EPA/600/A-96/118
Mercury Control in Municipal Waste Combustors and
Coal-fired Utilities
S.V. Krishnan*
Acurex Environmental Corporation
4915 Prospectus Drive, Durham, NC 27713
Brian K. Gullett*
National Risk Management Research Laboratory
Air Pollution Prevention and Control Division (MD-65)
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Wojciech Jozewicz
Acurex Environmental Corporation
4915 Prospectus Drive, Durham, NC 27713
*Author to whom correspondence should be addressed
+Currently with: BOC Group, 100 Mountain Avenue, Murry Hill, NJ 07974
-------
ABSTRACT
Control of mercury (Hg) emissions from municipal waste combustors (MWCs) and coal-
fired utilities has attracted attention due to current and potential regulations. Among several
techniques evaluated for Hg control, dry sorbent injection (primarily injection of activated carbon)
has shown promise for consistently removing high levels of Hg from MWC or coal flue gas.
However, the performance in terms of amount of Hg removed per amount of sorbent varies
widely between the MWC and coal-fired applications and from unit to unit.
In this study, we have conducted bench-scale experiments under conditions simulating
MWCs and coal-fired units to study Hg capture by dry sorbents. The effect of reaction
temperature on the capture of different Hg species [Hg° and Hg(II)] by various types of dry
sorbents was the focus of bench-scale tests. An attempt has also been made in this study to
compare the bench-scale results with results obtained from pilot studies and to explain disparities
in fuel- and unit-specific performance. Our investigations showed that the reaction temperature
and Hg species strongly affect Hg control. The results obtained in this suggested the two
following mechanisms for Hg capture:
i) Capture of Hg° by activated carbons is limited by sorption kinetics, and
ii) Capture of mercuric chloride (HgCl2) by activated carbons is limited by collision with
carbon particles.
-------
INTRODUCTION
Title III of the 1990 Clean Air Act Amendments (CAAA) requires the U.S. Environmental
Protection Agency (EPA) to submit a study on 189 hazardous air pollutants (HAPs). This study
would include emissions and a risk (to public health) assessment of the 189 HAPs. Among the
189 HAPs, mercury (Hg) has drawn special attention because of concerns that ingestion offish
containing certain forms of Hg (like methyl mercury) could be harmful to human health.
An EPA "mercury" report to Congress1 cites the largest emitters of Hg as coal-fired
utilities, medical waste incinerators (MWIs), municipal waste combustors (MWCs), chlor-alkali
plants, copper and lead smelters, and cement manufacturers. These sources account for 90 to 100
percent of anthropogenic Hg emissions. Utility boilers account for 24 percent of the total
anthropogenic Hg emissions of which more than 90 percent are attributable to coal-fired utility
boilers. MWIs and MWCs together account for nearly 60 percent of the anthropogenic Hg
emissions. Because of intended, proposed, or existing regulations, control of Hg emissions from
these sources needs to be studied.
There are several methods for controlling Hg emissions in MWCs and MWIs either in use
or currently being tested. These include a variety of chemical, adsorption, and absorption
techniques.2 Primarily, dry sorbent injection (DSI) followed by a fabric filter (FF) or electrostatic
precipitator (ESP), a spray dryer (SD) followed by a FF or ESP, and wet scrubbing (WS) have
been tested as Hg control methods. The sorbents tested in the DSI/FF process to control Hg
emissions include calcium-based sorbents and activated carbons. Varying levels of Hg control in
MWCs using these technologies have been reported.3"5 All tests showed that flue gas
desulfiirization (FGD) units (SD/FF and WS systems) with addition of DSI (injection of activated
carbon) lead to consistently high (> 90 percent) Hg removal in MWCs.
Adoption of control strategies from MWCs, primarily DSI (injection of activated carbon),
has not met with similar success in coal-fired utilities. A coal-fired pilot-scale study6 showed that
high Hg removal at 120 °C required a carbon:Hg ratio of 3,000:1. The amount of carbon:Hg for
similar levels of removal in MWCs has been shown to be an order of magnitude lower. A similar
result, showing the need for high carbon:Hg ratios in coal-fired utilities, was obtained from pilot-
scale studies7 at the University of North Dakota's Energy and Environmental Research Center
-------
(UNDEERC).
