MERCURY EMISSIONS FROM COMBUSTION SOURCES (U.S. EPA Report, 1997)

Mercury is often found as a trace contaminant in fossil fuels or waste materials. When these materials are combusted, the combination of the elevated temperature of the process and the volatility of mercury and mercury compounds results in mercury being emitted in the combustion gas exhaust stream. This section addresses mercury and mercury compound emissions from seven stationary source combustion processes:

Coal combustion

Oil combustion

Municipal waste combustion

Sewage sludge incineration

Hazardous waste combustion

Medical waste incineration

These seven processes fall into two general categories. The first three involve fuel combustion for energy, steam, and heat generation, while the last four are primarily waste disposal processes, although some energy may be recovered from these processes. A summary of the estimated emissions from each of the above categories is as follows:

Category	Emissions, Mg (tons)
Coal combustion	67.8 (74.6)
Oil combustion	7.4 (8.1)
Wood combustion	0.1 (0.1)
Municipal waste combustion	26 (29)
Sewage sludge incineration	0.86 (0.94)
Hazardous waste combustion	6.3 (6.9)
Medical waste incineration	14.5 (16)

The paragraphs below provide a general introduction to the two combustion categories. As part of this introduction, a summary of nationwide fuel usage is presented in detail. This information was used to develop nationwide emissions of mercury for different sectors and fuels.

In 1994, the total annual nationwide energy consumption in the United States was 3.584 x 10¹²megajoules (MJ) (88.789 x 10¹⁵ British thermal units [Btu]).³⁹ Of this total, about 54.889 x 10¹² MJ (52.077 x 10¹⁵ Btu) or 59 percent involved consumption of coal, petroleum products, and natural gas in nontransportation combustion processes. (No data were available on energy consumption for wood combustion from the U.S. Department of Energy.) Table 6-1 summarizes the 1994 U.S. distribution of fossil fuel combustion as a function of fuel type in the utility, industrial, commercial, and residential sectors. The paragraphs below provide brief summaries of fuel use patterns; additional details on fuel consumption by sector for each State can be found in "State Energy Data Report, Consumption Estimates, 1994"³⁹.

As shown in Table 6-1, at 22.129 x 10¹² MJ (20.995 x 10¹⁵ Btu) per year, the industrial sector is the largest consumer of fossil fuels. This sector uses a mixture of natural gas (46 percent), fuel oil (7 percent), other petroleum fuels (35 percent), and coal (12 percent). The other petroleum fuels that are used include primarily liquified petroleum gas, asphalt and road oil, and other non-classified fuels.

The utility sector is the second largest fossil fuel energy consumer at the rate of 22.020 x 10¹² MJ (20.892 x 10¹⁵ Btu) per year. About 81 percent of this energy was generated from coal combustion, with bituminous and lignite coal contributing substantially greater quantities than anthracite coal.

As shown in Table 6-1, substantially smaller quantities of fossil fuel are used in the commercial and residential sectors than are used in the utility and industrial sectors. The fuels used are primarily natural gas, fuel oil, and liquified petroleum gas (the "other petroleum fuels" in the residential category). Almost all States use a mixture of the fuels, but the distributions vary substantially, with some States like California and Louisiana using primarily natural gas and others like New Hampshire and Vermont using a much greater fraction of fuel oil. One unique case is Pennsylvania where anthracite coal is used in both the residential and commercial sectors.

TABLE 6-1. 1994 DISTRIBUTION OF FOSSIL FUEL CONSUMPTION IN THE UNITED STATES

Annual energy consumption, 1012 MJ (1015 Btu)
	Utilities	Industrial	Commercial	Residential	Total
Bituminous/lignite coal	17.760(16.850)	2.642(2.507)	0.076(0.072)	0.041(0.039)	20.519(19.468)
Anthracite coal	0.018(0.017)	--(--)	0.012(0.011)	0.017(0.016)	0.047(0.044)
Distillate oil	0.100(0.095)	1.169(1.109)	0.489(0.464)	0.928(0.880)	2.686(2.548)
Residual oil	0.893(0.847)	0.448(0.425)	0.184(0.175)	--(--)	1.525(1.447)
Other petroleum fuels	0.027(0.026)	7.710(7.315)	0.121(0.115)	0.485(0.460)	8.343(7.916)
Natural gas	3.222(3.057)	10.160(9.639)	3.139(2.978)	5.249(4.980)	21.770(20.654)
Total	22.020(20.892)	22.129(20.995)	4.021(3.815)	6.719(6.375)	54.889(52.077)

Source: Reference 39.

