Development of Mercury Speciation Factors for EPA's Air Emissions Modeling Programs,
Technical Support Document
Version 1
Contacts: Janice Godfrey, Eric Goehl, Madeleine Strum, and Art Diem
of the EPA's Office of Air Quality Planning and Standards (OAQPS)
April 2020
I.	Purpose
This document describes the development of mercury (Hg) speciation factors for the U.S.
Environmental Protection Agency's (EPA's) air emissions modeling programs. Atmospheric
mercury is a global environmental issue due to its toxicity, persistence, and long-range
transportability. Mercury speciation plays an important part in the toxicity and exposure of
mercury to living organisms. Speciation also has a considerable influence on the fate and
transport of mercury within and between environmental compartments, including the atmosphere
and oceans, among others. Moreover, speciation affects the controllability of mercury
emissions.1'2
II.	Background
Mercury is emitted in three main forms: elemental gaseous mercury (Hg°), gaseous
divalent mercury (Hg2+), and particulate-bound divalent mercury (Hgp). Oxidation Hg° toHg2+
occurs in the flue gas for certain industrial processes, with a portion of Hg2+ adsorbed onto fly ash
to form Hgp. Mercury removal and transformation also occurs when it passes through air
pollution control devices (APCDs). As described in more detail below, mercury speciation can
differ between sectors due to differences in fuel type, APCDs, operating procedures, and other
factors, some of which are not yet fully understood. This demonstrates the need for additional
data for a variety of point sources of emissions, such as smelters and waste incinerators;
especially as sources, fuels, feedstocks, and control techniques change.
Many factors affect mercury speciation. For example, coal type, and chlorine content are
important factors in the speciation and capture of mercury with different types of air pollution
control technologies.3 In the U.S., bituminous coals tend to have relatively high concentrations
of chlorine (CI). This can result in the oxidization of Hg° to Hg2+ (primarily HgCl2). The Hg2+
can be adsorbed onto fly ash carbon and captured in an electrostatic precipitator (ESP) or fabric
filter (FF). On the other hand, mercury in the exhausts of plants burning other types of coal tends
to be predominately Hg°. [The capture of mercury from the flue gas from these plants tends to be
1	AMAP/UNEP, 2002.
2	Streets, D. et al., 2019.
3	US EPA, 2004.
Page 1

-------
lower, whether the units are equipped with an ESP, FF, dry flue gas desulfurization (FGD)
scrubber, or wet FGD scrubber.4]
III. General Approach for Developing Speciation Profiles for Source Categories
We conducted a literature review to find available data and information regarding the
forms of mercury emitted from various sources. One of the main documents identified through
this search is the 2008 Global Atmospheric Mercury Assessment.5 In this assessment, the United
Nations Environment Programme (UNEP) cooperated with the Arctic Monitoring Assessment
Programme (AMAP) to produce the 2008 Global Atmospheric Mercury Assessment
(AMAP/UNEP 2008). The AMAP/UNEP review includes data from several mercury studies
from around the world and was prepared by expert groups engaged by AMAP and UNEP.
Information submitted by governments, intergovernmental and non-governmental organizations,
as well as other available scientific information, were used to prepare the report.
A more recent study (Muntean, 2018), compared speciated mercury gridded emissions
inventories with chemical transport models and concentration measurements to investigate the
effectiveness of mitigation measures and the mercury cycle in the environment.6 The authors
developed three emissions scenarios based on different hypotheses of the proportion of mercury
species in the total mercury emissions for each activity sector. The study used the GEOS-Chem
3-D mercury model to explore the influence of speciation shifts, to reactive mercury forms, in
particular, on regional wet deposition factors. Scenario 1 used speciation factors from
AMAP/UNEP, which reflects a generic power plant mercury profile (Hg° = 50%, Hg2+ = 40%,
and Hgp = 10%). Scenario 2 used speciation factors from AMAP/UNEP, but substituted factors
for the power generation sector with EPA data derived from an information collection request
(ICR) done in conjunction with the Clean Air Mercury Rule (CAMR). Unlike the AMAP/UNEP
speciation factors for power generation, the CAMR speciation factors are specific to the type of
fuel and APCD used. Scenario 3 used speciation factors gleaned from recent, small studies,
mainly carried out in Asia. The study shows that using the Scenario 2 emissions of reactive
mercury, can improve wet deposition estimates near sources and therefore give us more
confidence in the profiles developed from the AMAP/UNEP and CAMR ICR.
Although several recent smaller studies have been carried out in Asia, we decided, given
that most of these studies were focused on a small number of facilities in a limited region, to
generally use speciation factors derived for the AMAP/UNEP 2008 report. The exception to this
is where the EPA collected emission tests for mercury speciation for coal-fired power plants for
the CAMR, which was documented in the Mercury Air Toxics Standard (MATS), as well as,
4	AMAP/UNEP, 2002.
5	AMAP/UNEP, 2008.
6	Muntean, M. et al., 2018.
Page 2

