EPA/600/A-96/044
National Waste Processing Conference ASME 1996
CONTROL OF AIR POLLUTION EMISSIONS FROM
MUNICIPAL WASTE COMBUSTORS
James 0. Kilgroe
U. S. Environmental Protect ton Agency
National Risk Management Research Laboratory
Research Triangle Park, North Carolina
Anthony Licata
Licata Energy and Environmental Consultants, Inc.
Yonkers, New York
ABSTRACT
The November 1990 Clear Air Act Amendments (CAAAs)
directed EPA to establish municipal waste combustor (MWC)
emission limits for particulate matter, opacity, hydrogen
chloride, sulfur dioxide, nitrogen oxides, carbon monoxide,
dioxins, dibenzofurans, cadmium, lead, and mercury. Revised
MWC air pollution regulations were subsequently proposed
by EPA on September 20, 1994, and promulgated on
December 19, 1995. The MWC emission limits were based
on the application of maximum achievable control
technology (MACT), This paper provides a brief overview of
MWC technologies, a summary of EPA's revised air pollution
rules for MWCs, a review of current knowledge concerning
formation and control of polychlorinated dibenzo-/?-dioxins
and polychlorinated dibenzofurans, and a discussion of the
behavior and control of mercury in MWC flue gases.
INTRODUCTION
In June 1987, EPA announced its intention to develop new
air pollution rules for MWCs.' This decision was based, in
part, on a study of the potential environmental risk
associated with MWCs.2 Pollutants posing the highest risks
included polychlorinated dibenzo-p-dioxins (PCDDs).
polychlorinated dibenzofurans (PCDFs), and hazardous trace
metals. On December 20, 1989, EPA proposed New Source
Performance Standards (NSPS) for new MWCs and Emission
Guidelines (EGs) for existing MWCs.' NSPS and EGs for
MWCs larger than 225 Mg/day in capacity were promulgated
in February 1991.4
* This paper has been reviewed in accordance with the U.S
Environmental Protection Agency's peer and administrative
policies and has been approved for publication.
The November 1990 CAAAs directed EPA to establish MWC
emission limits for particulate matter (PM), opacity,
hydrogen chloride (HC1), sulfur dioxide (SO2), nitrogen
oxides (NOx), carbon monoxide (CO), PCDDs/PCDFs,
cadmium (Cd), lead (Pb), and mercury (Hg),^ Revised MWC
air pollution regulations were subsequently proposed by EPA
on September 20, 1994, and promulgated on December 19.
1995.These emission limits were based on the
application of MACT. For existing units. MACT is defined as
the best emission limit achieved by 12 percent of the
operating units in a category such as large or small units. For
new units, MACT is defined as the best emission limit
achieved by the best single unit in a category of units.-*
This paper provides a brief overview of MWC technologies,
a summary of EPA's revised air pollution rules for MWCs, a
review of current knowledge concerning formation and
control of PCDDs/PCDFs, and a discussion of the behavior
and control of Hg in MWC flue gases. The focus of the paper
is on the performance of combustion and flue gas cleaning
technologies used at MWC facilities in controlling emissions
of PCDDs/PCDFs and Hg.
MWC TECHNOLOGIES
Three major types of MWCs are commonly used in the U.S.;
field-erected mass bum incinerators, refuse-derived fuel (RDF)
combustors. and factory-constructed modular mass burn
incinerators.' The best combustion technologies ensure
adequate waste burnout and produce minimal products of
incomplete combustion (PICs) in the flue gas.
In the U.S.. dry flue gas cleaning technologies arc generally
used to control air pollution emissions.'' Wet scrubbing
systems arc seldom used. PM is typically collected in

-------
*
TABLE 1. BASIS FOR MACT PERFORMANCE REQUIREMENTS3
(DECEMBER 19, 1995)8-9
Proposed Requirements	Basis for Emission Limits^
Emissions Guidelines (EG) - Existing Plants
Small (>35 to 225 Mg/day)
Large (>225 Mg/day)
New Source Performance Standards (NSPS) - New Plants
Small (>35 to 225 Mg/day)	GCP + SD/FF + CI
Large (>225 Mg/day)	GCP + SD/FF + CI + SNCR
* Technologies which provide equivalent or better performance may also be used.
b GCP
= Good Combustion Practice
DSI
= Dry Sorbent Injection into the Combustor Furnace or Flue Gas Duct
ESP
= Electrostatic Precipitator
FF
= Fabric Filter Baghouse
CI
= Carbon Injection
SD/ESP
= Lime Spray Dryer Absorber and ESP
SD/FF
= Lime Spray Dryer Absorber and Fabric Filter Baghouse
SNCR
= Selective Non-catalytic Reduction
c No NOx control requirements for small MWC plants or large existing mass burn refactory eombustors
GCP + DSI + ESP (or FF) +Q
GCP + SD/ESP (or SD/FF) + CI + SNCRc
electrostatic precipitators (ESPs) or fabric filters (FFs). Most
trace metals such as Cd and Pb are solids at flue gas cleaning
temperatures and are efficiently collected in ESPs or FFs. Hg
is normally a vapor at flue gas cleaning temperatures, and
special methods must be used for its control.'^'"
Acid gases are controlled in dry or semi-dry scrubbers by
injecting either a calcium- or sodium-based reagent into the
flue gas to convert HCl and SO2 into solid compounds that
can be collected in a PM control device.""^ The most
commonly used reagents are quicklime (CaO), hydrated lime
[Ca(OH)2]. limestone (C2CO3), and sodium bicarbonate
[Na(C03>2]. Several methods may be used to inject and mix
the sorbent with the flue gas: lime spray dryers (SDs), dry
sorbent injection (DSI) of limestone or lime into the
combustor furnace or the flue gas duct. Selective non-
catalytic reduction (SNCR), the most advanced NOx control
technology being applied in the U.S. on MWCs, uses either
ammonia (NHj) or urea (CO^Hj^] as a reagent to reduce
NOx to nitrogen.1^
The chemistry involved in acid gas and NOx control and the
mechanisms involved in PM control are generally well
understood. The major problem associated with control of
these pollutants (PM, acid gases, and NOx) is the engineering
optimization of flue gas cleaning processes that are also
effective in the control of trace metals and trace organics.
The most difficult to control MWC pollutants of concern are
PCDDs/PCDFs and Hg. PCDDs/PCDFs can be formed in
MWCs as high temperature PICs, or they can be formed
downstream of the combustor by low temperature synthesis
reactions involving fly. ash.'"' Hg, typically a vapor at
flue gas cleaning temperatures, is difficult to collect in flue
gas cleaning equipment.
FEDERAL EMISSION CONTROL
REQUIREMENTS
On December 19, 1995, EPA promulgated revised NSPS for
new MWCs and EGs for existing MWCs7~ These revised
rules require the use of good combustion practice (GCP) and
MACT flue gas cleaning techniques to continuously limit
emissions of PCDDs/PCDFs, CO, PM. Cd, Hg, Pb, HCl, SOj.
and NOx. The control technologies on which the per-
formance requirements are based are summarized in Table I.
Alternatively, technologies which can provide equivalent or
better performance than those on which the standards are
based may also be used.
The control of PCDDs/PCDFs is based on the use of GCP in
combination with appropriate flue gas cleaning techniques.
The objectives of GCP are to maximize furnace destruction of
organics and minimize low temperature PCDD/PCDF
formation reactions.**'-18 purnace destruction of organics is
controlled by establishing technology-based emission limits
for CO (see Table 2). Downstream formation and stack
emissions are controlled by a limit on steam load for waste-
to-energy plants and a limit on PM control device operating
temperature. The load is a surrogate parameter used to limit
2

-------
TABLE 2. GOOD COMBUSTION PRACTICE FOR MWCs
(DECEMBER 19, 1995)8-9
1, CO Requirements
EG Limit, ppm NSPS Limit, ppm
(Ave. Time, h)	(Ave. Time, h)
Mass Bum Water Wall (MBWW)
100 (4)
100
(4)
Mass Bum Refractory Wall (MBRW)
100 (4)
100
(4)
Mass Burn Rotary Water Wall (RWW)
250(24)
100
(24)
Mass Bum Rotary Wall Refractory (RWR)
100 (24)
100
(4)
Refuse-Derived Fuel Stokers (RDF)
200 (24)
150
(24)
Fluidized Bed Combustors (FBC)
100 (4)
100
(4)
Modular Combustion Units (MCU)
50 (4)
50
(4)
Coal/RDF Mixed Fuel-Fired