FUEL OIL COMBUSTION

As shown in Table 6-1, based on energy consumption estimates by the U.S. Department of Energy, fuel oil use spans the four sectors of energy users. Distillate fuel oil is used in all sectors with the largest use in the residential (35 percent) and the industrial (43 percent) sectors, but also with amounts used in both the commercial (18 percent) and utility (4 percent) sectors. Residual oil is used primarily in the industrial (29 percent) and utility (59 percent) sectors. Because the oil combustion process is not complex, and control systems are not widely applied to oil-fired units, the discussion below will focus on fuel characteristics and on emissions from oil-fired units.³⁹

Fuel Oil Characteristics

The fuel oil characteristics of greatest importance for characterizing mercury emissions from fuel oil combustion are the heating value and the mercury content of the oil. The heating value is used for converting from emission factors with mass- or volume-based activity levels to those with activity levels based on heat input.

The term fuel oil covers a variety of petroleum products, including crude petroleum, lighter petroleum fractions such as kerosene, and heavier residual fractions left after distillation.⁴⁰ To provide standardization and means for comparison, specifications have been established that separate fuel oils into various grades. Fuel oils are graded according to specific gravity and viscosity, with No. 1 Grade being the lightest and No. 6 the heaviest. The heating value of fuel oils is expressed in terms of kJ/L (Btu/gal) of oil at 16?C (60?F) or kJ/kg (Btu/lb) of oil. The heating value per gallon increases with specific gravity because there is more weight per gallon. The heating value per mass of oil varies inversely with specific gravity because lighter oil contains more hydrogen. For an uncracked distillate or residual oil, heating value can be approximated by the following equation:

Btu/lb = 17,660 + (69 x API gravity)

For a cracked distillate, the relationship becomes:

Btu/lb = 17,780 + (54 x API gravity)

Table 6-9 provides an overall summary of the heating values of typical fuel oils used in the U.S., and Table 6-10 shows the variability in fuel oil heating values used in various regions of the country. Appendix B of Reference 40 provides additional details.

The data base for mercury content in fuel oils is much more limited than the coal mercury content data base. A number of petroleum industry associations were contacted, but none who responded have done any research on metals content in fuel oils. No single centralized data base is available, and the information presented below is based on limited data from individual studies.

TABLE 6-9. TYPICAL HEATING VALUES OF FUEL OILS

FUEL OIL GRADES
	No. 1	No. 2	No. 4	No. 5	No. 6
Type	Distillate	Distillate	Very light residual	Light residual	Residual	Crudeb
Color	Light	Amber	Black	Black	Black
Heating valuea
kJ/L	38,200	40,900	40,700	41,200	41,800	40,000-42,300
(Btu/gal)	(137,000)	(141,000)	(146,000)	(148,000)	(150,000)	(144,000-152,000)
kJ/kg	45,590-46,030	44,430-45,770	42,370-44,960	41,950-44,080	40,350-43,800	40,700-43,300
Btu/lb)	(19,670-19,860)	(19,170-19,750)	(18,280-19,400)	(18,100-19,020)	(17,410-18,900)	(17,500-18,600)

Source: Reference 40; and Reference 54.

^a The distillate samples, as well as the residual samples, analyzed for Btu/gal and Btu/lb heating values are different; therefore, the heating values presented do not directly correspond to one another.

^b These crude oil values are based on a limited number of samples from West Coast field sites presented in Reference 55 and may not be representative of the distribution of crude oils processed in the United States.

Concentrations of mercury in fuel oil depend upon the type of oil used. No comprehensive oil characterization studies have been done, but data in the literature report mercury concentrations in crude oil ranging from 0.023 to 30 ppmwt, while the range of concentrations in residual oil is 0.007 to 0.17 ppmwt. Because only a single mean value was found in the literature for mercury concentration in distillate oil, no conclusions can be drawn about the range of mercury in distillate oil. Table 6-11 lists typical values for mercury in oils, which were obtained by taking the average of the mean values found in the literature. The value for distillate oil is the single data point found in the literature and may not be as representative as the values for residual and crude oils.

TABLE 6-11. MERCURY CONCENTRATION IN OIL BY OIL TYPE

	Mercury concentration, ppmwt
Fuel oil type		Range	Typical value
Residual No. 6	-	0.002-0.006	0.004a
Distillate No. 2	-	-	<0.12b
Crude	46	0.007-30	>3.5c

Source: References 40, 50, 56.

