94
R.R. Turner and G.R. Southworth
plant site. Concentrations of mercury in fish in the receiving aquatic system also
decreased and the fish consumption advisory was lifted in March 1992. This site
is relatively unique in experiencing such a positive response to deliberate
corrective actions focused on reducing biological contamination, and is one of
few known sites of industrial Hg contamination in North America where a fish
consumption advisory was actually lifted. Additional investigative and remedial
actions are still underway.
2.3
Oak Ridge, Tennessee
Approximately 11 400 metric tons (Mg) of elemental mercury were shipped from
the US government's strategic stockpile to a nuclear weapons plant in Oak Ridge
in the early 1950S to support the enrichment of lithium-6, a stable isotope of
lithium, which was used in the production of hydrogen bombs. The enrichment
process entailed counterflowing liquid mercury and lithium hydroxide. The
lighter isotope of lithium, 6Li, partitioned preferentially to the mercury and was
thus concentrated. The process otherwise had many similarities to electrolytic
production of chlorine and caustic soda (chloralkali) using mercury cell
technology. Spills and leaks of elemental mercury occurred during process
development as well as during the production phase, which ended in 1963. One
step in the process, involving nitric acid cleaning of the mercury, was found to be
a significant source of environmental releases, and this step was altered in 1957 to
conserve the stock of elemental mercury, but not before nearly 100 Mg were
released into the headwaters of a local creek.
This site offers one of the few known situations where a mass balance for a
major historical industrial use of mercury is available. As shown in Table I,
approximately 8% of the mercury inventory charged to Oak Ridge was either lost
to the environment or could not be accounted for. Some of the unaccounted
mercury was most likely never received by the facility, as incoming flasks were
only counted, not weighed, with the presumption that each flask contained the
typical 35 kg. Some confirmation of the total measured losses to the local creek
have been provided by estimates of inventories of mercury in downstream
floodplain soils (70 Mg) and reservoir sediments (80 Mg). Similarly, the use of
dated sediment cores from the reservoir have independently confirmed the
chronology of aquatic Hg releases from the site (Turner et al. 1984), while the
annual growth rings of eastern red cedar U. virginiana) trees growing near the
plant site and floodplain recorded the atmospheric Hg releases (Turner and
Bloom 1994).
Ambient air concentrations of mercury (2 to 1200 ng m -3) have been
monitored continuously at the plant site since 1986 (Turner and Bogle 1993),
while ambient air monitoring on the floodplain (2 to 120 ng m -3; Barnett et al.
1996) was limited to the periods just before and during soil excavation associated
with remediation. In 1993, Lindberg et al. (1995) estimated emissions of mercury
from the 100-ha floodplain to the atmosphere using a micrometeorological
gradient approach at about 86 ng m -2 h -I. They concluded that while this
R.R. Turner and G.R. Southworth
plant site. Concentrations of mercury in fish in the receiving aquatic system also
decreased and the fish consumption advisory was lifted in March 1992. This site
is relatively unique in experiencing such a positive response to deliberate
corrective actions focused on reducing biological contamination, and is one of
few known sites of industrial Hg contamination in North America where a fish
consumption advisory was actually lifted. Additional investigative and remedial
actions are still underway.
2.3
Oak Ridge, Tennessee
Approximately 11 400 metric tons (Mg) of elemental mercury were shipped from
the US government's strategic stockpile to a nuclear weapons plant in Oak Ridge
in the early 1950S to support the enrichment of lithium-6, a stable isotope of
lithium, which was used in the production of hydrogen bombs. The enrichment
process entailed counterflowing liquid mercury and lithium hydroxide. The
lighter isotope of lithium, 6Li, partitioned preferentially to the mercury and was
thus concentrated. The process otherwise had many similarities to electrolytic
production of chlorine and caustic soda (chloralkali) using mercury cell
technology. Spills and leaks of elemental mercury occurred during process
development as well as during the production phase, which ended in 1963. One
step in the process, involving nitric acid cleaning of the mercury, was found to be
a significant source of environmental releases, and this step was altered in 1957 to
conserve the stock of elemental mercury, but not before nearly 100 Mg were
released into the headwaters of a local creek.
This site offers one of the few known situations where a mass balance for a
major historical industrial use of mercury is available. As shown in Table I,
approximately 8% of the mercury inventory charged to Oak Ridge was either lost
to the environment or could not be accounted for. Some of the unaccounted
mercury was most likely never received by the facility, as incoming flasks were
only counted, not weighed, with the presumption that each flask contained the
typical 35 kg. Some confirmation of the total measured losses to the local creek
have been provided by estimates of inventories of mercury in downstream
floodplain soils (70 Mg) and reservoir sediments (80 Mg). Similarly, the use of
dated sediment cores from the reservoir have independently confirmed the
chronology of aquatic Hg releases from the site (Turner et al. 1984), while the
annual growth rings of eastern red cedar U. virginiana) trees growing near the
plant site and floodplain recorded the atmospheric Hg releases (Turner and
Bloom 1994).
Ambient air concentrations of mercury (2 to 1200 ng m -3) have been
monitored continuously at the plant site since 1986 (Turner and Bogle 1993),
while ambient air monitoring on the floodplain (2 to 120 ng m -3; Barnett et al.
1996) was limited to the periods just before and during soil excavation associated
with remediation. In 1993, Lindberg et al. (1995) estimated emissions of mercury
from the 100-ha floodplain to the atmosphere using a micrometeorological
gradient approach at about 86 ng m -2 h -I. They concluded that while this