Control of Mercury Emissions from Coal Fired Electric Utility Boilers:
An Update
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
February 18,2005
-------
Glossary of Terms
ACT
APC
APPCD
B-PAC
CFD
CS-ESP
DOE
EPA
EPRI
ESP
E-3
FF
FGD
FGD (Norit)
HS-ESP
ICR
L/G
LSFO
MEL
NETL
NOx
ORD
PAC
PJFF
PM
PPPP
PRB
PS
R&D
RD&D
SCA
SCR
SDA
SDA/FF
SEA
STS
TBD
UBC
Activated Carbon Injection
Air Pollution Control
Air Pollution Prevention and Control Division
Brominated Powdered Activated Carbon (product name from Sorbent
Technologies Corp, Twinsburg, OH)
Computational Fluid Dynamics
Cold-side Electrostatic Precipitator
United States Department of Energy
United States Environmental Protection Agency
Electric Power Research Institute
Electrostatic Precipitator
Product name of Norit Americas' halogenated powdered activated carbon
Fabric Filter (baghouse)
Flue Gas Desulfurization
Norit FGD is the product name for an activated carbon produced by Norit
Americas
Hot-side Electrostatic Precipitator
Information Collection Request
Liquid-to-Gas ratio
Limestone Forced Oxidation scrubber
Magnesium Enhanced Lime scrubber
National Energy Technology Laboratory
Nitrogen Oxides
Office of Research and Development
Powdered Activated Carbon
Pulse Jet Fabric Filter
Particulate Matter
Pleasant Prairie Power Plant
Subbituminous coal mined in Wyoming's Powder River Basin
Particulate Scrubber
Research and Development
Research, Development, and Demonstration
Specific Collection Area
Selective Catalytic Reduction
Spray Dryer Absorber
Spray Dryer Absorber with downstream Fabric Filter
Sorbent Enhanced Additive
Sodium Tetrasulfide
To Be Determined
Unburned Carbon
-------
Control of Mercury Emissions from Coal Fired Electric Utility Boilers:
An Update
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
February 18,2005
INTRODUCTION
Coal-fired power plants in the U.S. are known to be the major anthropogenic source of domestic
mercury emissions.1 The Environmental Protection Agency (EPA) has recently proposed to
reduce emissions of mercury from these plants.2 In March 2005, EPA plans to promulgate final
regulations to reduce emissions of mercury from coal-fired power plants. To help inform this
regulatory effort, a White Paper on the status of mercury control technologies for electric utility
boilers was released in February 2004 by EPA's Office of Research & Development.3
Subsequently, much new information has become available on these technologies. This White
Paper has been prepared to document the current status of mercury controls and help inform the
upcoming regulatory action. As will be discussed, control of mercury emissions from coal-fired
boilers is currently achieved via existing controls used to remove particulate matter (PM), sulfur
dioxide (SO2), and nitrogen oxides (NOX). This includes capture of particulate-bound mercury in
PM control equipment and soluble mercury compounds in wet flue gas desulfurization (FGD)
systems. Available data also show that use of selective catalytic reduction (SCR) NOx control
enhances the concentration of soluble mercury compounds in flue gas from some coal-fired
boilers and results in increased mercury removal in the downstream wet FGD system. Controls
are also under development specifically for the purpose of controlling mercury emissions. This
White Paper will focus on the control options that have been, or are currently being, used/tested
at power plants.
THE U.S. POWER SECTOR
The U.S. fleet of coal-fired generating assets covers a range of coals and plant configurations.
The coal and plant characteristics impact the effectiveness of various mercury control methods at
these plants. The U.S. coal-fired power plants typically burn one of three types of fuel: (1)
bituminous coal (also referred to as "high rank" coal), (2) subbituminous coal, and (3) and lignite
(subbituminous coal and lignite are referred to as "low rank" coals). Some of the characteristics
of interest for these coal types are given in Table I.4
The current capacity of U.S. coal-fired power plants is just over 300 GW and includes a wide
range of combinations of installed air pollution control (APC) configurations. In response to
current and proposed NOX and SO2 control requirements, additional NOX control and flue gas
-------
desulfurization (FGD) systems are expected to be installed and more widely used in the future
(see Figures 1 and 2 below). Over half of the U.S. coal-fired capacity is projected to be equipped
with selective catalytic reduction (SCR) and/or FGD technology by 2020. Table 2 shows the
current and projected coal-fired capacity by APC configuration.5
BEHAVIOR OF MERCURY IN COAL-FIRED ELECTRIC UTILITY BOILERS
Mercury may be present in the flue gas in several forms. The specific chemical form - known as
the speciation - has a strong impact on the capture of mercury by boiler air pollution control
(APC) equipment.* Mercury may be present in the flue gas as elemental mercury vapor (Hg°), as
a vapor of an oxidized mercury species (Hg2+), and as particulate-bound mercury (Hgp).
Mercury is present in coal in trace amounts (approximately 0.1 ppm on average). Research by
the U. S. Geological Survey indicates that much of the mercury in coal is associated with pyrite.
Other forms of mercury that have been reported are organically bound, elemental, and in sulfide
and selenide minerals.6 During combustion the mercury is released into the exhaust gas as
elemental mercury vapor, Hg°. This vapor may then be oxidized to Hg2+ via homogeneous (gas-
gas) and heterogeneous (gas-solid, surface catalyzed) reactions.
The primary homogeneous reaction is that with gas-phase chlorine. As the combustion exhaust
gases exit the boiler and cool, thermodynamic equilibrium shifts to favor formation of HgCb
vapor. The temperature window where this transformation occurs varies, based upon coal
conditions, from about 620 °F to 1250 °F.7 At the temperature after the last heat exchanger,
normally around 300 °F, one would expect all of the mercury to be in the oxidized form if the
reactions went to equilibrium. However, gas-phase mercury oxidation is slow and highly
dependent upon the amount of chlorine8 in the flue gas and, in practice, the amount of oxidized
mercury in the flue gas can range from a few percent to over 90%. Therefore, the transformation
of elemental mercury to oxidized mercury is kinetically limited, i.e., the chemical reactions
associated with mercury oxidation do not go to completion.
Heterogeneous (gas-solid, surface catalyzed) mercury oxidation is more complex and depends
upon the availability of surfaces having electrophyllic groups that attract the electron-rich Hg°
atom. The heterogeneous reactions occur mostly on fly ash surfaces or boiler surfaces, especially
if the fly ash contains high amounts of unburned carbon. A proposed heterogeneous oxidation
mechanism indicates the chlorination of carbon by HC1 is a first step toward heterogeneous
oxidation of Hg° to HgCb, and adsorption of the mercury onto the carbon.9 The mercury that is
adsorbed onto solid surfaces, such as fly ash or unburned carbon, is the particulate-bound
mercury, Hgp, which can be captured by downstream PM control devices. Hence, fly ash
characteristics - especially carbon - as well as coal chlorine content play an important role in
mercury speciation and capture.
* In general it is thought that Hg° will not be removed by pollution control equipment without first converting it to
another form of mercury - either Hg2+ or Hgp However, there is also the possibility for interaction between a
charged surface and the elemental mercury vapor. These interactions may be in the form of electrostatic, van de
Waals, and polarization energies (elemental mercury vapor is polarizable).
-------
Other flue gas species - especially 863 and H2O - have also been shown to affect mercury
speciation, tending to suppress Hg° oxidation to Hg2+. This is due to competition for active sites
on the surface of carbon or other flue gas solids. In general, bituminous coals tend to have higher
chlorine contents and also tend to produce higher levels of unburned carbon (UBC) in the fly
ash. As a result, the flue gas from the burning of bituminous coals tends to contain higher
amounts of Hg2+ species while that of subbituminous and lignite coals tends to contain more Hg°
vapor.
MERCURY REMOVAL BY EXISTING CONTROLS
Mercury may be captured as a cobenefit of PM controls and SC>2 controls, as well as through
mercury-specific control technologies. The degree of this cobenefit can vary significantly
depending upon the type of coal being burned and the specific control technology configuration.
This native capture (i.e., mercury capture without add-on mercury-specific control technology) is
seen in Figure 3, which shows mercury removal rates from EPA's Information Collection
Request (ICR) for three different coal types and APC configurations in use at power plants.
There are some important trends in this figure.
• For the same APC configuration, the average mercury removal for bituminous coal was
greater than that for other coals.
• Mercury removal for a fabric filter (FF) was significantly higher than those for a cold-
side ESP (CS-ESP) or hot-side ESP (HS-ESP) for both bituminous and subbituminous
coals (no FF data for lignite coals).
• Average mercury removal for bituminous coal-fired boilers with Spray Dryer Absorber
and FF (SDA/FF) was very high (over 95%); for subbituminous coal-fired boilers with
the same control configuration mercury removal was considerably less (about 25%),
which was actually less than for a FF alone (about 75%).
• In several cases there was a high level of variability in capture efficiency.
The tendency for a higher native mercury capture from boilers burning bituminous coal is likely
a result of the higher chlorine content in the coal and of the tendency of these coals to produce
higher levels of unburned carbon in the flue gas. Both factors will contribute to greater levels of
mercury as Hg2+ and Hgp, which are easier to capture in existing air pollution control equipment
than Hg°.
The improved mercury capture for plants using FF as compared to those using ESPs can be
explained by the increased contact the gas experiences with fly ash and unburned carbon (UBC)
as those accumulate as a filter cake on the FF. The filter cake acts as a fixed-bed reactor and
contributes to greater heterogeneous oxidation and adsorption of the mercury.
The poor removal of mercury by SDA/FF on low rank coals can be explained by the fact that
much of the HC1 in the flue gas is captured by the SDA, leaving inadequate HC1 at the FF to
participate in the oxidation and capture of Hg°. For bituminous coals, usually having a higher
percentage of mercury as Hg2+ due to higher coal chlorine and UBC, this HC1 stripping effect
appears not to be important.
-------
The high variability of mercury capture for several situations indicates that for several cases
there are other important factors besides coal rank and APC configuration. For example, the
bituminous coal with CS-ESP data covers a range of coal chlorine, fly ash carbon, ESP inlet
temperature and coal sulfur levels - all of which can impact mercury capture efficiency. So, even
within any classification of coal or control technology, there may be a significant amount of
variability in the native mercury capture.
Mercury Capture in PM Controls
As seen earlier in the ICR data (Figure 3), a FF can be very effective for mercury capture,
especially for bituminous coals, but for subbituminous coal as well. However, this FF-only
configuration represents less than 5% of the U.S. coal burning capacity and is expected to
decline in the next 15 years (see Table 2).
The native mercury capture in plants having only CS-ESP or HS-ESP was shown to be much less
effective when compared to those with the FF-only configuration. This is because there is much
less contact between gaseous mercury and fly ash in ESPs. Also, HS-ESPs operate at higher
temperatures at which capture in fly ash is not effective. As with the FF-only configuration, the
ESP-only configuration is expected to become less common (though still approximately 20% of
the total capacity) in the next 14 years with the expected installation of NOX and SO2 controls.
Mercury capture in PM control devices becomes much more important with the injection of
sorbents to the flue gas stream. This is discussed in great detail later in this document.
Mercury Capture in FGD Systems
FGD systems typically fall into one of two broad categories. The wet FGD systems include the
common limestone forced oxidation (LSFO) scrubber and the magnesium-enhanced lime (MEL,
or "mag-lime") scrubber. The dry FGD systems are typically spray dryer absorbers (SDA),
which are usually installed in combination with a FF (SDA/FF).
Mercury Capture in Wet FGD Systems
Mercury in the oxidized state (Hg2+) is highly water soluble and thus would be expected to be
captured efficiently in wet FGD systems. Data from actual facilities has shown that capture of
over 90% Hg2+ can be expected in calcium-based wet FGD systems, though there are cases
where significantly less has been measured. It has been suggested that this is primarily a result of
scrubber equilibrium chemistry and good predictive capability for total mercury capture in wet
FGD systems using a thermochemical equilibrium model has been discussed.10 It has also been
shown that under some conditions Hg2+ will be reduced to Hg° and the mercury will be
reemitted.11 In some cases, the reduction of Hg2+ to Hg° and subsequent re-emission have been
abated with the help of sulfide-donating liquid reagent.7 So this limiting FGD scrubber chemistry
and reemission of mercury may result in Hg2+ capture that is significantly less than 90%.
Experience has shown that Hg2+ reduction and reemission may be more difficult to avoid in
magnesium-enhanced lime (MEL) scrubbers due to the much higher sulfite concentration in
these systems.12
-------
The effect of scrubber chemistry and operating conditions on mercury emissions exhibited in
Figure 4, which shows the measured mercury emissions as liquid-to-gas ratio (L/G) was varied
on a 100 MMBtu/hr pilot facility with inlet mercury concentration in the range of about 10-25
|ig/dsm.3 Higher L/G resulted in lower outlet mercury emissions which has implications for wet
FGD type - Limestone Forced Oxidation having higher L/G than Magnesium Enhanced
Limestone (MEL) wet FGD.
Figure 5 shows the mercury removal for various FGD systems reported by different sources,
including mercury removal with SCR in operation.12'13'14'15 All results, except those on the far
right of the figure, are from wet FGD systems. There isn't enough data in this figure to show
clear trends between various wet FGD system types. But, more detailed examination of scrubber
operating characteristics would likely reveal that the scrubber chemistry may be optimized to
achieve high mercury removal as well as high 862 removal.10' n'12 In any event, it is clear that
the use of SCR and FGD combination consistently yielded mercury removal of nearly 90% in
each of the applicable cases shown in Figure 5. In a comparable study partially funded by the
U.S. DOE, 6 boilers fired with bituminous coal and equipped with SCRs, ESPs, and wet FGDs
were shown to reduce total mercury emissions by 85%.16 SCR impact on mercury speciation is
discussed in the following section.
Oxidation ofHg° to Hg2+ by SCR Catalysts
Because Hg2+ can be captured much more effectively than Hg° in wet FGD systems, methods to
increase the amount of Hg2+ upstream of the wet FGD should improve mercury capture in the
wet FGD system. Under certain conditions, SCR catalysts have been shown to promote the
oxidation of Hg° to Hg2+, particularly for bituminous coal. The impact of SCR on mercury
oxidation is being investigated in two series of field tests: (1) EPRI-EPA-DOE sponsored field
tests15 and (2) DOE sponsored tests being conducted by CONSOL.13 The results of field test
programs suggest that oxidation of elemental mercury by SCR catalyst may be affected by the
following:15
• The coal characteristics, especially the chlorine content
• The amount of catalyst used to treat the gas stream
• The temperature of the reaction
• The concentration of ammonia
• The age of the catalyst
The above factors have significance regarding the potential benefits of SCR on mercury capture
for bituminous coals vis-a-vis subbituminous or lignite coals. A comparison of the effects of
SCR shows that oxidation of Hg° to Hg2+ is greater for bituminous coals than for subbituminous
coals (no data is available for lignite). In fact, in most cases the use of SCR resulted in about 85 -
90+% mercury in the oxidized form when firing bituminous coals. Figure 6 shows data from the
EPRI-EPA-DOE field test, the DOE-CONSOL field tests and from field tests conducted at
Dominion Resources Mount Storm Unit 2.12 In particular, the figure reflects the percent Hg2+
measured at the inlet of the CS-ESP for boilers equipped with SCR.12'13'15 Where data is
available with the SCR off-line, it is also shown. In every bituminous coal case except S3, the
percent Hg2+ increased. In the case of S3, a sampling artifact is suspected.11 In the case of the
PRB-fired unit, Hg2+ concentration remained very low. It should be noted that there may be
some uncertainty associated with speciated mercury measurements upstream of a PM control
-------
device. This is because PM in the extracted sample may cause oxidation of elemental mercury in
that sample.