^a Midpoint of the range of values.
^b Average of data from three sites.
^c Average of 46 data points was 6.86; if the single point value of 23.1 is eliminated, average based on 45

remaining data points is 1.75. However, the largest study with 43 data points had an average of

3.2 ppmwt. A compromise value of 3.5 ppmwt was selected as the best typical value.

Process Description

Fuel oils are broadly classified into two major types: distillate and residual. Distillate oils (fuel oil grade Nos. 1 and 2) are more volatile and less viscous than residual oils, having negligible ash and nitrogen contents and usually containing less than 0.1 weight percent sulfur. No. 4 residual oil is sometimes classified as a distillate; No. 6 is sometimes referred to as Bunker C. Being more viscous and less volatile than distillate oils, the heavier residual oils (Nos. 5 and 6) must be heated to facilitate handling and proper atomization. Because residual oils are produced from the residue after lighter fractions (gasoline and distillate oils) have been removed from the crude oil, they contain significant quantities of ash, nitrogen, and sulfur. Small amounts of crude oil are sometimes burned for steam generation for enhanced oil recovery or for refinery operations.^43,48

Oil-fired boilers and furnaces are simpler and have much less variation in design than the coal-fired systems described earlier. The primary components of the system are the burner, which atomizes the fuel and introduces it along with the combustion air into the flame, and the furnace, which provides the residence time and mixing needed to complete combustion of the fuel. The primary difference in systems that fire distillate oil and residual oil is that the residual oil systems must have an oil preheater to reduce the viscosity of the oil so that it can be atomized properly in the burner. Systems that fire distillate oil and residual oil also have different atomization methods.

The only source of mercury emissions from oil-fired boilers and furnaces is the combustion stack. Because the entire fuel supply is exposed to high flame temperatures, essentially all of the mercury and mercury compounds contained in the fuel oil will be volatilized and exit the furnace with the combustion gases. Unless these combustion gases are exposed to low-temperature air pollution control systems and high-efficiency PM control systems, which typically are not found on oil-fired units, the mercury and mercury compounds will be exhausted in vapor phase through the combustion stack.

Emission Control Measures

The three types of control measures applied to oil-fired boilers and furnaces are boiler modifications, fuel substitution, and flue gas cleaning systems.^40,48 Only fuel substitution and flue gas cleaning systems may affect mercury emissions. Fuel substitution is used primarily to reduce SO₂ and NO_x emissions. However, if the substituted fuels have lower mercury concentrations, the substitution will also reduce mercury emissions. Because PM emissions from oil-fired units are generally much lower than those from coal-fired units, high-efficiency PM control systems are generally not employed on oil-fired systems. However, the flue gas cleaning systems that are used on oil-fired units are described briefly below.

Flue gas cleaning equipment generally is employed only on larger oil-fired boilers. Mechanical collectors, a prevalent type of control device, are primarily useful in controlling PM generated during soot blowing, during upset conditions, or when a very dirty heavy oil is fired. During these situations, high efficiency cyclonic collectors can achieve up to 85 percent control of PM, but negligible control of mercury is expected with mechanical collectors.

Electrostatic precipitators are used on approximately one-third of the oil-fired power plants. Older ESP's may remove 40 to 60 percent of the PM, but negligible mercury control is expected. Newer ESP's may be more efficient, but no data are available for oil-fired power plants. Recent test data indicate mercury control efficiencies for ESP's controlling emissions from oil-fired utility boilers of 42 and 83 percent.⁴⁶ Scrubbing systems have been installed on oil-fired boilers to control both sulfur oxides and PM. Similar to systems applied to coal combustion (presented in Reference 40), these systems can achieve PM control efficiencies of 50 to 90 percent. Because they provide gas cooling, some mercury control may be obtained, but little data are available on their performance.

Emissions

The only substantive source of mercury emissions from fuel oil combustion operations is the combustion gas exhaust stack. Three types of information were used to develop emission factors for oil combustion. First, the data described above on fuel oil heating value and mercury content of fuel oils were used to develop emission factors by mass balance, assuming conservatively that all mercury fired with the fuel oil is emitted through the stack. Second, the emission factors developed in AP-42 for residual and distillate oil combustion and in Reference 47 for residual oil combustion were evaluated. Third, rated emission test data were evaluated and summarized. The paragraphs below first present the results generated from each of the three sources. Then, the relative merits of the emission factors generated via each of the procedures are discussed, and the best "typical" emission factors are identified.