-------
other sectors where sufficient industry data {i.e., test reports) or sector specific literature sources
are available. Industry data are used for the coal-fired power plant Portland Cement and Chlor-
Alkali production profiles. Sector specific literature sources are used for geothermal power and
mobile sources profiles. Where data are insufficient to develop sector-specific speciation factors,
we developed factors using an average of the source categories for which we have determined a
speciation profile.
Once a set of speciation factors are established, they are assigned a speciation profile
code to facilitate the speciation process for air emissions modeling. The speciation profile codes
are also added to EPA's SPEC!ATE database. Profile codes are summarized in Appendix A.
IV. Mercury Speciation for Specific Source Categories
Coal
As mentioned above, mercury speciation that occurs after emissions leave the boiler and
before they reach the APCD is mainly determined by coal properties, specifically chlorine,
mercury, and ash contents. For example, larger availability for chlorine in coal causes larger
formation of Hg2+ compared to H°. Higher concentrations of fine particles in exhaust gases
improve conditions for gas-to-particle conversion of gaseous mercury, resulting in higher
proportions of Hgp.7
In 2010, about 16% of global anthropogenic mercury emissions (unspeciated) were
attributed to commercial coal-fired power plants.8 In the 2014 National Emissions Inventory
(NEI), which covers the United States, coal combustion from electric generating units was the
largest mercury source category, contributing an estimated 44% of the estimated 52 tons
emitted.9 (Note that the NEI doesn't estimate mercury from fires but does include some nonpoint
sources, such as dental amalgam and fluorescent light breakage, as well as a small amount from
mobile sources.)
For the MATS rule, the EPA compiled documentation on mercury speciation for EGUs in
the United States. The speciation approach applied for the final MATS rule was developed as
part of the CAMR and based on a 1999 ICR. Eighty mercury emissions tests were collected from
69 facilities. These data include inlet and outlet speciated mercury concentration measurements
for Hgp, Hg2+, and Hg°. All emissions tests were conducted using the Ontario Hydro method and
included three sampling runs. Only the outlet data (measurements after the last control device)
were used to develop the mercury speciation profiles. Forty-three mercury speciation profiles
were developed from these data (see Table 1). Additional information on these profiles is
7	Pacyna and Pacyna, 2002.
8	AMAP2013.
9	2014 US EPA National Emissions Inventory (NEI).
Page 3

-------
available in the "Electric Generating Utility Mercury Speciation Profiles for the Clean Air
Mercury Rule"10 and the workbook "EGU_Hg_Speciation_Data_CAMR.xls," which are
available in the regulations.gov docket EPA-HQ-QAR-2009-0234. document	A
profile code is assigned to each bin as follows: "EGUBIN" concatenated with the bin number.
Petroleum Coke (pet coke) is an additional fuel combustion profile derived from the
average of bins 2, 23, 25, 38, and 43 {i.e., Hgp= 0.10, Hg2+ = 0.30, and Hg° = 0.60). The
resultant speciation factors are assigned the profile code HGPETCOKE. For coal-fired power
plants that were not a part of the MATS analysis, and for which we do not have sufficient
information to include in one of the MATS bins, we intend to use the speciation factors from
AMAP/UNEP for combustion emissions from power plants (see Table 2).
Table 1. Electric Generating Utility Mercury Speciation Profiles for the Clean Air
Mercury Rule
Bin
Type
Fuel, Boiler, Emission control device(s)
Percent Speciation
Particulate Hg
Oxidized Hg
Elemental Hg
0
Bituminous Coal, Coal Gasification
0.51%
8.47%
91.02%
1
Bituminous Coal, PC Boiler with ESP-CS
6.11%
68.20%
25.70%
2
Bituminous Coal and Pet. Coke, PC Boiler with
ESP-CS
1.17%
46.56%
52.27%
3
Bituminous Coal, PC Boiler with SNCR and
ESP-CS
20.32%
27.12%
52.56%
4
Bituminous Coal, PC Boiler with ESP-HS
4.90%
57.84%
37.26%
5
Bituminous Coal, PC Boiler with PM Scrubber
1.80%
19.51%
78.69%
6
Bituminous Coal, PC Boiler with Dry Sorbent
Injection and ESP-CS
0.16%
67.10%
32.74%
7
Bituminous Coal, PC Boiler with FF Baghouse
3.98%
62.58%
33.44%
8
Bituminous Coal, PC Boiler with SDA/FF
Baghouse
9.17%
28.86%
61.97%
9
Bituminous Coal, PC Boiler with SCR and
SDA/FF Baghouse
5.06%
46.04%
48.90%
10
Bituminous Coal, PC Boiler with ESP-CS and
WetFGD
0.22%
7.78%
92.00%
11
Bituminous Coal, PC Boiler with ESP-HS and
WetFGD
0.63%
20.68%
78.70%
12
Bituminous Coal, PC Boiler with FF Baghouse
and WetFGD
6.48%
33.00%
60.52%
13
Subbituminous Coal, PC Boiler with ESP-CS
0.16%
30.83%
69.01%
10 Bullock and Johnson, 2011.
Page 4

-------
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Fuel, Boiler, Emission control device(s)
Percent Speciation
Particulate Hg
Oxidized Hg
Elemental Hg
Subbituminous Coal, PC Boiler with ESP-HS
0.06%
12.52%
87.41%
Subbituminous Coal, PC Boiler with FF
Baghouse
1.49%
82.83%
15.68%
Subbituminous Coal, PC Boiler with PM
Scrubber
1.45%
5.11%
93.44%
Subbituminous Coal, PC Boiler with SDA/ESP
0.32%
3.82%
95.86%
Subbituminous Coal, PC Boiler with SDA/FF
Baghouse
0.99%
4.35%
94.67%
Subbituminous Coal, PC Boiler with ESP-CS
and Wet FGD
0.43%
2.94%
96.63%
Subbituminous Coal, PC Boiler with ESP-HS
and Wet FGD
1.17%
4.46%
94.37%
Lignite Coal, PC Boiler with ESP-CS
0.09%
3.62%
96.29%
Subbituminous Coal, Cyclone Boiler with PM
Scrubber
2.34%
5.75%
91.91%
Subbituminous Coal/Pet. Coke, Cyclone Boiler
with ESP-HS
0.93%
7.52%
91.55%
Lignite Coal, Cyclone Boiler with ESP-CS
0.04%
16.99%
82.97%
Bituminous Coal/Pet.Coke, Fluidized Bed
Combustor with SNCR and FF Baghouse
42.44%
27.87%
29.70%
Not Used
Bituminous Waste, Fluidized Bed Combustor
with FF Baghouse
2.12%
38.81%
59.07%
Lignite Coal, Fluidized Bed Combustor with
ESP-CS
1.37%
11.64%
87.00%
Lignite Coal, Fluidized Bed Combustor with FF
Baghouse
0.42%
71.18%
28.40%
Anthracite Waste, Fluidized Bed Combustor with
FF Baghouse
3.01%
37.30%
59.70%
Bituminous Coal, Stoker Boiler with SDA/FF
Baghouse
19.96%
17.94%
62.11%
Not Used
Lignite Coal, PC Boiler with ESP-CS and FF
Baghouse
0.19%
64.49%
35.32%
Lignite Coal, PC Boiler with SDA/FF Baghouse
0.36%
12.62%
87.02%
Lignite Coal, PC Boiler with PM Scrubber
0.16%
2.98%
96.86%
Lignite Coal, PC Boiler with ESP-CS and Wet
FGD
0.82%
13.45%
85.74%
Page 5