It would be desirable to increase the oxidation of mercury by SCR when firing subbituminous
coals to levels approaching the oxidation levels of bituminous coals. To investigate if this was
possible, Senior and Linjewile17 compared the results of thermochemical equilibrium
calculations of mercury species concentration to full-scale and pilot test results. There was good
correspondence between the results of the calculations and test results, suggesting that oxidation
of Hg° to Hg2+ with SCR when firing subbituminous coal is limited by equilibrium rather than by
kinetics. Hence, an improvement in catalytic oxidation of Hg° to Hg2+ with SCR on boilers firing
low-rank coals is not possible without a change in flue gas chemical composition (such as from a
higher chlorine in coal) or a lower catalyst temperature.
Senior and Linjewile also found that, when ammonia was injected, oxidation of Hg° to Hg2+
tended to drop somewhat.17 This suggests that the presence of ammonia may interfere with
mercury oxidation on the catalyst. Another concern regarding the use of SCR for mercury
oxidation is that of the catalyst age (i.e., as the SCR catalyst ages, the oxidation of mercury may
decline due to a loss in catalyst activity). Although field tests of catalyst oxidation between years
did not show a significant change in mercury oxidation, it is thought that the age of this catalyst
may not be adequate to show a significant change. Most SCR systems installed on U.S. facilities
have been operating for only a few years, so the effect of catalyst age on mercury oxidation may
not be apparent yet.
Mercury removal by SDA/FF systems
As shown in Figure 3, mercury is very efficiently removed by SDA/FF combinations when used
on bituminous coal-fired boilers - an average of approximately 95%. Mercury - mostly in the
form of Hg2+ at the inlet of the SDA with bituminous coals - is captured in the filter cake of the
FF. However, mercury capture in SDA/FF systems tends to be much less in low-rank coals. For
low-rank coals, the low capture of mercury by SDA/FF systems is believed to be a result of the
scrubbing of HC1 in the SDA, which makes oxidation and capture of mercury (mostly in the
form of Hg° for these coals) in the downstream FF less effective. In fact, Figure 3 shows higher
mercury capture by FF when firing subbituminous coal than mercury capture by SDA/FF. This is
believed to be a result of the SDA scrubbing effect in removing HC1 that could otherwise be
available to react on the FF.
Data/Science Gaps and Associated Recommendations
To meet regulatory time lines, research and development (R&D) efforts should be focused on
those areas that are likely to affect the largest number of boilers or are likely to significantly
impact the ability of a class of boilers to reduce mercury. The following are items (in no
particular order) where there is a shortage of data or where knowledge remains inadequate:
• Public information regarding mercury oxidation across SCR catalysts when lignite coal is
fired could not be found.
• The role ammonia plays in interfering with mercury oxidation should be studied in
greater detail. If ammonia interferes with mercury oxidation in the manner previously
-------
suggested, then SCR effectiveness in oxidizing mercury may be reduced as the catalyst
ages.
• The speciation of mercury at the inlet of the SCR, or inlet of the PM control device (in
the case where SCR is not installed), remains difficult to predict because the processes
that govern mercury speciation in the boiler are not yet adequately understood. A better
understanding of this would improve the ability to predict mercury removal performance
of FGD and other air pollution control equipment.
• The understanding of mercury oxidation across SCRs may best be conducted with
coordinated laboratory/pilot and field testing. Field testing alone does not provide
adequate control of conditions to understand this phenomenon.
• More information on the effectiveness of mercury control with wet scrubbers on lignite
or subbituminous fired boilers is needed.
• Reduction of oxidized mercury in wet FGD and subsequent re-emission requires
enhanced understanding.
• Parallel efforts to improve measurement and data collection accuracy and to reduce the
effort necessary for speciated measurements of mercury are critical. These are very
challenging, labor intensive measurements. Significant progress has been made over the
last few years to improve the reliability and accuracy of on-line measurement systems.
However, accurate and dependable measurement of mercury at very low concentrations
will be needed to prove the efficacy of mercury control systems.
• An improved understanding of the behavior of mercury in the boiler and air pollution
control system may offer insights to addressing operational variability. Modeling
supported by verification testing should be pursued to develop these capabilities.
• Efforts examining the potential for leaching of Hg and other metals (e.g., Se, As) from
coal combustion residues (fly ash, scrubber sludge, etc.) are ongoing. Based on limited
number of samples, results of these efforts have indicated that leaching of mercury from
flyash and flyash-sorbent mixtures does not appear to be of concern. However, the
potential for leaching of Hg and other metals should continue to be evaluated over a
range of coal combustion residue types and their management practices.
• The potential for release of mercury in processes involving beneficial use of coal
combustion residues (e.g., wallboard production and other high temperature processes)
should be evaluated.
APPROACHES FOR ENHANCING MERCURY CAPTURE BY SO2 OR PM
CONTROLS
Important factors that have been found to influence the mercury capture as a cobenefit of other
APC equipment include the coal characteristics, especially the chlorine content, and the carbon
content of the fly ash. Therefore, efforts to improve capture efficiency of existing equipment
have been directed primarily at several approaches: fuel blending, addition of oxidizing
chemicals, controlling UBC content of the fly ash, addition of a mercury-specific oxidizing
catalyst downstream of the PM control device. Many of these options are shown in Figure 7.
-------
Fuel Blending
Blending of small amounts of bituminous coal with subbituminous or lignite coal may provide
some benefit to capture of mercury by existing equipment. Coal blending has been shown to
affect UBC and mercury removal. For example, Sjostrom showed that at Holcomb Station, a 360
MW, PRBa-fired boiler equipped with SDA/FF for SC>2 and PM control, vapor phase mercury
capture across the SDA/FF system could be increased from less than 25% to nearly 80% by
blending small amounts of western bituminous coal with the PRB coal.18 Since mercury removal
of SDA/FF systems firing 100% bituminous coals has been shown to be about 90% or greater,
this permits mercury removal performance approaching that of bituminous coals while firing
mostly PRB. The effectiveness of this approach in improving cobenefit mercury controls is likely
to be very facility-specific. The long-term effects of blending fuels would need to be evaluated
for each application. Blending may change boiler slagging/fouling characteristics or the
performance of APC equipment.
Addition of Oxidizing Chemicals
The addition of chlorine to the fuel or injection into the flue gas is another approach that is being
tested for enhancing intrinsic capture of mercury. At Laskin 2 (firing PRB) and at Stanton 10
(firing North Dakota lignite), chlorine salts were added to the fuel to assess the impact of
increasing fuel chlorine on mercury oxidation and capture. Laskin 2 is equipped with a Particle
Scrubber (PS) and Stanton 10 with a SDA/FF. In both cases, mercury oxidation increased,
although for some salts the mercury capture did not increase. In the case of Laskin, opacity was
observed to increase as a result of salt addition and in the case of Stanton 10, pressure drop
across the FF increased. Long-term effects, such as corrosion, plugging, impacts on combustion
equipment could not be assessed during the short-term parametric tests.19 Therefore, the use of
coal additives offer some promise at improving mercury capture; however, they may have other
impacts that need to be evaluated.
Increasing UBC in Fly Ash
Carbon in the fly ash has been shown to be an important factor in mercury capture by PM control
equipment. For example, at Salem Harbor #1, which fires bituminous coal and has a CS-ESP,
mercury removal rates over 80% were measured on some occasions. These were attributed in
large part to the very high carbon content in the fly ash - on the order of 15% or more and
sometimes around 30%.20 As a result of this phenomenon, it is possible to optimize the trade off
between higher fly ash carbon and improved mercury control by adjusting combustion conditions
or fuels. Since unburned carbon is unavailable for the production of steam and electricity
following combustion there is a trade-off between the overall plant efficiency versus mercury
adsorption. In general, this approach may be employed at any plant with a CS-ESP or FF, but
may be more applicable to those with dry bottom boilers. However, the carbon content in the fly
ash may need to be kept within acceptable limits due to constraints with ESP or FF performance
or opacity. For plants with HS-ESPs, this approach is not expected to be effective because the
carbon is not very effective in capturing mercury at HS-ESP temperatures. Finally, other impacts
a PRB coal is a commonly used subbituminous coal that is mined in the Powder River Basin (PRB) in Wyoming.
10
-------
of increased UBC (e.g., waste disposal and byproduct use) could also pose additional constraints
on use of this mercury control approach.
Mercury-Specific Catalysts
Enhancing capture by wet FGD processes is possible if Hg° can be oxidized to Hg2+. Research
efforts are underway to evaluate catalysts that are installed upstream of the wet FGD or injection
of oxidizing chemicals upstream of the FGD. These approaches are undergoing full-scale
evaluation. Important concerns involve longevity of the catalyst and the long-term effects of
oxidizing chemicals on downstream equipment.
Improvement of Mercury Capture in Wet FGD
As mentioned earlier the amount of mercury that is captured in the wet FGD system is limited by
the amount of oxidized mercury entering the scrubber and by the scrubber equilibrium chemistry.
The mercury removal performance of the scrubber may also decrease if absorbed Hg2+ is reduced
and reemitted as Hg° vapor.
Evaluation oflCR Data
An analysis of the EPA ICR data was conducted in order to evaluate reemission from the FGD
system. The data for systems with CS-ESP/wet FGD and HS-ESP/wet FGD were evaluated. The
increase of the amount of elemental mercury across the entire system was assumed to be the
result of Hg2+ mercury that was absorbed in the wet scrubber and subsequently reduced and
reemitted as Hg° vapor. The data* was evaluated using the following equation:
%Hg° increase
For the CS-ESP/wet scrubber configuration there was a 5% increase in Hg° across the system for
bituminous (1 unit). The subbituminous (3 units) and lignite (2 units) plants showed a 4-50% and
12-21% decrease of Hg° across the system, respectively.
For the HS-ESP/wet scrubber configuration, however, there was a greater tendency for
increasing Hg° across the system. For bituminous plants (2 units) the Hg° increased by 2-12%,
while the subbituminous plants (4 units) ranged from a 30% decrease up to 67% increase (note: 3
of the 4 subbituminous units showed a net increase of elemental mercury across the APC
system). There were no lignite plants with the HS-ESP/wet scrubber configuration.
Field Testing
Babcock & Wilcox (B&W) and McDermott Technology, Inc. completed field tests at two
commercial coal-fired utilities with wet FGD systems. The work was funded by DOE/NETL, the
Ohio Coal Development Office, and B&W. The test sites were (1) Michigan South Central
Power Agency's (MSCPA) 55 MW Endicott Station and (2) Cinergy's 1300 MW Zimmer
* Only data where the inlet and outlet flows were within +/- 30% were used in this analysis. Also, cases where there
was a net increase of Hg2+ across the system were not used in this analysis.
11
-------
Station. Endicott Station uses a limestone forced oxidation (LSFO) wet FGD system, while
Zimmer Station using Thiosorbic® Lime (magnesium enhanced lime, MEL) and ex situ
oxidation. High-sulfur bituminous coal was burned at both locations. The results of the tests were
as follows:21
At the Endicott Station, total mercury removal across the wet FGD system during the 4-month
long tests ranged from 76% to 79%. Most of the oxidized mercury present in the flue gas was
removed in the scrubber system (approximately 96% removal). No increase in elemental
mercury concentration across the scrubber was observed. This indicated that the control
technology was successful in maintaining high levels of oxidized mercury removal and also
simultaneously suppressing mercury reemission.
Following the Endicott test program, a 15-day verification test was performed at the Zimmer
Station. The average total mercury removal across the wet FGD system averaged approximately
51% during this test. Lower removal of oxidized mercury was observed, as compared to that
measured at Endicott (87% vs. 97%). Additionally, the elemental mercury concentration
increased across the wet FGD system, by approximately 40%. This indicated that the technology
was less effective in removing oxidized mercury and ineffective in suppressing reemission of
captured mercury from the scrubber.
In order to evaluate the individual and cumulative role of SCR catalyst, ammonia injection and
chemical additive on the speciation and removal in a limestone forced oxidation (LSFO)
scrubber, testing was conducted at Dominion Resources Mount Storm power plant (Unit 2, 563
MW, firing medium-sulfur Eastern bituminous coal).12 Baseline mercury removal testing was
completed under several scenarios including full flue gas bypass of the SCR and flue gas flowing
through the SCR with and without ammonia injection. After baseline testing, a chemical additive
(sodium hydrosulfide, NaHS) was injected into the scrubber recirculation pumps to evaluate its
impact on oxidation and reemission of elemental mercury. The results of the tests are given in
Tables.
The results showed that during testing with the SCR bypassed and no injection of the additive,
the scrubber still captured greater than 90% of the oxidized mercury (71% total mercury).
However, there was reemission of Hg2+ as Hg° vapor as indicated by the approximately 15% net
increase of elemental mercury across the scrubber. Under the same conditions (SCR bypassed)
but with the chemical additive injected, there was again greater than 90% capture of the Hg2+ and
approximately 30% capture of Hg° (actually Hg° that was oxidized and retained as Hg2+). There
was also an increased total mercury removal, at 78%. In tests where the flue gas was directed
through the SCR, both with and without chemical additive injected into the scrubber
recirculation pumps, the removal of Hg2+ increased to greater than 95% and the total mercury
removal increased to greater than 90%.
These results indicate the effectiveness of the chemical additive (a NaHS injection technology
that has been patented by B&W) in suppressing the reemission of elemental mercury. The tests
also showed that the presence of the SCR catalyst significantly impacted the mercury speciation
profile at the inlet of the wet scrubber, causing oxidation of the remaining elemental mercury.
The oxidized mercury was effectively removed by the wet scrubber.12
12
-------
Data and Science Gaps and Associated Recommendations
For each of these methods used to enhance the removal of mercury by existing equipment, the
effectiveness under the range of operating conditions, and range of coals that a plant may operate
with, would need to be evaluated. Long-term impacts to the plant, particularly corrosion with
approaches that involve chemical additives to the flue gas or fuel, should be examined. Impacts
to other air pollution control devices must also be considered.
The effectiveness of oxidation catalysts over an extended period of exposure to flue gas may
need to be understood better before such technology can be commercially implemented.
MERCURY CONTROL BY SORBENT INJECTION
Unlike the technologies described earlier, where mercury removal was achieved as a cobenefit
with removal of other pollutants, mercury control via injection of sorbent materials into the gas
stream of coal-fired boilers is under development. Injection of dry sorbents, such as powdered
activated carbon (PAC), has been used for control of mercury emissions from waste combustors
and has been tested at numerous utility units in the United States. However, sorbent injection
experience on waste combustors may not be directly transferable to coal-fired electric utility
boilers due to the following reasons.
1. The concentration of mercury in the flue gas of waste combustors is an order of magnitude
higher than for coal-fired boiler systems. Consequently, the amount of mercury captured
per unit mass of carbon injected will, in general, be higher in waste combustors compared
to coal-fired boilers.
2. Typically, the flue gases of waste combustors have higher chlorine concentrations than
those found in flue gases of coal-fired utility boilers. Since performance of ACI in
situations with low levels of chlorine in the flue gas may be adversely affected, ACI
performance on waste combustors may, in general, not be equivalent to that on coal-fired
boilers.
3. In general, coal-fired power plants are much larger in size compared to waste
combustors. For example a large municipal waste combustor may be about the same size
as a small, 40-50 MW, coal-fired plant. Accordingly, duct dimensions, generally, are
much larger in coal-fired plants compared to those at waste combustors. Since mixing of
injected AC and flue gas in the duct affects mercury capture performance, design of AC
injection systems may, in general, be more involved for coal-fired boilers.
