Thomas W. Burke and Michael G Hauptmann
Brookhaven National Laboratory

Jim Van Vliet
Idaho National Engineering and Environmental Laboratory

Bhavesh R. Patel
Department of Energy, Brookhaven Group

Elevated levels of tritium, a radioactive isotope of hydrogen, were first detected in December 1996 at concentrations above the 20,000 pCi/L drinking water standards (DWS) in newly installed groundwater monitoring wells located directly down gradient of Brookhaven National Laboratory’s (BNL’s) High Flux Beam Reactor (HFBR). Strong concerns and fears were immediately voiced by local stakeholders over the announcement that radioactive contaminants had entered the sole source aquifer which serves as the drinking water supply for neighboring communities. In response, the Department of Energy (DOE), on February 1997, committed to implement an accelerated interim remedial action to ensure protection of public health and the environment. An immediate aggressive campaign of well drilling, tritium analysis, engineering evaluation, source investigation, numerical modeling, remedial design, construction, and community outreach was undertaken. The accelerated response was performed under the Comprehensive Environmental Response Compensation and Liability Act (CERCLA) process and in accordance with the Interagency Agreement between DOE, U.S. Environmental Protection Agency (US EPA), and New York State Department of Environmental Conservation (NYSDEC).

Horizontal and vertical delineation of the tritium plume was accomplished through the installation of forty five (45) Geoprobes and sixty five (65) hollow stem auger vertical profile wells that were sampled at several discrete depth intervals. The plume, at concentrations above DWS, was determined to extend approximately 2,500 feet south of the HFBR, at depths ranging from 40 to 150 feet below land surface and a maximum tritium concentration of 660,000 pCi/L. In response, a removal action was taken which involved pumping groundwater from the leading edge of the tritium plume and recharging it further north on the BNL property at an existing recharge basin to allow additional time for radiological decay and dispersion prior to reaching the site boundary. Aqueous granulated active carbon (GAC) treatment units were used to treat the groundwater for VOCs prior to discharge. In addition, a comprehensive source investigation determined that the HFBR spent fuel pool (SFP) is the principal source of tritium and has an estimated leak rate of approximately 6 to 9 gallons/day.

Tritium, a radioactive isotope of hydrogen, was detected in groundwater adjacent to the High Flux Beam Reactor (HFBR) facility in late 1996. Tritium was first found south of the HFBR during routine drilling and sampling of a groundwater monitoring well. In response to the discovery, the Department of Energy (DOE) and Brookhaven National Laboratory (BNL) initiated an interim response to ensure protection of public health and environment. Accelerated actions are allowed under the Comprehensive Environmental Response, Compensation and Liability Act (CERCLA) to execute rapid environmental response actions, without first going through the protracted Remedial Investigation/Feasibility Study (RI/FS) process.

An aggressive campaign of well drilling, tritium sampling, engineering evaluations, and numerical modeling was initiated. The resulting information was used to determine the source, the extent of the tritium groundwater plume, and the design of the groundwater interception system to ensure that the HFBR tritium plume does not migrate off of the BNL site above the drinking water standard. The end result of the immediate response activities has been to mitigate any real or perceived impacts from the tritium release.

Data and characterization information gathered during the immediate response action will be incorporated into the ongoing RI/FS of BNL Operable Unit (OU) III. The HFBR spent fuel pool (SFP) tritium plume has been designated area of concern-29 (AOC-29) in OU III and a final remedial action will be determined in the CERCLA RI/FS process.

Elevated levels of tritium in groundwater were first discovered down gradient of the HFBR in monitoring wells. Samples revealed tritium in the groundwater in excess of the Environmental Protection Agency (EPA) drinking water standard [20,000 picocuries per liter (pCi/L)] near the reactor. The highest tritium concentration detected in groundwater (660,400 pCi/L) was from a Geoprobe^TM sample collected just south of the reactor.

The HFBR is a 30 mega watt heavy water moderated and cooled research reactor used principally for basic experimental research requiring external neutron beams. The reactor began operation in October 1965 and has been in service since that time. The reactor produces neutrons to support research in scattering experiments in nuclear and condensed matter physics and structural biology and chemistry.

Typically, the reactor is operated for approximately 30 days, shut down for refueling for 7 days, and then restarted. Because the reactor is heavy water moderated and cooled, tritium is produced by neutron absorption in the coolant and is a by-product of operation. Tritium concentrations in the primary reactor coolant are maintained in the range of 1.5 to 2.7 Ci/L and tritium concentrations in the Spent Fuel Pool (SFP) have ranged from 30 to 140 microcuries per liter (mCi/L) (1 Ci = 10⁶ mCi).