At a carbon:Hg ratio of 3,000:1 and an activated carbon cost of $ 1.125/kg, our estimates
indicate that the material cost would be approximately $ 500,000 per year for a 500 MW coal-
fired power plant. Chang et al. (Reference 6) report an annual cost of carbon injection of $
100,000 - $ 1 million for Hg control in a 500 MW coal-fired plant. Another estimate8 suggests
that controlling 50 percent of the Hg emitted by U.S. utility power plants could range from $ 1
billion to $ 10 billion per year. These estimates can vary depending on equipment, installation,
and disposal costs. The significant additional costs associated with Hg removal in coal-fired
utilities emphasize the need for process optimization.
The concentration and speciation of Hg found in flue gas from a coal-fired utility is quite
different than that of a MWC. Typically, the total Hg concentration in coal systems is at least an
order of magnitude lower than that found in a MWC flue gas. Differences in the flue gas
composition [particularly the higher hydrogen chloride (HC1) concentration in a MWC] lead to
different equilibrium speciation of Hg between elemental mercury (Hg°), mercuric chloride
(HgCy, and mercuric oxide (HgO),9 These differences lead to uncertainties in applying Hg
removal experience in MWCs to Hg capture in coal-fired utilities. Therefore, in order to optimize
the Hg removal process in coal-fired utilities, the effects of Hg concentration, Hg species, and
other process conditions must be understood.
In this paper, we discuss our bench-scale results obtained for Hg control under MWC and
coal-fired conditions. In the former case, capture of roughly 30 ppb of inlet Hg by different types
of sorbents was studied at 100 and 140 °C. In the latter case, capture of lower concentrations (1-
2 ppb Hg) of Hg by the same sorbents was studied. Capture of different species of Hg (Hg° and
HgCI2) was studied in both the MWC and the coal-fired situations. Finally, the understanding of
Hg° and HgCI2 capture obtained from our bench-scale efforts is extended to interpret the pilot-
scale studies of UNDEERC.
SORBENTS
Activated carbons PC-100 and FGD (both manufactured by American Norit Company,
Inc.) were the sorbents used in this study to test capture of Hg° and HgCl2. The precursors for
PC-100 and FGD are bituminous coal and lignite, respectively. These activated carbons have
-------
been used in field studies10. The significant differences in their physical characteristics are their
surface areas ( approximately 1,000 m2/g for PC-100 versus approximately 500 m2/g for FGD)
and particle diameters (17,55 um for PC-100 versus 6.8 um for FGD). The details of their pore
size distributions are given in an earlier study.11
EXPERIMENTAL PROCEDURES
A fixed-bed arrangement was used in all the sorption tests for both Hg° and HgCl2. There
were, however, minor differences in the operational procedures and analysis in the sorption tests.
Hg° sorption (MWC simulation)
Figure 1 is the schematic of the bench-scale apparatus used to study capture of Hg° by
PC-100 and FGD, Approximately 100 mg of either PC-100 or FGD was placed in a glass reactor
(inside the furnace) and heated to the desired reaction tremperature of 100 or 140 °C. Pure
nitrogen (N2) carried the Hg° vapors through the By-Pass where the inlet Hg° concentration was
measured by an on-line detector (Ametek, model 400). The effect of flue gas constituents, such
as HC1, sulfur dioxide (SO2), and water (H2O), on the capture of Hg species should also be
assessed; however, this work was limited in scope to the effect of Hg concentration and species in
an inert atmosphere. Upon establishing a baseline of 30 ppb, the Nj/Hg0 stream was switched to
flow through the reactor. The percentage Hg° captured was obtained by recording the drop in
Hg° concentration measured by the on-line analyzer.
Hg° sorption (Coal-fired simulation)
The significantly lower concentrations of Hg° in the coal-fired simulation (2-3 ppb) were
near the detection limit of the on-line Hg° analyzer (Figure 1). Therefore, the Hg° analyzer was
used to provide only an estimate of the Hg° concentration in the gas stream. Since accurate on-
line analysis of Hg° in the gas stream was not possible, the sorbents were removed and analyzed
for total Hg by x-ray fluorescence (XRF), Therefore, batch experiments were performed
(different exposure times of sorbent to Hg°).
Typically 100 mg of activated carbon was placed (stage 1) in the reactor (see Figure 1)
followed by two 100 mg plugs of PC-100 which served as breakthrough traps (stages 2 and 3).