-------
Bin
Type
Fuel, Boiler, Emission control device(s)
Percent Speciation
Particulate Hg
Oxidized Hg
Elemental Hg
37
Bituminous Coal, Cyclone Boiler with
Mechanical Collector
18.75%
42.74%
38.51%
38
Bituminous Coal/Pet. Coke, Cyclone with ESP-
CS and Wet FGD
0.07%
11.30%
88.63%
39
Lignite Coal, Cyclone Boiler with SDA/FF
Baghouse
9.95%
17.07%
72.98%
40
Subbituminous Coal, Fluidized Bed Combustor
with SNCR and FF Baghouse
0.27%
3.42%
96.32%
41
Subbituminous Coal/Bituminous Coal, PC Boiler
withESP-CS
0.88%
42.82%
56.30%
42
Subbituminous Coal/Bituminous Coal, PC Boiler
withESP-HS
2.86%
49.11%
48.03%
43
Bituminous Coal/Pet. Coke, PC Boiler with FF
Baghouse
2.20%
78.41%
19.39%
44
Bituminous Coal/Subbituminous Coal, PC Boiler
with FF Baghouse
5.95%
42.10%
51.95%
Portland Cement
The EPA received mercury speciation data for four U.S. Portland Cement companies (7
facilities) for its 2018 Residual Risk and Technology Review.11 In the summary analysis, there
was a wide range of values for the elemental (Hg°) to oxidized (Hg2+) split of mercury emissions
- from a low of 28% to as high as 97% Hg°. Industry stated that this wide range reflects the
reality that multiple factors determine the mix of mercury species emitted by a kiln. These
factors include the concentration of mercury in the raw materials (namely limestone), operating
stack temperature, and type of APCDs used. The data support an estimate that, on average, 66%
of mercury emissions from kilns is Hg° and 34% is Hg2+.12 These data reflect the use of
continuous emission monitoring systems (CEMS) and the presence of controls for mercury
emissions. Mercury control devices include activated carbon injection, wet scrubbers, or both.
We assign this mercury profile to any Portland cement operation having kiln exhaust. The profile
code for the Portland cement profile is HGCEM.
Mercury emissions from the kiln that were in the particulate form (not measured by
CEMS) were considered inconsequential and were not included in the mercury speciation for
kilns. Because mercury is usually completely volatilized in the kiln, mercury is typically not
emitted from the clinker cooler. However, the data on metal content of clinker showed a small
amount of mercury. For the purposes of speciating mercury emissions from the clinker cooler13
and other post-kiln Portland cement product storage and handling operations, it is assumed
mercury emissions are 100 percent in the form of particulate emissions and are assigned the
11	US EPA, 2018.
12	Portland Cement Association, 2016.
13	US EPA, 2018
Page 6

-------
profile code HGCLI. For raw material operations that take place before the kiln, any mercury
emissions are presumed to be in elemental form and are assigned the profile code HGELE.
A separate source monitoring study conducted in Florida in the year 2000 determined
emission profiles that support the Portland cement industry speciation profiles. The study can be
found in the EPA Technical Report EPA/600/R-00/102. The study found that Hg2+ represented
less than 25 percent of the total mercury emissions from Portland cement kilns. The ratio of Hg2+
to Hg° was just under 1:3. The analytical procedures for determination of mercury from stack
sampling followed EPA guidelines in 40 CFR 60, Appendix A, Method 29. Since the study only
included limited monitoring from one kiln, we cannot conclude these numbers are representative
of all Portland cement facilities. Regardless, the study may be helpful as supplementary
information.
Other Anthropogenic Source Categories
As discussed above, given that most of the more recent studies were focused on a small
number of facilities in a limited region (Asia), we concluded that it was more representative to
use the AMAP/UNEP 2008 mercury speciation factors generally for several source categories (as
shown in table 2). However, for future modeling assessments for any given source category, if
the EPA obtains additional source category specific mercury speciated data, it would probably be
appropriate to use the new data in lieu of, or in combination with, the speciation factors
presented in this technical support document. Nevertheless, it is important to note that there is
uncertainty in using one set of speciation factors for a given industry as mercury speciation can
vary due to differences in fuel type, control devices, operating procedures, and other factors. The
estimates presented in Table 2 were determined on the basis of data on mercury in exhaust gases
from various sources collected by Pacyna et. al.14 over two decades prior to their publication in
2002 entitled "Global Emission of Mercury from Anthropogenic sources in 1995".
Table 2. Speciation Profiles from the 2008 Global Atmospheric Mercury Assessment
Sector
Hg°
Hg2+
HgP
Combustion emissions from power plants
0.5
0
4
0.1
Combustion emissions from residential heating
0.5
0
4
0.1
Combustion emissions from industrial/commercial/residential boilers
0.5
0
4
0.1
Iron and steel production
0.8
0
15
0.05
Non-ferrous (Cu, Zn, Pb) metal production
0.8
0
15
0.05
Large-scale gold production
0.8
0
15
0.05
Waste incineration
0.2
0
6
0.2
Cremation emissions
0.8
0
15
0.05
Artisanal and small-scale gold mining*
1.0
0
0
0.0
14 Pacyna and Pacyna, 2002.
Page 7

