N-Nitrosodimethylamine (NDMA) is present in groundwater and drinking water from industrial, agricultural, water treatment, and military/aerospace sources. NDMA is a suspected human carcinogen and an emerging groundwater contaminant that has been detected at a number of Department of Defense (DoD) and National Aeronautics and Space Administration (NASA) sites involved in the production, testing, and/or disposal of liquid propellants containing unsymmetrical dimethylhydrazine (UDMH).  NDMA was a common contaminant in UDMH-containing fuels (e.g., Aerozine-50) and is also produced when these fuels enter the environment through natural oxidation processes.  Currently, the most effective treatment technology for NDMA in groundwater is pump-and-treat with ultraviolet irradiation (UV).  However, this approach is expensive because it requires high energy input to effectively reduce the levels of NDMA to meet regulatory requirements.  The objective of this ESTCP project was to demonstrate and validate the application of in situ propane biosparging for treatment of NDMA in groundwater. 

Previous laboratory studies revealed that natural propanotrophs in many aquifers are capable of biodegrading NDMA to low ng/L concentrations while growing aerobically on propane.  During this in situ demonstration, propane gas and oxygen were added to groundwater via three biosparging wells to stimulate this process. The demonstration was performed at the Aerojet Superfund Site (Aerojet) in Rancho Cordova, CA in a location downgradient of a site where liquid rocket engines were developed and tested. The groundwater in this area has NDMA concentrations ranging from ~ 2,000 to > 30,000 ng/L.  Currently, the groundwater in this region is captured by a groundwater extraction & treatment (GET) system and NDMA is removed by ultraviolet irradiation. 

To evaluate effectiveness of biosparging, NDMA concentrations in groundwater were monitored in a series of performance monitoring wells (PMWs) placed within a Test Plot Area (TPA), three of which (PMW-2, PMW-3, PMW-4) were within or slightly downgradient of the expected zone of influence of three biosparge wells (BW-6, BW-7, PMW-1).  It should be noted that PMW-1 was used as both a biosparge well and a performance monitoring well throughout the demonstration. Monitoring wells PMW-5 and PMW-6 were downgradient of the plot and expected to be influenced later in the demonstration, as treated water reached this region. Well BMW-1, which was side-gradient (~ 75 ft west of the center of the biosparge zone) was used as a control well to monitor NDMA concentrations outside of the treatment zone.  The biosparging system was operated for a period of 374 days from start-up to shut-down. Full rounds of groundwater sampling were conducted on 12 occasions. This included two baseline sampling rounds on Day -84 and -70, nine performance sampling events during active sparging (Days 42, 84, 161, 185, 213, 241, 287, 311, and 353) and two rebound events after biosparging ceased (Day 385 and 430). The variables that were adjusted and optimized throughout the demonstration included (1) the percentage of propane in the air-propane feed; (2) the length of sparging cycles; (3) the number of sparging cycles per day; and (4) the breakdown of the sparge cycle, which was composed of an initial air sparge, and period of combined air-propane sparging, and then a final air sparge to clear the sparge lines of propane gas.  When the system was optimized, the percent propane in the sparge gas set at 40% of the LEL, (which equated to ~ 0.84% propane in the feed gas) and the system was operated for 12 cycles per day with propane being added for 40 minutes during each cycle.  The amount of propane added to the TPA after optimization was ~ 1.83 lbs/day, and a total of approximately 475 lbs of propane was injected throughout the demonstration. 

The biosparging approach was highly effective for the removal of NDMA from the aquifer.  From baseline sampling (average concentrations from Day -70 and Day -84) to the final day of sampling during active biosparging (Day 353), concentrations of NDMA declined by 99.7 % to > 99.9 % in the four PMWs within the zone of influence of the biosparge system (PMW-1 to PMW-4).  Baseline concentrations of NDMA, which averaged 25,000 ± 6000 ng/L (7 test plot monitoring wells, two baseline events) declined to between 2.7 and 72 ng/L by Day 353 (mean value 40 ± 30 ng/L). The NDMA concentration at well PMW-2 was below 3 ng/L on Day 353.  By comparison, the NDMA concentration in the side-gradient control well (BMW-1) averaged 36,000 ng/L during baseline sampling and was 31,000 ng/L on Day 353, a decline of only 14 %. Concentrations of NDMA in the far downgradient wells PMW-5 and PMW-6 began to show measurable declines near the end of the demonstration, presumably as treated water from the biosparge plot began to reach this region of the aquifer.  NDMA in PMW-5 declined to 5,400 ng/L on Day 430 (from an initial average of 26,000 ng/L) and NDMA in PMW-6 fell to 13,000 ng/L on Day 430 (from an initial average of 22,500 ng/L). The rate of NDMA biodegradation in the TPA was calculated in wells PMW-2, PMW-3 and PMW-4. First-order rate constants were determined using data from Day 84 to Day 353.  The degradation rates were 0.019 day -1 for PMW-3 (R2 = 0.95), 0.031 day -1 for PMW-4 (R2 = 0.82) and 0.037 day -1 for PMW-2 (R2 = 0.68).  These rates equate to NDMA half-lives ranging from 19 to 36 days. 

This biosparging technology is ready for full scale application. The expected cost drivers for installation and operation of a full-scale propane biosparging delivery system for the remediation of NDMA-contaminated groundwater, and those that will determine the cost/selection of this technology over other options include the following:

•       Depth of the plume below ground surface;

•       Width, length, and thickness of the plume;

•       Aquifer lithology and the presence or absence of impervious layers that would impede sparging;

•       Regulatory/acceptance of alternatives to sparging that include groundwater extraction and re-injection;

•       Length of time for clean-up (e.g., necessity for accelerated clean-up);

•       The presence of indigenous propanotrophic bacteria capable of degrading NDMA;

•       Presence of co-contaminants such as chloroform, chlorinated ethenes, and chlorinated ethanes;

•       The radius of influence that can be achieved via sparging; and

•       O&M costs.

Based on a cost analysis for treatment of a shallow groundwater plume (~ 10 – 40 ft bgs) of ~ 400 ft in width, a propane biosparge barrier was determined to be the most cost effective option compared to current alternatives, which included pump-and-treat with either ultraviolet (UV) or biological (via  fluidized bed bioreactor) removal of NDMA.  Under this scenario, and assuming a 30 year operational period with equivalent costs for groundwater monitoring, the in situ barrier approach was more than 40% less expensive than either of the ex situ alternatives. The primary cost difference between the alternatives was the high capital cost of building an ex situ water conveyance and treatment facility, which is required for the UV or FBR system, but not for the in situ biosparge barrier. The capital costs for the ex situ options were ~ 3 times those for the in situ biobarrier.  

In summary, the data from this ESTCP field test clearly indicate that propane biosparging can be an effective approach to reduce the concentrations of NDMA in a groundwater aquifer by 3 to 4 orders of magnitude, and that concentrations in the low ng/L range can be achieved with continuous treatment. These results are consistent with data achieved in pure culture studies as well as with various bioreactor tests.  Moreover, for many applications, a propane biosparging system is expected to be significantly less expensive to install and operate than a conventional pump-and-treat system for NDMA removal from groundwater.