Similar to testing in the MWC simulation, the N^g0 stream is switched through the reactor after
establishing a baseline of roughly 1 ppb. At the end of a fixed duration (8, 12, or 24 h), the Nj/Hg0
-------
stream is diverted away from the reactor, and the three stages are analyzed separately for total Hg
content using XRF.
The percentage Hg° capture by PC-100 or FGD after exposure to Hg° for time t, 0, was
calculated as:
0, = [0,7(0,+ B2 + 03)]*100 (1)
where 0,, B2, and B3 are the total Hg measured (ppmw using XRF) in each of the three sequential
reactor stages. In all of our experiments, the fraction of total Hg capture in stage 3, 63/(B, + 02 +
B3), was less than 0.1, indicating little, if any, Hg° breakthrough.
HgCl2 sorption (MWC and Coal-fired situation)
The schematic for studying HgCl2 capture shown in Figure 2 is similar to that used for
studying Hg° capture. However, in the present study, there is no on-line analyzer for HgCl2.
Therefore, the procedure for studying HgCl2 sorption is identical to the coal-fired Hg° simulation.
The fraction of HgCl2 captured by FGD or PC-100 is obtained as described earlier [from equation
(1)].
The methodology and accuracy of the XRF technique used to measure total Hg in FGD
and PC-100 have been given earlier.9'12
RESULTS
MWC simulation
Figure 3 shows the instantaneous capture of HgCl2 [30 ppb in N2 (MWC simulation)] by
100 mg of activated carbons FGD and PC-100 at two temperatures - 100 and 140 °C. Both
sorbents show relatively high capture (> 80 percent) of incoming HgCl2 at 100 °C. Even up to 5
h of reaction time, PC-100 captures more than 80 percent of the incoming HgCl2 at 100 °C, and
FGD approximately 85 percent. Higher temperatures reduce the percentage capture of incoming
HgCl2 for both the sorbents, although the effect of temperature on HgCl2 capture by FGD is more
pronounced.
The capture of 30 ppb of Hg° (MWC simulation) by 100 mg of FGD and PC-100 at the
two temperatures — 100 and 140 °C — is shown in Figure 4. PC-100 captures nearly all of the
incoming Hg° (at both temperatures) in the initial time period (less than 2 h of exposure to the
Hg°/N2 stream). FGD captures far lower percentages of incoming Hg° at both temperatures when
-------
compared to PC-100. Similar to the results seen in Figure 3, increasing the temperature causes a
reduction in the capture of Hg° by FGD, but has very little effect on the capture of Hg° by PC-
100.
Depending on the manufacturing process [activation with nitrous oxide (N2O), ammonia
(NH3), or ZnCl2-NH4Cl-CO2; or heat-treated at 900 °C], activated carbon acquires properties of a
solid base.13 Reactions of gaseous oxygen (O2) with the surface of active carbon at temperatures
below 100 °C produce O2 complexes which, on hydration, can form hydroxyl or other basic
groups.14 The large internal surface areas of PC-100 and FGD may provide sufficient basic active
sites to capture the acidic HgCl2.
From Figures 3 and 4, higher percentages of incoming HgCl2 compared to incoming Hg°
are captured by FGD and PC-100 at both temperatures. In a previous investigation,9 we showed
the likelihood of different active sites for the capture of Hg° and HgCl2 in activated carbons.
Assuming similar access to sorption sites for Hg° and HgCl2 and based on our discussion above, it
appears that the sites for Hg° capture are far fewer than those for HgCl2 capture or the sorption of
Hg° requires a higher activation energy.
Coal-fired simulation
Figure 5 shows capture of HgCl2 (approximately 1 ppb HgCl2 in N2) by 100 mg of FGD
and PC-100 at 100 and 140 °C. High amounts (> 80 percent) of incoming HgCl2 are captured by
PC-100 at both temperatures. FGD captures lower amounts at both temperatures when
compared to PC-100. Increasing the temperature has very little effect on the capture of HgCl2 by
PC-100; whereas, it causes a decrease in the capture of HgCl2 by FGD.