-------
AMAP (2008)
* We are not aware of any artisanal or small-scale gold mining in the U.S., but we have included it here for
completeness.
Additionally, the EPA has some limited test data available for the Iron and Steel
Production Sector, the Electric Arc Furnace Sector, mercury-cell chlor-alkali plants and waste
incineration that are discussed below.
Discussion of Source Categories Other than Electric Generating Units and Portland Cement
Below we discuss briefly how speciation occurs within the flue gas from each source
category. We also mention shortcomings in the use of the AMAP/UNEP 2008 numbers, as well
as other possible considerations when applying speciation factors. There are several source
categories within the EPA's Risk and Technology Review program that do not fall into one of
the categories in Table 2 (e.g., paper mills). Therefore, we developed a default industrial profile
to use where the mercury emissions are not a result of fuel combustion and for which we do not
identify any other appropriate surrogate profile. This profile is an average of the non-fuel
combustion categories (i.e., Hg° = 0.73, Hg2+ = 0.22, and Hgp= 0.05). This profile was derived
by averaging the following source categories: Iron and Steel Production; Non-Ferrous (Cu, Zn,
Pb) Metal Production; Large-Scale Gold Production; Chlor-Alkali Industry (caustic soda
production), Hydrochloric Acid (HC1) Production facilities; Waste Incineration; Cremation
Emissions; Portland cement; Electric Arc Furnaces; and Elemental Mercury (See Appendix A for
profile details). The profile code for this default industrial profile is HGIND.
Combustion Emissions from Industrial/Commercial/Residential Boilers
Mercury speciation for industrial, commercial, and residential boilers may differ
significantly from power plants due to varying fuel types, control devices, and operating
procedures. For example, the APCDs applied for industrial coal combustion generally have a
lower Hg2+ removal efficiency than those applied for coal-fired power plants.15 The speciation
profile for combustion emissions from industrial/commercial/residential boilers is composed of
the AMAP/UNEP 2008 mercury speciation factors for combustion and is assigned the profile
code HGCMB.
Iron and Steel Production and Coke Ovens
Mercury is vaporized in high-temperature facilities, including coke oven, sintering
machine, blast furnace, and convertor facilities. Mercury in the flue gas is oxidized
homogeneously and heterogeneously. A portion of the mercury is removed in dust collectors
(e.g., ESPs and baghouses) and flue gas desulfurization devices, and the remaining mercury in
15 Zhang, L. et al., 2016.
Page 8

-------
flue gas is emitted into the atmosphere. The EPA has very limited test data available for this
category. The Michigan Department of Environmental Quality required mercury speciation tests
at the Severstal facility in Dearborn, Michigan. Tests performed by NTH Consultants, Ltd.
showed that the average speciation factors for this iron and steel facility were Hg° = 80%, Hg2+ =
15% and Hgp = 5%.16 Tests were performed with both a basic oxygen furnace (BOF) with an
ESP and BOF with a bag house. This test compares well with the AMAP/UNEP speciation
profile for Iron and Steel Production. The speciation profile code used for iron and steel
production and coke ovens is HGMET.
Electric Arc Furnaces (EAFs)
EAFs produce steel from metal scrap, which is melted and refined by passing an electric
current between electrodes and through the scrap. Much of the mercury in this source category
comes from the elemental mercury in automobile switches built before 2004. The National
Vehicle Mercury Switch Recovery Program is a partnership between the EPA, environmental
organizations, the Environmental Council of the States, and several industry trade associations.
Two trade associations—the American Iron and Steel Institute and the Steel Manufacturer
Association—are among the participating entities. The goal of this initiative is the removal of
mercury-containing light switches from scrap vehicles before the vehicles are flattened,
shredded, and melted to make new steel. Begun in 2006, the national initiative will help cut
mercury air emissions by up to 75 tons over the 15-year span of the project.17
Due to the characteristics of the raw materials, we determined that the source category
most like EAFs is the Iron and Steel Production source category. Therefore, we intend to use the
Iron and Steel Production AMAP/UNEP 2008 speciation profile for the EAF category, as it
aligns with what we would expect from the EAF source category given that most of the mercury
originates from mercury-containing light switches in the scrap metal from automobiles. Test data
for speciation of the mercury from the EAF source category is limited. Eagle Mountain Scientific
performed an annual compliance test including speciated mercury for Gerdau Ameristeel in St.
Paul, Minnesota. Test reports show most of the mercury is elemental with an average speciation
profile of Hg° = 76%, Hg2+ = 24%, and Hgp = 0.0%.18 This test lends support to the use of the
AMAP/UNEP 2008 speciation profile for Iron and Steel Production for the EAF category, as
76% of the mercury is in the form of Hg°, which is similar to the profile described above for Iron
and Steel Production (i.e., 80% elemental mercury). The speciation profile code used for EAFs is
HGMET.
Non-Ferrous Metal Smelters
16	NTH Consultants, Ltd, 2009.
17	https://archive.epa.gov/mercnre/archive/web/titml/index-4.hfinL
18	Eagle Mountain Scientific, Inc., 2008.
Page 9