The literature on fuel oil combustion suggests that essentially all mercury in the fuel oil is vaporized in the combustion zone and exhausted as a vapor in the combustion gas stream. Using the assumption that 100 percent of the mercury in fuel oil leaves the boiler or furnace in the exhaust gases, the data in Tables 6-9 and 6-11 were used to calculate uncontrolled emission factors for No. 2 distillate and No. 6 residual oil. Data presented in Reference 52, which show average crude oil heating values of 42,500 kJ/kg (18,300 Btu/lb) and 41,300 kJ/L (148,000 Btu/gal), can be combined with the mercury content data in Table 6-11 to calculate uncontrolled emission factors for crude oil combustion. The results of these calculations are presented in Table 6-12.

The calculated emission factors in Table 6-12 were compared to the available emission factors for fuel oil combustion from AP-42. The AP-42 presents emission factors for No. 2 and No. 6 fuel oils; no emission factors are developed for crude oil in AP-42.⁵³ The AP-42 emission factor for residual oil (No. 6) combustion is based on emission tests from 15 sites conducted from April 1990 through April 1994. The average emission factor reported for mercury emissions is 1.13 E-04 lb/10³ gallons(0.73 lb/10¹² Btu). This emission factor is rated C. The comparable calculated emission factor for residual oil in Table 6-12 based on the mercury content in the oil is 3.3 E-05 lb/10³ gallons (0.21 lb/10¹² Btu).

The AP-42 emission factor for distillate oil (No. 2) combustion (3.0 lb/10¹² Btu) is actually based on the average concentration of mercury in residual oil.⁴⁰ It is not based on any emission test data and is rated E. Additionally, the residual oil mercury concentration data used to develop this estimate are somewhat dated. The comparable calculated emission factor for distillate oil in Table 6-12 is 6.2 lb/10¹²

TABLE 6-12. CALCULATED UNCONTROLLED MERCURY EMISSION FACTORS

FOR FUEL OIL COMBUSTION

Calculated mercury emission factors
	kg/1015J	lb/1012 Btu	g/Mg fuel oil	10-3 lb/ton fuel oil	g/103L fuel oil	lb/106 gal fuel oil
Residual No. 6a	0.092	0.21	0.004	0.008	0.0039	0.033
Distillate No. 2a	2.7	6.2	0.12	0.24	0.01	0.86
Crudeb	82	190	>3.5	7.0	3.4	28

^aBased on typical heating values in Table 6-9 and mercury concentrations in Table 6-11.

^bBased on average crude oil heating values in Reference 54 and mercury concentrations in Table 6-11.

Btu and is based on the average of the mercury concentration measured in distillate oil samples at three sites as part of the California AB2588 study.⁵⁰

Reference 40 contains some mercury emission test data for the combustion of residual oil, distillate oil, and a 1:1 mixture of residual/crude oil. All of these data were developed from 1979 through 1981 and were presented in the previous mercury L&E. In an effort to eliminate mercury emission test data collected using older, less reliable emission test methods, EPA elected to utilize only post-1990 emission test data. This approach is consistent with the approach utilized in EPA’s Utility HAP Study. Therefore, the emission test data from Reference 40 are not utilized here; instead, more recent test data are presented.

Table 6-13 presents the results of a series of emission tests for the combustion of residual oil reported in Reference 47. As part of this test program, residual oil mercury concentrations were also measured; these data are also presented in Table 6-13. The data show that the mercury emissions from residual oil combustion are highly variable and that in most cases, the measured stack emissions are higher than the inlet fuel levels. Because these data are not normally distributed and appear to be log normal, a geometric mean was calculated to better represent the range of the data (References 47 and 56). The geometric mean for these data is 0.46 lb/10¹² Btu. Data are not available for distillate or crude oil combustion in Reference 47.

In summary, three mercury emission factors are presented for residual oil combustion: the 0.73 lb/10¹² Btu factor from AP-42, 0.46 lb/10¹² Btu from EPRI, and 0.21 lb/10¹² Btu from the EPRI residual oil analyses. Because the 0.46 lb/10¹² Btu emission factor is essentially the midpoint of the range of the three values, this factor was selected as the best "typical" emission factor for residual oil combustion. Because there are no emission test data for distillate oil combustion, the mass balance approach was used to estimate the best "typical" emission factor for distillate oil combustion.