In the MWC simulation (see Figure 3), PC-100 shows a high capture (> 90 percent) of
incoming HgCl2 at both temperatures for initial time periods (less than 2 h), but shows a steady
decline in the capture percentage with increasing exposure time. On the other hand in the coal-
fired simulation (see Figure 5), such a decline with exposure time is not evident for PC-100. This
may be on account of the faster rate of depletion of basic active sites for HgCl2 capture in the
MWC simulation.
The capture of low concentration Hg° (approximately 1 ppb in N2) by 100 mg of PC-100
and FGD at 100 and 140 °C is shown in Figure 6. Here too, PC-100 captures a higher fraction of
-------
incoming Hg° than FGD at both temperatures. Temperature is seen to have a marginal effect on
the capture of low concentration Hg° by PC-100, similar to the other cases discussed before.
From Figures 5 and 6, it is clear that PC-100 captures a higher fraction of incoming HgCl2
than Hg° at both temperatures. This is similar to the MWC simulation (see Figures 3 and 4). On
the other hand, FGD is seen to capture similar levels of HgCl2 and Hg° for the coal-fired
simulation (see Figures 5 and 6), unlike the MWC simulation (Figures 3 and 4).
DISCUSSION
Some of the results obtained in our study are summarized in Tables 1 and 2. The results
shown in Tables 1 and 2 are for percent Hg capture by 100 mg of sorbent (PC-100 and FGD,
respectively) exposed to a flow of Hg (Hg° or HgCl2) in 300 cnrVmin N2 for 6 h.
A few of the main conclusions that can be drawn from Tables 1 and 2 are:
i) PC-100 captures more Hg° than FGD at both temperatures and concentrations (i.e., MWC
and coal-fired simulations) during 6 h of exposure.
ii) A larger percentage of incoming Hg° is captured in a coal-fired situation by both activated
carbons at both temperatures.
iii) Lower percentages of HgCl2 are captured in a coal-fired simulation (compared to capture
under a MWC simulation) by both activated carbons at both temperatures.
iv) In the MWC simulation HgCl2 is removed more easily than Hg°. In the lower Hg
concentration coal simulations, the reverse is true: equal or greater amounts of Hg° than
HgCl2 were removed.
In a past study,9 the capture of roughly 30 ppb Hg° or 30 ppb HgCl2 by fresh PC-100 was
compared with PC-100 exposed to heat (140 °C). This study showed that the capture of Hg° by
heat-treated PC-100 was roughly 50 percent lower than that captured by fresh (or untreated) PC-
100. Since heat treatment causes loss of internal pore structure and surface area, these
experiments suggest that Hg° capture is limited by capture on surface-bound active sites. For
HgCl2, there was no difference in the capture, suggesting that capture is not mediated by active
sites. Since PC-100 has roughly twice the specific surface area of FGD and likely twice the
number of active sites, it follows that PC-100 will be more effective at Hg° capture than FGD, as
per conclusion i).
8
-------
The coal-fired simulation (with l/30th of the Hg° concentration of the MWC simulation
but the same amount of activated carbon) has an activated carbon active site/Hg° ratio 30 times
higher than that of the MWC simulation. Since Hg° capture is limited by the availability of active
sites this should result in a significantly higher capture of Hg° in the coal-fired simulation. Tables
1 and 2 show higher capture percentages at all temperatures, but not as high as expected if
reaction is limited by the number of available sites. This suggests that capture of Hg° in the coal-
fired simulation may be limited by an additional rate limiting step other than site availability.
From Tables 1 and 2, both PC-100 and FGD capture similar amounts of HgQ2 in the
coal-fired simulation despite large differences in their surface areas and, likely, active site
population. This suggests that there must be sufficient active sites for HgCl2 capture on both PC-
100 and FGD such that site availability does not limit sorption. These sites are likely basic in
nature; previous work9 has shown that activated carbons and lime-based sorbents both have high
capture of acidic HgCl2 despite large differences in surface area. These abundant basic sites are
most likely functional groups including oxygen as the proton acceptor.
With the number of active sites not limiting capture, a possible controlling mechanism for
HgCl2 capture may be mass transport of HgCl2. This mechanism involves transport of HgCl2 from
the flowing bulk fluid to the surface of the carbon particle, and from the surface to the active sites
via diffusion.
At the ultra-low Hg concentrations tested in the coal-fired simulation, the probability of
Hg in a flowing stream contacting a stationary carbon particle increases with Hg concentration.