- Spreader Stokers (Coal-RDF/SS)
250 (24)
150
(24)
- Pulverized Coal (Coal-RDP/PC)
150 (4)
150
(4)
2.	Load not to exceed maximum load demonstrated during most recent PCDD/PCDF compliance tests.
3.	PM control device inlet temperature not to exceed a temperature 17 C above the maximum temperature demonstrated during most
recent PCDD/PCDF compliance tests
4.	Chief facility operator, shift supervisors, and control room operators must meet training and certification requirements.
the relative amount of PM carried out of the combustor with
flue gas (PM carryover). The purpose of the inlet temperature
limit is to control PCDD/PCDF synthesis reaction rates, and
solid- and vapor-phase partitioning of PCDDs/PCDFs in the
PM control device. Solid-phase PCDD/PCDF emissions are
limited by the use of efficient PM control equipment such as
high performance ESPs or FFs.
All MWC plants must comply with an Hg emission limit of
80 ug/dscm or an 85 percent reduction in Hg emissions.** The
Hg emission limit is based on the use of powdered activated
carbon in conjunction with dry scrubbing. The activated
carbon adsorbs gas-phase Hg and is collected in the PM
control device. Acid gases (HC1 and SO2) and metal (Cd,
Hg, and Pb) emission limits require equipment and operating
conditions that are also needed for PCDD/PCDF control.
Acid gas sorbents may reduce PCDD/PCDF formation rates
nd allow for reductions in the PM control device operating
temperature. The use of activated carbon for Hg control
improves PCDD/PCDF control.
Hie 1995 EPA emission requirements for new and existing
sources are summarized in Tabic 3. The rules have different
requirements for new and existing plants, and for small (> 35
to 225 Mg/day) and large (> 225 Mg/day) plants. Emission
limits are expressed either in mass concentration per dry
standard cubic meter (dscm) or parts per million on a dry
volumetric basis (ppmv), corrected to 7 percent Oj and
standard conditions at 20 C (68 *F) and 101.3 kPa (14.7 psi)
The PCDD/PCDF limits are: 125 ng/dscm for small existing
plants, 60 ng/dscm for large existing plants with ESP-based
air pollution control systems, 30 ng/dscm for existing non-
ESP-based systems, and 13 ng/dscm for large and small new
plantsA7 Each MWC will be subject to annual PCDD/PCDF
compliance tests unless they qualify for less frequent testing.
MWC plants are allowed to conduct PCDD/PCDF performance
tests on only one unit per year if all units achieve emission
levels for 2 consecutive years of 30 ng/dscm for small
existing plants, 15 ng/dscm for large existing plants, and 7
ng/dscm for all new plants.^
Although PCDD/PCDF emissions cannot be continuously
monitored, operating and emission parameters which
correlate with PCDD/PCDF emissions will be continuously
monitored and controlled. These continuous compliance
parameters include CO emission limits, boiler steam load, PM
control device inlet temperature, and activated carbon hourly
feed rates. Opacity and SOj are to be continuously monitored
to guarantee proper operation of the flue gas cleaning
equipment. The use of these continuous monitoring and
compliance parameters will ensure that each MWC operates at
3

-------
TABLE 3. EMISSION LIMITS FOR MWCsa
(DECEMBER 1995)8-9
Guideline Limits * Existing Plants NSPS Limits - New Plants (or
	(or % Reduction)	% Reduction)	

Small
Large
Large and Small
Pollutant/
>35 to 225 Mg/Day
>225 Mg/Day
235 Mg/day
Measurement



ng/dscm [ J*5
125 {30]
60c[15]
13 [7]
PM,mg/dscm
69
27
15
Opacity*1, %
10
10
10
Cd, mg/dscm
0.10
0.04
0.01
Pb, mg/dscm
1.6
0.50
0.10
Hg, mg/dscm
0.08 (85)
0.08 (85)
0.08 (85)
HCl, ppmv
250 (50)
35 (95)
25 (95)
S02, ppmve
80 (50)
35 (75) '
30 (80)
NOx, ppmv15
Hone
200 -250d,f
150d-8
* All emissions corrected to 7 percent Oj.
k Average of three stack tests using EPA Method 23. Values are weight of total tetra- through octa- congeners. Values in brackets
for [emission limits to qualify for less frequent testing]
c Emission limit for ESP-based air pollution control systems. Non ESP-based systems must comply with a 30 ng/dscm limit or
the "less frequent testing" requirement.
d EPA Method 9. Limit for 6-minute averages,
e 24-hour averaging lime.
f 200 ppmv for MB WW. 250 ppmv for RWW, 250 ppmv for RDF, 240 ppmv for FBC. no NOx control requirement for MBRW.
and 200 ppmv for others.
8 Applies to large plants only. 150 ppmv, except 180 ppmv is allowed for the first year of operation.
conditions necessary to control emission of PCDDs/PCDFs,
Hg, and other regulated pollutants.
OIOXIN AND MERCURY CONTROL METHODS
Control technologies in the U.S. for PM and acid gas
control also provide a degree of control for PCDDs/PCDFs
and Hg. Both PCDDs/PCDFs and Hg may be in a vapor phase
at flue gas temperatures, and collection of these pollutants is
primarily dependent on sorption on fly ash particles or flue
gas cleaning sorbent particles with subsequent collection to a
PM control device. The capture efficiency of PCDDs/PCDFs
and Hg in MWC air pollution control equipment typically
used in the U. S. depends primarily on the amount and
properties of carbon in the fly ash, the amounts and
properties of sorbents injected into the flue gas, and the
operating temperature of the PM control device.
Field tests have shown that in semi-dry scrubbing systems,
PCDDs/PCDFs and Hg emissions decrease with increasing fly
ash carbon content. Variations in combustion conditions
affect the amounts of carbon in fly ash and its capacity for
adsorbing semi-volatile trace organics and Hg. Increased
combustion efficiencies needed to maximize destruction of
organics reduce the absorption capacity of fly ash and its
ability to capture semi-volatile pollutants in PM control
devices. The formation and emission of PCDDs/PCDFs and
emission of Hg are affected by the PM control device
operating temperature as the sorption of semi-volatile
pollutants is reduced with increasing operating temperatures.
Methods which can be used to enhance control of
PCDDs/PCDFs and Hg in MWCs equipped with conventional
dry and semi-dry flue gas cleaning systems include:
Enhanced PCDDs/PCDFs control
	Good combustion practices,
 Injection of activated carbon, and
	Injection of specialized multipollutant sorbents.
Enhanced control of Hg
	Injection of sodium sulfide,
	Injection of activated carbon, and
	Injection of specialized multipollutant sorbents.
PCDD/PCDF FORMATION MECHANISMS
There are three primary routes for PCDD/PCDF formation:
(1) gas-phase reactions involving chlorinated precursors such
as chlorobenzcnes (CBs). chlorophcnols (CPs), or
polychlorinated biphenyis (I'CBs); (2) surface reactions
involving gas-phase precursors and fly ash; and (3) solid-
phase reactions on the surface of fly ash involving metal
4

-------
chlorides and fly ash carbon.1'"24 Formation route (3),
which involves the reaction of unhurncd carbon in fly ash in
the presence of oxygen and water vapor (o form
PCDDs/PCDFs. is called de novo synthesis 20,2' The
reactions associated with formation routes (2) and (3), which
involve both gas- and solid-phase reactions, arc called
heterogeneous reactions
Gas-phase precursors can originate as waste thermal
decomposition products or as high temperature PICs. Low
temperature oxidation reactions involving fly ash carbon can
also produce CPs or other precursor compounds that in turn
react to form PCDDs/PCDFs by surface mediated reactions
(condensation, absorption-desorption, etc.).2^
De novo synthesis consists of low temperature carbon
oxidation reactions which provide the biaryl ring structures
for PCDD/PCDF formation and metal ion catalyzed reactions
which provide the necessary chlorine (CI) for PCDD/PCDF
formation. Low temperature carbon oxidation reactions may
be catalyzed by metal ions or carbon structures similar to
activated carbon.26.27	fQT (oxyJ-cblorinalion
reactions can be provided from either metal chlorides in the
fly ash or CI in the flue gas.27-28
De novo synthesis reactions generate a variety of chloro-
organic compounds including CPs, chloro-benzonitriles,
-thiophenes, -benzofurans, -benzothiophenes. PCDDs,
-naphthalenes, PCDFs, and -benzenes (see Figure 1).23
Laboratory experiments (see Figure 2) show that tie novo
reactions occur at temperatures ranging from approximately
250 to 600C with maximum tetra- to ocla-PCDD formation
rates near 300C. Maximum tetra- to octa-PCDF formation
rates also occur at 300C with a lower peak near 450C.23
Maximum PCDD/PCDF formation rates are typically reported
lo occur near 300^.20-23 Al temperatures above 600C,
chloro-organics are rapidly destroyed, and at temperatures
below 2508C, reaction rates result in minimal formation.
PCDD/PCDF CONTROL AT MWCs
Stack emissions of PCDDs/PCDFs from MWCs have been
found to range from < 1.0 to > 20,000 ng/dscm depending on
combustion and flue gas cleaning conditions (sec
Figure 3).29"32 Although the stack emissions are dependent
on combustion conditions, the highest emissions are
generally obtained with MWCs equipped with only ESPs
followed in order of decreasing emissions by DSI/ESP,
SD/ESP, DSI/FF, and SD/FF.
Factors affecting the formation and subsequent emission of
PCDDs/PCDFs from MWCs include the:
 Composition and properties of waste,
~	Combustion Conditions,
*	Composition of flue gas,
	Amount of entrained PM,
	Flue gas time/temperature profile,
	PM control device operating temperature, and
*	Methods of acid gas and PM control.
Composition end Properties pf Waste
Rapid changes in waste composition or properties may
cause combustion upsets and lead to PCDD/PCDF formation.
Although PCDDs/PCDFs are formed during steady state
combustion conditions, the amounts formed are believed to
increase substantially during combustion upsets associated
with improper feed conditions. It is important to blend or
mix waste prior to combustion to reduce variations in heating
content, volatility, and moisture content. 16-18
While waste composition probably affects the amounts of
specific organics formed during combustion, there is no
conclusive scientific evidence that specific solid waste
components, such as polyvinyl chloride (PVC), are primarily
responsible for the formation and emission of
PCDDs/PCDFs.'2.33 Sufficiant CI is available from other
waste components to account for observed levels of
PCDDs/PCDFs in MWCs.
Combustion Conditions
PCDDs/PCDFs contained in the waste are believed to be
destroyed in active flame zones or the high temperature
pl*ndb
hefuofltirilc*
tilKtfiKnoC |
twrn&iurafi* |
bcturthkiftaEK*
OApfulttlencs
dihcnitifurafSK
haiyfnr*
Oi- to Octa-
CWoco-
organic*
Fly Aj$i Healed
in Moii* Air
5a >0 J0000 15000 20000 25000 30000
Organic* Rcmcd, og/g
FIGURE!. CHLORO-ORGANICS FORMED DURING
DENOVO SYNTHESIS
5