As a part of the previous L&E study, two test reports prepared as a part of the California "Hot Spots" program were reviewed.^54,57 The emission factors generated from these three reports are summarized in Table 6-14. Each of the reports contained the data on fuel oil characteristics needed to calculate mercury input rates, so Table 6-14 contains both calculated emission factors based on mercury input levels and measured emission factors based on stack tests. Because mercury levels in all of the fuel oils tested were below detection limits, all calculated emission factors are reported as "less than" values. Note that only one of the two tests showed mercury levels above the detection limit in the stack. That test showed measured emissions to be substantially greater than mercury input to the process, making the results suspect. These discrepancies may be a function of the analytical problems that have been reported for mercury methods applied to combustion sources. These problems are discussed in more detail in Section 9. On balance, these data provide little information for emission factor development.

The available information on uncontrolled mercury emissions from crude oil combustion is ambiguous. The limited test data presented in Table 6-14 show measured factors that range from less than 0.05 to 15 kg/10¹⁵ J (<0.12 to 34 lb/10¹² Btu), a range of almost three orders of magnitude. The calculated emission factor of 84 kg/10¹⁵ J (190 lb/10¹² Btu), which is based on limited fuel composition and heating value data, expands the range even further. Because these data are quite sparse and the relative quality of the data is uncertain, the midpoint of the range was selected as the best "typical" emission factor.

TABLE 6-13. MERCURY CONCENTRATIONS IN RESIDUAL OIL AND

MERCURY EMISSION FACTORS FROM RESIDUAL COMBUSTION

Unit name	Residual oil mercury concentration, ppmw	Mean mercury emission factor, lb/1012 Btu
117	0.0023	0.60
118	0.0040	0.98
112	0.0060	1.3
13	<0.040	0.23
103	<0.090	<3.6
106	<0.10	<5.0
107	<0.10	<37
104	<0.10	12
105	<0.10	<4.7
108	<0.10	<32
109	<0.90	1.8
13	<0.030	0.16
118	0.0040	0.50
112	0.0060	0.24
13	<0.040	<0.066
117	0.0023	0.49
Source: Reference 47.

TABLE 6-14. MERCURY EMISSION FACTORS FOR CRUDE OIL COMBUSTION

GENERATED FROM CALIFORNIA "HOT SPOTS" TESTS

Calculated mercury emission factorsa
		kg/1015J	lb/1012 Btu	g/Mg fuel oil	10-3 lb/ton fuel oil	g/103L fuel oil	lb/106 gal fuel oil
Pipeline/ process heaterb	Crude	<2.4	<5.6	<0.10	<0.20	<0.097	<0.81
Generatorc	Crude	<2.4	<5.6	<0.10	<0.21	<0.10	<0.83

Measured mercury emission factorsaa
	kg/1015J	lb/1012 Btu	g/Mg fuel oil	10-3 lb/ton fuel oil	g/103L fuel oil	lb/106 gal fuel oil
Pipeline/ process heaterb	<0.052	<0.12	<0.0022	<0.0044	<0.0021	<0.018
Generatorc	14.7	34.1	0.62	1.2	0.61	5.1

Source: Reference 54; Reference 57.

^aEmission factors were based on assumed crude oil heating value of 42,500 kJ/kg (18,300 Btu/lb) and density of 0.97 kg/L (8.1 lb/gal).

^bMercury detection limit is 0.1 mg/kg.

^cMercury detection limit is 0.1 mg/L.

The uncontrolled emission factors for distillate, residual, and crude oil are presented in Table 6-15. Data are insufficient to develop controlled emission factors for fuel oil combustion. There is considerable uncertainty in these emission factor estimates due to the variability of mercury concentrations in fuel oil, the incomplete data base on distillate oil, and the uncertainty in sampling and analysis for detecting mercury. Therefore, these estimates should not be used to determine emissions from specific oil-fired units.

TABLE 6-15. BEST TYPICAL MERCURY EMISSION FACTORS FOR FUEL OIL COMBUSTION

Typical mercury emission factors
	kg/1015J	lb/1012 Btu	g/Mg fuel oil	10-3 lb/ton fuel oil	g/103L fuel oil	lb/106 gal fuel oil
Residual No. 6	0.020	0.46	0.009	0.017	0.0085	0.071
Distillate No. 2	2.7	6.2	0.12	0.24	0.01	0.86
Crude	41	95	1.7	3.5	1.7	14

Total 1994 mercury emissions from oil combustion (utility, industrial, and commercial/residential) are estimated to be 7.6 Mg (8.4 tons); see Appendix A for details.