Dry sorbent may be injected into the ductwork upstream of a PM control device - normally
either an ESP or FF. In some cases, an FGD (dry or wet) system may be downstream of the
sorbent injection point. Usually the sorbent is pneumatically injected as a powder. The injection
location will be determined by the existing plant configuration and whether additional
downstream PM control equipment, such as a FF, is retrofit. For example, to segregate collected
fly ash from collected sorbent it may be beneficial to retrofit a pulse-jet FF (PJFF) downstream
13
-------
of an existing ESP and inject the sorbent between the ESP and the PJFF. This type of particulate
removal configuration is called a Compact Hybrid Particle Collector (COHPAC™) and when
combined with sorbent injection is called Toxic Emission Control (TOXECON™). Therefore,
for boilers currently equipped only with PM control devices, implementing sorbent injection for
mercury control would likely entail either:
• injection of powdered sorbent upstream of the existing PM control device (ESP or FF); or
• injection of powdered sorbent downstream of the existing ESP and upstream of a retrofit
PJFF, the TOXECON™ option; or
• injection of powdered sorbent between ESP fields (TOXECON-II™ approach).
Above powdered sorbent injection approaches might also be employed in combination with
existing SO2 control devices. For example, powdered sorbent might be injected prior to the SO2
control device or after the SO2 control device, subject to the availability of a means to collect the
powdered sorbent downstream of the injection point.
In general, factors that appear to impact the performance of any particular sorbent include:
• injection concentration of the sorbent measured in Ib/MMacl*;
• flue gas conditions, including temperature and concentrations of HC1 and sulfur trioxide
(S03);
• the air pollution control configuration;
• the characteristics of the sorbent; and
• the method of injecting the sorbent.
These factors are discussed in more detail in the following sub-section.
Mercury Control by Conventional PAC Injection
The most widely tested sorbent for mercury control at utility boilers is PAC. Initial work focused
on use of PAC because it is a material that is currently available for other uses (e.g., water
treatment). PAC has been evaluated for mercury control in several pilot- and full-scale tests.
More recently, field tests have been performed with other powdered sorbent materials such as
enhanced PAC and silica-based sorbents, which are specifically formulated for controlling
mercury emissions from power plants. As will be shown in the following sections, these
specially-formulated sorbents appear to offer advantages over standard PAC in certain situations.
Numerous field tests have been undertaken to evaluate the use of powdered sorbent, especially
PAC, on capture of mercury from power plants. These tests have been sponsored by the
Department of Energy/National Energy Technology Laboratory (DOE/NETL), the Electric
Power Research Institute (EPRI) and utility companies. Table 4 shows the test programs that
b Sorbent injection rate is expressed in Ib/MMacf, i.e., pounds of sorbent used for each million actual cubic feet of
gas. For a 500 MW boiler, a sorbent rate of 1.0 Ib/MMacf will correspond to approximately 120 Ib/hour of sorbent.
14
-------
have either evaluated, or are evaluating, standard PAC injection for mercury control. As shown
in the table, the test programs cover a variety of configurations and fuel types.
Short-term PAC injection field tests
In the 2001-2003 time period, the Department of Energy/National Energy Technology
Laboratory (DOE/NETL), the Electric Power Research Institute (EPRI) and a group of utility
companies funded relatively short-term field test projects to evaluate the use of ACI as
summarized in Table 5. The first four projects reflected in the table were DOE/NETL Phase I
projects. Experience gained in these relatively short-term projects added to insights on factors,
mentioned above, that appear to impact the mercury capture performance achieved via PAC
injection. Some of this experience is described below.
Figure 8 shows the performance of a commonly used PAC product, Norit FGD, in full-scale
parametric tests conducted in four of the projects.22 Results in this figure reveal that, in general,
injection of PAC at increasing amounts tends to increase mercury removal efficiency. However,
in some cases, a limiting value of removal efficiency may be reached above which additional
injected carbon will not provide additional mercury removal.
Temperature is known to impact the adsorption capacity of PAC, and therefore plays a very
significant role. In most cases, the gas temperature at the available injection point upstream of
the PM control device is around 300 °F and PAC has been shown to work effectively at this
temperature. However, at temperatures approaching 350 °F or more, the effectiveness of
standard PAC drops off rapidly.23 This was verified in testing at Salem Harbor, where increasing
the ESP inlet temperature from 300°F to 350°F reduced mercury removal from approximately 90
% to the 10-20 % range. Fortunately, cases where gas temperatures are this high are less
common; such high temperatures, however, are found in lignite-fired boilers. Enhanced PACs or
other sorbents may offer the capability to operate at much higher temperatures.
As shown in Figure 8, mercury removal close to 90% was achievable at injection rates
approaching 20 Ib/MMacf at Brayton Point 1. On the other hand, at Pleasant Prairie Power Plant
(PPPP) #1 it was not possible to achieve greater than about 65% mercury removal regardless of
the PAC injection rate. An important difference between these two CS-ESP sites was the amount
of unburned carbon in the fly ash. The PPPP fly ash typically has very low levels of UBC in the
fly ash (<1%) while the Brayton Point unit is equipped with low-NOx burners and typically has
higher levels of UBC in the fly ash. Another significant difference between these two sites was
the relatively high level of chlorine in the coal used at Brayton Point (around 2000-4000 ppm)
versus the relatively low chlorine level in the PPPP coal (around 8 ppm).23 Moreover, the
alkaline fly ash at PPPP was believed to have neutralized some of the HC1 in the flue gas,
thereby leaving relatively little chlorine available for mercury chemistry. The large difference in
the amount of chlorine available in the flue gas for mercury chemistry is believed to contribute to
the high level of Hg2+ at Brayton Point and the low level of Hg2+ at PPPP. Adequate chlorine in
the gas stream is believed to be necessary for capture of Hg° by standard PAC. Because the
mercury in the PPPP gas was almost entirely Hg°, the low chlorine content limited the capture
possible by the PAC. Nevertheless, test results showed that removal efficiency of Hg° and Hg2+
was about the same at PPPP.23
15
-------
It is believed that acid gas components such as 863 compete with mercury for the active sites on
PAC, and thereby can impact mercury capture performance of the PAC. In the tests conducted at
Abbot, the best performance, 73% mercury capture, was achieved by injection of fine FGD at
13.8 Ib/MMacf at an ESP inlet temperature of 341 °F. The high sulfur flue gas appeared to
impair the performance of the PAC.
APC configuration can have a very significant impact on PAC performance. Gaston Unit 3 is a
270 MW bituminous coal fired boiler with a HS-ESP and a FF installed downstream of the air
preheater in a COUP AC™ arrangement. The downstream FF in the COUP AC arrangement is
smaller than a FF sized for the full ash loading. As shown in Figure 8, at Gaston, mercury
removal rates around 90% attributable to the PAC were possible at injection rates less than 5
Ib/MMacf. It is noted, however, that flue gas temperatures at COHPAC™ inlet, about 270 °F,
were lower than those typically found at the inlet of PM controls at power plants. Such low
temperatures may have enhanced capture of mercury in the carbon sorbent. Figure 8 also shows
that the performance of PAC injection upstream of pilot FF when firing a subbituminous coal is
similar to the performance demonstrated at Gaston.23 The intimate contact between the PAC and
the gas stream that occurs in a FF is believed to contribute to much higher removal than is
possible through "in-flight" capture alone, such as for injection of PAC upstream of an ESP.
Although there may be less chlorine present in the subbituminous coal, the high contact between
the PAC and the gas in the FF facilitates the heterogeneous oxidation and adsorption of Hg° onto
the PAC.
For all of the results shown in Figure 8, the sorbent used was a product called Norit FGD. In the
case of PAC injection upstream of a CS-ESP, such as at PPPP and Brayton Point, the choice of
sorbent made a significant difference in the performance. For example, at PPPP using treatment
rates in the narrow range of 2.2 to 2.3 Ib/MMacf the mercury removal ranged from 37% to 51%
depending upon the sorbent selected. And, at Brayton Point for a treatment rate of 10 Ib/MMacf
the mercury removal ranged from 55% to 73%, depending upon the sorbent selected.22 However,
when the sorbent was captured in a FF, such as at Gaston, no significant difference in
performance was observed between standard PAC sorbents (due to increased contact time, as
discussed above). Therefore, it is likely that the PAC properties such as particle size make a
significant difference in performance when most of the mercury capture is in-flight.
Speciated mercury capture data collected in the above projects is presented in Table 6. These
data indicate that concentrations of Hgp in flue gas at the outlet of PM controls were less than
0.35 |ig/Nm3 for conditions without and with sorbent injection. This simply indicates that PM
controls were operating reasonably. With regard to gaseous mercury (both Hg° and Hg2+), the
following observations are made:
• Hg° concentrations at the outlet of PM controls were below 0.51 |ig/Nm3 under sorbent
injection conditions at Salem Harbor and Gaston. However, the same was not the case at
Pleasant Prairie, where such concentrations amounted to about 4.3 |ig/Nm3. This may be
attributed to most of the mercury in the flue gas at Pleasant Prairie being Hg° and
mercury removal efficiencies being about 70% in contrast to the much higher removal
efficiencies achieved at Salem Harbor and Gaston.
16
-------
• Hg2+ concentrations at the outlet of PM controls were below 1 |ig/Nm3 under sorbent
injection conditions at all plants.
• ACT was quite efficient in decreasing Hg2+ concentrations in flue gas at PM control
outlets as seen in results from Pleasant Prairie and Gaston. At Pleasant Prairie, without
ACT, Hg2+ concentration at ESP exit was more than 6 |ig/Nm3, which was reduced to less
than 0.5 |ig/Nm3 with ACL Similarly, Hg2+ concentration at COHPAC exit was reduced
from 10.4 |ig/Nm3 under no-ACT conditions to 0.8 |ig/Nm3 with ACL
Incidentally, at Pleasant Prairie, without ACT, Hg° concentration at ESP exit was more
than 9 |ig/Nm3, which was reduced to less than 4.5 |ig/Nm3 with ACT; that is a reduction
of approximately 50%. This reduction is lower than the greater than 90% reduction
measured for a corresponding decrease in Hg2+ concentration.
The above observations indicate that ACT appears to be quite effective in controlling emissions
of Hg 2+. However, this needs to be substantiated as more data is available.
Cremer et al. discuss the results of a program to evaluate a model for assessing the injection,
mixing, and associated mass transfer effects of injecting sorbent into the gas stream for mercury
capture.24 In this program the PPPP PAC injection system was modeled using Computational
Fluid Dynamic (CFD) software. Modeling suggested that mixing played a significant role in the
in-flight mercury capture and, combined with test data, it suggested that some of the mercury
removal was from PAC that had fallen out of the gas stream and deposited on surfaces. Modeling
also suggested that only the smaller PAC particles (< 20 um in diameter) were contributing
significantly to mercury capture. Other programs to evaluate the importance of injection system
design and particle size characteristics have also confirmed that these mass-transfer factors are
significant in influencing performance of in-flight mercury capture by sorbent injection.25 As a
result, the factors that govern mass transfer of the sorbent must be well controlled for good
performance. And, comparison of test results from different field test programs should consider
if the mixing and distribution effects are comparable before drawing conclusions regarding the
effects of other test conditions.
Longer term field test results with PAC injection26
Following the short-term testing at Gaston described above, some longer term tests were
undertaken. The main objectives of these tests were to further evaluate the potential for the
mercury capture performance seen in short-term tests and to test higher permeability FF bags
with lower pressure drop. In full-load (270 MW) tests, PAC was injected nearly continuously in
the period June 26-November 25, 2003, at rates less than 0.7 Ib/MMacf to maintain acceptable
FF cleaning frequency for FF operation at an air-to-cloth ratio of 8.0 ft/min. This resulted in
weekly mercury removal between 80- 90%, with an average of 86%, for the test period as shown
in Figure 9. To achieve 90% removal, additional lower load (195 MW), or lower throughput,
tests were conducted with FF operating at an air-to-cloth ratio of 6.0 ft/min. In these tests, PAC
was injected for 2 weeks in November 2003 and greater than 90% mercury removal was
achieved at injection rates of 2 Ib/MMacf or more. Subsequently, in December 2003 full-load
(270 MW) tests were conducted with higher permeability bags0 mounted in the FF, which
0 The higher permeability bags were 7.0 denier versus the original bags, which were 2.7 denier (denier is a measure
of the linear density of a fiber and provides an indication of the cross section or thickness of the fibers). The
17
-------
operated at a air-to-cloth ratio of 8.0 ft/min. Greater than 90% mercury removal was achieved at
injection rates of 0.8 Ib/MMacf or more with acceptable FF cleaning frequency. The results of
these tests indicate that if the FF is designed properly to accommodate the carbon loading, 90%
or greater removal of mercury with relatively modest PAC injection rates may be possible in a
TOXECON™ arrangement.
Potential mercury capture constraints with PAC injection
In general, the efficacy of mercury capture in standard PAC increases with the amount of Hg2+ in
flue gasd, the number of active sites6 in the PAC, and lower temperature. The amount of Hg2+ in
flue gas is directly influenced by the amount of chlorine present in the flue gas. Based on these
factors, standard PAC injection appears to be generally effective for mercury capture on low-
sulfur bituminous coal applications, but less effective for the following applications:
• Low-rank coals + ESP with current capacity of greater than 150 GW and not expected to
grow significantly in the future. Lower chlorine and higher calcium contents in coal lead
to lower levels of chlorine in flue gas, which results in reduced oxidation of mercury and,
therefore, lower Hg2+ in flue gas;
• Low-rank coals + SDA + FF with current capacity of greater than 10 GW and expected to
grow significantly in the future. Similar effect as above, except lime reagent from the
SDA scavenges even more chlorine from flue gas;
• High-sulfur coal, current capacity with wet FGD of approximately 100 GW and likely to
grow to more than 150 by 2015. Relatively high levels of SOs compete for active sites on
PAC, which reduces the number of sites available for mercury. Generally, plants will use
wet FGD and, in many cases, SCR; PAC injection may be needed as a trim application;
and
• Hot-side ESPs with capacity greater than 30 GW and not likely to grow. Weak (physical)
bonds get ruptured at higher temperatures resulting in lower sorption capacity.
Mercury Control by Halogenated PAC Injection
Some situations, described above, may not have adequate chlorine present in the flue gas for
good mercury capture by standard PAC. Accordingly, halogenated PAC sorbents have been
developed to overcome some of the limitations associated with PAC injection for mercury
control in power plant applications.27'28 Two different halogenated PAC sorbents have been
tested in field tests. They are Sorbent Technologies Corp. brominated-PAC (B-PAC) and Norit
America's halogenated PAC (E-3).
Halogenated PACs offer several potential benefits. Relative to standard PAC, halogenated PAC
use: (1) may expand the usefulness of sorbent injection to many situations where standard PAC
permeability of the bags was 130 cfm/ft2 @ 0.5" H2O versus the nominal 30 cfm/ft2 @ 0.5"H2O for the original
bags.
Standard PAC binds mercury via physical (i.e., weak) bonds, which are formed more easily with Hg2+. There have
been results that show a similar removal for both elemental and oxidized mercury. However, the results do not
account for surface oxidation/sorption on the carbon.
e These are collection of atoms/radicals such as oxygen, chlorine, hydroxyls, which provide binding sites.
18
-------
may not be as effective; (2) may avoid installation of downstream FF, thereby improving cost-
effectiveness of mercury capture; (3) would, in general, be at lower injection rates, which
potentially will lead to fewer plant impacts and a lower carbon content in the captured fly ash;
(4) may result in somewhat better performance with low-sulfur (including low-rank) coals
because of less competition from SCh; and (5) may be a relatively inexpensive and attractive
control technology option for developing countries as it does not involve the capital intensive FF
installation.
As shown in Table 7, halogenated PACs have been tested at full-scale for many different
combinations of coal type and PM controls. In each of these tests, relatively high levels of
mercury removal were achieved with relatively modest injection rates when compared to results
from tests with non-halogenated PAC injection in similar plant configurations.