-------
3000
a 1000

0
PCDD
PCDF
Teua-toOcia-
PCDD/PCDF
Fly Ash Heated
in Moist Air
250 " 300 * 350 " 400 ' 470 ' '
Temperaiurc, C
FIGURE 2. EFFECT OF TEMPERATURE ON DENOVO FORMATION
RATES OFTETRA - TO OCT A- PCDD/PCDF

c
c
o
"1
c
ffi
o
c
o
o
LL
a
O
0;
Q
O
O
CL
100000
10000 :
1000
100
10
0.1
o
o,
o
,o
o
200
X
C
!+ o
cm o


I
3
oo
ft
300
Temperature, F
i w
^ "ft
. iv y
400
x
x
X
X
s
W
X *x
X	M8WW/ESP
A	M8WWSO-ESP
X	nwtwa-csp
0 HWR/0-SP
A	ROF/S&CSP
V	noF/sse
I	HOF/FSrtSP
4-	M8RW/DSIESP
4	M8WYWS1>PF
O	MCXVSO^F
O	ROF/SOfF
500
600
100
150	200
Temperature, C
250
300
FIGURE 3	DEPENDENCY OF PCDD/PCDF EMISSIONS ON TECHNOLOGY
AND PM CONTROL DEVICE OPERATING TEMPERATURE
6

-------
regions of MWCs 10-16 However, the waste composition
and combustion conditions determine the availability of
specific precursors needed for PCDD/PCDF formation. These
precursor compounds are formed by the thermal destruction,
oxidation, and synthesis reactions which occur in the burning
waste bed, the active flame region above the bed, and the high
temperature regions of the furnace. Combustion conditions
and the time/temperature profile in the cooling zones
downstream of the combuslor determine the amounts of
PCDDs/PCDFs precursor material entering flue gas cleaning
devices and the potential for formation within the devices.
Organics that have been imp]cated in PCDD/PCDF formation
include: CBs, CPs, PCBs, and the carbon in fly ash.25*28
GCP can be used to maximize the furnace destruction of
organics and minimize the downstream formation of
PCDDs/PCDFs by controlling the amounts of PM carried out
of the furnace with flue gas. Furnace destruction of organics
must include destruction of both gas- and condensed-phase
organics. Field test experiments have shown that formation
of PCDD/PCDF and other trace organics correlates with the
CO and the amounts of PM carried out of the combustor with
flue gases (see Figures 4 and 5)16,17.34-36 jn extensjve
tests at a RDF combustor, CO and total hydrocarbons (THCs)
were found to be the two best parameters for predicting
PCDDs, PCDFs, CBs, CPs. and PAHs at the SD inlet.36
Waste and its associated thermal decomposition products
must be exposed to elevated temperatures for a sufficient time
to completely destroy their organic components. Time scales
required for destruction arc typically measured in milliseconds
for gaseous compounds in active flame zones. Combustion
reaction times in the order of seconds to minutes may he
required for the complete destruction of small solid particles.
Combustion temperatures and residence times of 980*C
(1800F) and I to 2 seconds, respectively, are generally
believed sufficient to thermally destroy gas-phase
compounds.10 However, even at temperatures of 980C or
higher, residence times needed for the complete combustion
of entrained solid particles may be insufficient, and residual
unbumed carbon in fly ash may lead to reactions which form
PCDDs/PCDFs.16
The amount of air used for combustion must be high enough
to minimize the existence of fiiel-rich pockets and low
enough to avoid quenching of combustion reactions.1'16
The distribution of combustion air is also important.
Burning refuse beds contain drying, devolatization,
combustion, and burnout zones. Each zone requires a different
amount of combustion air. State-of-the-art MWCs often use
zoned underfire air supplies to provide proper air distribution
to the refuse bed and overfire air to complete combustion of
unbumed material leaving the bed.'
Poor mixing increases the amount of organic material
available for the formation of PCDDs/PCDFs. It may result in
local stoichiometrics that are insufficient for the complete
oxidization of gas- and solid-phase organics. Poor mixing
may also lead to the formation of difficult-to-destroy soot
particles. Methods of achieving good mixing include the use
of furnace configuration and overfire air jets,'*16
Composition of Flue Gas
The effects of flue gas composition on PCDDs/PCDFs are
complex.24.25 Oxygen is required for the low temperature
carbon oxidation reactions that are associated with de novo
synthesis. Oxygen and H20 are also required for the Deacon
process reaction which provides CI for the chlorination of
tn tsoo
a>
c
u.
o
o
a
o
D
O
Dl
1000
500
ia^o.70




0

.


0
0
0







O Good operation
 Poor operation

0
0
0

%	






20
403
600
00
tooo
CO at 12% CO2, ppm
FIGURE 4	RELATIONSHIP BETWEEN CO AND PCDD/PCDF CONCENTRATIONS
AT SD INLET, MID-CONNECTICUT MWC
7

-------
O!
2
cn
3
T3
O)
u.
03
CO
3
Q>
cc
u!~
Q
o
CL
Q
o
o
CL
R2 =0.57.1
FIGURE 5.
	8	10	12	"
Uncontrolled Ash/Refuse Fed, kg/Mg
RELATIONSHIP OF PM CARRYOVER TO PCOD/PCDF
EMISSIONS, QUEBEC CITY MWC
PCDDs/PCDFs or their precursor compounds. The CI for the
Deacon process reaction may come from either (he inorganic
chlorides in fly ash or the HQ in flue gas. Some bench and
pilot scale experiments show that PCDD/PCDF formation
increases with increasing levels of HC1 or Clj in combustion
gates.Although the flue gas concentrations of Oj and
H2O may affect the rate of PCDD/PCDF formation, the waste
moisture content and excess air levels determine the
concentrations of these two constituents. The amount of
excess air must be fixed at levels needed to obtain good
combustion, and it is generally not practical to control the
moisture content of wastes to the extent needed to control
formation of PCDDs/PCDFs.
Entrained Particulate Matter
The entrainment and carryover of fly ash into the cooler
regions of MWCs may lead to the formation of PCDDs/PCDFs
and other trace organics. Metal ions or fly ash carbon can
catalyze condensation formation reactions, and fly ash
carbon can serve as the source of organics for the de novo
synthesis of PCDDs/PCDFs. PM carryover is determined by
the aerodynamic properties of particles, the method of waste
combustion, and combusior gas flow characteristics.
Methods of limiting PM carryover include proper furnace
design, control of the underfire-to-overfire air ratio, the
amount of excess air, and load (refuse burn rate).'6.18
The correlation between PM carryover and PCDD/PCDF
emission rates from tests at the Quebec City, Canada, mass
burn combustor is shown in Figure 5.34 These data, which
are measured in the stack downstream of an ESP, show the
effects of changing combustion conditions on emission of
PCDDs/PCDFs. Changes in the PM carryover rate can be
attributed primarily to changes in the flue gas flow rate and
the ratio of underfire-to-overfire air. At higher loads,
increased gas flow rates increase entrainment and carryover of
PM. Higher underfire air flow rates also increase PM
entrainment. During the low combustion temperature tests at
Quebec City, Canada, excess air rates were increased to lower
the combustion temperature. This had the effect of increasing
volumetric flow rates and particle entrainment. During the
poor combustion air distribution tests, the underfire-to
-overfire air ratio was increased, thereby increasing PM
carryover. Increasing emission rates of PCDDs/PCDFs with
increasing amounts of fly ash is consistent with
heterogeneous formation theories.
Timeflemperature Profile
Pilot scale experiments have shown that the concentration
of PCDDs/PCDFs in flue gas downstream of the furnace
depends on the time/temperature profile in the cooling
sections of combustion systems.37"39 The time/temperature
profile is determined by the time required by flue gas and
suspended PM to pass through the heat extraction regions of
boilers, superheaters, and economizers. High flue gas
temperature quench rates reduce the time that gas-phase
organics and entrained particles spend in the temperature
range associated with high PCDD/PCDF formation rates.
PM Control Device Operating Temperature
ESPs and FFs can function as chemical reactors that
generate and emit PCDDs/PCDFs,* A large fraction of the
PCDDs/PCDFs entering PM control devices is commonly
associated with collectible fly ash. However, the large mass
of particles within the device can serve as a source for the
synthesis of PCDDs/PCDFs. Limiting the temperature at
which PM control devices are operated is important in
controlling the formation and emission of PCDDs/PCDFs.'^"
18,40,41
The ESP operating temperature is perhaps the most
important variable affecting the formation and emission of
chloro-organics in ESP systems. At PM control device
operating temperatures above 250C, de novo synthesis
reaction rates become significant and the partitioning of
8

