The chemical and physical interactions of radioactive
compounds are key to understanding how they can contaminate the environment and, more importantly,
how best to remove them.
From 1952 to 1989, the Rocky Flats Nuclear Weapons
Plant, located about 24 km northwest of Denver, Colorado, made components for the nation's nuclear
arsenal using various radioactive materials, including plutonium and uranium; toxic metals
such as beryllium; and hazardous solvents, degreasers, and other chemicals. The key component
produced at Rocky Flats was the plutonium pit, commonly referred to as the trigger for a nuclear
weapon. The pit provides energy to fuel the explosion.
In 1989 the Federal Bureau of Investigation
and the Environmental Protection Agency abruptly halted nuclear production work to investigate
environmental and safety concerns, and the site was added to the EPA's Superfund list later that
year. In 1993 the secretary of energy announced the end of the nuclear production mission, and the
area became known as the Rocky Flats Environmental Technology Site (RFETS) in 1995. Nearly 40 years
of nuclear weapons production left behind contaminated facilities, soils, and surface and ground
water.
More than 2.5 million people
live within an 80-km radius of the site, and 300 000 of those live in the Rocky Flats watershed.
The sudden shutdown left large quantities of plutonium and other hazardous substances in various
stages of processing and storage. Because plutonium is dangerous to human health, even in minute
quantities, the cleanup of plutonium-contaminated materials is highly complex, tedious, and
labor intensive.
In March 1995 the US Department
of Energy estimated that the cleanup for Rocky Flats would cost in excess of $37 billion and take
70 years to complete. By 1996, DOE and independent contractor Kaiser-Hill Co had initiated a massive
effort that eventually resulted in a credible plan to accelerate the closure of Rocky Flats by 31
December 2006 at a contracted cost of $7 billion. After a troubled start, Kaiser-Hill completed
the task nearly a year ahead of schedule.
What led to the turnaround?
Without question, an incentive-laden contract, strong support and stable funding from Congress,
high-level DOE support that mobilized the entire DOE complex to assist in the cleanup, technological
innovation, and improved scientific understanding all contributed. Much has been made of the
contractor's fee of more than $500 million, but less has been said about the role that scientific
understanding played in guiding key cleanup decisions and facilitating good project management.
The impetus to understand
the science behind plutonium contamination gained momentum in 1995 when intense rainfall and
wet springtime conditions raised concerns about the mobility and dispersal of plutonium and americium.
To account for increased concentrations of plutonium at various surface water-monitoring locations,
researchers hypothesized that plutonium was soluble in surface and ground water. But modeling
efforts at the time predicted very little movement of plutonium. The discord between the data and
predictions prompted DOE and Kaiser-Hill in 1995 to establish the Actinide Migration Evaluation
(AME) advisory group. The idea was to solicit advice and technical expertise on how elements such
as plutonium, uranium, and americium are likely to behave in the air, surface water, ground water,
and soil (see box 1).
Supported by scientific
measurements, the group found that plutonium and americium form insoluble oxides and colloids
that adhere to small organic and mineral particles in soil. The particles can migrate throughout
the Rocky Flats environment by wind and surface water; particles are lifted from some location,
suspended in air or water, and then redeposited as sediment somewhere else. This understanding
showed that soluble transport models were, in fact, not appropriate to describe the transport
of plutonium and americium and led to the adoption of erosion and sediment-transport models. And
it provided the basis for how best to negotiate a cleanup agreement and settle on an allowable standard
of 50 picocuries per gram of soil. The relevant measure of plutonium and americium concentration
is how much radiation is given off per unit volume or mass.
Did it save a lot of taxpayer
dollars? That's difficult to determine. What everyone agrees on is that scientific understanding
provided clarity and focus on the real issues surrounding plutonium and americium in the RFETS
environment. The clarity and focus in turn allowed for good project management, guided remediation
efforts, and most certainly helped shave decades and billions of dollars off the initial cleanup
estimate.