Mercury control performance with halogenated PACs
Figure 10 shows comparative results from PPPP (full scale parametric data with PAC and pilot-
scale data with B-PAC), St. Clair (full scale parametric test results with PAC and with B-PAC),
and measured full-scale in-flight B-PAC mercury removal at Stanton 10 (S10), and parametric
test data from Brayton Point.22'27 In all cases, injection was upstream of a CS-ESP or in-flight
removal upstream of the FF was measured. Also, except for Brayton Point, in all cases the HC1
content was expected to be very low due to the coal type (subbituminous in the case of PPPP and
St. Clair, and lignite in the case of S10). As shown, PAC was unable to remove more than about
70% of the mercury at PPPP or St. Clair. In contrast, with B-PAC about 90% removal was
achieved at around 5 Ib/MMacf at both St. Clair and with the PPPP pilot. Also, for all cases
where B-PAC was tested, similar results in terms of percent removal vs. injection rate were
achieved. It should be noted that the ESP associated with B-PAC testing at St. Clair has a larger
than typical SCA (700 ft2/kacfm)29 and this may potentially have contributed to high levels of
mercury removal with modest B-PAC injection rates. On the other hand, since B-PAC injection
rates were modest, they likely did not require a larger collection surface. As shown in Table 7,
70% mercury removal with a B-PAC injection rate of 4 Ib/MMacf was obtained at the Lausche
plant, which fires a high-sulfur bituminous coal and has a mid-sized ESP with SCA (370
ft2/kacfm).27 It is believed that mercury capture at Lausche was affected by the high sulfur
content in coal; 863 in flue gas is believed to compete with mercury for active sites in the
sorbent. At Cliffside (Low-sulfur bituminous coal with a hot side-ESP, and B-PAC), mercury
reductions of 80% were measured at minimal load, and 40% at full load, during short term (2-
weeks parametric) tests.30 It is noted, however, that additional HS-ESP test programs with
longer-term (30-day continuous) tests are planned/underway. These programs are summarized
later in this document. Based on above observations, for boilers not firing high-sulfur coals
and/or not using hot side-ESPs, injection of halogenated PAC, such as B-PAC, appears to have
the potential to provide high levels of mercury removal under conditions that are challenging for
PAC.
Figure 11 shows comparative test results from programs including PAC and B-PAC sorbents
where a downstream FF was used. In these tests, B-PAC performance consistently was superior
in terms of percent reduction for a given injection concentration. For the bituminous coal cases,
the Valley pilot and the Gaston full-scale data correspond almost exactly. Similarly, the Stanton
10 and Valley pilot data correspond almost exactly, although Stanton 10 is a lignite-fired boiler
19
-------
with SDA/FF and Valley is bituminous coal-fired. A comparison of the Stanton 10 data with
PAC injection versus Stanton 10 data with B-PAC injection shows how dramatic the difference
in performance can be.
Performance of a halogenated sorbent such as B-PAC appears to be relatively consistent
regardless of coal type and appears to be mostly determined by whether or not the capture is in-
flight (as in upstream of a CS-ESP) or on a fabric filter. This is shown in Figure 12 where B-
PAC results from parametric tests are shown for in-flight and upstream of a fabric filter for
various coal types. This is a significant development because performance of standard PAC is
impacted by coal type as well as equipment.
Finally, available data shown in Figure 13 reflect that performance of B-PAC and E-3 for
Western coal (PRB, lignite) is similar to that of TOXECON™ (using conventional PAC) for
Eastern bituminous.
All of the data described above in this section are from shorter term parametric tests. The
purpose of these tests was to determine conditions for longer-term testing with sorbents of
interest. In these parametric tests, a sorbent was injected continuously for a few hours and
mercury measurements were made. Subsequently, an operational parameter (usually the sorbent
injection rate) was adjusted and the effect on mercury capture was measured. Each of the data
points in Figures 10, 11 and 12 represents such a measurement. The results from these shorter
term parametric tests were used to determine the condition s for longer-term tests, which are
discussed below.
Longer term tests with halogenated PAC injection
As shown in Table 7, halogenated PACs have been tested for periods of 10 to 30 days at three
plants. The results shown in Figures 14, 15, and 16 reflect relatively high levels of mercury
removal were maintained for most of the testing periods with modest injection rates.
Speciated mercury capture data collected in the above projects is presented in Table 8. These
data indicate that concentrations of Hgp, Hg°, and Hg2+ in flue gas at the outlet of PM controls
were less than 0.6 jig/Nm3 each with halogenated sorbent injection. Thus, in general,
halogenated PAC injection appears to be quite effective in controlling emissions of each of Hgp,
Hg°, and Hg2+. Note, however, that data from Pleasant Prairie indicated that injection of standard
PAC may be limited in controlling Hg°.
Additional testing with halogenated PAC injection
A significant numbers of field tests are planned or ongoing over the next two years to further
evaluate halogenated PACs for power plant applications. These are summarized below.
DOE Phase II Round 1 Projects
• B-PAC: 1-week parametric testing at Duke Energy, Cliffside 2, 40 MW, low-sulfur
bituminous, HS-ESP; December 2004
• B-PAC: 30-day parametric and 30-day continuous testing at Duke Energy, Buck 5, 140
MW, low-sulfur bituminous, HS-ESP; Spring 2005
20
-------
• Norit E-3: 30-day continuous testing at AEP Conesville 6, 500 MW, bituminous, ESP,
wet FGD; June 2005
• Norit E-3: 30-day continuous testing at Detroit Edison Monroe 4, 375 MW,
PRB/bituminous blend, ESP; November 2005
DOE Phase II Round 2 Projects (expected for 2006)
• B-PAC: 30-day parametric and 30-day continuous testing at Progress Energy, Lee 1, 80
MW, low-sulfur bituminous, ESP
• B-PAC: 2-week parametric testing at Progress Energy, Lee 2, 75 MW, low-sulfur
bituminous, HS-ESP
• B-PAC: 30-day parametric and 30-day continuous testing at Midwest Generation,
Crawford 7, 100 MW, subbituminous, ESP, sell ash for concrete
• B-PAC: 2-week parametric testing at Midwest Generation, Will County 3,130 MW,
subbituminous, HS-ESP, sell ash for concrete
• Norit E-3: Testing at TXU Big Brown Steam Electric Station, lignite-subbituminous
blend
• Norit E-3: Testing at 2 facilities with TOXECON II™ configurations; one bituminous
and one PRB; locations TBD
Other Advanced Sorbents and Additives
Aside from standard and halogenated PACs, other advanced sorbents and additives, designed to
overcome shortcomings of PAC in certain power plant applications, are being developed and
tested. These are briefly described below.
About 15% of the power plant fly ash in the United States, or about 10 million tons, is sold as a
cement additive. PAC, when added to fly ash, however, adsorbs the Air Entraining Admixtures
(AEA) that are used to provide the proper amount of fine air bubbles in concrete for good
strength.31 It is possible to segregate the fly ash from the PAC by the use of a TOXECON™
system. However, this entails significant capital expenditure and operating expense associated
with the retrofit FF. An alternative is to modify the sorbents. Silica-based sorbents and specially-
formulated PACs are under development to overcome some of the difficulties associated with
use of PAC on facilities that sell their fly ash as a cement additive. Silica-based sorbents are
being introduced by a company called Amended Silicates. These sorbents have not yet been
tested for mercury removal at the full scale, but will be tested in 2005 at the Miami Fort unit 6, a
175 MW bituminous coal fired boiler with three small ESPs in series.32'33'34 Specially-
formulated B-PAC has been shown to avoid high foam index and offer similar mercury removal
performance as B-PAC in a laboratory test facility.31 However, additional development and
testing of such specially-formulated sorbents is needed.
Sodium Tetrasulfide (STS) is a liquid chemical reagent that is injected into the gas stream in the
same location as PAC. STS has been used on municipal waste combustors to remove mercury. It
has the benefit of reacting to form a stable mercury compound (cinnabar) that can safely be
added to concrete or disposed of. This technology has been tested for coal power plant use at the
pilot scale.35
21
-------
Efforts are underway to develop lower cost sorbents than commercial PAC sorbents. Sources
include capture of devolatilized char from the furnace for use as a low-cost sorbent or oxidizing
catalyst.36 This carbon, known as the "Thief Process", was tested in a TOXECON™ slipstream
at the Pleasant Prairie Unit 1. The mercury reduction provided by the Thief carbon was not as
high as for conventional PAC injected at the same rate, however, if the cost is significantly less it
may justify the higher injection rate. Testing of several other low-cost sorbents such as corn char
or treated fly ash, which may provide cost-effective benefits in some cases, has been
undertaken.37 These sorbents may be most useful in a TOXECON™ arrangement, which will
both segregate the sorbent from the ash and provide improved gas-sorbent contact so that even
low capacity sorbents can provide good performance without impacting fly ash quality. Another
possibility is to use a low-cost sorbent in combination with an additive to enhance sorbent
capacity.
As discussed in the previous sections, chlorine plays an important role in facilitating the capture
of Hg° on sorbent. For low-rank fuels in particular, there may not be adequate chlorine to achieve
high Hg° capture efficiency with PAC. Use of halogenated PAC sorbents comprises one
approach. Another is to inject a chemical into the fuel, or into the gas stream, to provide the
halogens needed for high sorbent capacity. At the 220 MW Leland Olds plant firing lignite coal
and equipped with CS-ESPs, a Sorbent Enhancement Additive (SEA) has been tested. SEA, a
fuel additive, was added between the feeders and pulverizers. For the field tests, average baseline
mercury removal was 18% and the objective was to determine the treatment regimen to reach
55% removal. A one month test with SEA and PAC showed that a 63% mercury removal rate
was achieved over the period.38 Therefore, as an alternative to more expensive sorbents, it may
be possible to use standard PAC, or even less expensive sorbents, along with an additive, when
such additives are broadly available.
Potential Plant Impacts Related to Sorbent Injection
Sometimes a technology may be very useful in reducing the pollutant required, but result in other
adverse side effects that could significantly impact the plant reliability. The following are some
issues that have been considered as possible concerns for sorbent injection.
Impact of PAC injection on downstream PM collector
Calculations reveal that the increase in PM loading to the ESP or FF due to PAC injection would
be relatively modest (see Figure 17). For example, at a standard PAC injection rate of 10
Ib/MMacf, with an ESP arrangement, about a 4% or less increase in total ash loading will be
expected. For most applications, halogenated sorbents will likely be injected at rates less than 5
Ib/MMacf and incremental PM loading in flue gas will be limited to about 2 percent. Thus,
change in PM loading from PAC or halogenated PAC injection may be less than the loading
change expected from routine fuel or fuel batch changes at a power plant. Consequently,
concerns related to impact on ESP performance generally should not relate to increased mass
loading with injection of these sorbents. In fact, in full-scale testing on large (SCA > 400) and
small (SCA-140) ESPs, no adverse impact was shown on PM removal performance.23'39 Results
from Brayton Point shown in Figure 18 reflect that ESP operation did not change with increase
in PAC injection rate, even when this rate was as high as 20 Ib/MMacf.23
22
-------
Balance-of-plant impacts resulting from standard carbon injection are under investigation at
Georgia Power's Yates Units 1 and 2, which are equipped with CS-ESPs with SCA's 173 and
144 ft2/kacfm, respectively. During parametric ACT tests on these units, sorbent was injected at
rates ranging from 2 Ib/MMacf to greater than 12 lb/MMacf.40 It was observed that at both units,
ESP arc rates within all fields increased from 0 to 1 arcs/min to greater than 10 arcs/min. In some
instances arc rates as high as 35 arcs/min were observed, but the arcing was very sporadic in
nature.41 Since arcing can degrade ESP PM capture performance, this was addressed in the
subsequent one-month duration longer-term testing conducted on Unit 1 in the period mid-
November through December 2004. Observations from longer-term tests are pending. In the
longer-term testing, however, it was noted that sorbent injection rate greater than 4.5 lb/MMacf
did not result in increased mercury removal, which ranged from 71-96% across the system.42
Thus it appears to be unclear at this time if increased ESP arcing will occur at practical injection
rates. Results from the longer-term testing at Yates Unit 1 should provide valuable information
in this regard and this issue needs further investigation.
Measurements of PM concentrations at the outlets of ESPs at Yates Units 1 and 2, taken during
parametric and longer-term testing, reflected that 70% of these measured concentrations were
within, or less than, those measured during baseline testing. PM concentrations greater than
baseline values were measured when sorbent injection rates were in excess of 4.5 lb/MMacf It
was observed that injection rates greater than 4.5 lb/MMacf did not appear to provide additional
mercury removal beyond that obtained at 4.5 lb/MMacf. Thus, it is not clear if unacceptable
excursions in ESP outlet PM concentration will take place at practical injections rates. It was also
observed that the highest ESP outlet PM concentration (0.3 Ib/MMBtu) occurred at a sorbent
injection rate of 17 lb/MMacf42 Based on these observations, the potential for unacceptable PM
emissions occurring due to sorbent injection at plants with small ESPs needs further
investigation.
Yates Unit 1 is equipped with a Jet Bubbling Reactor FGD. During longer-term testing, samples
of the scrubber slurry were taken periodically. During the period of 25 November through 10
December the scrubber slurry was observed to be either black or dark in color, while the carbon
injection rate typically ranged from 4-6 lb/MMacf (with a few, brief periods at higher rates).
Between November 25 and 29, the scrubber slurry was its darkest color, with its color slowly
lightening over this time period. Prior to this time period, the scrubber slurry did not exhibit any
visual evidence of carbon contamination. In the subsequent time period, the carbon injection rate
was as high as 12 lb/MMacf, yet no further darkening was observed. Furthermore, while
occasionally high PM concentrations were measured at the ESP outlet, no visible sign of carbon
was noted on any of the particulate measurement train (Method 17) filters. From this limited set
of data, it was observed that the breakthrough of carbon to the scrubber did not appear to be
directly related to the magnitude of the carbon injection rate. Slurry samples will be analyzed to
further evaluate this issue.42
TOXECON™ results from Gaston have indicated, however, that retrofit FF will need to be sized
properly to maintain acceptable FF pulsing frequency. Also, longer -term (24 days) halogenated
carbon (E-3) injection tests on the SDA/FF system at Stanton Unit 1043 revealed that the cleaning
frequency of the FF increased to every 3 to 4 hours, as compared to 6 to 8 hours during baseline
operation without sorbent injection. However, during sorbent injection tests, the slurry feed to
23
-------
the SDA was varied to accommodate variations in coal sulfur content, which did not occur
during the baseline testing. Since slurry feed rate can affect the frequency of FF cleaning, it was
not possible to quantify the contribution of the sorbent injection to the changes in FF cleaning
cycle. However, it was estimated that sorbent injection increased PM loading by a small amount,
nominally 0.2% at an injection rate of 1 Ib/MMacf and this small increase likely did not
influence FF cleaning cycle. In addition to a change in cleaning frequency, a 4 - 6% increase in
opacity was also observed for a very short time period (< 5 minutes) immediately after each FF
cleaning step.
Potential for increase in fine PM emissions
Typically, the median of the particle size distribution for PAC or halogenated PAC sorbents is
about 20 um, and less than 5% of the sorbent particle mass is fine PM (i.e., less than 2.5 um in
size).44 Calculations suggest that the increase in fine PM loading (i.e., fine PM mass added with
sorbent injection relative to total PM mass in the flue gas) with sorbent injection rate of 10
Ib/MMacf would be less than 0.2%. Further, fine PM removal efficiency of ESPs has been noted
to be about 96%.45 Accordingly, sorbent injection would be expected to increase direct fine PM
emissions by less than 0.01%. These indications, however, need to be substantiated with
measurements.