-------
PCDDs/PCDFs into a vapor phase increases with increasing
temperature.
The results of tests to evaluate the effects of ESP operating
temperature on the formation and emission of PCDDs/PCDFs
sue depicted in Figure 6. These tests were conducted on a mass
burn refractory MWC with a water spray flue gas quench
chamber and an ESP located in Montgomery County, Ohio.''
Quench water flow rates were adjusted to obtain nominal ESP
inlet temperatures of 300 , 200 , and 150 C, The test
conditions included: normal and poor combustion (low
temperature) tests at 300 *C inlet temperature; normal
combustion with and without furnace injection of CaCOj at
200'C inlet temperature; and normal combustion with furnace
injection of CaCOj and duct injection of Ca(OH)2 at 150C
ESP inlet temperature.
The flue gas concentrations of PCDDs/PCDFs were higher at
the ESP outlet than at the inlet for all tests, indicating
PCDD/PCDF formation within the ESP. Under normal
combustion conditions at the high ESP inlet temperature (300
*C), ESP inlet concentrations of PCDDs/PCDFs averaged 200
ng/dscm, while stack concentrations averaged 17,000
ng/dscm. Reducing the ESP inlet temperature to 200C
without sorbent injection reduced average stack emissions to
870 ng/dscm. For tests with furnace injection of CaCO-j,
stack emissions of PCDDs/PCDFs averaged 1480 ng/dscm at
an ESP inlet temperature of 200aC and 670 ng/dscm at 150C
inlet temperature. The lowest emissions (57 ng/dscm) were
obtained using duct injection of Ca(OH)2. Operating at the
lowest practical ESP operating temperature is critically
important in minimizing PCDD/PCDF emissions.
Methods of Acid Gas and FM Control
The methods of acid gas and PM control are major
determinants of PCDD/PCDF emissions. Acid gas controls
modify the chemistry of the PCDD/PCDF formation
environment, affect flue gas quench rates, and allow operation
of PM control devices at low temperatures. Low operating
temperatures are necessary to minimize de novo synthesis
rates and partitioning of trace organics between solid and
vapor phases. Efficient collection of PM is necessary to
collect solid-phase organics. Vapor-phase organics can be
absorbed onto the surface of PM as it passes through the fly
ash filter cake in FFs. DSI and SD can be used to reduce acid
gases, modify PCDD/PCDF formation chemistry, and allow
for lower PM control device operating temperatures.
Experiments in a 14.7 kW (63,000 Btu/hr) pilot scale
combustor have shown that injection of Ca(OH)2 into the flue
gas at temperatures greater than 800C significantly reduces
the formation of PCDDs/PCDFs,^ The reduction in
PCDD/PCDF yield appears to result from both a reduction in
the HC1 content of flue gas and an inhibitory effect on fly ash
surface reactions. DSI can also be used to remove HC1 and
reduce heterogeneous formation rates in the PM control
device, but substantial amounts of PCDDs/PCDFs may be
O
E
<->
m
5
O!
o
0
01
Q
O
o
L
100000 7
toooo
1000
100
INJECTION
IN OUT LOCATION SORBENT
None
Furnace
Ouct
CaCOj
Ca(OH|,
350
ESP Inlet Temperature, "C
FIGURE 6.	EFFECT OF ESP INLET TEMPERATURE ON PCDD/PCDF
FORMATION. MONTGOMERY COUNTY MWC
9

-------
formed upstream of the sorbent injection locations near the
inlet to the PM control device. Most of the PCDDs/PCDFs
formed upstream will be retained in the PM control device if
the PM control device operating temperatures are low enough.
SD systems combine the advantages inherent in high flue gas
quench rates, sorbents which probably modify PCDD/PCDF
synthesis reactions, and low PM control device operating
temperatures.
Flue gas concentrations of PCDDs/PCDFs can either
decrease or increase across an ESP.^ Increases are associated
with high rates of PCDD/PCDF synthesis within m PM control
device. PM control devices containing collected fly ash can
function is reactors for formation of PCDDs/PCDFs,* The
amount formed will depend on the temperature, the mass of
PM within the control device, the composition of the fly ash,
the composition of the flue gas, and the residence time of PM
within the device. Maximum de novo synthesis reaction rates
occur at temperatures near 300C.^-* Lowering PM control
device operating temperatures to less than 250*C results in
low PCDD/PCDF formation rates and alters the partitioning
of vapor- and solid-phase PCDDs/PCDFs. In ESPs, operating
temperature is critical, and it may be necessary to scrub acid
gases from the flue gas to permit lowering of the PM control
device operating temperature to a level where acid gas
corrosion of the device is not a problem.
PM control device collection efficiencies >99 percent are
probably necessary to adequately control PCDD/PCDF
emissions. At the lower PM control inlet device temperatures
typically employed by dry scrubbing systems, formation
rates are greatly reduced and PCDDs/PCDFs are predominantly
retained on captured fly ash. Although substantial quantities
of trace organics may be formed during the combustion
process, most of the semi-volatile organics are effectively
collected in dry scrubbing systems. This is especially true of
SD/FF systems.
MWCs equipped with SD/ESP systems are less effective in
controlling PCDD/PCDF emissions than similar MWCs
equipped with SD/FF.3-4^ The latter devices often have
better PM control efficiencies, and the flue gas is passed
through a filter cake where PCDDs/PCDFs can be absorbed on
fly ash and sorbent particles. A review of data from eight
different combustion units equipped with SD/ESPs indicated
PCDD/PCDF emissions ranging from 9 to 173 ng/dscm
(mass). All but one of these units had emissions of less than
75 ng/dscm, and the average for the range of typical
emissions was 38.1 ng/dscm.4^
Hie performance of SD/ESP systems for controlling
PCDDs/PCDFs in mass burn MWCs can also be improved by
the injection of activated carbon into the flue gas at the
entrance to the SD system or by use of a carbon-enhanced
lime-based sorbent in place of lime.43^ During three EPA-
sponsored tests at the Camden County, New Jersey, MWC,
stack concentrations of PCDDs/PCDFs without activated
carbon averaged 46.8 ng/dscm. During three tests in which
360 mg/dscrn of dry activated carbon was injected into the
flue gas upstream of the SD/ESP, stack concentration of
PCDD/PCDF averaged 5.6 ng/dscm.44'45 Tests in Europe on
MWCs equipped with SD/FF and SD/ESP systems have also
shown the effectiveness of carbon injection for reducing
PCDD/PCDF emissions,43
An EPA review of data from 20 different combustion units
equipped with SD/FF systems indicated PCDD/PCDF
emissions ranging from 1 to 22 ng/dscm. Nineteen units had
emissions <12 ng/dscm, and the average for these units was
6.6 ng/dscm. Two large MWCs with DSI/FF systems had
PCDD/PCDF emissions of 5 and 18 ng/dscm.4^
CONTINUOUS CONTROL OF PCDDs/PCDFs
There is currently no feasible method for continuously
measuring PCDD/PCDF emissions. Continuous control of
PCDDs/PCDFs is of concern from regulatory and risk
assessment perspectives. EPA's strategy to ensure
continuous compliance with PCDD/PCDF emission limits is
to place limits on CO concentrations, steam load, PM control
device operating temperature, carbon feed rate, and opacity.
The first four parameters are used to limit formation and
partitioning of PCDDs/PCDFs, and the final parameter
(opacity) will ensure proper operation of the PM control
device which is needed for effective collection of solid-phase
PCDDs/PCDFs, The effectiveness of these parameters for
continuously controlling PCDD/PCDF emissions is verified
by periodic compliance tests on each MWC that is subject to
regulation.
CONTROL OF OTHER ORGANICS
Field test measurements show that a wide variety of trace
organics are formed during combustion of municipal solid
waste. Trace organics formed include semi- volatile organics
such as PCDDs/PCDFs, CBs, CPs, PCBs, PAHs, and a variety
of volatile organics. Analysis of data from the Quebec City
MWC mass burn tests and the Mid-Connecticut RDF
combustor tests show that flue gas concentrations of
PCDDs/PCDFs, CBs, CPs, and PAHs all correlate with each
other and with flue gas concentrations of CO and THC.3*3
Measures taken to control emission of PCDDs/PCDFs are also
effective in controlling CBs, CPs, and PAHs. (It is expected
that these measures will also control PCBs, but the
concentrations of PCBs in MWC flue gases are generally so
low that strong statistical correlations with other pollutants
are not found.)
Combustion controls are effective in controlling all trace
organics. Dry and semi-dry scrubbers with FFs and
temperature control are effective in controlling semi-volatile
organics, but the degree of control is probably related to the
amounts of carbon in the fly ash. Activated carbon or
multipollutant sorbent containing calcium and carbon (lime
enhanced sorbents) should also enhance capture of all semi-
volatile organics in MWCs equipped with dry and semi-dry
scrubbing systems. Dry and semi-dry scrubbing systems are
probably not effective for controlling volatile organics, and
the best method for controlling these pollutants is the use of
GCP as a preventive measure.
10