For instance, consider two situations where the concentration of Hg in the gas phase are 2 and 4
molecules per unit area. Further, assume that the probability of a Hg molecule colliding with the
carbon particles in that unit area is 0.1. Our calculations show that the probability of at least one
molecule colliding with a carbon particle is 0.19 and 0.344, for the lower and higher
concentrations, respectively. Increasing the number of molecules to six per unit area increases the
probability of at least one molecule colliding with a carbon particle to 0.47. That is, as the
concentration of the Hg species increases in the gas stream, the probability of collision increases
asymptotically. Since the sites available for HgCl2 capture are abundant and assuming that
intraparticle diffusion resistance is linear with concentration, a likely controlling mechanism for
-------
HgCl2 capture is the rate of gas/solid contact. Conclusion iii) of our study, which states that less
HgCl2 is captured at the lower HgCl2 concentration, supports this argument.
COMPARISON WITH FIELD RESULTS
The field studies at the University of North Dakota10 include capture of Hg° and HgCl2 by
lignite activated carbon (lignite AC), FGD, an iodated activated carbon (iodated AC), and PC-
100. Here, the particle size of FGD and iodated AC are approximately the same. Three types of
coal were fired in three separate cases which led to different Hg speciation in the flue gas. Table 3
summarizes the Hg speciation results of this study.10 The following discussion highlights the three
coal cases, extends our findings to these field results, and compares our capture results for PC-
100 and FGD with theirs.
Absaloka Case
Miller et al. (Reference 10) found that FGD at 3000:1 carbon;Hg ratio (by weight)
performed similarly as iodated AC at a weight ratio of 1200:1 (carbon:Hg). Absaloka coal flue
gas has more Hg° than Hg++ (see Table 3). Hg"" represents the oxidized form of Hg and, based on
our thermodynamic calculations,9 HgCl2 is the most likely form. As per our conclusions (from
bench-data), iodated AC, which has special sites for Hg° capture, should perform better than FGD
at similar carbon:Hg ratios and perhaps perform equally as well as FGD at the lower carbon:Hg
ratio. Their speciation data also showed that Hg° was removed far better by iodated AC than FGD
at the lower carbon:Hg ratios, in agreement with our findings.
Comanche Case
The results10 for the Comanche case show higher capture of total Hg at 3000:1 FGD:Hg
ratio than 1200:1 iodated AC:Hg ratio. Here most of the Hg is oxidized and therefore, according
to our bench-scale findings, capture should be limited by collision of Hg*+ with carbon. At the
lower ratio for iodated AC, the probability of collision decreases and hence the lower capture
seen. Also, iodated AC is seen to capture a higher fraction of Hg° than FGD, confirming the
hypothesis of sorption kinetics limiting the capture of Hg°.
Blacksville Case
At 121 °C, FGD at a 3000:1 carbon:Hg ratio captures nearly 90 percent of total Hg.10 At
the same temperature, lignite AC at 1200:1 ratio captures less than 80 percent of total Hg.10
10
-------
Blacksville coal-fired flue gas has more Hg+* than Hg° (see Table 3). The larger amounts (or
number) of FGD particles (due to similar particle sizes and assuming similar densities) would
increase the probability of collision with the Hg++ molecules leading to higher capture. The results
of Miller et al. (Reference 10) also show that, even at the lower carbon:Hg weight ratio, iodated
AC captures a higher fraction of Hg° than FGD, Also, FGD captures a higher fraction of incoming
Hg++, in agreement with our conclusions.
PC-100 versus FGD
Field data10 show that, at an identical carbonrHg weight ratio (3000:1), FGD captures a
greater percentage of Hg++ than PC-100 and PC-100 captures a higher fraction of Hg° than FGD.
At equivalent Hg concentrations, the total mass feed rate of carbon is the same and the number
of FGD particles is higher than PC-100 due to FGD's smaller particle diameter (about 5 um
versus 20 um for PC-100). The number of FGD to PC-100 particles in the field study is
approximately equivalent to the inverse of their volumetric ratio, or 64:1 (assuming similar
densities for the two carbons). The larger number of FGD particles increases the probability of
collision with Hg++ and offers an explanation for the higher Hg** capture observed by Miller et al.
(Reference 10).