The confinement steel dome structure for the HFBR has the general shape of a half sphere and an outside diameter at the base of nearly 180 feet. Beneath the confinement structure, the reactor is built on a solid concrete mat foundation with steel reinforcement, which has a uniform thickness of 5 feet and a diameter of 140 feet. The HFBR SFP has been determined to be the principal source of tritium and is estimated to be leaking 6 to 9 gallons/day.

The SFP has inside dimensions of 43 by 8 feet and has a 8 by 10 foot bay on the north side. Most of the pool is 20 feet deep except for a 30-foot deep, 8 by 10 foot well on the north side. The concrete walls of the SFP are 3 feet thick.

Identification of the tritium plume source was a major focus of the BNL tritium remediation team. On the basis of a number of factors, BNL reached the conclusion that the source of tritium must be from the HFBR building. The following facts support identification of the HFBR as the principal source of the tritium plume (1) results of the groundwater sampling indicate high tritium concentrations down gradient of the HFBR, (2) low concentrations occur immediately up gradient of the HFBR, (3) contamination is very high near the top of the water table in the immediate vicinity of the reactor, (4) no unusual levels of tritium are detected outside of the groundwater flow path from the reactor, (5) tritium plume concentration data are consistent with a long-term continuous source, and (6) leak tests confirm that the SFP leak rate is 6 to 9 gallons/day [1].

Two separate leak tests have confirmed that the SFP is leaking tritiated water at a rate of 6 to 9 gallons per day [2]. Additionally, historical concentrations of tritium in the pool have been nominally 40 mCi/L and increased to about 140 mCi/L in 1995. The modeling of these concentrations are consistent with the tritium concentrations observed in the plume emanating from the HFBR building.

An analysis of the sanitary sewer system near the HFBR was conducted using the available information and the conclusion was drawn that the sanitary system is not a major source of contamination for the tritium plume. The strongest line of evidence to support this assertion is that peak tritium concentrations measured in the sanitary system have historically been less than measured tritium concentrations in the wells near the HFBR. The highest tritium activity measured in the sanitary system was as high as 200,000 pCi/L, but usually the concentration was less than 60,000 pCi/L. Measured tritium in groundwater immediately down gradient of the HFBR exceeds 200,000 pCi/L, and is over 600,000 pCi/L in at least two samples. The sanitary sewer lines therefore could not be a source for the measured values in excess of 200,000 pCi/L.

The hydrogeologic system underlying BNL is comprised of unconsolidated sediments (sand, gravel, and clay) and older Cretaceous sediments lying upon nearly impermeable crystalline bedrock.

Precipitation, in the form of rain or snow, forms the source of recharge for the Nassau/Suffolk aquifer system. Average yearly precipitation at BNL is approximately 48 inches, of which approximately 23 inches is lost to evapotranspiration, less than an inch flows directly to surface water, and approximately 25 inches infiltrates to the aquifer [3]. As a result of the high rate of infiltration, groundwater seepage to streams from the aquifer accounts for more than 95 percent of the total stream flow throughout the area.

Water table elevation contours and general flow directions within the Upper Glacial aquifer are known. A groundwater divide runs east west down the approximate centerline of Long Island and near the northern boundary of BNL. In general, groundwater flows to the south or south east beneath BNL carrying any dissolved contaminants in that direction. Production well pumping or water discharge to ponds may alter the water table gradient locally. The average groundwater flow velocity at BNL is about 0.8 feet/day.

Delineation of the tritium plume was accomplished through the installation and collection of groundwater samples from 45 Geoprobe^TM and 77 vertical profile boreholes. A total of approximately 1,900 groundwater samples were collected during the immediate response, of which approximately 1,500 were selected for tritium analysis and approximately 400 were archived for possible analysis at a later date. As the immediate response proceeded, available groundwater analytical results were used to continuously update the known extent of the plume and to guide the placement of new wells and the collection of additional groundwater samples. In addition to the Geoprobe^TM and vertical profile boreholes, 27 permanent monitoring wells and piezometers were initially installed to provide long term plume and water level monitoring, and two horizontal wells were installed beneath the HFBR building to aid in tritium source characterization. An additional 50 piezometers and monitoring wells have been installed.

Geoprobe^TM samples were collected by driving a sturdy well screen with a hydraulic percussion hammer to the desired depth below the water table and pumping the temporary well to retrieve the water sample. By retrieving water samples from different depth intervals, several samples could be collected from the same test hole. The Geoprobe^TM is limited to about 100 feet below ground surface. Approximately 230 samples were collected and analyzed from the 45 Geoprobe^TM boreholes.