-------
MERCURY CONTROL AT MWCs
The capture of Hg in flue gas cleaning devices depends on
the
*	Waste composition,
	Flue gas properties (temperature, gas composition,
moisture, etc.),
	Hg form [speciation and phase (vapor or solid)],
	Fly ash and sorbent properties, and
*	Type of control device.
The waste composition and the relative amounts of Hg in
each component determine the concentration of Hg in MWC
flue gas, Hg mass balances from Environment Canada's
Quebec City mass burn test project indicated that more than
96 percent of Hg In the MWC output streams was in either the
collected ESP ash or the flue gas, indicating the volatile
nature of Hg in MSW.34
MERCURY SPECIATION AND CONTROL
MECHANISMS
The form of Hg in flue gas depends on the flue gas
composition and temperature. Thermal-chemical calculations
indicate that Hg is converted to elemental mercury (Hg) in
the high temperature regions of the combustor, As the flue
gas cools, some or all of the Hg is converted to other Hg
species.46.47 por combustion systems containing
substantial amounts of CI in the waste (or fuel), the two
predominant forms of Hg at flue gas cleaning temperatures (<
300C) are believed to be ionic mercury (Hg2+) and Hg.4"
4' Thermochemical equilibrium calculations indicate that the
49
Hgz+ will be predominantly mercuric chloride (HgClj)
Most metals condense to form solid particles as flue gas is
cooled so that they can be collected as PM. However, both
Hg" and HgCl2 arc vapors [Hg(v)J at typical flue gas cleaning
control device operating temperatures (300 to !40C), and
special methods must be devised for their capture.12-13
Hg in MWC flue gases can be captured if it is in the solid
phase [Hg(s)] or is adsorbed on fly ash or special sorbents
such as activated carbon or enhanced lime-based sorbents
containing activated carbon. Hg(v) capture without the use of
special sorbents depends on the amount and properties of
carbon in the fly ash. Well-designed and well-operated mass
burn combustors have little carbon in their fly ash and, even
when equipped with SD/FFs or SD/ESPs, they often exhibit
Hg control levels below 50 percent. Conversely. RDF
combustors contain relatively high amounts of carbon in the
fly ash (> 2 percent). And. they can exhibit control
efficiencies above 80 and 90 percent when equipped with
SD/ESPs and SD/FFs, respectively. Figure 7 shows the
distribution of Hg stack concentrations for RDF and non-RDF
MWCs equipped with SD/FF and SD/ESP systems.4*
Two techniques employed for Hg capture in dry flue gas
cleaning systems are the use of activated carbon and sodium
sulfide (Na^S).12-13 NajS, a crystalline solid, is dissolved
in water and injected upstream of the flue gas cleaning
equipment. Hg(v) and Hg2+(v) are converted to a solid form
of Hg [mercury sulfide (HgS)J that can be collected in a PM
control device, NajS has been used at MWC facilities in
Europe and Canada for Hg control. Na^S test results from
European facilities show Hg emissions ranging from 40 to 70
fig/dscm and removal efficiencies from 65 to 90 percent.1
NajS is not currently being used in the U.S. Activated carbon
can be used in three ways to control Hg(v) emissions: (1) it
can be injected as a powder in dry or semi-dry scrubbers to
E
o
ai
5
o>
n
c
o
c
m
o
c
o
O
cn
X
1000
800
600
400
200
Hon-RDF
D

FIGURE 7.	HG CAPTURE IN MWCS WITH SD/FF AND SD/ESP
LARGE PLANTS (>225Mg/d)
1 1

-------
absorb Hg(v) for subsequent collection in an ESP or FF,'2'1^
(2) it can be used in a sorbent system combining both lime
(A
and carbon in a single multipolluuinl sorbent. and (3) flue
gas can be filtered through a carbon bed polishing filter
downstream of other air pollution control devices to enhance
removal of Hg(v) and other pollutants.5' Carbon bed filters
are currently being applied to European MWCs where they are
primarily used to improve emission control of
PCDDs/PCDFs, Hg. and other trace metals.51
CARBON CAPTURE MECHANISMS
The adsorption of Hg and organics by activated carbon is
controlled by the properties of both the carbon and the
adsorbate, and by the conditions under which they are
contacted. This phenomenon is generally believed to result
from the diffusion of vapor-phase molecules into the pore
structure of carbon particles. These molecules are retained at
the surface because of intermolecular Van der Waals forces.
As the temperature falls, or as the partial pressure of the
vapor above the carbon rises, the average time that a
molecule resides on the surface increases. So does the
fraction of the available surface covered by the adsorbate.
However, the carbon surface is not uniform and consists of
sites whose activities vary. More active sites will become
occupied first and, as the activity of the remaining available
sites decreases, the adsorption energy will change.
The physical structure of activated carbon is not known in
detail, but it is believed to contain randomly distributed pores
in the carbon, between which lies a complex network of
irregularly, interconnected passages. Pores range in diameter
down to a few angstroms, and provide an internal surface area
from 300 to 1,000 m2/g of carbon. The volume of pores at
each diameter is an important variable that directly affects
carbon performance.
Since adsorption takes place at the carbon/gas interface, the
surface area of the carbon is one of the most important factors
to consider. Another important factor is the pore radius.
Bench scale tests have shown that increasing the surface area
and adding sulfur compounds to the sorbent result in higher
adsorption rates of some Hg species. Until recently, most of
the laboratory work on carbon adsorption has been done on
Hg, not with the Hg compounds normally found in MWC flue
gas.
The actual adsorption capacity of carbon is affected by:
	Gas temperature,
	Inlet concentrations of Hg,
	Species of Hg,
	Type of carbon and surface area,
	Contact time,
	Flue gas moisture.
Acid content of flue gas, and
	Concentration of compounds which
compete with Hg for sorption sites .
method, carbon properties, flue gas temperature, and PM
control method. Performance tests in the U.S. and Europe
have been limited primarily to the application of carbon
injection to mass bum MWCs equipped with SD/FF or SD/ESP
systems.I3.42.43.52
EPA has sponsored two major field tests on the injection of
powdered activated carbon for Hg control and has selected this
technology as the basis for Hg emission control requirements
on MWCs.44,45,52"54 Both tests were conducted at
facilities with conventional mass-burn waterwaU combustors.
The first was conducted on an MWC in Stanislaus County,
California, equipped with a SD/FF system. The second was
conducted on two MWCs in Camden County, New Jersey,
equipped with SD/ESP systems. These tests show that stack
concentrations of Hg depend on the Hg concentration in the
flue gas (SD inlet concentration), the carbon content of fly
ash, and the operating conditions of the carbon injection
system.
Mercury concentrations in MWC flue gas are highly
variable with time. In MWCs, the total solid-phase flue gas
carbon (carbon in fly ash plus the carbon injection rate)
appears to be the key determinant of Hg capture. In the
absence of carbon injection, the amount of Hg captured
depends on the amount and properties of carbon in the fly
ash. When the fly ash carbon content is low or when Hg
concentrations are high, poor removal efficiencies are
obtained. When the fly ash carbon content is high and the Hg
concentrations are low, stack emissions of Hg are low.
Powdered activated carbon can be injected into the flue gas
to increase solid-phase carbon concentrations and improve
Hg capture. Increasing the carbon injection rate reduces both
the average and variability of emissions. At high carbon
injection rates, there is generally sufficient carbon to capture
low or high levels of Hg. The amount of excess carbon
needed for continuously high levels of capture will depend on
the variation of Hg concentration in the flue gas. Highly
variable Hg inlet concentrations will require high excess
carbon injection rates to ensure continuous Hg capture.
In SD/FF tests at the Stanislaus County MWC, Hg capture
without carbon injection ranged from 16 to 46 percent.
Outlet Hg concentrations for these tests ranged from 311 to
538 ng/dscm. Hg capture increased with increasing carbon
injection rates and, at the highest injection rates of
approximately 70 to 100 mg/dscm, Hg outlet concentrations
ranged from 17 to 77 fig/dscm (see Figure 8)52-54 jpoutlet
temperatures at Stanislaus County normally ranged between
136 and 145 C.
During the Camden County SD/ESP carbon injection test
project, Hg capture without carbon injection ranged from 18
to 92 percent.44*'45 When dry carbon injection rates
exceeded 150 mg/dscm and the ESP inlet temperature was 132
"C, stack emissions of Hg were generally less than 80
Hg/dscm (see Figure 9). The dependency of Hg reduction
efficiency on the total solid carbon concentration in the flue
gas during the Camden County tests is shown in Figure 10.
PERFORMANCE OF ACTIVATED CARBON
SYSTEMS
The performance of activated carbon systems depends
primarily on the carbon injection rate, carbon injection
12