Fly ash sale for concrete manufacture
As discussed above, there is some concern about the impacts to marketing of fly ash for
beneficial reuse, especially when the ash is used as a cement additive. This is more of a concern
for units with ESPs where standard PAC treatment rate could be high enough to cause a problem
with fly ash sales. But, it may also be a concern in some plants equipped with a FF. At this time
there are a few technical approaches that may mitigate such concerns. One is segregating the fly
ash with a TOXECON™ system. This, however, would entail higher capital costs. TOXECON-
II™ may offer a lower cost approach, but it is under early testing. Finally, the last and the lowest
capital cost approach - specially formulated sorbents - is under development as discussed above.
Adverse effects of PAC or additives on downstream equipment
So far, none of the test programs have shown any significant adverse impact to downstream
equipment from the injection of PAC. Additives used to enhance PAC performance or other
sorbents that may be used, however, might cause adverse effects on plant components. The
potential for such impacts would have to be evaluated through long-term field testing conducted
over time periods of several months or more.
Other Considerations Associated with Broad-Scale use of Sorbent Injection Systems
An assessment of the potential for broad-scale use of sorbent injection systems for mercury
control has to take in to consideration retrofitting time frames that may be necessary to install the
equipment and the availability of sorbents and associated hardware. These considerations are
described below.
Retrofitting time frames and availability of hardware
A dry sorbent injection system is comprised of a storage silo, metering valves, blower and
pneumatic material handling equipment, piping to the ductwork where injection occurs, and
24
-------
associated control hardware. Injection ports are typically installed on low-pressure ductwork. All
of this equipment is standard hardware that is readily available. EPA estimated that about 15
months would be needed from initial engineering review of technologies by the owner to
completion of control technology testing for a PAC injection system without a retrofit FF.46 The
outage could be a week or so due to the minor changes to the ductwork that would be needed for
the PAC injection system. Retrofitting a PAC injection system becomes more complicated if a
TOXECON™ arrangement is chosen. In that case, about 26 months would be needed from initial
engineering review of technologies by the owner to completion of control technology testing for
the PAC injection system.
Since advanced sorbents will generally be injected at lower rates and specialized formulations,
currently under development, may not adversely impact the value of the fly ash for beneficial
reuse, there is a reduced likelihood that addition of a downstream FF or other PM control device
modification will be needed. Therefore, retrofitting time frames for injection systems utilizing
advanced sorbents should be the same as those for standard PAC systems without retrofit FF.
Availability of Sorbents
According to Norit, there is adequate excess capacity in the standard PAC market to absorb
significant growth from current demand levels. However, if the regulations for mercury are
passed, they would envision that plant(s) to supply this market would be built within 2-3 years.47
In general, halogenated PAC sorbents are made from standard PAC materials and halogens.
Based on indications from vendors, supply of PACs may not be a constraint in manufacturing
adequate amounts of halogenated PACs to provide significant reductions if this technology is
used.
B-PAC is currently available from a facility capable of producing 1500 tons/year of sorbent.
According to the vendor, production can be expanded in a six month period to meet market
demand.27
Data/Science Gaps and Associated Recommendations
A number of chemical effects appear to be significant; however, they are not understood well
enough to permit accurate prediction of sorbent performance under all circumstances, which may
be desirable for design and selection of a sorbent injection system. Some of these effects include:
• The impact that chlorine, or HC1, has on sorbent capacity to adsorb Hg° is recognized but
not understood in a quantitative way. This is a particular problem for coals with low
chlorine levels that produce mercury mostly in the form of Hg°.
• 863 is known to interfere with mercury capture. But, like chlorine, a quantitative
understanding is lacking. This is important for high sulfur boilers or those with flue gas
conditioning.
• Mercury concentration and speciation may influence the capture effectiveness of the
sorbent. However, quantitative data on this effect is lacking.
25
-------
Improved understanding of the chemical effects may be addressed through continued lab and
pilot programs in parallel with the full-scale demonstration programs.
Mass transfer plays a critical role in influencing the effectiveness of a sorbent in capturing
mercury, particularly for an application where mercury removal is largely in-flight. Some
specific issues include:
• The role of sorbent physical characteristics. Reference 24 suggested that particle size
distribution plays an important role.
• The relative importance of macro-scale (mixing of sorbent with the gas) versus micro-
scale (diffusion at sorbent-gas interface) mass transfer processes under different
conditions will guide sorbent injection system design.
• The extent to which performance can be improved by improvements in sorbent injection
and distribution methods should be evaluated.
To improve understanding of mass-transfer effects, further development and validation of
models, such as those described in Reference 24, should continue in combination with testing.
There are a significant number of boilers with HS-ESPs. Advanced sorbents may have the
potential to dramatically reduce the control cost for these units that might otherwise need to add
a downstream FF at high capital cost. Although HS-ESP test programs are planned, none of the
demonstration programs appear to address HS-ESPs on low rank fuels. This data gap may need
to be addressed, especially with advanced sorbents.
Efforts evaluating leaching of mercury and other metals from by products of PAC injection have
generally shown that leaching of mercury does not appear to be a concern.48'49 The EPA's Office
of Research and Development also has a program to evaluate the potential for leaching of metals
from the management of mercury-enriched coal combustion residues (CCRs). To date, an
evaluation of CCR's from five coal-fired power plants has been completed - three facilities with
activated carbon injection (ACT) and two with scrubbers. This includes analysis and
quantification of the leaching potential for mercury and non-mercury metals (arsenic and
selenium). Results have also been completed to evaluate mercury leaching potential for 5
additional facilities representing 9 units.
The findings to date indicate that, for most management practices, leaching of mercury from fly
ash does not appear to be of concern for land disposal of CCRs from facilities with activated
carbon injection (for both regular PAC and brominated PAC). The limited results from scrubber
sludge samples suggest that further evaluation is warranted. Leaching results for arsenic and
selenium do suggest a potential concern and warrant further evaluation. Efforts are underway to
obtain additional CCRs from a wider-range of coal types and air-pollution control
configurations. In addition, better information on CCR management practices is being obtained
to help clarify and document the fate and transport of mercury and other metals. This is an on-
going research program within the EPA's ORD.
Some concern has been raised that the presence of brominated carbon may facilitate the
formation of chlorinated and/or brominated organic air toxics (dioxins/furans) in the flue gas.
26
-------
In October 2004, the Department of Energy (DOE) sponsored an approximately 30-day
continuous B-PAC injection demonstration at Detroit Edison's St. Clair Power Plant. During this
30-day demonstration, an EPA-contracted sampling team collected flue gas samples from Unit
#1, a 160 MW boiler that was burning PRB coal. For the measured chlorinated and brominated
dioxins/furans, only one test (of three) showed a discernible difference in the amount measured
as compared to that from the corresponding blank sample. All chlorinated and brominated
dioxin/furans measured from the testing at DTE St. Clair were at least an order of magnitude
lower than the limit of 60 ng/dscm for large municipal waste combustor (MWC) units with an
ESP-based APCS (there are currently no limits for coal-fired utility boilers). For this reason,
ORD does not expect there to be any significant increase in the emission of either the chlorinated
or then brominated dioxins/furans from a coal-fired boiler with the addition of B-PAC or other
similar brominated activated carbon upstream of a CS-ESP, assuming typical operating
temperatures and conditions.
If dioxin/furan compounds were formed in the flue gas, it is likely that they may actually be
captured by unburned carbon in the fly ash and/or on the injected PAC. Samples of the St. Clair
fly ash were subjected to an organic extraction procedure to determine if there was enough
brominated organic material to pursue a leaching test. The results were all in the single digit |ig/g
range. Based on this analysis, it was determined that any brominated compound that may leach
from the fly ash would be at or below the analytical detection limit.
MERCURY CONTROL COSTS ASSOCIATED WITH SORBENT INJECTION
The cost of applying any technology is comprised of annualized capital costs, variable operating
costs and fixed operating costs. The costs of a sorbent injection system are usually very small
compared to other air pollution control equipment if addition of a PJFF or other major PM
control device retrofit is not performed. According to Reference 50, capital costs for sorbent
injection systems may be in the range of about $5/KW - sometimes less. Being simple pieces of
equipment, the fixed operating costs for these systems are also relatively low. So, the major costs
associated with a sorbent injection system are the cost of the sorbent and the disposal of
additional material.
Figure 19 shows estimates of the cost of sorbent and disposal of sorbent for sorbent injection
upstream of a CS-ESP. Estimates are made using halogenated PAC sorbent cost of $1.00/lbf and
PAC sorbent cost of $0.50/lb; disposal is estimated at a cost of $25/ton. As shown in Figure 19,
halogenated PAC is estimated to provide up to about 90% removal at a cost of sorbent and
disposal under 1 mill/kWhr (1 mill/kWhr = $1.00/MWhr). Costs for standard PAC are estimated
to be greater than those for halogenated PAC due to the significantly higher injection rate that is
necessary. At a capital recovery factor of 13.3% and a capacity factor of 80%, levelized capital
charge is approximately 0.1 mills/kWhr, which is significantly less than the variable operating
cost associated with sorbent injection and disposal. A potential cost that is not included here is
the cost of fly ash disposal in the event that fly ash is currently being sold for beneficial reuse but
f Vendors claim that halogenated PACs cost about $0.75/lb today. However, considering the more developmental
nature of these sorbents and the fact that these are aimed at a relatively narrow market, conservatively $1.0/lb was
assumed in this analysis. It is recognized, however, that if these sorbents are available at $0.75/lb under future
market conditions, costs associated with their use will only be less than indicated.
27
-------
must be disposed of if contaminated with carbon. Since most plants do not currently sell their fly
ash, this is not an incremental cost for them. However, for the minority of plants that sell their fly
ash, the incremental costs are estimated to be in the range of 0.38 to 1 mill/kWhr, depending
upon heating value and ash content of the coal and the heat rate of the unit and assuming a
differential between fly ash revenues and disposal cost of $30/ton.50
Figure 20 shows the results of similar cost calculations for PAC injection upstream of a FF. The
cost advantage of halogenated PAC over standard PAC when injected upstream of a FF is not
expected to be as great as when injected upstream of a CS-ESP. Regardless of the sorbent, 90%
removal appears to be possible at sorbent and disposal costs well below 0.50 mills/KWhr when
this technology is available (see a discussion of the outlook for technology availability in a
subsequent section). For facilities that sell their fly ash for concrete, if the fly ash is rendered
unmarketable, the differential cost to dispose of the fly ash is similar to that described above for
CS-ESP.
EXPECTED RESULTS IN NEXT TWO YEARS
The ongoing research activities into control of mercury are proceeding at a rapid pace. Much has
been learned in the last year, and more progress is anticipated over the next two years. There are
several test programs planned over the next year that will explore:
• Mercury capture by FGD and the impact of SCR on FGD capture.
• Use of advanced sorbents in difficult configurations, such as HS-ESP, and sorbents that
are formulated to work with concrete additives when ash is disposed.
• Start up of the first commercial full-scale TOXECON™ system designed expressly to
accommodate sorbent addition.
• Testing of advanced sorbents at units with HS-ESPs without a downstream FF.
• Methods to enhance capture of mercury by existing equipment by fuel additives,
oxidizing chemicals, or oxidation catalysts upstream of FGD.
• Methods to enhance capture of mercury by standard PAC or low cost sorbents using fuel
additives, oxidizing chemicals, or oxidation catalysts upstream of FGD.
• Further evaluation of the fate of mercury in wall-board produced from FGD by-product
to assure that it does not become volatilized and released.
Data available from above testing should help in advancing the development of a broad suite of
viable mercury control approaches.
OUTLOOK FOR TECHNOLOGY AVAILABILITY
In order to examine the status of technology and the factors affecting its availability, Table 9
summarizes the available information to reflect potentially viable mercury control approaches for
various boiler configuration and coal type combinations. As can be appreciated from the
information included in the table, technology availability will: (1) vary by boiler configuration
and coal type; (2) depend on available data (direct and relevant) with much more data expected
in the next 2 years; and (3) depend on regulatory framework, i.e., a spectrum from minimum risk
to technology forcing. The principle concerns relating to broad-scale use of mercury controls are
28
-------
the reliability of mercury reductions possible and the risks of adverse side effects. To the extent
that required mercury reductions are within the capabilities of the technology with minimum
risks of side effects, mercury controls could be considered available. However, as discussed in
this paper, there remain some questions regarding their performance relative to broad-scale use.
These questions are being investigated in ongoing efforts.
In the February 2004 ORD White Paper, it was projected that PAC injection technology would
be available after 2010 for commercial application on all key combinations of coal type and
control technology and that mercury removal levels in the 70% to 90% range could be
achievable. In the same white paper, it was projected that enhanced multipollutant control
systems (PM controls + dry FGD, PM controls + wet FGD, or SCR + PM controls + dry or wet
FGD) would be available after 2010 to provide removal levels between 60 and 90 %. Finally, the
February 2004 White Paper also projected that optimized multipollutant control would be able to
provide 90-95% mercury removal for all coals after 2015. These projections assumed the funding
and successful implementation of an aggressive, comprehensive RD&D program.
As discussed in this document, since the release of the earlier White Paper, additional data,
mostly from short-term tests, have become available on mercury control approaches for power
plants. Also, a broad and aggressive RD&D program is underway, which will yield experience
and data in the next few years. Accordingly, ORD continues to believe that PAC injection and
enhanced multipollutant controls will be available after 2010 for commercial application on
most, if not all, key combinations of coal type and control technology to provide mercury
removal levels between 60 and 90%. However, considering the progress made with halogenated
PAC sorbents and other chemical injection approaches, it is now believed that optimized
multipollutant controls may be available in the 2010-2015 timeframe for commercial application
on most, if not all, key combinations of coal type and control technology to provide mercury
removal levels between 90 and 95%. Such optimized controls could include less expensive use
of sorbent (PAC or halogenated PAC) injection with enhanced SCR and/or enhanced FGD
systems.
A national retrofit program can be initiated after the technology is available. However, full
implementation of such a retrofit program would take a number of years to accomplish and
achieve emission reductions, since large numbers of utilities would need time to order, design,
fabricate and test such units. Based on EPA experience with coal-fired utility boiler retrofit
technologies, we estimate that once a utility has signed a contract with a vendor, installation on a
single boiler could be accomplished in the following timeframes:
• Sorbent injection upstream of an existing ESP or FF could be installed with
commissioning complete in six months to 1 year;
• Sorbent injection upstream of a retrofit fabric filter (e.g., COHPAC) could be retrofitted to
an existing ESP in under 2 years;
• A new SCR/FGD/PM/mercury control system could be retrofitted in 2-3 years dependent
on the retrofit difficulty8; and,
8 It is important to note that due to the high capital cost of SCR and FGD, these technologies are not expected to be
installed solely on the basis of mercury removal, but primarily on the basis of controlling other pollutants, because
much less costly mercury removal methods are available.
29
-------
• Existing SCR or FGD to enhance mercury control could be retrofitted in about one to two
years.
The installation timeframes described above include the time periods associated with control
technology fabrication, delivery, construction, and testing; approval of construction permit; and
modification of operating permit.
SUMMARY
Although the potential mercury emissions from coal-fired U.S. utility boilers are calculated to be
75 tons per year based on the mercury content in coal, the actual current emissions are estimated
to be 48 tons per year due to mercury capture with pollution controls for PM and 862. NOx
controls can augment mercury capture in PM and SC>2 control equipment. And, with increased use
of NOx and SC>2 pollution controls in the coming years as shown in Table 2, the co-benefit
removal of mercury from these controls is expected to be significantly increased. Furthermore, it
may be possible to operate these pollution controls in a way to improve mercury capture without
sacrificing control of other pollutants. RD&D efforts aimed at enhancing mercury removal of
existing controls in the next year should provide information on strategies to maximize mercury
co-benefits.