As sampling and plume delineation progressed, the zone of highest tritium concentration was found to increase with depth below the water table in the direction of groundwater flow. Near the HFBR, the highest concentrations were found within 10 feet of the water table (50 to 60 feet below land surface). Near Brookhaven Avenue, approximately 1000 feet from the HFBR, the highest tritium concentrations were found approximately 35 to 55 feet below the water table (80 to 100 feet below land surface). Use of the Geoprobe^TM sampling system was discontinued where the depth of the plume exceeded the operational limit of the Geoprobe^TM (100 feet).

As an alternative to the Geoprobe^TM sampling method, BNL used vertical profiles to collect deeper groundwater samples in the southern portion of the plume. Vertical profiles are drilled using conventional hollow stem auger drilling techniques. The typical profile consists of 2-inch diameter steel casing with a 2-foot long stainless steel screen. Once drilled to the desired depth, the well is sampled, then pulled upward through the aquifer at either 5 or 10-foot intervals. Thus, in a manner similarly to the Geoprobe^TM system, a vertical profile is obtained. As many as 25 depth-discrete samples were collected from a single vertical profile.

A total of 77 vertical profiles have been completed from which approximately 1,700 samples were collected and 1,280 submitted for tritium analysis. The locations of vertical profiles are shown on Figure 1.

Twenty-seven new piezometers and monitoring wells were installed to support groundwater-modeling efforts. An additional 50 permanent monitoring wells and piezometers will be installed. The piezometers and monitoring wells will be drilled and installed using hollow stem augers. Piezometers consist of a 2-inch diameter riser pipe with, typically, a 5-foot well screen placed at a discrete monitoring depth. The monitoring wells used for the collection of water samples and water level elevations are typically completed with 4-inch diameter casing and a five-foot screen at the intended monitoring depth. In addition to the new wells, previously installed wells are available for monitoring the nature and extent of the tritium plume. Both types of wells will be used for long-term monitoring of groundwater levels, collection of groundwater samples, and to evaluate the effectiveness of the tritium plume remediation.

The combined results of the Geoprobe^TM, vertical profile, and monitoring wells sampling were used to define the horizontal and vertical extent of the tritium plume. The plume is located entirely within the boundaries of BNL. The portion of the plume which exceeds the drinking water standard for tritium (20,000 pCi/L) extends approximately 2,500 feet south of HFBR at depths ranging from 40 to 150 feet below land surface. Tritium concentration in the plume ranges from a highest detected concentration of 660,400 pCi/L immediately south of HFBR to approximately 6,440 pCi/L at a point 3,585 feet south of the HFBR. As is evident on the vertical profile of the plume, concentrations increase in depth below the water table down gradient of the HFBR. This is due to downward vertical groundwater flow within the aquifer from precipitation recharge. The decrease in concentration further from HFBR is due to natural radioactive decay, which decreases the tritium concentration in half every 12.3 years, and dispersion within the aquifer.

Considering groundwater flow, radioactive decay, dispersion, and groundwater modeling, it is estimated that tritium concentrations above 20,000 pCi/L may never cross the BNL boundary from the HFBR tritium plume. This will be verified by on going monitoring.

The tritium plume, source, and rate of migration have been investigated using techniques that range from simple mass balance calculations to analytical models and more complex numerical models. Despite the varying assumptions inherent in the different approaches, all of these techniques have provided basically similar results. These analyses, in combination with other activities such as characterization, investigation of potential sources, and the SFP leak test, confirm that the HFBR SFP is the principal source of the tritium plume.

The main conclusions are: (1) the HFBR SFP is the principal source of the tritium plume, (2) the plume is in, or close to, equilibrium, (i.e., the plume is not migrating significantly), (3) drinking water standards for tritium will never be exceeded at the site boundary from the HFBR, and (4) pumping from three wells at Princeton Avenue at 40 gpm each, as part of the immediate response action, will not adversely affect the highly contaminated portions of the tritium plume.

Three paths for remediation were chosen: (1) removal of spent fuel from the pool and installation of a stainless steel liner, (2) eliminating other potential sources of leakage by bringing the HFBR into compliance with Suffolk County Department of Health Services Article 12, and (3) groundwater pumping at the leading edge of the plume. In response to other groundwater plumes, BNL had previously installed potable public water to the residences and businesses down gradient of the site.