Since PC-100 has a larger surface area than FGD (approximately 1000 and 500 m2/g,
respectively) it has more active sites to adsorb Hg°. Despite the larger number of FGD particles,
PC-100 performs better for Hg° capture,10 confirming our theory of sorption limitation for Hg°.
These similarities between our bench-scale results and field data10 lend support to use of
bench-scale methods as a means of predicting full-scale performance. The effect of varying flue
gas constituents on Hg speciation and sorbent performance may further enhance the predictive
capability of bench-scale simulations.
SUMMARY AND CONCLUDING REMARKS
Bench-scale tests were performed in this study to assess the capture of Hg° and HgCl2 by
two types of activated carbons, PC-100 and FGD. The test conditions included variation of Hg
(Hg° and HgCl2) concentration and reaction temperature. The higher Hg concentration
corresponded to conditions in a MWC and the lower Hg concentration to conditions prevalent in
a coal-fired flue gas. Hg° and HgCl2 capture by the two activated carbons was studied at 100 and
11
-------
140 °C.
Among our main findings were:
i) PC-100 captures more Hg° than FGD at both temperatures and concentrations (i.e., MWC
and coal-fired simulations).
ii) A larger percentage of incoming Hg° is captured in a coal-fired simulation by both
activated carbons at both temperatures.
iii) Lower fractions of HgQ2 are captured in a coal-fired simulation (compared to capture
under MWC simulation) by both activated carbons at both temperatures.
iv) Unlike the MWC simulation, a higher fraction of incoming HgCl2 is not necessarily
removed (compared to Hg° capture) for the coal-fired simulation. That is, HgCl2 may not
necessarily be removed with greater ease than Hg° by activated carbon injection in a coal-
fired simulation.
These results along with our past efforts9 suggested two mechanisms for Hg capture:
i) Capture of Hg° by activated carbons is limited by sorption kinetics, and
ii) Capture of HgCl2 by activated carbons is limited by collision with carbon particles.
The conclusions arrived at from this study were compared with and are in agreement with field
data.10 Control of Hg in coal-fired combustors is challenging due to the low concentrations of Hg
species in the flue gas. Specifically, control of Hg° is limited by reactivity of the sorbent (that is,
type and number of active sites for Hg° capture) and control of HgCl2 is limited by collision
probability with the sorbent. A control strategy would therefore need accurate speciation of Hg
in the flue gas. If HgCl2 is the dominant species, total Hg control may be achieved by injecting
larger amounts of an inexpensive sorbent with 'basic' sites. On the other hand if Hg° is the major
Hg species, then the injected sorbent must have sufficiently active sorptive sites for Hg° capture.
ACKNOWLEDGEMENTS
This work has been supported in part by the Electric Power Research Institute (EPRI).
The technical advice of Ramsay Chang (EPRI) is greatly appreciated. The authors would also like
to acknowledge the contributions of Richard E. Valentine (EPA/APPCD) for equipment support,
and Lisa Adams and Hamid Bakhteyar (Acurex Environmental Corp.) for experimental assistance.
12
-------
REFERENCES
1. U. S. Environmental Protection Agency, "Mercury Study Report to Congress," Book 1 of
2, External Review Draft, U.S. EPA, Environmental Criteria and Assessment Office, EPA/600/P-
94/002a (NTIS PB95-167334), Cincinnati, OH (Jan. 1995).
2. Brna, T.G., and Kilgroe, J.D., "The Impact of Paniculate Emissions Control on the Control
of Other MWC Air Emissions," J. Air & Waste Mgt. Assoc., 40(9), 1324 (1990).
3, Getz, N.P., Thompson, I.B., and Amos, C.K., Jr., "Demonstrated and Innovative Control
Technologies for Lead, Cadmium and Mercury from Municipal Waste Combustors," presented at
the 85th Annual Meeting and Exhibition, Air & Waste Management Association, Kansas City,
MO (June 21-26, 1992).
4. Guest, T.L., "Mercury Control in Canada," presented at the 86th Annual Meeting and
Exhibition, Air & Waste Management Association, Denver, CO (June 13-18, 1993).
5. White, D.M., Kelly, W.E., Stucky, M.J., Swift, J.L., and Palazzolo, M.A., "Emission Test
Report: Field Test of Carbon Injection for Mercury Control, Camden County Municipal Waste
Combustor," U.S. Environmental Protection Agency, EPA-600/R-93-181 (NTIS PB94-101540),
(Sept. 1993).