-------
w 220

-------
\r " 			'i 				|	"	iiibw*i  in i	j 	^
e	loo	200	ma	*to	yn	o*.
Totat Carbon, mg/dsctn at 7% O 2
63 Dry Carbon at 132*C A Wet Carbon al 132'C	Ory Carbon at I77C
FIGURE 10. EFFECT OF TOTAL CARBON ON MERCURY CONTROL
IN SD/ESP, CAMDEN COUNTY MWC
The performance of activated carbon in adsorbing Hg is
dependent on temperature. The temperature at the inlet to the
PM control device is normally used as a parameter in
evaluating the performance of the device in collecting
condensed or adsorbed pollutants. The PM control device
inlet temperature for SD/FF and SD/ESP systems on MWCs is
normally between 135 and 145 C. Tests on the Camden
County SD/ESP system at 177 "C indicated only moderate
reduction in Hg capture relative to similar tests at 132 C (see
Figure 10).Temperature variations over the normal
operating range of SD/FFs and SD/ESPs can be expected to
have only minor effects on Hg capture. Similar tests on
European MWCs have shown that carbon injection can be
used to reduce Hg emissions in SD/FF and SD/ESP systems to
less than 80 ng/dscm.*^
ESPs and wet scrubbers are commonly used to control
emissions from European MWCs. Some European plants
have installed activated carbon beds downstream of the
primary air pollution control devices to act as polishing
filters for the control of metals, dioxins. and acid gases. The
use of activated carbon filter beds in combination with
conventional control equipment has demonstrated Hg
reductions exceeding 99 percent and Hg outlet concentrations
of less than 1 |ig/dscm.^ *
PERFORMANCE OF ENHANCED LIME-BASED
SORBENTS
Several enhanced lime-based sorbents are commercially
available for use in MWCs. One enhanced sorbent (Scansorb)
is designed for improved HC1 and SO2 control in dry injection
systems.^ Another sorbent (Sorbalit) is a multipollutant
sorbent designed for control of acid gases, Hg, and trace
organics.^ Sorbalit is produced by mixing lime, surface-
activated carbon, a sulfur-based compound, and other
additives. The Sorbalit carbon content can range from 4 to 65
percent depending on the technical and economic
requirements of each project.
Comparative performance tests between Sorbalit and a
typical activated carbon on a mass-burn MWC equipped with
an SD/FF system have indicated equivalent or slightly higher
capture of Hg by the enhanced lime-based sorbent. 50 He
lime-based sorbent captured 87.7 percent of the total Hg,
while the conventional activated carbon captured 84.2
percent. Analysis of the sampling train data indicated that it
captured 83.2 percent of the vapor-phase Hg, while the
conventional carbon captured 77.6 percent. Additional
performance testing is being planned to evaluate the
effectiveness of Sorbalit in controlling emissions from
MWCs with other flue gas cleaning equipment
configurations.
FORM OF MERCURY EMISSIONS
In conducting risk assessments, it is important to estimate
the form and cpeciation of Hg stack emissions. The
transport, deposition, and environmental uptake of Hg are
dependent on the form and speciation of Hg. Several studies
estimate the speciation of Hg in MWC flue gases. Metzger
and Braun estimate that nearly all Hg in MWCs at flue gas
cleaning temperatures is in the form of mercury chlorides.4^
Lindqvist and Schager estimate that the speciation of Hg in
raw flue gases is approximately 20 percent Hg , 60 percent
Hg2+, and 20 percent Hg(s).48 Pacyna estimates that Hg
14

-------
emissions from European waste incinerators consist of 10
percent Hg* , 85 percent Hg2+. and 5 percent Hg(s).^6
There is currently no validated U.S. method for determining
the speciation of Hg in stack gas. However, information on
the chemical behavior of Hg and the distribution of Hg in
EPA's multimetal sampling train (Method 29) can be used to
estimate the form and speciation of Hg in the MWC stack gas
These estimations are valid only for measurements
downstream of the air pollution control devices. The phase
and speciation of Hg at inlet sample locations can be
significantly affected by PM collected in the Method 29
probe and filter. For outlet samples, Hg found in the probe
and filter can be assumed to have been either vapor-phase Hg
adsorbed onto PM or a solid-phase Hg compound. Both are
associated with PM as designated by Hg(PM). HgClj is
soluble in water and should be removed in the nitric
acid/peroxide (HNOj/HjC^) impingers. Hg found in the
downstream permangenate/sulfuric acid (KMnO^/HjSO^)^
impingers was probably Hg0(v), The distribution of
multimetal train samples collected during the activated carbon
injection tests at the Camden County MWC and Stanislaus
County tests is shown in Figure 11.
For the Camden and Stanislaus County tests, the fraction of
Hg(PM) was generally below 5 percent and exceeded 10
percent for only one test. Tests with stack concentrations of
Hg >100 ng/dscm represent tests without carbon injection or
low carbon injection feed rates. For these tests, Hg ranges
from 2 to 26 percent of total Hg. As carbon injection rates
and Hg capture increase, the percentage of Hg" as a fraction of
total Hg increases. This implies that Hg^+ is more easily
captured by activated carbon than Hg*. For stack
concentrations of Hg <50 ng/dscm, the fraction of Hg"
ranges from approximately 14 to 72 percent.
From the results of these tests it can generally be concluded
that, for MWCs equipped with SD/ESPs and SD/FFs, the stack
emission of Hg(PM) is negligible unless further Hg is
80.0-
f/i
5
absorbed on the surface of PM between the stack sampling
location and the stack exit.
At low levels of control, the stack concentration of Hg is
probably 15 to 30 percent Hg(v) and the rest is Hg2+(v) and
HgfPM). In MWCs with SD/FF or SD/ESP systems, stack
concentrations of Hg(PM) are probably less than 5 percent.
At high levels of control, Hg^+(v) is selectively removed.
Increasing the relative concentration of Hg (v). The relative
concentration of Hg (v) may be 50 percent or higher.
CONCLUSIONS
Dioxin formation is predominantly associated with
heterogeneous reactions involving fly ash. These low-
temperature synthesis reactions can occur downstream of the
combustor at temperatures ranging from approximately 250
to 600C.
Spray dryers and FFs can be used to continuously reduce
PCDD/PCDF emissions to less than 20 ng/dscm. Activated
carbon, which is needed for Hg control in many MWCs, will
provide additional PCDD/PCDF control.
Spray dryers and ESPs can be used to reduce typical
PCDD/PCDF emissions from mass burn combustors to less
than 75 ng/dscm. Injection of activated carbon, to control
Hg emissions, can be used to further reduce PCDD/PCDF
emissions to 60 ng/dscm or less.
The Hg in MSW is volatilized during combustion and
converted to elemental and ionic Hg. Hg in MWC flue gas is
believed to be predominantly HgClj and Hg". Both are
predicted to be in a vapor phase at stack gas temperatures.
In dry flue gas cleaning systems, Hg(v) can be absorbed
onto the surface of particles for collection in PM control
devices. Hg(v) can be adsorbed either on the residual carbon
in fly ash, on activated carbon, or on enhanced lime-based
sorbents which have been injected into the flue gas.
RDF combustors have relatively high amounts of carbon in
their fly ash (>2 percent), and those equipped with SD/FFs
: v















iu-







: a
"I*










A




-liVi!
*

A,
*
A
	1
1	


& Elemental Hg
 Solid-Phase Hg
100 200 300 400 500 600 700 800
Slack Concentration of Hg, |ig/dscm
FIGURE 11. DISTRIBUTION OF Hg IN EPA METHOD 29 SAMPLING TRAIN,
CAMDEN COUNTY AND STANISLAUS COUNTY CARBON
INJECTION PROJECTS
15