RD&D activities relating to sorbent injection for mercury removal over the last year have shown
significant advances in this technology. Although most test programs have taken place over a few
days, some have run for several weeks, and one has run for 6 months. These longer term tests
have confirmed the performance of short term tests, and have also provided insights related to
operating concerns expressed by plant personnel. At this time no serious adverse effects on the
plant have been found with sorbent injection for tests up to one month in duration. Moreover,
control approaches have been tested at the full scale and pilot scale under conditions that were
previously shown to be difficult for standard PAC. These approaches, involving advanced
sorbents, fuel blending, and fuel additives, appear to provide effective and less expensive means
to address some of the mercury control difficulty previously identified with low rank coals.
Halogenated PAC sorbents have shown particularly impressive results; they have been tested on a
number of different facilities over a range of fuels and appear to capture mercury effectively for a
given configuration regardless of fuel type. Also, no balance-of-plant impacts have been reported
in field tests conducted thus far.
Some key observations are as follows:
1. Projections reflect that current and future NOx and SO2 emission reduction requirements are
expected to result in growing use of SCR and scrubber systems at coal-fired utility boilers.
Ongoing RD&D has the potential to provide the basis for enhanced mercury removal in such
systems. Assuming sufficient RD&D of representative technologies, new and existing systems
installed to control NOX and SO2 (e.g., SCR+FGD+FF) have the potential to achieve 90+%
control of mercury on bituminous coal-fired boilers. Subbituminous and lignite systems
appear to require mercury oxidation technology and/or additional advanced sorbents to
achieve these levels.
30
-------
2. While co-benefit mercury removal from the flue gas of boilers firing low rank coals is
generally lower than for bituminous coal, methods to improve co-benefit removal - such as
fuel blending and fuel additives - have shown great promise. If successful, these approaches
will narrow the difference between the various coals.
3. While it is more difficult to remove mercury from the flue gas of boilers firing low rank coals
with standard PAC injection, new halogenated sorbents appear to offer a very effective and
less expensive alternative that can deliver higher removals than possible with standard PAC
alone. However, longer-term demonstrations will be beneficial in that they will provide
additional experience and data, which will build confidence in use of these new sorbents.
4. Development and testing of new sorbents or additives has proceeded at a very rapid pace, and
will likely continue to do so. Sorbent injection systems can be installed quickly and new
sorbents or furnace additives can be tested relatively easily. This is in contrast to more capital
intensive technologies like SCR or FGD, which require a long time to engineer and install the
equipment or to make any significant changes to it, such as a catalyst change. Therefore, it is
anticipated that the rapid pace of technology development for new sorbents and chemical
additives will continue.
5. There is increased appreciation for the importance of mass transfer in the sorbent capture
process. Therefore, there is likely to be some improvement in performance due to improved
injection methods, especially for in-flight removal. Such enhancements have the potential to
improve performance for both standard and new sorbents.
6. Halogenated PAC injection with a CS-ESP has the potential to achieve 90% mercury control.
Standard PAC injection with an ESP (HS or CS) and a retrofit fabric filter, or a fabric filter
alone, has the potential to achieve over 90% mercury reduction. Proper design and
consideration of operational and residue impacts would need to be incorporated.
7. The one application that remains particularly challenging is units with HS-ESP. This
represents about 12% of current coal-fired capacity. Currently, retrofit of a downstream FF
after the air preheater and sorbent injection prior to the FF is the only approach that has
proven to be effective for these units at all load conditions. New sorbents have been tested for
achieving mercury removal upstream of a HS-ESP. Promising results were achieved with
these sorbents at part load conditions. With focused RD&D perhaps these sorbents can be
improved to be effective at all load conditions.
8. Cost estimates will vary depending upon specific conditions including regulatory structure.
Nevertheless, for most units, it is projected that the mercury removal would add no more than
about 2 mills/kWh to the annualized cost of power production. For many applications utilizing
halogenated PAC sorbents, costs are projected to be generally lower than 1 mill/kWh. For the
minority of plants that sell their fly ash and this sale is impacted due to mercury control, the
incremental costs, associated with waste disposal and lost revenue due to impact on fly ash
sale, are estimated to be in the range of 0.38 to 1 mill/kWhr. Control by enhancing/optimizing
cobenefit mercury removal in FGD and SCR systems has the potential to reduce costs
substantially, since optimized systems may require little additional investment and/or
operational costs, especially for bituminous coals.
9. It is believed that PAC injection and enhanced multipollutant controls will be available after
2010 for commercial application on most, if not all, key combinations of coal type and control
technology to provide mercury removal levels between 60 and 90%. Also optimized
multipollutant controls may be available in the 2010-2015 timeframe for commercial
application on most, if not all, key combinations of coal type and control technology to
-------
provide mercury removal levels between 90 and 95%. Such optimized controls could include
less expensive use of sorbent (PAC or halogenated PAC) injection with enhanced SCR and/or
enhanced FGD systems. It is noted, however, that broad scale commercial application of
control technology to remove mercury will be possible after the technology is available.
Therefore, initiation of a potential national retrofit program could take place after the
technology is available and such a program would take a number of years to fully implement.
Based on EPA experience with coal-fired utility boiler retrofit technologies, we estimate that
once a utility has signed a contract with a vendor, installation on a single boiler could be
accomplished in the following timeframes:
• Sorbent injection upstream of an existing ESP or FF could be installed with
commissioning complete in six months to 1 year;
• Sorbent injection upstream of a retrofit fabric filter (e.g., COFtPAC) could be retrofitted
to an existing ESP in under 2 years;
• A new SCR/FGD/PM/mercury control system could be retrofitted in 2-3 years
dependent on the retrofit difficulty; and,
• Existing SCR or FGD to enhance mercury control could be retrofitted in about one to
two years.
The installation timeframes described above include the time periods associated with control
technology fabrication, delivery, construction, and testing; approval of construction permit;
and modification of operating permit.
32
-------
Table 1. General Characteristics of Coals Burned in U. S. Power Plants
Coal
Bit
Subbit.
Lignite
Mercury
ppm (dry)
Range
0.036 -
0.279
0.025 -
0.136
0.080 -
0.127
Avg
0.113
0.071
0.107
Chlorine
ppm (dry)
Range
48-
2730
51 -
1143
133-
233
Avg
1033
158
188
Sulfur
% (dry)
Range
0.55-
4.10
0.22-
1.16
0.8-
1.42
Avg
1.69
0.50
1.30
Ash
% (dry)
Range
5.4-
27.3
4.7-
26.7
12.2-
24.6
Avg
11.1
8.0
19.4
HHV*
BTU/lb (dry)
Range
8646-
14014
8606-
13168
9487-
10702
Avg
13203
12005
10028
Higher Heating Value
33
-------
Table 2. Projected Coal-Fired Capacity by APC Configuration
APC Configuration
Cold-side ESP
Cold-side ESP + Wet Scrubber
Cold-side ESP + Wet Scrubber + ACI
Cold-side ESP + Dry Scrubber
Cold-side ESP + SCR
Cold-side ESP + SCR + Wet Scrubber
Cold-side ESP + SCR + Dry Scrubber
Cold-side ESP + SNCR
Cold-side ESP + SNCR + Wet Scrubber
Fabric Filter
Fabric Filter + Dry Scrubber
Fabric Filter + Wet Scrubber
Fabric Filter + Dry Scrubber + ACI
Fabric Filter + SCR
Fabric Filter + SCR + Dry Scrubber
Fabric Filter + SCR + Wet Scrubber
Fabric Filter + SNCR
Fabric Filter + SNCR + Dry Scrubber
Fabric Filter + SNCR + Wet Scrubber
Hot-side ESP
Hot-side ESP + Wet Scrubber
Hot-side ESP + Dry Scrubber
Hot-side ESP + SCR
Hot-side ESP + SCR + Wet Scrubber
Hot-side ESP + SNCR
Hot-side ESP + SNCR + Wet Scrubber
Total Existing Units
New Builds of Coal Steam Units
Fabric Filter + SCR + Wet Scrubber
Total All Units
Current
Capacity, MW
111,616
41,745
-
2,515
45,984
27,775
-
7,019
317
11,969
8,832
4,960
-
2,210
2,002
805
267
559
932
18,929
8,724
-
5,952
688
684
474
304,955
Current
Capacity, MW
-
304,955
2010 Capacity,
MW
75,732
34,570
379
3,161
35,312
62,663
11,979
4,576
2,830
10,885
8,037
4,960
195
2,950
2,601
805
267
557
932
11,763
10,509
538
3,233
6,864
1,490
474
298,263
2010 Capacity,
MW
221
298,484
2020 Capacity, MW
48,915
33,117
379
5,403
22,528
98,138
13,153
2,534
6,088
7,646
9,163
4,960
195
1,330
4,422
2,363
345
557
1,108
10,160
10,398
538
1,847
9,912
1,334
627
297,161
2020 Capacity, MW
17,292
314,453
Note: IGCC units are not included as part of this list.
Note: Current capacity includes some SCR and FGD projected to be built in 2005 and 2006
Note: 2010 and 2020 is capacity projected for final CAIR rule
Note: IPM projects some coal retirements and new coal in 2010 and 2020
34
-------
Table 3. Scrubber testing with SCR and/or additive to reduce re-emission of Hg° and improve mercury
capture.
Test Site Information
Test Site
Dominion
Resources
Mount Storm
Power Station
Unit 2, 563 MW
Coal
Medium-
sulfur
Eastern
bituminous
Controls
SCR, Cold-side
ESP,
limestone forced
oxidation FGD
Mercury Capture Across
Scrubber, %
Elemental
-15%
30%
0
0
Oxidized
> 90%
> 90%
> 95%
> 95%
Total
71%
78%
> 90%
> 90%
Test/Observation
4-day; SCR by-passed; no
additive/Hg(0) re-emission
4-day; SCR by-passed;
with additive/no re-
emission of Hg(0)
6-day; SCR on-line; no
additive/no re-emission of
Hg(0) and higher Hg
removal
7-day; SCR on-line; with
additive/no re-emission of
Hg(0) and higher Hg
removal; no effect from
additive
35
-------
Table 4. PAC test programs on power plant flue gases.
APC
Configuration
CS-ESP
HS-ESP
FF
SDA-FF
PS
Coal Type
Bituminous
Hudsonp
Brayton Point 1
Salem Harbor 1
Yates 1, 2
Valley"
Lausche
Miami Fort
Allen
Conesville
Gaston 3 (TOXECON)
Cliffside
none
none
none
Subbituminous
Pleasant Prairie 1
Powertonp
(TOXECON)
St. Clair 1
Meramec
Miller"
Nantikoke
none
Comanche 2P
Pleasant Prairie 1, 2P
TOXECON
Presque Isle 7-9
TOXECON
Holcomb
Laskin 2
Lignite
Leland Olds 1
Coal Creek
Stanton 1
Poplar River 1, 2
none
Spruce
Stanton 10
Antelope Valley 1
none
p: denotes pilot test
Italic denotes program not yet completed
Table 5. Relatively short-term PAC injection field test projects.
Test Site Information
Test Site
PG&E NEG
Brayton Point, Unit 1
PG&E NEG
Salem Harbor, Unit 1
Wisconsin Electric
Pleasant Prairie, Unit
2
Alabama Power
Gaston, Unit 3
University of Illinois
Abbott Station
Coal
Low-sulfur
Bituminous
Low-sulfur
Bituminous
Subbituminous
Low-sulfur
Bituminous
High-sulfur
Bituminous
Particulate
Control
Two CS-ESPs
in Series
CS-ESP
CS-ESP
HS-ESP +
COHPAC
CS-ESP
Mercury Capture, %
Baseline
90.8
90
5
0
0
ACI Test
Results
94.5
94
65
25-90
73
"Long-term" Test
Duration*
ACI for two 5 -day
periods
ACI for one 4-day
period
ACI for one 5 -day
period
ACI for one 9-day
period
* At these plants, both short-term parametric and "long-term" continuous tests were conducted.
In each of these tests the "long-term" testing lasted less than 10 days.
36
-------
Table 6. Speciated mercury results from ACT field tests.
Pleasant Prairie (Baseline)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
Pleasant Prairie (ACI, 11 Ibs/MMacf)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
Salem Harbor (Baseline)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
Salem Harbor (ACI, 10 Ibs/MMacf)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
Gaston (Baseline)
COHPAC Inlet
COHPAC Outlet
Removal Efficiency (%)
Gaston (ACI injection, 1.5 Ibs/MMacf)
COHPAC Inlet
COHPAC Outlet
Removal Efficiency (%)
Participate
(Mg/dNm3)
1.97
0.01
99.49
1
0
100
10.15
<0.34
>96
4.9
<0.09
>98
0.1
0
100
0.2
0.1
50
Elemental
(Mg/dNm3)
12.22
9.8
19.8
14.7
4.3
70.75
<0.27
<0.50
*
<0.27
<0.51
*
5.5
3.1
43.64
4.2
<0.1
>97
Oxidized
(Mg/dNm3)
2.51
6.01
*
1.7
0.4
76.47
0.09
0.41
*
0.07
0.02
71.43
8.8
10.4
*
5.9
0.8
86.44
Total
(Mg/dNm3)
16.71
15.82
5.27
17.4
4.7
72.99
< 10.51
<1.25
s88
<5.24
<0.62
= 88
14.4
13.5
6.25
10.3
1
90.29
* Efficiency calculation not appropriate because outlet value is greater than inlet value.
Table 7. Full-scale halogenated PAC testing.
Coal Type
Sub-bit Blend
Sub-bit Blend
Sub-bit
BitHigh-S
Bit Low-S
Lignite
Lignite
Lignite
PM
Control
CSESP
SD/FF
CSESP
CSESP
HSESP
SD/FF
*
SD/FF
Hg
Removal
(%)
94
93
80-90+
70
>80
40
95
70
95
Sorbent
B-PAC
E-3
E-3
B-PAC
B-PAC
B-PAC
B-PAC
B-PAC
E-3
Injection
Rate
(Ib/Mmacf)
3.0
1.2
4.0-4.5
4.0
6.4
5.7
1.5
1.5
1.5
Plant
St. Clair
Holcomb
Meramec
Lausche
Cliffside (minimum load)
Cliffside (full load)
Stanton 10
Stanton 10
Stanton 10
Testing
Period
Oct-04
Jun-04
Oct-04
Jan-03
Sep-03
do
early 2004
early 2004
early 2004
Duration
30 day continuous
30 day continuous
10 day continuous
2 - 3 hr tests
2 wks parametric
do
2 hr parametric
2 hr parametric
2 hr parametric
37
-------
Table 8. Speciated mercury results from halogenated PAC injection field tests.
Meramac (Baseline)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
Meramac (E-3 injection, 3 Ibs/MMacf)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
Holcomb (Baseline)
Met
Outlet
Removal Efficiency (%)
Holcomb (E-3 injection, 1.2 Ibs/MMacf)
Met
Outlet
Removal Efficiency (%)
St. Clair (Baseline)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
St. Clair (B-PAC injection, 3 Ibs/MMacf)
ESP Inlet
ESP Outlet
Removal Efficiency (%)
Participate
(Mg/dNm3)
3.54
0
100
3.2
0
99.9
0.47
0.01
97.87
0.01
0
100
0.11
<0.01
>90.9
0.26
<0.1
>61
Elemental
(Mg/dNm3)
9.03
6.57
27.2
3.8
0.4
89.47
7.71
10.75
*
8.49
0.39
95.4
7.18
6.99
2.65
7.11
0.38
94.66
Oxidized
(Mg/dNm3)
2.2
2.73
*
3.1
0.2
93.55
2.38
0.47
80.25
0.9
0.1
88.89
0.61
4.24
*
0.97
0.55
43.3
Total
(Mg/dNm3)
14.77
9.3
37.03
10.1
0.6
94.06
10.56
11.23
*
9.4
0.49
94.8
7.9
11.23
*
8.08
0.93
87.65
* Efficiency calculation not appropriate because outlet value is greater than inlet value.
