Policy Analysis Estimate of Mercury Emissions to the Atmosphere from Petroleum S. MARK WILHELM* Mercury Technology Services, 23014 Lutheran Church Road, Tomball, Texas 77375
An estimate of the contribution of mercury to the atmospheric environment from petroleum processed in the United States is constructed from recent data. The estimate is based on a mass balance approach for mercury in crude oil, in refined products, and in waste streams (air, water, solid waste) from refineries. Although there are insufficient data at present to have a high degree of confidence in the mean amount and range of mercury concentrations in crude oil or in refined products, the framework of the estimate should assist direction for the acquisition of additional data. On the basis of selected data that put the estimated mean concentration of total mercury in crude oil close to 10 ppb, it is calculated that the total amount of mercury in U.S. petroleum processed yearly is slightly over 8000 kg/yr. Of this amount, approximately 6000 kg/yr is estimated to be emitted to the atmosphere from combustion of liquid hydrocarbon fuels, which represents about 10% of the U.S. yearly emission rate of atmospheric mercury from coal combustion. The material balance predicts that the amount of mercury in air emissions from all U.S. refineries is on the order of 1500 kg/yr based on the assumption that fugitive mercury emissions from refineries are negligible. Atmospheric emissions of mercury from fuel oil burned in the United States are estimated in the U.S. EPA Mercury Report to Congress to be approximately 10 000 kg/yr, and this estimate may be in error on the high side by a factor of 3-10. If the mean amounts of mercury in U.S. distillate and residual fuel oils are in the range of 5-15 ppb, as suggested by more recent data, then U.S. fuel oil combustion should contribute no more that about 1000-3000 kg/yr (emission ratio ) 1) of mercury to the atmospheric burden.
Introduction The U.S. EPA recently announced (December 2000) that regulation of mercury and other hazardous air emissions from oil- and coal-fired electric utility steam-generating units is necessary and appropriate. The EPA decision derived from the mandate given to the EPA by Congress under the Clean Air Act and followed numerous studies of the hazards to the public health anticipated to occur as a result of hazardous air emissions by electric utility steam-generating units. The studies that were the basis of the EPA decision are summarized in the Report to Congress (1) and included examination of mercury emissions from both coal and oil combustion. While the concentration of mercury in U.S. coals is * Corresponding author phone: (281)255-3775; fax: (281)357-0721; e-mail:
[email protected]. 4704
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001
TABLE 1. THg in Crude Oil by INAA (7) source
amount (ppb)
SD (ppb)
notes
California California California California California Libya Libya Libya Louisiana Wyoming mean
114 81 88 29 688 78 2 079 62 75 23 77 3 200
2.8 1.9 3.0 103.9 2.4 11.9 5.1 1.7 1.8 3.4
detection limit ) 4 ppb Cymric
range 23-30 000 ppb
known with some certainty, the EPA admitted that the amounts of mercury in crude oil and fuel oils were much less certain. Under the Emergency Planning and Community RightTo-Know Act (EPCRA), the U.S. EPA requires certain facilities processing or using listed toxic chemicals to report the annual quantity of such chemicals entering the environment (2). The reporting requirements are termed the toxic release inventory or TRI. In 2000, the reporting threshold for mercury was revised from 10 000 lb to 10 lb. The current (June 2001) U.S. EPA estimate (for TRI guidance purposes) proposes that the average concentration of mercury in crude oil is 1.5 ppm, based on data on mercury concentrations in crude oil compiled by Wilhelm and Bloom (3). The newly proposed guidance estimate replaces a prior 6 ppm estimate proposed in an earlier (November 2000) draft of the guidance document. The estimate of 6 ppm proposed earlier by the EPA was based on data compiled by Brooks (4). The U.S. EPA estimates (4-6) of the mean concentration of mercury in crude oil are likely in error for reasons to be discussed, but they stem mainly from reliance on data from anomalous sources that skew the sample ensembles used in previous compilations. The data and analysis provided herein also cast some doubt on the EPA estimates for mercury concentrations in fuel oil and the resulting average U.S. emissions estimates. Using more recent data, an attempt is made to construct a better framework for estimating atmospheric mercury emissions from petroleum.
EPA Estimates of Mercury in Crude Oil The U.S. EPA estimates of the mean concentration of total mercury (THg) in crude oil, both historic (4-6) and present (1), are based primarily on data compiled by Brooks (4). Brooks’ compilation consisted mostly of data published by Shah, Filby, and various colleagues (7, 8) who were the first to apply instrumental neutron activation analysis (INAA) to U.S. crude oils. Shah et al. (7) reported concentrations for 10 crude oils as shown in Table 1 in 1970. The data of Shah et al., in its several reported forms, are the basis for U.S. EPA estimates (4-6) of parts per million levels for the mean amount of mercury in crude oil. The exercise (by EPA) to arrive at a mean amount involved averaging the mean or median of the range of concentrations from the early studies of Shah and Filby (7, 8). This method to arrive at a mean concentration is questionable in the sense that the average was constructed from several papers and reports that contain the same data but were considered as separate crude oil source studies (4). 10.1021/es001804h CCC: $20.00
2001 American Chemical Society Published on Web 10/31/2001
TABLE 2. THg in Alberta Crude Oils (9) stratigraphic era
no. of samples
no. above DL
high (ppb)
low (ppb)
Upper Cretaceous Lower Cretaceous Jurassic Triassic Carboniferous Devonian total
21 18 3 4 8 36 86
11 7 0 2 4 13 38
202 138 1 6 19 399 399
DL DL DL DL DL DL DL
a
meana (ppb)
median (ppb)
17.6 17.1 1 3 5 36 21.9
2.5 1 1 2.5 1.5 1 1
SD
DL (ppb)
46.0 38.1
2 2 2 2 2 2
2.4 6.3 92.5 63.6
Calculated assuming