6. Chang, R., Bustard, C. J., Schott, G., Hunt, T., Noble, H., and Cooper, J., "Pilot Scale
Evaluation of Carbon Compound Additives for the Removal of Trace Metals at Coal-Fired Utility
Power Plants," presented at the Second International Conference on Managing Hazardous Air
Pollutants, Washington, D.C. (July 13-15, 1993).
7. Miller, S.J., Laudal, D.L., Chang, R., and Bergman, P.D., "Laboratory-scale Investigation
of Sorbents for Mercury Control," presented at the Air & Waste Management Association Annual
Meeting, Paper No. 142, Cincinnati, OH (June 20-24, 1994).
8. Chang, R., and Often, G. R., "Mercury Emission Control Technologies: An EPRJ Synopsis,"
Power Engineering, pp. 51-57 (Nov. 1995).
9. Krishnan, S.V., Gullett, B.K., and Jozewicz, W., "Mercury Control by Injection of
Activated Carbon and Calcium-Based Sorbents," paper presented at the Conference on Solid
Waste Management: Thermal Treatment & Waste-to-Energy Technologies, U.S. EPA/AEERL &
AWMA, Washington, DC (April 18-21, 1995).
10. Miller, S.J., Laudal, D.L., Dunham, G.E., Chang, R., and Bergman, P.D., "Pilot-Scale
Investigation of Mercury Control in Baghouses," EPRJ/DOE Third International Conference on
Managing Hazardous and Paniculate Air Pollutants, Toronto, Canada, August 15-18, 1995.
13
-------
11. Jozewicz, W., Krishnan, S.V., and Gullett, B.K., "Bench-scale Investigation of
Mechanisms of Elemental Mercury Capture by Activated Carbon," presented at the EPRI/DOE
Second International Conference on Managing Hazardous Air Pollutants, Washington, B.C. (July
13-15, 1993).
12. Krishnan, S.V., Gullett, B.K., and Jozewicz, W., "Control of Mercury Emissions from
Coal Combustors," EPRI/DOE Third International Conference on Managing Hazardous and
Paniculate Air Pollutants, Toronto, Canada (August 15-18, 1995).
13. Tanabe, K., "Solid Acids and Bases, Their Catalytic Properties," Academic Press, New
York (1970).
14. Smisek, M., and Cerny, S., "Active Carbon. Manufacture, Properties, and Application,"
Elsevier, New York (1970).
14
-------
Table 1. Hg capture (%) comparison by activated carbon PC-100 at 6 h exposure and two
isothermal reaction temperatures (from results shown in Figures 3, 4, 5, and 6).
Hg Species
Hg«
HgCl2
30 ppb (MWC simulation)
100 °C
67%
90%
140 °C
70%
90%
1 ppb (coal-fired simulation)
100 °C
85%
50%
140 °C
85%
45%
Table 2. Hg capture (%) comparison by activated carbon FGD at 6 h exposure exposure and two
isothermal reaction temperatures (from results shown in Figures 3, 4, 5, and 6).
Hg Species
Hg«
HgCl2
30 ppb (MWC simulation)
100 °C
35%
85%
140 °C
20%
55%
1 ppb (coal-fired simulation)
100 °C
50%
50%
140 °C
30%
25%
Table 3. Flue gas Hg speciation for the three types of coal used by Miller et al, (Reference 10). The
sampling temperatures are about 121 °C.
Hg
Species
Hg«
Hg"
Absaloka
Cone, (ppb)
0.16
0.08
Fraction
0.67
0.33
Comanche
Cone, (ppb)
0.16
0,36
Fraction
0.31
0.69
Blacksville
Cone, (ppb)
0.08
0.44
Fraction
0.15
0.85
15
-------
tt
Hood
i
Main Trap
Hg° Generator
N
v
On-line Hg°
Analyzer v
N
Reactor
(3 sorbent stages for coal-fired case)
Chart
Recorder
Personal
Computer
Figure 1. Schematic of bench-scale apparatus used to study Hg° capture (v: valve).