-------
Mercury speciation for non-ferrous metal smelters is mostly estimated from the smelting
and roasting stage, where the majority of mercury is emitted. However, mercury emissions also
come from the slag dehydration and volatilization stages. Elemental gaseous mercury (Hg°)
tends to be the principal form in the flue gases emitted from non-ferrous metal smelters.
Catalytic metallic components and high particulate matter (PM) concentrations in flue gases are
the two primary causes. Mercury reclaiming towers in non-ferrous metal smelters preferentially
release Hg° to downstream flue gases. The speciation profile code used for non-ferrous metals
smelters is HGMET.
As with the other categories, there are a number of uncertainties in Hg speciation for this
category. More recent studies in China have shown speciation splits for Hg2+ to be closer to 50 to
60 percent, thereby reducing the splits for Hg° and Hgp.19 However, these data are from a limited
number of facilities in a limited geographical area. Studies from across the world suggest that the
deviation in speciation for non-ferrous metal production to be on the order of + 30 percent.20 It
should also be noted that in the previous (2002) AMAP the speciation for non-ferrous metals was
Hg° = 60%, Hg2+ = 30%, and Hgp = 10%.21
Chlor-Alkali Industry
Major points of mercury release in the mercury cell process of chlor-alkali production
include: byproduct hydrogen stream, end box ventilation air, and cell room ventilation air.
Typical devices/techniques for removal of mercury in these points are: (1) gas stream cooling to
remove mercury from hydrogen stream, (2) mist eliminators, (3) scrubbers, and (4) adsorption on
activated carbon and molecular sieves.22 Mercury cell chlor-alkali plants (MCCAPs) were
estimated to be the largest non-combustion anthropogenic sources of atmospheric mercury in the
U.S. in the early 1990s, emitting 6.5 megagrams/year from the 14 operating plants in 1994-1995
(US EPA, 1997). However, most facilities have closed or converted to non-mercury processes
since that time. Furthermore, the few remaining facilities have reduced emissions pursuant to the
MACT standards promulgated in 2003 (68 FR 70904, December 19, 2003). The mercury
emissions from this source category today are substantially lower than they were in the 1990s.
A February 2000 industrial source monitoring study at a chlor-alkali plant in Augusta,
Georgia found a high fraction of mercury emitted as Hg° gas in the production of chlorine and
caustic soda (mercury-based chlor-alkali production). This 9-day field sampling campaign study
is available from the EPA Technical Report EPA/600/R-02-007a. The results of this study
showed an average Hg° emission rate of 0.36 grams/minute from the roof ventilator. A separate
19	Zhang, L. et al., 2015.
20	Pacyna and Pacyna, 2002.
21	AMAP/UNEP, 2002.
22	Pacyna and Pacyna, 2002.
Page 10

-------
report23 based on this same monitoring study stated that the measurements in the cell building
roof vent of Hg2+ constituted approximately 2.1±0.7% of the concurrently measured Hg°. The
percent Hg2+ was substantially lower than the 30% Hg2+ estimate utilized by the EPA to model
the impact of MCCAPs for the 1997 Mercury Report to Congress and shown in Table 2.
Therefore, because these studies, based on measurements of mercury at a U.S. facility, are likely
to be a better representation of mercury speciation from the remaining domestic chlor-alkali
sources in the U.S. than the 2008 AMEP/UNEP report speciation, we are using a mercury
speciation of 97.3% Hg° and 2.7% Hg2+ for this industry. The 2.7% Hg2+ value was computed by
conservatively considering the error bound about the estimate from this measurement. The
profile code for chlor-alkali production is HGHCL.
Waste Incineration
Mercury speciation from waste incineration is another category where fluctuations can
depend on the type of waste burned, APCDs, and operational practices. The major incineration
types are municipal solid waste (MSW) incineration, medical waste incineration and industrial
and/or hazardous waste incineration. Based on a study by Park et al, a significant proportion of
mercury (80 to 96 percent) in the MSW releases from the incinerator into the flue gas is in the
form of Hg° at 850-1000°C24. High chlorine content in the waste results in higher Hg2+
proportion in the flue gas. Limestone slurry or powder sprayed in dry flue gas deacidification
(SD-FGD or D-FGD) absorbs a large amount of Hg2+ and activated carbon adsorbs a large
amount of both Hg° and Hg2+. Particles from SD-FGD and activated carbon injection (ACI) are
captured by the downstream FF. Hgp is removed by all types of dust controllers. The high Hg2+
formation rate due to the oxidative condition in flue gas and the high Hg2+ removal rate by
APCDs (especially SD-FGD, FF and ACI) cause significant variation in mercury speciation
profiles for incinerators.25 The profile code for waste incineration is HGINC.
An industrial source monitoring study conducted in Florida in the year 2000 found
emission profiles that support the fractions displayed in Table 2. This study can be found in the
EPA Technical Report EPA/600/R-00/102. For waste incineration, this study found nearly all
mercury emissions to be in the form of Hg2+ gas from a medical waste incinerator (MWI) and a
resource recovery incinerator (RRI). The analytical procedures for determination of mercury
from stack sampling followed EPA guidelines in 40 CFR 60, Appendix A, Method 29. Hg2+
was the dominant form of mercury at the RRI and the MWI, accounting for more than 75
percent and 98 percent, respectively, of the total mercury (somewhat higher than the split for
Hg2+ shown in Table 2). The ratio of Hg2+ to Hg° for RRI was 3.4:1 and for MWI was 51.6:1.
Since this study only included limited monitoring from two sources, we cannot conclude these
23	Landis, M. et al., 2003
24	Park, K. S. et al., 2008
25	Zhang, L. et al., 2016.
Page 11

-------
numbers are representative of all waste incineration. Regardless, the study may be helpful as
supplementary information.
Cremation Emissions
Extremely large uncertainties exist in this sector due to the diversity of mercury content
in human body and dental amalgam. Sectors that include cremation emissions should be further
examined for specific combustion practices. Cremation emissions are currently not a regulated
source of hazardous air pollutant emissions in the United States.26 The HGCRE profile code is
assigned to this category.
Mercury Production
Mercury has not been produced as a principal mineral commodity in the United States
since 1992.27 In 2018, mercury was recovered as a byproduct from processing gold-silver ore at
several mines in Nevada; however, production data were not reported.28
Artisanal and Other Gold Mining
No U.S. emissions reported.
Other Mining
The EPA will evaluate other source categories within the category of mining (e.g.,
taconite), as needed. Where possible we intend to use source specific data. Where no information
is available, we plan to use the AMAP/UNEP Non-Ferrous (Cu, Zn, Pb) Metal Production
profile (HGMET).
Geothermal Power Plants
Mercury is present in geothermal vents used for electric generation. Geothermal power
production equipment types that emit mercury to the atmosphere include off-gas steam ejectors
and cooling tower exhaust. Mercury speciation from these equipment types are documented in
the paper titled Mercury Emissions from Geothermal Power Plants29 EPA averaged the 3
speciation values from ejector off-gases and the 2 speciation values from cooling tower exhaust
26	Mari, M and Domingo, J., 2009.
27	USGS Minerals Yearbook, 2019.
28	USGS Minerals Yearbook, 2019.
29	Robertson, D.E. et al., 1977
Page 12