38
-------
Table 9. Potentially most-suitable mercury control approaches for various boiler configuration and coal
type combinations.
Coal/Technology,
(Capacity, %)h
Subbit + CS-ESP -
w/orw/oSCR
(-66 GW, 22.3%)
Low-S bit + CS-ESP
-w/orw/oSCR
(~ 55 GW, 18.6%)
Incremental
Mercury
Control
Hal-PAC
Hal-PAC
Direct
Experience'
> 90% St.
Clair (80 MW
tested); 80-
90%
Meramec (70
MW tested)
None
Relevant
Experience"
None
> 90% St.
Clair (80 MW
tested); 80-
90% Meramec
(70 MW
tested)
Remaining Issues
a) Direct
experience with
longer-term
testing; testing
with larger duct
sizes; air toxics;
ESP impacts,
residue impacts;
long-term
corrosion; sorbent
supply.
Same as a)
Ongoing Efforts
DOE Phase II
Round 1: 30-day
testing with E-3 at
DTE Monroe 4
(375 MW),
November 2005
DOE Phase II
Round 2: 30-day
continuous testing
withB-PACat
Midwest Gen
Crawford 7 (100
MW), TBD
DOE Phase II
Round 2: testing
with Alstom's
enhanced sorbent
at PacificCorp.
Dave Johnston
plant, TBD
DOE Phase II
Round 2: 30-day
continuous testing
withB-PACat
Progress E. Lee 1
(80 MW), TBD
DOE Phase II
Round 2: testing
with Alstom's
enhanced sorbent
at Reliant En.
Portland unit, TBD
These numbers were generated using the following approach: (1) current capacity numbers are from information in Table 2; (2)
breakdown of coal production was taken from DOE Energy Information Administration, Annual Coal Report 2003, Table 6,
available at (http://www.eia.doe.gov/cneaf/coal/page/acr/table6.html). This reflected production of bituminous, subbituminous,
and lignite in 2003 of 51.5%, 41.3%, and 7.7%, respectively; (3) coal production pattern was considered to be similar to usage in
electric utilities; (4) coal usage pattern was approximated to 50% bituminous, 40% subbituminous, 5% ND lignite, and 5% TX
lignite; and (5) 1/3 of bituminous coal consumed in power plants was considered to be high-sulfur and the remaining 2/3 was
assumed to be low-sulfur.
Direct experience is where experience is available with the exact combination of coal type, existing control technology, and
mercury technology.
J Relevant experience is where applicable experience appears to be available even though not with the exact combination of coal
type, existing control technology, and mercury technology.
39
-------
Table 9. Potentially most-suitable mercury control approaches for various boiler configuration and coal
type combinations; (continued).
Coal/Technology,
(Capacity, %)
Hi-S bit + CS-ESP +
wet FGD (-42 GW,
14.2%)
Coal + HS-ESP
(-35 GW, 12.0%)
Hi-S bit + CS-ESP +
SCR + wet FGD
(-28 GW, 9.4%)
Incremental
Mercury
Control
Hal-PAC
TOXECON;
possibly Hal-
PAC
Hal-PAC
(trim), if
needed. SCR
+ FGD should
provide high
removal.
Direct
Experience
None
>90% Gaston
(135 MW
tested)
None
Relevant
Experience
70% Lausche
(18 MW
tested)
None
>80% various
data on wet
FGD + SCR;
70% Lausche
(18 MW
tested)
Remaining Issues
Same as a) +
impact on FGD
b) Direct
experience with
longer-term
testing; FF
impacts, residue
impacts; sorbent
supply.
Same as a) +
impact on FGD
Ongoing
Efforts
DOE Phase II
Round 1: 30-day
testing with E-3 at
AEP Conesville
(500 MW), June
2005
DOE Phase II
Round 1: 30-day
continuous testing
with B-PAC at
Duke En. Buck 5
(140 MW), low-S
bit, spring 2005
DOE Phase II
Round 2: 2 -week
parametric test
with B-PAC at
Progress En. Lee
2 (75 MW), low-S
bit, TBD
DOE Phase II
Round 2: 2-week
parametric test
with B-PAC at
Midwest Gen Will
County 3 (130
MW), TBD
40
-------
Table 9. Potentially most-suitable mercury control approaches for various boiler configuration and coal
type combinations; (continued).
Coal/Technology,
(Capacity, %)
Hi-S bit + CS-ESP -
w/ or w/o SCR (-27
GW, 9.3%)
ND Lignite + CS-
ESP - w/ or w/o
SCR
(~8 GW, 2.8%)
TX Lignite + CS-
ESP - w/ or w/o
SCR
(~8 GW, 2.8%)
Low-S bit + FF - w/
or w/o SCR (~ 7
GW, 2.4%)
Subbit + FF - w/ or
w/o SCR
(~6 GW, 2.0%)
Incremental
Mercury
Control
TOXECON
Hal-PAC
Hal-PAC
PAC
PAC
Direct
Experience
None (coal
effects less
significant
with FF)
None
None
None (high
intrinsic
removal
expected)
Comanche 2
pilot (600
acfm tested) -
>90%
possible.
Good
performance
expected with
FF.
Relevant
Experience
ICR data and
> 90% Gaston
(135 MW
tested)
> 90% St.
Clair (80 MW
tested); 80-
90% Meramec
(70 MW
tested)
> 90% St.
Clair (80 MW
tested); 80-
90% Meramec
(70 MW
tested)
ICR data and
> 90% Gaston
(135 MW
tested)
ICR data; >
90% Gaston
(135 MW
tested)
Remaining Issues
Same as b)
Same as a)
Same as a)
Same as b)
Same as b)
Ongoing Efforts
DOE Phase II
Round 2: testing
with Alstom's
enhanced sorbent at
Basin Electric
Leyland Olds, TBD
DOE Phase II
Round 2: 30-day
testing with PAC
and advanced
sorbent (TBD) at
TXU Big Brown
(300 MW),
TOXECON
configuration, TBD
41
-------
Table 9. Potentially most-suitable mercury control approaches for various boiler configuration and coal
type combinations; (continued).
Coal/Technology,
(Capacity, %)
Medium-S bit +
SD+FF - w/ or w/o
SCR (~6 GW, 1.9%)
Subbit + SD+FF -
w/ or w/o SCR
(~5 GW, 1.5%)
ND Lignite + FF -
w/ or w/o SCR
(-0.7 GW, 0.2%)
ND Lignite +
SD+FF - w/ or w/o
SCR
(-0.5 GW, 0.2%)
TX Lignite + FF -
w/ or w/o SCR
(-0.7 GW, 0.2%)
TX Lignite + SD+FF
-w/ or w/o SCR
(-0.5 GW, 0.2%)
Incremental
Mercury
Control
PAC (trim), if
needed
Hal-PAC
PAC
Hal-PAC
PAC
Hal-PAC
Direct
Experience
None (high
intrinsic
removal
expected)
> 90%
Holcomb (360
MW tested)
None. Good
performance
expected with
FF.
> 90%
Stanton 10 (60
MW tested)
None. Good
performance
expected with
FF.
None. Good
performance
expected with
FF.
Relevant
Experience
ICR data and >
90% Gaston
(135 MW
tested)
None
Comanche 2
pilot (600
acfm) - >90%
possible
ICR data; >
90% Gaston
(135 MW
tested)
None
Comanche 2
pilot (600 acfm
tested) - >90%
possible
ICR data; >
90% Gaston
(135 MW
tested)
> 90% Stanton
10 (60 MW
tested)
Remaining Issues
Same as b)
Same as a); duct
size may not be an
issue for FF
applications
Same as b)
Same as a); duct
size may not be an
issue for FF
applications
Same as b)
Same as a); duct
size may not be an
issue for FF
applications
Ongoing Efforts
42
-------
Figure 1. Projected coal-fired capacity with FGD
Current 2010
QCAIR Retrofits
H Existing or CAA Title IV/State Retrofit
2015
Figure 2. Projected coal-fired capacity with SCR
200
180
Current 2010
HCAIR Retrofits
• Existing or NOx SIP Call/State Retrofit
2015
43
-------
Figure 3. Mercury removal rates measured for various coal types and air pollution control configurations
(from EPA ICR data, 1999).
Bituminous Coal
Sub-bituminous Coal
Mercury Removal (%)
en
n
f
f
^
T
^
-1-
Q.
(/)
U,
(/)
O
Q
O
H:
Q.
^
OT
O
Q
O
H:
Q.
OT
OT
I
en
n -
I
1
4
>
4
»
4
Lignite
Q. Q.
(/) (/)
U, u,
(/) (/)
O I
Q
O
H:
Q.
W
O
Q
O
H:
Q.
Af\
n
Z 4
4
H:
Q.
O
<
Q
Q
O
St
U. O Q.
U. (O
LU
Figure 4. Effect of liquid-to-gas ratio on mercury emission at common operating pH values
51
4.0 -r
• pH =5
DpH = 5.4
ApH = 5.9
1> 3.0 -
§
'8 2.0 -
1
I 1.0 -
-2 no
o u.u
" I
5 T
D
"tt"
ft
1 1 1
50 100
L/G Ratio, gal/1000 ft3
150
44
-------
Figure 5. Mercury removal of various FGD systems
100 T
QO
80
70
— * fin
E 50
E
o 40
(£ tu
T 30
20
10
0 -
MEL Lime LSFO JBR LSNO SDA
V Cl ILUI I *
•
•
:
PI
—
:
-
i
i
i
•
£iDliDlu-» VJ
Ll_ —1
CO ja
+1 o
CL
CO
LLI
• :
I
ID
O
_l
CL
CO
LU
t!
O
O
T3
LU
-t-
CR+ESP-
CO
E
o
CO
5
CO CO
O
TJ
O
o
O
O)
O
Q
CO
O
Q
CO
c\i
O
45
-------
Figure 6. Percent oxidized mercury into ESP.
Hg(2+) - no SCR
Hg (2+)-with SCR
Note: there is a possible sampling artifact with S3
46
-------
Figure 7. Options for enhancing mercury capture in existing air pollution controls.
Coal & Air
Oxidizing Wet Scrubber Stack
Catalysts
SCR PM Control
47
-------
Figure 8. Parametric test results of mercury removal by standard PAC- all data using Norit FGD.
O Gaston, Bit, COHPAC
D PPPP, Subbit, CS-ESP
A Brayton, Bit, CS-ESP
X PAC-Pilot-Subbit-FF
0.0
5.0 10.0 15.0
Injection Rate (Ib/MMacf)
20.0
25.0
Figure 9. Results of relatively long-term PAC testing at Gaston
52
7/1803 B/7/03 6/27/03 9/16.03 10/6/03 10/26/03 11/15/03
48
-------
Figure 10. Comparison of mercury removal by B-PAC to that by standard PAC for injection upstream of
CS-ESP; standard PAC data is with Norit FGD.
O
t
I
S
O)
ctn
yn
en
An
on
9n J
m -1
I
n*
D
y(^
7 ^r
^ To /
I /
\\ 1
\l
0
0
Q^^
s6
-%K~ o
o
D BPAC-PPPP ESP Pilot
-A- BPAC S10, Lignite (in flight after SDA)
_X— BPAC St Clair Sub ESP
-X- PAC St. Clair, Sub, ESP
O PAC at PPPP
O Brayton, Bit, CS-ESP
10 15 20
Injection Rate (Ib/MMacf)
25
30
35
Figure 11. Comparison of mercury removal by B-PAC to that by standard PAC for injection upstream of
FF; standard PAC data is with Norit FGD.
OGaston.Bit, COHPAC
XPAC-Pilot-Subbit-FF
D PAC, Valley, Bit, Pilot FF
A BPAC, Valley, Bit, Pilot FF
OBPAC,S10,Lig, SDA/FF
XPAC.S10, Lig,SDA/FF
0.0 1.0 2.0 3.0 4.0
Injection Rate (Ib/MMacf)
5.0
6.0
7.0
49
-------
Figure 12. Performance of B-PAC in parametric tests are shown for in-flight and upstream of a fabric
filter for various coal types.
- 40
o
01
30
20
10 -
OH-
0.0
1.0
OBPAC-PPPP ESP Pilot
D BPAC S10, Lignite (in flight after SDA)
XBPACStClair, Sub, ESP
OBPAC,S10,Lig,SDA/FF
A BPAC, Valley, Bit, Pilot FF
2.0 3.0
Injection Rate (Ib/MMacf)
4.0
5.0
Figure 13. Performance of halogenated PACs versus that with standard PACs.5
100
EEHC mt?e: con
E. Bit./TOXECON
E. Bit./ESPs*
Lignite Coal/ESP
Lignite SDIFF - B-PAC & E-3, Stanton 10
Lignite in-flight - B-PAC, Stanton 10
PRB ESP - B-PAC, St. Clair
PRB SDIFF - E-3, Holcomb
* DOE Field Data
10 15 20
injection Concentration, Ib/MMacf
50
-------
Figure 14. Results from DOE-sponsored 30-day full-scale testing at DTE St. Clair 80 MW boiler with
CS-ESP, firing PRB and an 85:15 PRB:bituminous blend and using B-PAC.54
Detroit Edison St. Clair Plant - Total Hg Removal
First 25 Days - Average 94%
9/24 9/26 9/28 9/30 10/2 10/4 10/6 10/8 10/10 10/12 10/14 10/16 10/18
Figure 15. 10-day results from DOE-sponsored full-scale testing at Ameren Meramec 70 MW Unit 2
with CS-ESP, firing PRB and using Norit E-3.55
100
o
(Q
O
O
•10/13/0412:00 10/15/0412:00 10/17/0412:00 10/19/0412:00 10/21/0412:00 10/23/0412:00
51
-------
Figure 16. Results from DOE-sponsored 30-day full-scale testing at Holcomb 360 MW boiler with
SD+FF, firing PRB blend and using Norit E-3. "
o
E
o
o:
D)
I
IUU
on
ou
on
zu
n
*-""* *^ys* ^ -\^-^A^VV^/V-|
** Average Removal
Injection Concentration:
^••^^ V-NJV^-
: 93%
1.2lb/MMacf
"v*v
I
7/7
7/12
7/17
7/22
7/27
8/1
8/6
8/11
Figure 17. Calculated incremental particulate loading in the flue gas from sorbent injection
•D
(0
O
|
o
4.50
4.00
3.50
3.00
2.50
1.50
1.00
0.50
0.00
OH bit.
PRB
ND Lignite
0123456789 10 11 12
Sorbent Injection Rate (Ib/MMacf)
52
-------
Figure 18. ESP power behavior at Brayton Point with increase in PAC injection rate.
56
250
HI
7/13/02 QOO
7/1^020:00
7/1^02 QOO
7/22/020:00
53
-------
Figure 19. Estimated sorbent and disposal costs for sorbent injection upstream of a CS-ESP.
100
OBPAC In Flight, Parametric
DPAC In Flight Sub/Lig
XPAC In Flight Bit
0.5 1.0 1.5 2.0
Sorbent and Sorbent Disposal Cost (mills/KWhr)
Figure 20. Estimated sorbent and disposal costs for sorbent injection upstream of an FF.
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
Sorbent and Sorbent Disposal Cost (mills/KWhr)
0.45 0.50
54
-------
References
1 Study of Hazardous Air Pollutant Emissions from Electric Utility Steam Generating Units
Final Report to Congress., Volume 7, Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, Research Triangle Park, NC; EPA-453R-98-
004a, February 1998; (NTIS PB98-131774).
2 Proposed National Emission Standards for Hazardous Air Pollutants; and, in the
alternative, Proposed Standards of Performance for New and Existing Stationary
Sources: Electric Utility Steam Generating Units; Proposed Rule, Federal Register, 40
CFR Parts 60 and 73, January 30, 2004.
