorley, J.A., Wendt, J.O.L., Dunn, J.E., 1987.
On the occurrence of transient pu€s in a rotary kiln
incinerator simulator, II. Contained liquid wastes on
sorbent. J. Air Pollut. Control Assoc. 37, 934.
·
An organic sample is collected and averaged over at
least a 2-h period that smooths out the transient
peaks of organics emissions during batch-fed opera-
tion. For emissions that are associated with long-
term potential health impacts, this is an appropriate
sampling method. However, if a compound has a
short-term potential health impact, it may be impor-
tant to measure the time-resolved emissions of the
compound.
Radian Corporation, 1992a. Health Risk Assessment for the
Aerojet Propulsion Division Propellant Thermal Processor
RD&D Test Burns No. 1 and 2. vol. I (of II), April.
Radian Corporation, 1992b. Emissions from the Propellant
Thermal Processor RD&D Test Burn No. 2. March, 1992.
US EPA, 1992. Interim Draft EPA Requirements for Quality
Management Plans. EPA QA/R-2, Washington, DC, 20460,
July.
US EPA, Interim Final Guidance for Planning for Data
Collection in Support of Environmental Decision Making
Using the Data Quality Objectives Process. EPA QA/G-4,
Washington, DC, 20460.
References
American Society for Quality Control, 1994. Draft Speci®ca-
tions and Guidelines for Quality Systems for Environmental
Technology Programs. ANSI/ASQC E4-19xx, January
1994.
US EPA, 1993. EPA Requirements for Quality Assurance
Project Plans. EPA QA/R-5, Draft Final Report, July.