-------
air for profile code HGGEO, resulting in mercury speciation percentages of 87.0 percent for Hg°,
13.0 percent for Hg2+, and 0.0 percent for Hgp.
Mobile Sources
Mercury speciation of mobile sources is not addressed in the AMAP/UNEP 2008 report.
However, small amounts of mercury from mobile sources are included in the EPA's NEI. In the
2014 NEI mobile sources contribute less than 2 percent of the total mercury emissions. The
speciation factors for mobile sources are derived from source testing done at the EPA. Onroad
mobile source emission factors are documented in the EPA Technical Report EPA-420-R-16-
016, titled Air Toxic Emissions from On-road Vehicles in MOVES2014. Emission factors for
Hg°, Hg2+, and Hgp were obtained from a 2005 test program at the EPA's National Exposure
Research Laboratory (NERL). In this program mercury samples in raw exhaust were collected
from 14 light-duty gasoline vehicles and two heavy-duty diesel vehicles. Data collected included
elemental and total gas-phase mercury and particulate mercury. The emission rates were
computed in grams/mile (g/mi). Gaseous divalent mercury was computed by subtracting Hg° from
total gas-phase mercury. The emission factors for gasoline vehicles are 1.1 xlO"7 g/mi Hg°, 9.9 xlO"9 g/mi
Hg2+and 4 xlO"10g/m Hgp, resulting in mercury speciation percentages of 91.5, 8.2 and 0.3 for
Hg°, Hg2+, and Hgp, respectively. The emission factors for diesel vehicles are 6.2xl0"9 g/mi Hg°,
3.2 xlO"9 g/mi Hg2+and 1.6 xl0"9g/m Hgp, resulting in mercury speciation percentages of 56, 29
and 15 for Hg°, Hg2+, and Hgp, respectively. The mobile gasoline profile is assigned the code
HGMG, and the mobile diesel profile is assigned the code HGMD.
For nonroad engines, emission factors are documented in the EPA Technical Report
EPA-420-R-18-011, titled Speciation Profiles and Toxic Emission Factors for Nonroad Engines
inMOVES2014b. Nonroad gasoline and diesel vehicle emission factors for mercury (all phases)
were derived from onroad air toxics factors discussed above (EPA-420-R-16-016), converting gram
per mile emission factors to gram per gallon emission factors using fuel economy estimates.
Elemental Mercury Categories
Dental Alloy Production30, Bench Scale Reagents, Fluorescent Lamp Breakage31, and
Portland cement raw material handling operations prior to kilning32 are assumed to be entirely
elemental gaseous mercury (Hg°). These categories are assigned the profile code HGELE.
30	Goodrich, J. et al., 2016.
31	Johnson, N. etal., 2008.
32	US EPA, 2018 (p. 16 of 199, and Table 6)
Page 13

-------
V. Concluding Remarks
Based on our analysis, we conclude the speciation factors shown in Table 2 should
generally be used as the default factors for most source sectors in the absence of source category
specific data. However, for electric generating power plants that were included in the MATS
analysis, we recommend the factors used in the Mercury Air Toxics Standard (MATS) analyses,
and for other sectors where sufficient source category specific data are available (e.g., Portland
Cement) those data should be used to determine the speciation splits. For source categories not
specifically listed in Table 2 or in Appendix A, and for which insufficient information is
available to assign a surrogate source category profile, we recommend the use of an average of
the source categories for which we have determined a speciation profile, as described above and
listed in Appendix A.
Nevertheless, as mentioned above, there are uncertainties and data gaps. Future studies
on mercury speciation would be helpful for gaining a better understanding of the environmental
and human health risks from mercury. Conventional mercury measurement methods must be
carefully performed to effectively determine the speciation distribution. In addition, CEMS
intended to provide a direct determination of either total Hg° and/or Hg° and Hg2+ are currently
under development and evaluation in the field. The factors presented here reflect limited insight
and generally should not be used when more specific data is available.
VI. References
AMAP/UNEP, 2002. Technical Background Report to the Global Mercury Assessment. Artie
Monitoring and Assessment Programme/UNEP Chemicals Branch (2002).
http://www.eurocbc.org/final-assessment-report-25nov02.pdf (last accessed April 2020)
AMAP/UNEP, 2008. Technical Background Report to the Global Mercury Assessment. Artie
Monitoring and Assessment Programme/UNEP Chemicals Branch (2008).
https://www.amap.no/dociiments/doc/technical-backeroimd-report4o4he-global-atmosph.eric-
mercurv-assessment-2008/755 (last accessed April 2020)
Arctic Monitoring and Assessment Programme/United Nations Environment Programme, 2013.
Technical Background Report for the Global Mercury Assessment. AMAP, Oslo, Norway/UNEP
Chemicals Branch, Geneva, Switzerland.
https://www.arnap.no/docurnents/doc/techmcal-background-report-for-the-global-rnercurv-
assessment-2013/848
Bullock, D., Johnson, S., 2011. Electric Generating Utility Mercury Speciation Profiles for Clean
Air Mercury Rule, EPA-454/R-11-010. EPA.
https://www.regulations.gov/document'?D=EPA-HQ-< IXK-2009-0234-19910
Page 14