3 Control of Mercury Emissions from Coal-fired Electric Utility Boilers, Air Pollution
Prevention and Control Division, National Risk Management Research Laboratory,
Office of Research and Development, U.S. Environmental Protection Agency, Research
Triangle Park, NC, February 26, 2004.
4 An Assessment of Mercury Emissions from U.S. Coal-Fir ed Power Plants, EPRI
Technical Report 1000608, October 2000.
5 Krolewski, M., U.S. EPA (Clean Air Markets Division-Washington, DC), personal
communication with Srivastava, R.K., U.S. EPA (NRMRL-RTP, NC), November 2004.
6 Mercury in U.S. Coal - Abundance, Distribution, and Modes of Occurrence., USGS Fact
Sheet FS-095-01, September 2001.
7 Control of Mercury Emissions from Coal-Fired Electric Utility Boilers: Interim Report,
Air Pollution Prevention and Control Division, National Risk Management Research
Laboratory, Office of Research and Development, U.S. Environmental Protection
Agency, Research Triangle Park, NC; EPA-600/R-01-109, April 2002.
8 Livengood, C.D., Mendelsohn, M.H., Lani, B.W., "The Chemistry of Mercury
Oxidation", DOE/NETL Mercury Control Technology R&D Program Review,
Pittsburgh, PA, July 14-15, 2004
9 Niksa, S., Fujiwara, N., "Predicting Complete Hg Speciation Along Coal-Fired Utility
Exhaust Systems", Joint EPRI DOE EPA Combined Utility Air Pollution Control
Symposium, The Mega Symposium, Washington, D.C., August 30-September 2, 2004.
10 Niksa, S., Fujiwara, N., "The Impact of Wet FGD Scrubbing On Hg Emissions From
Coal-Fired Power Stations", Joint EPRI DOE EPA Combined Utility Air Pollution
Control Symposium, The Mega Symposium, Washington, D.C., August 30-September 2,
2004.
11 Nolan, P., Downs, W., Bailey, R., Vecci, S., "Use of Sulfide Containing Liquors for
Removing Mercury from Flue Gases", US Patent # 6,503,470, January 7, 2003.
12 Renninger, S., Farthing, G., Ghorishi, S.B., Teets, C., Neureuter, J., "Effects of SCR
Catalyst, Ammonia Injection and Sodium Hydrosulfide on the Speciation and Removal of
Mercury within a Forced-Oxidized Limestone Scrubber", Joint EPRI DOE EPA
Combined Utility Air Pollution Control Symposium, The Mega Symposium,
Washington, D.C., August 30-September 2, 2004.
13 Winberg, S., Winthum, J., Tseng, S., Locke, J., "Evaluation of Mercury Emissions from
Coal-Fired Facilities with SCR-FGD Systems", DOE/NETL Mercury Control
Technology R&D Program Review, Pittsburgh, PA, July 14-15, 2004.
55
-------
14 Nolan, P., Reddinger, K., Amrhein, G., Kudlac, G. "Mercury Emissions Control in Wet
FGD Systems", Air Quality III, Crystal City, VA, September 9-12, 2002.
15 Chu, P., Laudal, D., Brickett, L., Lee, C.W., " Power Plant Evaluation of the Effect of
SCR Technology on Mercury", Joint EPRI DOE EPA Combined Utility Air Pollution
Control Symposium, The Mega Symposium, Washington, D.C., May 19-22, 2003.
16 Cabral, N.J., "Comments of American Coal for Balanced Mercury Regulations," EPA
docket item OAR-2002-0056-5478, December 2004.
17 Senior, C.L, and Linjewile, T., " Oxidation of Mercury Across SCR Catalysts in Coal-
Fired Power Plants", DOE/NETL Mercury Control Technology R&D Program Review,
Pittsburgh, PA, July 14-15, 2004.
18 Sjostrom, S., Starns, T., Amrhein, J., Bustard, J., Durham, M., Penrod, W., Linville, C.,
O'Palko, A., Chang, R., "Full-Scale Evaluation of Mercury Control by Injecting
Activated Carbon Upstream of a Spray Dryer and Fabric Filter", Joint EPRI DOE EPA
Combined Utility Air Pollution Control Symposium, The Mega Symposium,
Washington, D.C., August 30-September 2, 2004.
19 Richardson, C., Machalek, T., Marsh, B., Miller, S., Richardson, M., Chang, R., Strohfus
M., Smokey, S., Hagley, T., Juip G., Rosvold, R., "Chemical Addition for Mercury
Control in Flue Gas Derived from Western Coals" Joint EPRI DOE EPA Combined
Utility Air Pollution Control Symposium, The Mega Symposium, Washington, D.C.,
May 19-22,2003.
20 DOE National Energy Technology Laboratory Mercury Field Evaluation - PG&E NEG
Salem Harbor Station - Unit 1, Project No. 00-7002-76-10, Field Evaluation Summary
Report, January 2003.
21 McDonald, D. K., Amrhein, G. T., Kudlac, G. A., Madden Yurchison, D., "Full-Scale
Testing of Enhanced Mercury Control Technologies for Wet FGD Systems", Final
Report, U. S. Department of Energy Contract DE-FC26-OONT41006, May 7, 2003.
22 Data provided by Jean Bustard, ADA Environmental Services, September 16, 2002.
23 Durham, M., "Field Test Program to Develop Comprehensive Design, Operating and
Cost Data for Mercury Control", DOE/NETL Mercury Control Technology R&D
Program Review, Pittsburgh, PA, August 12, 2003.
24 Cremer, M., Senior, C., Chiodo, A., Wang, D., Valentine, J., "CFD Modeling of
Activated Carbon Injection for Mercury Control in Coal Fired Power Plants", Joint EPRI
DOE EPA Combined Utility Air Pollution Control Symposium, The Mega Symposium,
Washington, D.C., August 30-September 2, 2004.; Note: the presented results supersede
the results discussed in the paper, and the author should be contacted for the presentation
rather than using the conference paper provided in the conference proceedings. Contact
seni or@reacti on-eng. com
25 Madsen, J., O'Brien, T., Rogers, W., Cugini, A., "Computational Modeling of Mercury
Capture by Activated Carbon Injection", DOE/NETL Mercury Control Technology R&D
Program Review, Pittsburgh, PA, July 14-15, 2004.
26 Berry, M., Irvin, N., Monroe, L., Bustard, J., Lindsey, C., Brignac, P., Taylor, T.,
Schlager, R., Sjostrom, S., Starns, T., Chang, R., O'Palko, A., "Field Test Program for
Long-Term Operation of a COHPAC® System for Removing Mercury from Coal-Fired
Flue Gas", Joint EPRI DOE EPA Combined Utility Air Pollution Control Symposium,
The Mega Symposium, Washington, D.C., August 30-September 2, 2004.
56
-------
27 Nelson, S., Landreth, R., Zhou, Q., Miller, J., "Accumulated Power-Plant Mercury-
Removal Experience with Brominated PAC Injection", Joint EPRI DOE EPA Combined
Utility Air Pollution Control Symposium, The Mega Symposium, Washington, D.C.,
August 30-September 2, 2004.
28 Nelson, S., "Advanced Utility Sorbent Field Testing Program", DOE/NETL Mercury
Control Technology R&D Program Review, Pittsburgh, PA, July 14-15, 2004.
29 McCoy, M., Rogers, W., Landreth, R., Brickett, L., "Full-Scale Mercury Sorbent
Injection Testing at DTE Energy's St. Clair Station", Joint EPRI DOE EPA Combined
Utility Air Pollution Control Symposium, The Mega Symposium, Washington, D.C.,
August 30-September 2, 2004.
30 Carpenter, T., McMurry, R., Potter, S., McGinnis, D., Corey, Q., Nelson Jr., S., Landreth,
R., Miller, J., "Mercury Sorbent Results for a Hot-Side ESP at the Cliffside Plant," EPA
docket item OAR-2002-0056-5627, January 14, 2005.
31 Zhou, Q., Zhang, Y., Nelson, S., "Evaluations of Concretes Containing Mercury
Sorbents", Joint EPRI DOE EPA Combined Utility Air Pollution Control Symposium,
The Mega Symposium, Washington, D.C., August 30-September 2, 2004.
32 Butz, J., "Demonstration of Amended Silicates™ for Mercury Control at Miami Fort
Unit 6 - DOE Project 41988", DOE/NETL Mercury Control Technology R&D Program
Review, Pittsburgh, PA, July 14-15, 2004.
33 Butz, J., Lovell, J., Broderick, T., Sidwell, R., Turchi, C., "Evaluation of Amended
Silicate~ Sorbents for Mercury Control", Joint EPRI DOE EPA Combined Utility Air
Pollution Control Symposium, The Mega Symposium, Washington, D.C., May 19-22,
2003.
34 Butz, J., "Demonstration of Amended Silicates™ for Mercury Control at Miami Fort
Unit 6 - DOE Project 41988", DOE/NETL Mercury Control Technology R&D Program
Review, Pittsburgh, PA, July 14-15, 2004.
35 Monroe, L., Berry, M., Irvin, N., Gale, T., Merritt, R., Heaphy, R., Chang, R., Taylor, B.,
Licata, A., Beittel, R., "An Evaluation of Activated Carbon and Sodium Tetrasulfide for
the Removal of Vapor Phase Mercury from PRB Flue Gas at Pilot-Scale", Joint EPRI
DOE EPA Combined Utility Air Pollution Control Symposium, The Mega Symposium,
Washington, D.C., August 30-September 2, 2004.
36 Freeman, M., Pennline, H., Granite, E., Hargris, R., O'Dowd, W., "A Technique to
Control Mercury From Flue Gas: The Thief Process", DOE/NETL Mercury Control
Technology R&D Program Review, Pittsburgh, PA, July 14-15, 2004.
37 Ley, T., Ebner, T., Fisher, K., Slye, R., Patton, R., Chang, R., "Assessment of Low-Cost
Novel Sorbents for Coal-Fired Power Plant Mercury Control", Joint EPRI DOE EPA
Combined Utility Air Pollution Control Symposium, The Mega Symposium,
Washington, D.C., August 30-September 2, 2004.
38 Thompson, J., Pavlish, J., Holmes, M., Bush, C., Brickett, L., "Enhancing Carbon
Reactivity for Mercury Control: Field Test Results from Leland Olds", Joint EPRI DOE
EPA Combined Utility Air Pollution Control Symposium, The Mega Symposium,
Washington, D.C., August 30-September 2, 2004.
39 Dombrowski, K., Richardson, C., Machalek, T., Chapman, D., Chang, R., Monroe, L.,
Berry, M., Irvin, N., McBee, K., Sjostrom, S., "Sorbent Injection for Mercury Control
Upstream of Small-SCA ESPs", Joint EPRI DOE EPA Combined Utility Air Pollution
57
-------
Control Symposium, The Mega Symposium, Washington, D.C., August 30-September 2,
2004.
40 Richardson, C., "Sorbent Injection for Small ESP Mercury Control in Bituminous Coal
Flue Gas," DOE/NETL Mercury Program Review Meeting, Pittsburgh, PA, July 14-15,
2004.
41 Barton, R., McRanie, R., "The Potential Effect of Activated Carbon Injection on Power
Plant Operation," EPA docket item OAR-2002-0056-5485 Southern Att l.pdf, December
2004.
42 Richardson, C., "Sorbent Injection for Small ESP Mercury Control in Low-Sulfur Eastern
Bituminous Coal Flue Gas," Quarterly Technical Progress Report October 1 - December
31, 2004, prepared for DOE/NETL, January 2005.
43 Machalek, T., Richardson, C., Dombrowski, K., Ley, T., Ebner, T., Fisher, K., Brickett,
L., Chang, R., Strohfus, M., Smokey, S., Sjostrom, S., "Full-Scale Activated Carbon
Injection for Mercury Control in Flue Gas Derived from North Dakota Lignite," Joint
EPRI DOE EPA Combined Utility Air Pollution Control Symposium, The Mega
Symposium, Washington, D.C., August 30-September 2, 2004.
44 McCann, B., Norit Americas, Inc., personal correspondence with Hutson, N., U.S. EPA
(NRMRL-RTP, NC), Februrary 16, 2005.
45 AP-42, Fifth Edition, Volume 1 Chapter 1: External Combustion Sources., available at
http://www.epa.gov/ttn/chief/ap42/chO 1/fmal/cO 1 sO 1 .pdf.
46 Engineering and Economic Factors Affecting the Installation of Control Technologies for
Multipollutant Strategies, Air Pollution Prevention and Control Division, National Risk
Management Research Laboratory, Office of Research and Development, U.S.
Environmental Protection Agency, Research Triangle Park, NC; EPA-600/R-02-073,
October 2002.
47 Thomas, B., Norit Americas, Inc., personal correspondence with Hutson, N., U.S. EPA
(NRMRL-RTP, NC), September 6, 2001.
48 Senior, C., Bustard, C.J., Baldrey, K., Starns, T., Durham, M., "Characterization of Fly
Ash From Full Scale Demonstration of Sorbent Injection For Mercury Control On Coal
Fired Power Plants", Joint EPRI DOE EPA Combined Utility Air Pollution Control
Symposium, The Mega Symposium, Washington, D.C., May 19-22, 2003.
49 Aljoe, W., Feeley, T., Murphy, J., and Brickett, L., "The Fate of Mercury in Coal
Utilization By-Products - DOE/NETL's Research Program", Joint EPRI DOE EPA
Combined Utility Air Pollution Control Symposium, The Mega Symposium,
Washington, D.C., August 30-September 2, 2004.
50 Performance and Cost of Mercury and Multipollutant Emission Control Technology
Applications on Electric Utility Boilers, U.S. Environmental Protection Agency, Office of
Research and Development, National Risk Management Research Laboratory, Research
Triangle Park, NC; EPA/600/R-03/110, October 2003.
51 Nolan, P., Farthing, G., Yurchison, D., Holmes, M., "Development of Mercury Emissions
Control Technologies for the Power Industry", Joint EPRI DOE EPA Combined Utility
Air Pollution Control Symposium, The Mega Symposium, Atlanta, GA, August 16-20,
1999.
52 Berry, M., Irvin, N., Monroe, L., Bustard, J., Lindsey, C., Brignac, P., Taylor, T.,
Schlager, R., Sjostrom, S., Starns, T., Chang, R., O'Palko, A., "Field Test Program for
58
-------
Long-Term Operation of a COHPAC® System for Removing Mercury from Coal-Fired
Flue Gas", Paper #118, Joint EPRI DOE EPA Combined Utility Air Pollution Control
Symposium, The Mega Symposium, Washington, D.C., May 19-22, 2003.
53 Original figure from Testimony of Dr. Steven A. Benson, University of North Dakota,
Energy & Environmental Research Center (EERC) to the U.S. Senate, Committee on
Environment and Public Works, Subcommittee on Clean Air, Climate Change, and
Nuclear Safety, June 5, 2003. The figure modified by Sid Nelson of Sorbent
Technologies Corp. to overlay data from demonstrations of halogenated carbon injection,
November 27, 2004.
54 Nelson, S., Sorbent Technologies Corp., personal communication with Hutson, N., U.S.
EPA (NRMRL-RTP, NC), October 17, 2004.
55 Starns, T., "Full-Scale Evaluation of Mercury Control by Injecting Activated Carbon
Upstream of a Spray Dryer and Fabric Filter", POWERGEN Conference, Orlando, FL,
November 30,2004.
56 Durham, M., Bustard, J., Starns, T., Sjostrom, S., Lindsey, C., Martin, C., Schlager, R.,
Chang, R., Renninger, S., Monroe, L., Berry, M., Johnson, D. "Full-Scale Results of
Mercury Control by Injecting Activated Carbon Upstream of ESPs and Fabric Filters",
POWERGEN Conference, Las Vegas, NV, December, 2003.
59
------- |