-------
Hood
Main Trap
Diffusion Vial
By-pass
N
2 V
*r HgCI Generator
Furnace
Furnace
N
X
Reactor
(3 sorbent stages for all cases)
Figure 2. Schematic of bench-scale apparatus used to study HgCL capture (v: valve).
-------
CD
o
LU
100
90
80
70
o
tf 60
TO
50
40
0
10 15
TIME (h)
PC-100
100°C
PC-1 00
140 °C
— B—
FGD
100°C
FGD
140 °C
20
Figure 3. MWC case: Capture of HgCI (30 ppb in N) at 100 and 140 °C by 100 mg of FGD and PC-100
-------
100
^ 80
o
LJJ
60
O
o
TO
40
20
0
4 6
TIME (h)
PC-100
100°C
PC-100
140°C
—B—
FGD
100°C
FGD
140 °C
8
Figure 4. MWC case: Capture of Hg° (30 ppb in N ) at 100 and 140 °C by 100 mg of FGD and PC-100
-------
ro
o
100
^ 80
60
a.
O 40
D
LLJ
CM
O
O)
20
0
10
15 20
TIME (h)
PC- 100
100°C
PC-100
140 °C
FGD
100°C
FGD
140 °C
- 0 -
25
30
Figure 5. Coal-fired case: Capture of HgCI (1 ppb in N ) at 100 and 140 °C by 100 mg of FGD and PC-100
-------
IS3
100
^ 80
o
LU
60
O
o
O)
40
20
0
10
15
20
TIME (h)
PC-100
100°C
PC-100
140 °C
—B—
FGD
100°C
FGD
140 °C
-e-
25
30
Figure 6. Coal-fired case: Capture of Hg° (1 ppb in N ) at 100 and 140 °C by 100 mg of FGD and PC-100
-------
NRMRL-RTP-P-090
TECHNICAL REPORT DATA
(Pleate read Inunctions on the reverse before complete
. REPORT NO.
EPA/600/A-96/118
2.
3. F
4. TITLE AND SUBTITLE
Mercury Control in Municipal Waste Combustors and
Coal-fired Utilities
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
'.AUTHORISI s. V.Krishnan (Acurex), B. K. Guilett (EPA),
and W. Jozewicz (Acurex)
B. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME AND ADDRESS
Acurex Environmental Corporation
4915 Prospectus Drive
Durham, North Carolina 27713
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-D5-0005
12, SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Published paper; 10/94-2/96
14. SPONSORING AGENCY CODE
EPA/600/13
«.SUPPLEMENTARY NOTES APPCD project officer is Brian K. Gullett, Mail Drop 65, 919/541
1534. Presented at AlChE Spring National Meeting, New Orleans, LA, 2725-29/96.
16. ABSTRACTThe paper gj_ves resu].ts of a s tudy of bench-scale experiments under condi-
tions simulating municipal waste combustors (MWCs) and coal-fired utilities to study
mercury (Hg) capture by dry sorbents. The effect of reaction temperature on the
capture of different Hg species--Hg° and Hg(n)—by various types of dry sorbents
was the focus of bench-scale tests. An attempt was also made in this study to com-
pare the bench-scale results with results obtained from pilot studies and to explain
disparities in fuel- and unit-specific performance. Our investigations showed that
the reaction temperature and Hg species strongly affect Hg control. The results ob-
tained in this study along with our past efforts suggested two mechanisms for Hg cap-
ture: (a) capture of Hg° by activated carbons is limited by sorption kinetics; and (b)
capture of HgC12 by activated carbons is limited by collision with carbon particles.
(NOTE: Control of Hg emissions from MWCs and coal-fired utilities has attracted
attention due to current and potential regulations. Among several techniques evalua-
ted for Hg control, dry sorbent injection — primarily injection of activated carbon--
shows promise for consistently removing high levels of Hg from MWC or coal flue
gas. However, the performance in terms of amount of Hg removed per amount of
sorbent varies widely between MWC/coal-fired applications and from unit to unit.)
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Mercury
ombustion
Wastes
Utilities
oal
Sorbents
Pollution Control
Stationary Sources
Municipal Waste Com-
bustors
13 B
07B
21B
H4G
21D
11G
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CUASS (ThisReport)
Unclassified
21. NO. OF PAGES
2O. SECURITY CLASS (This page I
Unclassified
22. PRICE
EPA Form 2220-1 (9-73J
------- |