Nearly the size of a small city with its
own fire department, medical offices, cafeteria, and water- and sewage-treatment plants, Rocky
Flats comprised more than 800 structures on a 1.6-square-kilometer industrial area surrounded
by approximately 24 square kilometers of controlled open space (see Figure 1). The open space continues
to serve as a buffer between Rocky Flats and the nearby, growing communities and is home to many species
of animals and plants.
Water at Rocky Flats and
the surrounding area is distributed among surface water, shallow ground water, and deep ground
water.1 A series of detention ponds had been constructed along creeks to manage plant
waste and surface water runoff. Shallow ground water refers to water within the alluvium and weathered
bedrock and is found to a depth of 30 m. Water from the surface filters downward, recharging the shallow
ground water, which in turn recharges the stream channels at certain times of the year. Beneath
the alluvium is highly impermeable bedrock that inhibits vertical flow. As a result, shallow ground
water flows laterally and either discharges into the streams or emerges as hillside springs and
seeps. Deep regional ground water flows about 200–300 m below the surface. Because of the
intervening bedrock, that regional ground water aquifer is hydrologically isolated from the
Rocky Flats surface and shallow ground water and from actinide contaminants.2
Winds at RFETS predominantly
flow from the northwest to the southeast. They can periodically become so strong and gusty—exceeding
160 km/h—that they shatter the windshields of vehicles parked on the site. The wind is an important
factor in the dispersal of soil and actinides. Indeed, air monitoring and subsequent calculations
of the actinide loads showed that air transport was a dominant actinide migration pathway, before
and during cleanup.
When it was operating, the Rocky Flats
nuclear plant generated a huge volume of waste contaminated with radionuclides and other hazardous
substances. The majority was shipped offsite, but improper disposal, ruptured or leaking pipes,
fires, and faulty storage units resulted in local soil and water contamination. By far the largest
source of plutonium and americium contamination in soils emanated from chemical drums stored
in an area known as the 903 Pad. Between 1958 and 1969, an estimated 19 000 liters of tainted
lathe coolant (about 86 g or 5.3 curies of plutonium) leaked into the ground; wind
and surface-water erosion then carried plutonium and americium in a pattern that tracks roughly
with the prevailing winds to the east and southeast, at low levels past the eastern site boundary
(see Figure 2).
Plutonium and americium
generally exhibited the same spatial distribution in surface soils, with wide variations in radioactive
activities occurring throughout the site. Approximately 90 percent of the radioactive inventory
was in the top 12 cm of the soil.3 The concentrations ranged up to several picocuries
per liter in streams and ponds, and up to a few nanocuries per gram in soils and sediments.
Chemical reactions, particularly
redox reactions in soil and ponds, are often hypothesized to explain actinide mobility. At one
extreme, the actinides may react with surrounding materials to create soluble and mobile compounds.
At the other extreme, the actinides might remain unchanged at the molecular scale and become bound
to natural organic and mineral materials. These natural materials themselves may undergo chemical
reactions to form mobile components, thereby carrying the actinides along with them.
The contrast between actinide
solubilities—the solubility of plutonium and americium is very low, whereas that of uranium
is relatively high—drove researchers' consideration of colloidal and particulate transport
processes and prompted the AME advisory group to carefully evaluate evidence that could distinguish
solubility and colloidal and particulate results. For example, actinide chemists have long known
that under environmental conditions plutonium is most stable as oxides,4 and colloid-sized
materials,5 but detailed knowledge of their reactivity in the environment is limited
to concentrations of picocuries per liter in water and picocuries to nanocuries per gram in soil.
Under natural environmental
conditions, plutonium solubility is limited by the formation of amorphous plutonium hydroxide
[Pu(OH)4] or polycrystalline plutonium oxide (PuO2). Formation of
these compounds provides an upper limit on the amount of dissolved—that is, ionic or molecular—plutonium
that can be present. Plutonium oxide's measured solubility range5 of 10−10
to 10−13 mol/L
is limited by the formation of Pu(OH)4. Due to that very low solubility and the tendency
of compounds of Pu(IV), the fourth oxidation state of plutonium, to adhere to organic and mineral
particles, the primary path of plutonium transport is through the migration of fine particles.