-------
Eagle Mountain Scientific, Inc., 2008. Results of the Annual Compliance Test Including
Speciated Mercury, Prepared for Gerdau Ameristeel, Report #902508.
Goodrich, J. et al., 2016. Exposures of Dental Professionals to Elemental Mercury and
Methyl mercury, Expo Sci Environ Epidemiol., 26(1), 78-85.
Johnson, N. et al., 2008. Mercury Vapor Release from Broken Compact Fluorescent Lamps and
In Situ Capture by New Nanomaterial Sorbents, Environmental Science & Technology 42(15),
5772-5778.
Landis, M. et al., 2003. Divalent inorganic reactive gaseous mercury emissions from a mercury
cell chlor-alkali plant and its impact on near-field atmospheric dry deposition, Atmospheric
Environment, 38 (2004), 613-622.
Mari, M. and Domingo, J., 2009. Toxic emissions from crematories: A review, Environment
International 36 (2010), 131-137.
Muntean, M. et al., 2018. Evaluating EDGARv4.tox2 speciated mercury emissions ex-post
scenarios and their impacts on modelled global and regional wet deposition patterns,
Atmospheric Environment, 184 (2018), 56-68.
NTH Consultants, Ltd, 2009. Performance Test BOF ESP and BOF Secondary Baghouse
Severstal North America, prepared for Michigan Department of Environmental Quality, Project
No. 74-080094-08.
Pacyna, E., Pacyna, J., 2002. Global emission of mercury from anthropogenic sources in 1995,
Water Air Soil Pollut., 137, 149-165.
https://doi.org/10.1023	2430561
Park, K. S. et al., 2008. Emission and speciation of mercury from various combustion sources,
Powder Technol., 180, 151-156.
Portland Cement Association, 2016. Email from Elizabeth Horner, Portlandtabl Cement
Association, to Sharon Nizich and Keith Barnett, OAQPS, U.S. EPA. "Mercury Speciation,"
Additional Requested Changes to Default Assumptions in Portland Cement RTR Modeling File,
October 28, 2016. EPA-HQ-OAR-2016-0442-0116.
https://www.regulations.gov/documeiit''O EPA~H.€M >A iOjoaM l.v
Robertson, D.E. et al., 1977. Mercury Emissions from Geothermal Power Plants, Science, Vol.
196, No. 4294 (Jun. 3, 1977), 1094-1097.
Page 15

-------
https://doi.org/10.1126/science.860 i 3 I
Streets, D. et al., 2019. Global and regional trends in mercury emissions and concentrations,
2010-2015.^tmosphericEnvironment, 201 (2019), 417-427.
US EPA, 1997. Mercury Study Report to Congress, Vol. 3, EPA-452/R-97-003. Office of Air
Quality Planning and Standards, Office of Research and Development, Washington, DC.
(Chapter 4).
US EPA Technical Report EPA/600/R-00/102. 2000. South Florida mercury monitoring and
modeling pilot study.
US EPA Technical Report EPA/600/R-02-007a. 2002. Characterization of mercury emissions at
a chlor-alkali plant.
US EPA Technical Report EPA-600/R-04/147. 2004. Selective Catalytic Reduction Mercury
Field Sampling Project.
US EPA, 2018. Residual Risk Assessment for the Portland Cement Manufacturing Source
Category in Support of the 2018 Risk and Technology Review Final Rule, Office of Air Quality
Planning and Standards, EPA-HQ-OAR-2016-0442-0222.
https://www.regiilations.gov/dociimeiit?D=EPA-HO-t	* _> 0 s 
-------
Appendix A
The following list includes the speciation codes given to each of the categories used in
the EPA's air emissions modeling programs.
Profile Code
I Gl lilNOO
l (,I lilNO 1
Mil'lil\02
EGUBIN03
I Gl lil\04
K(illil\05
I Gl lil\06
EGIIBIN07
I Gl lilNOS
EGUBIN09
KG I lil MO
EGUBIN11
KG I lil M 2
I Gl lilM3
KG I lil M 4
EGLBLM5
KG I lil M 6
I Gl lilM7
KG I lil MS
EGUBIN19
KG I'lil\20
Profile Description
Bituminous Coal. Coal Gasification
Bituminous Coal. PC Boiler with l-SP-CS
Bituminous Coal and Pel. Coke. PC Boiler wilh
i:sp-cs
Bituminous Coal, PC Boiler with SNCR and
ESP-CS
Bituminous Coal. PC lioilcr wilh l-SP-IIS
Bituminous Coal. PC Boiler wilh PM Scrubber
Bituminous Coal. PC lioilcr wilh Dry Sorbenl
Injection and l-SP-CS
Bituminous Coal, PC Boiler with FF Baghouse
Bituminous Coal. PC Boiler with ND.\ IT'
Baghouse
Bituminous Coal, PC Boiler with SCR and
SDA/FF Baghouse
Bituminous Coal. PC Boiler with l-SP-CS and
Wet I (il)
Bituminous Coal, PC Boiler with ESP-HS and
Wet FGD
Bituminous Coal. PC Boiler wilh IT" Baghouse
and Wet I (il)
Subbiluminous Coal. PC Boiler with l-SP-CS
Subbiluminous Coal. PC Boiler with l-SP-IIS
Subbiluminous Coal, PC Boiler wilh FF
Baghouse
Subbiluminous Coal. PC Boiler with PM
Scrubber
Subbiluminous Coal. PC Boiler with Sl).\ l-SP
Subbiluminous Coal. PC Boiler wilh Sl).\ IT
Baghouse
Subbituminous Coal, PC Boiler with ESP-CS
and Wet FGD
Subbituminous Coal. PC Boiler wilh l-SP-IIS
and Wet I (il)
Profile (Hg°, Hg2+, Hgp)
o iJ|i)2. ii 0X47. i) i)()5 I
ii 257ii. 0 (iX2i>. ii no I I
i) 5227. i) 4o5(\ () i)| 17
0.5256, 0.2712, 0.2032
() 372(\ i) 57S4. o.i)4^o
() 7X(->1). () N5I. i) i) I So
o 3274. o (-.71 o. i) oo 1 o
o 3144^06258700398
o MlJ7. 0 2XX(\ o o^| 7
0.4890, 0.4604, 0.0506
o iJ2oo. o 077S. o oo22
0.7870, 0.2068, 0.0063
o (1052. o 3300. o oMX
o (ii)o |. o 30X3. o oo | (i
o S741. o 1252. o oooo
o 1568, U.8283, U.U149
o 1)344. o 05 | |. o o 145
o i)5X(\ o 03X2. o oo32
o i)4(i7. o 0435. o now
0.9663. 0.0294. 0.0043
o lM37. o o44(\ oo||7
Page 17