The spent fuel elements have been removed from the SFP and shipped to the Savannah River Site for storage and final disposition. The equipment in the pool (such as control rod blades, fuel storage racks, strike-plate, spent fuel retard chute, etc.) have been transferred for storage/disposal. Water from the pool has been pumped to storage tanks via double walled piping.

To eliminate the pool as the source of actual, potential, or future tritium contamination in the groundwater, an impervious liner, with leak detection and collection capability, is to be added to the SFP.

The BNL has designed, constructed, tested, and put into operation an interim system to intercept the tritium plume. The system uses a pump-and recharge system to ensure that tritium above the U.S. EPA drinking water standard of 20,000 pCi/L will not leave the Brookhaven site. The interim groundwater extraction system provides a level of redundancy because the current understanding of the tritium plume and groundwater flow suggests that tritium greater than the drinking water standard will never cross the BNL boundary from the HFBR tritium plume.

Three groundwater extraction wells have been installed approximately 3,500 feet south of the HFBR near Princeton Avenue, where the maximum tritium concentration is 6,500 pCi/L. Groundwater is being pumped from a depth of about 150 feet below land surface and piped 3,300 feet northward to an existing recharge basin within the BNL site. Each well pumps tritiated groundwater from the aquifer at a rate of about 40 gallons per minute.

Prior to entering the recharge basin, the water passes through activated carbon filters to remove chemical contamination that is also present in groundwater in the area due to other past BNL activities (Figure 2). The carbon vessel, based on the current total VOC concentrations in the extracted water, is projected to last at least twelve months before change out is required. Discharge water samples are collected and analyzed weekly as per New York State Department of Environmental Conservation permits. When the water enters the infiltration basin, the maximum tritium concentration is expected to be approximately 2,500 pCi/L, well below the drinking water standard. Samples are analyzed on a regular basis to determine the tritium concentrations being recharged. Evaporation of tritiated water from the infiltration basin has been calculated and shown not to pose a risk to human health or the environment. Air monitoring stations are measuring tritium concentrations in air on a regular basis.

Once the water reenters the ground, it will flow southward taking approximately 19 years to reach the BNL site boundary. By that time, natural decay and dilution will have diminished tritium levels to nearly undetectable levels. Monitoring wells located at the Laboratory boundary will provide further insurance that tritium, above the drinking water standards, will not leave the BNL site.

The pump-and-recharge remediation is being conducted as an immediate action to ensure that tritium above the drinking water standards does not migrate across the BNL boundary. It also gives BNL and DOE time to study alternative remediation technologies and prepare a plan, if necessary, to address the high levels of tritium found immediately south of the HFBR.

A pumping test of the system was completed on May 9, 1997 and the system went on line on May 12, 1997. The long-term remediation of the plume will be determined in the OU III RI/FS as AOC-29.

The following remedial activities are:

1. Implement a comprehensive groundwater plume monitoring plan and install additional groundwater monitoring wells as needed for long-term monitoring.
2. Perform a detailed analysis of tritium transport from the source under transient conditions of recharge and pumping.
3. The Feasibility Study will determine if any additional actions are required (i.e., additional extraction wells and/or source containment).
4. On an ongoing basis, re-contour the tritium plume using the most recent tritium data and evaluate implications on the pump-and-recharge system.
5. Continue as planned to construct and install a new pool liner for the SFP and implement other upgrades.

Over the past half year an aggressive technical effort by BNL has mitigated a potential or perceived health risk from the tritium recently discovered in groundwater south of the HFBR. Steps are underway to prevent additional releases of tritium from the reactor. A groundwater extraction-and-recharge system has been put into operation, which will limit further migration of the tritium contamination until a permanent remedy for the tritium plume is established.

Numerical modeling of the HFBR tritium plume based upon the detailed hydrogeological and contaminant distribution information collected to date suggests that, even without the extraction-and-recharge system, tritium levels above the drinking water standard will never cross the BNL boundary. Now that the major short-term objectives of the Tritium Remediation Project have been achieved, the restoration efforts have been returned to the standard CERCLA process for completion.

1. Idaho National Engineering and Environmental Laboratory, "High Flux Beam Reactor Tritium Source Identification Volume I, Technical Report," Brookhaven National Laboratory, Upton, New York, July, 1997.
2. D. PORTS, "High Flux Beam Reactor Fuel Canal Leak Test Report, Leak Test #2", Brookhaven National Laboratory, Upton, New York, March, 1997
3. Geraghty & Miller, "Regional Groundwater Model," Brookhaven National Laboratory, Upton, New York, November 1996.

BACK