Indeed, when concentrations of plutonium above fallout levels have been investigated in detail,
the plutonium has been linked to colloids and particulates.6
Synchrotron radiation studies
Although researchers at Rocky Flats
suspected that plutonium contamination in the environment was in a particle form—most likely
the very insoluble PuO2—definitive proof did not exist to verify its chemical
form and oxidation state.7 Los Alamos National Laboratory researchers led by Steven
Conradson performed x-ray absorption fine structure spectroscopy at the Stanford Synchrotron
Radiation Laboratory to determine the chemical form of plutonium in RFETS soils and concretes.8
Although not well suited for the extremely dilute samples typical of the RFETS environment, the
technique successfully identified the chemical fingerprints of select, higher-concentration
samples through a careful tuning of the spectroscopy data collection—that is, a judicious
choice of which absorption region to analyze, together with long data-collection times.
X-ray absorption near-edge
structure analysis identified the oxidation state of plutonium in soils and concretes as Pu(IV).
An analysis of the extended x-ray absorption fine structure in the spectra—the spectral
oscillations in the region beyond the absorption edge—unambiguously identified the chemical
form of plutonium in soil and contaminated concrete around the site as the relatively insoluble
hydrous oxide PuO2·xH2O
(see box 2).
Judging from the geochemical
characteristics found using x rays, one might conclude that insoluble oxides of plutonium and
americium would be trapped in the ground and remain immobile. That's true to a point. A growing number
of field studies, however, document the movement of low concentrations of low-solubility radionuclides
in surface and ground waters.6,9,10 Those small concentrations can be transported
in surface water and soils by particles of sizes typically ranging from a nanometer to several microns.11,12
Because the particles remain suspended in ground water, they can move in the natural watershed
and settle into the series of ponds around RFETS.
Ultrafiltration studies
From 1998 until 2001, Texas A&M
University's Peter Santschi and coworkers examined 239Pu, 240Pu, and
241Am concentrations in the field and through laboratory studies at RFETS.10
Since the environmental forms of actinides in the surface waters were in the concentration range
of 10−3
to 10−1 pCi/L,
filtration and tangential-flow ultrafiltration were the only methods suitable to separate and
analyze the different phases (see box 3). Measurements of total 239Pu, 240Pu,
and 241Am concentrations in storm runoff and pond discharge samples collected during
spring and summer from 1998 to 2000 demonstrated that most of the 239Pu, 240Pu,
and 241Am transported from contaminated soils to streams occurred in the particulate
(roughly larger than 0.45 µm)
and colloidal (roughly between 2 nm to 0.45 µm)
phases.
In general, most of the
Pu and Am in RFETS water was found in the particulate phase, with most of the material that passed
a 0.5 µm filter being
colloidal. Based on graphite-furnace atomic absorption spectroscopy, transmission electron
microscope, and energy dispersive x-ray microprobe images, colloids were primarily composed
of clay and organic matter.
So-called isoelectric
focusing experiments of radiolabeled colloids from RFETS soils revealed that colloidal Pu formed
in the tetravalent state and was mostly associated with a negatively charged organic colloid having
a molecular weight of 10–15 kilodaltons, rather than with the more abundant inorganic colloids
made up of iron oxide and clay. Santschi's evidence strongly argued against the presence of mobile
colloidal microparticles mainly in the form of PuO2, but suggested that PuO2
is imbedded in, or attached to, organic matter containing some iron.
Each of these complementary
studies provided evidence that the low levels of Pu and Am in surface water at RFETS are transported
by the colloidal and particulate fraction of the water, not by the dissolved fraction.
Modeling actinide transport
Understanding that Pu and Am exist in
the form of insoluble particles clarified that the initial models of contaminant transport—ones
based on soluble forms of Pu—were flawed and indefensible. To best fashion the range of possible
remediation and management scenarios, AME advisers needed the ability to predict how the radioactive
material moved under existing conditions.
AME chose the Water Erosion
Prediction Project (WEPP) model,13,14 a state-of-the-art process-oriented computer
model that simulates hillside erosion processes and estimates the spatial and temporal distributions
of soil erosion and sediment deposition in stream channels and impoundments. Because it accounts
for enrichment of transported sediment in fine particles, the WEPP model is well suited for contaminant
transport calculations.