-------
Profile Code
Profile Description
Profile (Hg°, Hg2+, Hgp)
EGUBIN21
rx;i iiiN22
Lignite Coal, PC Boiler with ESP-CS
Subbiluminous Coal. Cyclone Boiler with I'M
Scrubbier
0.9629, 0.0362, 0.0009
0 | w|. 0 0575. o.o234
EGUBIN23
rx;i liiN24
Subbituminous Coal/Pet. Coke, Cyclone Boiler
with ESP-HS
1.ignite Coal. Cyclone Boiler with l-SP-CS
0.9155, 0.0752, 0.0093
i) K2l)7. 0 16iw. 0 oo()4
EGUBIN25
I.Gl m\26
Bituminous Coal/Pet. Coke, Fluidized Bed
Combustor with SNCR and FF Baghouse
Not I 'sell al this lime
0.2970, 0.2787, 0.4244
EGUBIN27
I.Gl m\28
EGIBLN29
KGl IJIN30
Bituminous Waste, Fluidized Bed Combustor
with FF Baghouse
1 .ignite Coal. I'luiilizeil Bed Combuslor with
i :sp-( s
Lignite Coal, Fluidized Bed Combustor with FF
Baghouse
Anthracite W aste. I'luiilizeil Bed Combustor
with IT Baghouse
0.5907, 0.3881, 0.0212
0.X700. 0 | k-,4 o o|37
0.2840, 0.7118, 0.0042
o 5l)7o. 0.3730. o 0301
EGUBIN31
KGl lil.N32
Bituminous Coal, Stoker Boiler with SDA/FF
Baghouse
Not used al this time
0.6211, 0.1794, 0.1996
EGUBIN33
KGl IJIN34
KG I HI N35
KGl IJIN36
Lignite Coal, PC Boiler with ESP-CS and FF
Baghouse
1 .ignite Coal. I'C Boiler with Sl).\ IT Baghouse
1 .ignite Coal. PC Boiler with I'M Scrubber
Lignite Coal. PC Boiler with l-SP-CS and W et
1 (il)
Bituminous Coal, Cyclone Boiler with
Mechanical Collector
Bituminous Coal Pel Coke. Cyclone with l-SP-
CS and Wei K.I)
0.3532, 0.6449, 0.0019
0 X7o2. 0.12(->2. 0 003(1
0 i)(iX(i. 0 o2l)X. 0 001 (1
0 S574. 0 1345. 0 oos2
EGIBLN37
I.GlI5IN3S
0.3851, 0.4274, 0.1875
0 88(1.1. 0 1 13o. 0.0007.
EGUBIN39
KG! BIN40
Lignite Coal, Cyclone Boiler with SDA/FF
Baghouse
Subbiluminous Coal. I'luiilizeil Bed Combustor
with SNCR and IT Baghouse
0.7298, 0.1707, 0.0995
0 lKi32. 0 0342. 0 oo27
EGUBIN41
I.GlI5IN42
Subbituminous Coal/Bituminous Coal, PC
Boiler with ESP-CS
Subbiluminous Coal Bituminous Coal. PC
Boiler with l-SP-l IS
0.5630, 0.4282, 0.0088
0 4S03. 0 41>I 1. 0 o2K(i
Page 18

-------
Profile Code
Profile Description
Profile (Hgu, Hg2 , Hgp)
EGUBIN43
I.Cl m\44
IK,(KM
11(,(II
IK,(Ml}
IK,(KI
HGELE
IIGGEO
IK,(.11)
IK.IKI
IIGINC
ll(.IM)
ll(.MI)
IK.MII
IIGMG
IIGMW I
Bituminous Coal/Pet. Coke, PC Boiler with FF
Baghouse
liituminous Coal Suhhimmiiiolis Coal. PC
Boiler with IT Baghouse
Portland Cement Kiln l-\haust
('email Clinker Cooler
I'nel Combustion
Cremation (humans and animals)
Elemental (dental alloy, reagents, fluorescent
lamp breakage, Portland cement raw material
handling operations, artisanal scale gold mining)
(ieolheinial power plain (non-binary)
Large-scale gold production
Chlor-Alkali Plants
Waste Incineration
Industrial (a\ erage of non-comb prollles)
Mobile Diesel
Metal Production (iron and steel production,
non-ferrous metal production)
Mobile (iasolinc
Medical Waste Incineration
HGPETCOKE
Petroleum Coke Combustion
0.1939, 0.7841, 0.0220
() 5 1o 421 ii. i).i)5l>5
I) I) 34. DO
II <1. IIII. I .I)
0 50. 0 40. 0 | o
o So. o I 5. o 05
1.0. 0.0. 0.0
iK7.o|;voo
) so. 0 I 5. 0 05
) l>73. i) i)27. i)i)
) 2'). i) (•>'). i) 2')
) 73. i) 22. i) i)5
) 5(\ i) 2l). i) I 5
) x<). o | 5. o 05
) i) | 5. o os2. i) di)3
) 2'). i) (•>'). i) 2d
0.60. 0.30. 0.10
Page 19

-------