-------
may attain Hg removal efficiencies of >90 percent due to
adsorption of Hg(v) onto the fly ash carbon. Other types of
MWCt, such as mass-burn water-wall combustors, may require
the Injection of activated carbon or enhanced lime-based
sorbents to obtain efficient Hg(v) control.
Nearly all of the uncontrolled Hg in MWC stack gas is in a
vapor form. Method 29 sampling train data suggest that
carbon absorption methods collect Hg^+(v) more efficiently
than Hg(v). As Hg removal efficiencies increase, HgClj (v)
is preferentially removed and the proportion of Hg that is
Hg*(v) increases. In MWCs equipped with SD/FFs and
SD/ESPs, the relative amount of Hg(PM) in stack gas will
generally be less than 5 percent.
ACKNOWLEDGMENTS
The major portion of the work covered in this paper was
supported by the U. S. Environmental Protection Agency's,
Municipal Waste Combustion Research Program. Testing of
the enhanced lime-based sorbent (Sorhalii) was partially
supported by the Dravo Lime Company.
REFERENCES
 1J U.S. EPA, Assessment of Municipal Waste Combustor
Emissions Under the Clean Air Act, Advance Notice of
Proposed Rulemaking, 52 CFR 399, July 7, 1987.
[2J Cleverly, D. H., Municipal Waste Combustion Study:
Assessment of Health Risks Associated with Municipal Waste
Combustion Emissions, EPA/530-SW-87-02Ig (NTIS PB87-
206132), September 1987.
[3J U.S. EPA, Air Pollution Standards of Performance for
New Stationary Sources; Rule and Proposed Rules, 40 CFR
Parts 60, 51, and 52, December 20. 1989.
[4] U.S. EPA, Standards of Performance for New Stationary
Sources and Final Emission Guidelines; Final Rules, 40 CFR
Parts 51,52, and 60, February 11, 1991.
{5] U.S. Congress, Clean Air Act Amendments of 1990,
P.L. 101-549, Washington, EC, November 15, 1990.
[6]	U.S. EPA, Standards of Performance for New Stationary
Sources: Municipal Waste Combustors, and Emission
Guidelines: Municipal Waste Combustors, Proposed Rules,
40 CFR Part 60, September 20, 1994.
[7]	U.S. EPA, Standards of Performance for New Stationary
Sources and Emission Guidelines for Existing Sources:
Municipal Waste Combustors, 40 CFR Part 60, December 19,
1995.
[8]	U.S. EPA, Fact Sheet: New Municipal Waste
CombustorsSubpart Eb Standards, Office of Air Quality
Planning and Standards, Research Triangle Park, NC,
November 3, 1995.
[9]	U.S. EPA, Fact Sheet: Existing Municipal Waste
CombustorsSubpart Cb Standards, Office of Air Quality
Planning and Standards, Research Triangle Park, NC,
November 3, 1995
[10]	Seeker, W. R., W. S. Lanier, and M. Heap, Municipal
Waste Combustion Study: Combustion Control of Organic
Emissions, EPA/530-SW-87-021c (NHS PB87-206090),
June 1987,
[11]	Kiser, J. V. L. The IWSA Municipal Waste
Combustion Directory: 1993 Update of U. S. Plants,
Integrated Waste Services Association, Washington, DC,
1993.
[12]	Bma, T. G., Toxic Metal Emissions from MWCs and
Their Control, In Proceedings: 1991 International
Conference on Municipal Waste Combustion, Volume 3,
EPA-600/R-92-209c (NTIS PB93-124196), pp 23-39,
November 1992.
[13]	White, D. M.. K. Nebel, and M. Johnston, Municipal
Waste Combustors: A Survey of Mercury Emissions and
Applicable Control Technologies, In Proceedings: 1991
International Conference on Municipal Waste Combustion,
Volume 3, EPA-600/R-92-209c (NTIS PB93-124196), pp
247-257, November 1992.
[14]	Bma, T. G. and J. D. Kilgroe, Particulate Emissions
Control and Its Impacts on the Control of Other Air Pollutant
Emissions from Municipal Waste Combustors, Journal of Air
& Waste Management Association, 40 (9): 1324, 1990.
[15]	White, D. M. et al NO, Control Technologies
Applicable to Municipal Waste Combustion, EPA-6O0/R-94-
208 (NTIS PB95-144358), Air and Energy Engineering
Research Laboratory, Research Triangle Park, NC, December
1994,
[16]	Kilgroe, J. D. et al.. Combustion Control of Organic
Emissions from Municipal Waste Combustors, Combustion
Science and Technology, 74 (1990), 223-244.
[17]	Schindler, P. J. and L. P. Nelson, Municipal Waste
Combustion Assessment: Technical Basis for Good
Combustion Practice, EPA-600/8-89-063 (NTIS PB90-
154949), August 1989.
[18]	Kilgroe, J. D W. S. Lanier, and T. R. von Alten,
Development of Good Combustion Practice for Municipal
Waste Combustors, Proceedings of National Waste
Processing Conference, American Society of Mechanical
Engineers, Book No. 100328, New York, NY, 1992.
[ 19] Shaub, W. M. and W. Tsang, Dioxin Formation in
Incinerators, Environmental Science and Technology, 17
(1983), 721-730.
[20]	Vogg, H. and L. Stieglitz, Thermal Behavior of
PCDD/PCDF in Fly Ash from Municipal Waste Incinerators,
Chemosphere. 15 (1986), 1373-1378.
[21]	Hagenmaier, H. et al.. Catalytic Effects of Fly Ash
from Waste Incineration Facilities on the Formation and
Decomposition of Polychlorinated Dibenzo-p-dioxins and
Polychlorinated Dibenzqfurans, Environmental Science and
Technology, 21 (1987), 1080-1084.
[22]	Stieglitz, L, and G. Vogg, Formation and
Decomposition of Polychlorodibenzodioxins and Furans in
Municipal Waste, Report KFK 4379. Laboratorium fur
Isotopentechnik, Institut fur Heize Chemi,
Kernforechungszentrum Karlsruhe. February 1988.
[23]	Schwarz, G., L, Steiglitz, and W. Roth, Formation
Conditions of Several Polychlorinated Compound Classes on
Fly Ash of a Municipal Waste Incinerator, Organohalogen
Compounds; Hutzinger, O ; Fiedler, H.; Eds.; ECO Informa
Press: Bayreuth, Germany, Vol. 3,1990, pp 169-172.
[24]	Altwicker, E , Fundamental Aspects of Dioxins
(PCDD) from Combustion, New York State Energy Research
and Development Authority, Report 92-3, 1992.
16