To estimate stream channel
sediment erosion and deposition, output from the WEPP model was routed into yet another—the
US Army's Hydrologic Engineering Center sediment-transport model, HEC-6T, which can accommodate
up to 100 tributaries flowing into a main channel.15 The combination was crucial to
modeling the RFETS watersheds and using soil data to predict surface-water actinide concentrations.
AME applied the soil-erosion
and sediment-transport models to the hillslopes and channel systems at RFETS and compared the
results with monitoring data to parameterize, initialize, and calibrate the models. The coupled
models could then be used to simulate storm events and the transport of 239Pu, 240Pu,
and 241Am contaminants, estimate the amount of contaminated sediment in surface
water, and analyze which hillslopes and drainages the contamination moved along. Finally, the
coupled models were used with data on climate and soil contamination to predict rates of sediment
and contaminant transport under various management scenarios designed to handle the cleanup.
As part of the modeling
process, the predicted soil erosion—that is, the mass eroded per unit area—was combined
with actinide soil-concentration data to generate a map of actinide mobility predicted for a specific
storm event. Surprisingly, the results of those maps revealed that the largest Pu and Am loads delivered
to surface water do not necessarily originate from areas with the highest concentrations of Pu
and Am in the soil. The combination of topography, vegetative cover protecting the soil, soil erodibility,
and actinide concentration determines the rate of erosion and contaminant transport.
The area east of the 903
Pad, for example, generally contains the highest levels of Pu and Am in the area (see Figure 3). The
area around the 903 Pad, however, is relatively flat, with slopes of only about 1%. Consequently,
that area suffers far less soil erosion by water than other, steeper parts of the watershed—with
a corresponding reduction in the amount of Pu and Am transported.
Cleanup
The scientific understanding developed
through the integrated studies described above clarified the issues surrounding Pu and Am migration
in the RFETS environment. Once Kaiser-Hill, DOE, the EPA, the Colorado state and local governments,
and concerned citizens' groups reached a common appreciation of the technical issues, the different
groups could then reach long-sought agreements on how to proceed with cleanup. Realizing that
Pu and Am existed primarily in particulate forms led to an understanding of their movement at the
site via wind and water. That set the stage for discussing the potential risks to human health and
the environment, possible remediation efforts, specific soil-removal technologies, and ways
to best reconFigure the landscape.
Site operators responded
with a major shift of emphasis to soil erosion and the need to control it. The most poignant illustration
of that shift was a management directive distributed to every employee from Kaiser-Hill president
Nancy Tuor; the directive discussed preventing the dispersal of contaminants during remediation
efforts and reducing the transport of Pu and Am to nearby stream channels or locations off site.
Such measures allowed site remediation to proceed rapidly and thus meet or beat deadlines.
In 1996 the maximum allowable
radionuclide action level was 651 pCi/g. In 2002, armed with improved understanding of
Pu behavior, DOE, the Colorado Department of Public Health, and the EPA released a series of reports
that formed the basis for a new maximum surface-soil action level of 50 pCi/g; that standard
was based on risk analysis and was the result of huge community involvement. Because the Pu contamination
was generally confined to surface soils, the greatest public health risk came from the forces of
wind and water. In actual decontamination, demolition, and remediation, workers therefore set
up large tents at the 903 Pad to insulate work in progress from wind, rainfall, and erosion. The work
focused on removing soil contaminated at the more aggressive standard, down to one meter below
the surface, and replacing it with fresh soil; soil contaminated at depths greater than one meter
was allowed to remain in place, even at higher concentrations. To decontaminate the concrete walls
of buildings, workers used a variety of techniques, including pressure washing of the top layers
to remove the radioactive particles. They then used the clean concrete as backfill around the site.
Operators developed a
storm-water pollution-prevention plan, designed to minimize the erosion, sedimentation, and
runoff of water across the site. Erosion-control measures included straw bales and wattles, straw
crimping, silt fences, mats, hydromulch and crimped synthetic fibers (Flexterra), and riprap
lining of drainage channels. Some new wetland areas were also prepared.
As a result of the cleanup
activities and control measures, surface water and air monitoring stations at the site boundary
have actually shown a decrease in actinide migration. Several of those measures are expected to
work only for a few months to a few years, and will require regular maintenance until the region stabilizes
and the vegetation is reestablished.
A new paradigm?
Superfund sites, such as RFETS, represent
important environmental problems of national significance. So it is important that our best science
is applied to improve the technical basis for decision making.16 A confluence of several
fortunate factors made the RFETS cleanup successful: the willingness of Kaiser-Hill to seek outside
scientific advice; the acceptance, down to the project level, of the value of that advice in avoiding
pitfalls and improving operations; and stakeholders' acceptance, albeit more gradual, of the
independence and veracity of the AME scientific advisers. This willingness and acceptance helped
DOE, the integrating contractor, regulators, and the involved community to focus on specified
goals and objectives.
Establishing particle-transport
mechanisms as the basis of Pu and Am mobility, rather than aqueous sorption–desorption
processes, provided a successful scientific foundation for understanding the scope and nature
of the problem and how best to solve it using erosion control technology. The understanding prompted
contractors to rapidly apply soil-erosion and sediment-transport models. That, in turn, led
to the design and sitewide use of erosion control technology to mitigate the transport of radioactive
particles. Moreover, a scientific understanding of the problems helped define a clearer endpoint
and led to the most extensive cleanup in the history of Superfund legislation. Consequently, the
project finished one year ahead of schedule, saved taxpayers billions of dollars, and removed
an annual liability of more than $600 million from the DOE budget.
We thank Christine
Dayton, Ian Paton, and the Actinide Migration Evaluation advisory group. We are grateful to Kaiser-Hill
Co and the US Department of Energy for their support of AME studies, and thank the Stanford Synchrotron
Radiation Laboratory and DOE's Office of Basic Energy Sciences and Office of Biological and Environmental
Research for their support of actinide science that assisted the cleanup activities at Rocky Flats.
David Clark and David Janecky are technical staff members at Los Alamos National Laboratory in New Mexico. Leonard Lane is a consultant with L. J. Lane Consulting, Inc, in Tucson, Arizona, and was a hydrologist with the US Department of Agriculture's Agricultural Research Service.
2.R. T. Hurr, Hydrology of a Nuclear-Processing Plant Site, Rocky Flats, Jefferson County, Colorado, rep. no. 76-268, US Geological Survey, Denver, CO (1976).
3.M. I. Litaor, G. R. Barth, E. M. Zika, J. Environ. Qual.25, 671 (1996).
4.R. A. Harnish, D. M. McKnight, J. F. Ranville, Particulate, Colloidal, and Dissolved-Phase Associations of Plutonium and Americium in a Water Sample from Well 1587 at the Rocky Flats Plant, Colorado, rep. no. 93-4175, US Geological Survey, Denver, CO (1994).
5.R. Knopp, V. Neck, J. I. Kim, Radiochim. Acta86, 101 (1999).
11.D. J. Shaw, Introduction to Colloid and Surface Chemistry, 4th ed., Butterworth-Heinemann, Boston (1992).
12.W. Stumm, Chemistry of the Solid-Water Interface: Processes at the Mineral-Water and Particle-Water Interface in Natural Systems, Wiley, New York (1992).
13.J. M. Laflen, L. J. Lane, G. R. Foster, J. Soil Water Conserv.46, 34 (1991).
14.US Department of Agriculture, Agricultural Research Service, The WEPP Model, USDA-ARS, Washington, DC (3 March 1996). For more details on the WEPP model, see [LINK].
15.H. E. Canfield et al., Catena61, 273 (2005).
16.National Research Council, New Strategies for America's Watersheds, National Academy Press, Washington, DC (1999).
17.J. C. Myers, Geostatistical Error Management: Quantifying Uncertainty for Environmental Sampling and Mapping, Van Nostrand Reinhold, New York (1997).
Featured Jobs