-------
[25]	Altwicker, E. et al., Formal ion of Precursors to
Chlorinated Dioxin/Furans Under Heterogeneous Conditions,
Combustion Science and Technology, 83 (1992) 1-19.
[26]	Dickson, L. C., D. Lenoir, and O, Hutzinger,
Quantitative Comparison of de Novo and Precursor Formation
of Polychlorinaled Dibenzo-p-dioxins under Simulated
Municipal Solid Waste Incinerator Post Combustion
Conditions, Environmental Science and Technology, 26
(1992) 1822-1828,
[27]	Luijk, R. et al.. Mechanism of Formation of
Pofychlorinated Dibenzo-p-dioxins and Dibenzofurans in
Catalyzed Combustion of Carbon, Environmental Science and
Technology, 28 (1994), 312-321.
[28]	Bruce, K. R L. O. Beach, and B. K. Gullett, The Role
of Gas-Phase CI2 in the Formation ofPCDD/PCDF During
Waste Combustion, Waste Management, Vol. 11, 97-102,
1991.
[29]	Schindler, P., Municipal Waste Combustion Study:
Emission Data Base for Municipal Waste Combustors,
EPA/530-SW-87-012b (NHS PB87-206082), June 1987.
[30]	U.S. EPA, Municipal Waste Combustors. Background
Information for Proposed Standards, Post-Combustion
Technology Performance, Volume 3, EPA-450/3-89-27c
(NTIS PB90-154865), Research Triangle Park. NC, August
1989.
[31]	Rigo, H. G., Personal Communication, Rigo and Rigo
Assoc., Berea, OH, July 5, 1994.
[32]	National Renewable Energy Laboratory, Polyvinyl
Chloride Plastics in Municipal Solid Waste Combustion,
NREL/TP-430-5518, Golden, CO, April 1993.
[33]	Rigo, H. G., A. I. Chandler, and W. S. Lanier, The
Relationship Between Chlorine in Waste Steams and Dioxin
Emissions from waste Combustor Stacks, American Society
of Mechanical Engineers, Report CRTD-Vol.36, New York,
NY, 1995.
[34]	Environment Canada, The National Incinerator
Testing and Evaluation Program: Environmental
Characterization of Mass Burning Combustor Technology at
Quebec City, Volume D, Main Report, Ottawa, December
1987.
[35]	Kilgroe, J..D. and A. Finkelstein, Combustion
Characterization of RDF Incinerator Technology: A Joint
Environment Canada-United States Environmental Protection
Agency Project, In Proceedings: 1989 International
Conference on Municipal Waste Combustion, Volume 2,
EPA-600/R-92-052b (NTIS PB92-174671), pp 5A-67 Thru
5A-84, March 1992.
[36]	Environment Canada and U.S. EPA National
Incinerator Testing and Evaluation Program: The
Environmental Characterization of Refuse-derived Fuel (RDF)
Combustion Technology, Mid-Connecticut Facility,
Hartford, Connecticut, Summary Report, EPS 3/UP/7 (EPA-
600/R-94-140), December 1994.
[37]	Gullett, B. K., P. M. Lemieux. and J. E. Dunn, Role of
Combustion and Sorbent Parameters in Prevention of
Polychlorinaled Dibenzo-p-dioxin and Polyddorinated
Dibenzofuran Formation During Waste Combustion,
Environmental Science and Technology, 28 (1994), 107-
118.
[38]	Fangmark, I. et al.. Influence of Combustion
Parameters on the Formation of Polychlorinaled Dibenzo-p-
dioxins, Dibenzofurans, Benzene and Biphenyls and
Polyaromatic Hydrocarbons in a Pilot Incinerator, Environ.
Sci. Tech., 27 (1993), 1602-1610.
[39]	Fangmark, I. et al.. Influence of Postcombustion
Temperature Profiles on the Formation of PCDDs, PCDFs,
PCBzs, and PCBs in a Pilot Incinerator, Environmental
Science and Technology, 28 (1994) 624-629.
[40]	Kilgroe, J. D., W. S. Lanier, and T. R. von Alten,
Montgomery County South Incinerator Test Project:
Formation, Emission, and Control of Organic Pollutants, In
Proceedings: 1991 International Conference on Municipal
Waste Combustion, Volume 1, EPA-600/R-92-209a {NHS
PB93-124170), pp 161-175, November 1992.
[41]	Konduri, R. and E. Altwicker, Analysis of Tune Scales
Pertinent to Dioxin/Furan Formation on Fly Ash Surfaces in
Municipal Solid Waste Incinerators, Chemosphere, 1994
(28,1), 23-45.
[42]	Nebel, K. and C. Kane, Large and Small (MWC) Plant
Performance, Memorandum to W. Stevenson (June 22, 1994),
EPA Docket A-90-45, Washington, EXT.
[43]	 Brown, B. and K. S. Felsvang, Control of Mercury and
Dioxin Emissions from United Slates and European Municipal
Solid Waste Incinerators by Spray Dryer Absorption
Systems, In Proceedings: 1991 International Conference on
Municipal Waste Combustion, Volume 3, EPA-600/R-92-
209c (NTIS PB93-124196), pp 287-317, November 1992.
[44]	White, D. M. et al., Emission Test Report, Field Test
of Carbon Injection for Mercury Control, Camden County
Municipal Waste Combustor. EPA-600/R-93-181 (NTIS
PB94-101540), September 1993.
[45]	Kilgroe, J. D. et al., Camden County MWC Carbon
Injection Test Results, Presented at the 1993 International
Conference on Municipal Waste Combustion, Williamsburg,
VA, March 30 to April 2, 1993.
[46]	Hall, B., Reactions of Mercury with Rue Gas
Components, Statens Energiverk, National Energy
Administration, Sweden, (STEV-FBT-91-18). 1991.
[47]	Hall, B., O. Lindqvist, and D. Ljungstrom, Mercury
Chemistry in Simulated Flue Gases Related to Waste
Incineration Conditions, Environmental Science and
Technology, 24 (1990), 108-111.
[48]	Lindqvist, O. and P. Schager, Continuous
Measurements of Mercury in Flue Gases from Waste
Incinerators and Combustion Plants, VDI Berichte, (NR.
838), 1990, pp. 401-421.
[49]	Metzger, M. and H. Braun. In-situ Mercury Speciation
in Flue Gas by Liquid and Solid Sorption Systems,
Chemosphere, 16 (1987), 821-832.
[50]	Licata, A., M. Babu, and L-P. Nethe, An Economic
Approach to Controlling Acid Gases, Mercury and Dioxin
from MWCs, 16th ASME Biennial National Waste Processing
Conference & North American Waste-to-Energy Conference,
Boston, MA June 1994.
[51]	Hartenstein, H. U., Activated Carbon Filters for Flue
Gas Polishing of MWls, In Proceedings: 1993 Conference
on Municipal Waste Combustion, Air & Waste Management
Association (VIP-32), Pittsburgh, PA, 1993, pp 87-105.
[52]	Nebel, K. L. et al.. Emission Test Report OMSS Field
Test on Carbon Injection for Mercury Control, EPA-600/R-
92-192 (NTIS PB93-105518), Air and Energy Engineering
Research Laboratory, Research Triangle Park, NC, September
1992.
17

-------
[53]	White, D. M., et al Parametric Evaluation of
Activated Carbon Injection for Control of Mercury Emissions
from * Municipal Waste Combustor, Paper No. 92-40,06,
1992 Annual Meeting, Air & Waste Management
Association, Kansas City, MO, June 1992.
[54]	Bma, T. G., J. D. Kilgroe, and C. A. Miller, Reducing
Mercury Emission from Municipal Waste Combustion with
Carbon Injection into Flue Gas, ECO World "92 Conference,
Washington, DC, June 1992.
[55]	Felsvang, K, S. and O. Helvind, Results of Full Scale
Dry Injection Tests at MSW Incinerators Using a New Active
Absorbent, In Proceedings: 1991 International Conference
on Municipal Waste Combustion, Volume 3, EPA-60Q/R-92-
209c (NTIS PB93-124196), pp 229-243. November 1992.
[56] Pacyna, J. M., Anthropogenic Mercury Emission in
Europe, Water, Air and Soil Pollution, 56 (1991), 51-61.
This paper is contained in the proceedings of the 1996 ASME
Solid Waste Conference. The Conference was held in Atlantic
City, New Jersey, March 31 to April 3,1996
18

-------
DTD -o not TECHNICAL REPORT DATA
iNKMKlj~JrCl Jr-ir*-Uol (Please read Instructions on the reverse before complel
	
1. REPORT NO. 2.
EPA/600/A-96/044
3.
4. TITLE AND SUBTITLE
Control of Air Pollution Emissions from Municipal
Waste Combustors
5. REPORT DATE
S. PERFORMING ORGANIZATION CODE
7. AUTHORISi
James D. Kilgroe (EPA) and Anthony Licata
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Licata Energy and Environmental Consultants, Inc.
345 Concord, Eoad
Yonkers, New York 10710
10. PROGRAM ELEMENT NO.
It. CONTRACT/GRANT NO.
NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Published paper; 1/95 - 1/96
14. SPONSORING AGENCY CODE
EPA/600/13
is. supplementary notes APPCD project officer is James D. Kilgroe, Mail Drop 65, 919/
541-2854. Presented at 17th Biennial Waste Processing Conference and NAWTEC-IV
Waste-to-Energy Conference, 3/31-4/3/96, Atlantic City, NJ.
is. abstractThe paper provides a brief overview of municipal waste combustor (MWC)
technologies, a summary of EPA's revised air pollution rules for MWCs, a review
of current knowledge concerning formation and control of polychlorinated dibenzo-p-
dioxins and polychlorinated dibenzofurans. and a discussion of the behavior and con-
trol of mercury in MWC flue gases. (NOTE: The 1990 Clean Air Act Amendments
directed EPA to establish MWC emission limits for particulate matter, opacity, hy-
drogen chloride, sulfur dioxide, nitrogen oxides, carbon monoxide, dioxins, dibenzo-
furans, cadmium, lead, and mercury. Revised MWC air pollution regulations were
proposed by EPA on September 20, 1994, and promulgated on December 19, 1995.
The MWC emission limits were based on the application of maximum achievable con-
trol technology.)
17. KEY WORDS AND DOCUMENT ANALYSIS
i. DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Combustion
Wastes
Flue Gases
Mercury
Furans
Pollution Control
Stationary Sources
Municipal Waste Com-
bustors (MWCs)
Dioxins
13B
21B
14G
07B
07C
18, distribution statement
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)

-------