Autonomous Screening of Groundwater Remediation Technologies in the Subsurface Using the In Situ Microcosm Array (ISMA)

Abstract

The groundwater remediation industry continues to progress towards less expensive, more sustainable in situ remedies. However, in situ treatment requires site-specific performance data that can be difficult or impossible to obtain using conventional laboratory microcosm studies. To improve the representativeness of laboratory scale treatability studies, and aid in remedial technology implementation, we developed the In Situ Microcosm Array (ISMA). This autonomous diagnostic device enables the deployment of 10 flow-through sediment columns within a standard 10-cm groundwatermonitoring well. Suspended at the desired aquifer depth, the fully encapsulated ISMA meters groundwater directly from the aquifer to microcosms containing competing remedial technologies. Field demonstrations of the instrument were performed in two aquifers contaminated, respectively, with trichloroethene and hexavalent chromium, and with perchlorate. A cost assessment positions ISMA deployment costs within the range of conventional laboratory treatability studies. Results demonstrate the ISMA’s utility to perform cost-effective, high-throughput, screenings of multiple intervention strategies in the field, without impacting the subsurface environment examined.

Graphical abstract

1. Introduction

The U.S. Environmental Protection Agency (U.S. EPA) has forecasted that remediation of an estimated 294,000 contaminated sites in the U.S. will cost $200B over the next 30 years [1]. Since 1986, treatment selection, on-site containment, and off-site disposal have remained relatively static [2]. Recently there has been momentum to depart from containment strategies, and towards technologies that will help to expedite cleanup goals. The most promising technologies include those applied in situ, which include the application of chemical or biological reagents to the subsurface to drive contaminant transformation and/or sequestration [2, 3]. The U.S. government supports this objective through the Environmental Security Technology Certification Program (ESTCP), a demonstration and validation initiative designed to promote commercial translation of innovative environmental technologies [4].

Although in situ technology applications receive support, there is a higher implementation risk than ex situ remediation strategies, which is largely related to site specificity. Consequently, rigorous treatability studies are necessary to determine if in situ remediation is feasible to meet cleanup goals at any given location [5]. Treatability studies commonly employ laboratory batch bottle microcosms, laboratory flow-through columns, and/or field pilot-scale studies [6]. Both laboratory batch bottle and column microcosms test potential remediation strategies with site sediment and site groundwater, however, batch microcosms are favored by most guidance documents because of design simplicity and low implementation costs [7-11]. Laboratory flow-through sediment columns are more costly and consequently less common, however these are ideal to study transport phenomena that are key to the design and implementation of in situ remedial strategies [12]. Pilot field-scale implementation is the most expensive treatability testing variant, and there is an inherent risk due to the potential for unforeseen, adverse environmental impacts. Consequently, pilot-scale testing generally occurs after lab-scale treatability studies have been completed, thereby delaying site remediation progress.

Environmental remediation practitioners have acknowledged that results generated from laboratory studies are “quantitatively, even qualitatively, different from the same determination if it could be done in situ” [13]. The shortcomings of laboratory microcosm studies have been recognized by many and resulted in calls for the development of alternatives [9, 14]. By removing site groundwater from the subsurface environment, changes in chemical and microbial parameters are introduced, such as the out-gassing of carbon dioxide, volatilization of organic compounds, precipitation of metals and salts, or the deactivation/removal of microorganisms. Changes of this kind distort a legitimate extrapolation of results obtained from laboratory treatability studies [15, 16].

To improve the representativeness of lab-scale remedial studies and overcome the limitations to implementation of in situ remediation technologies, new approaches are needed [17]. We sought to address this need by creating autonomous instrumentation designed to conduct multiple flow-through treatability studies within a groundwater monitoring well. This manuscript describes the design and capabilities of the In Situ Microcosm Array (ISMA), and includes data from two field deployments.

2. Experimental section

2.1. Development of the In Situ Microcosm Array (ISMA)

A submersible, fully encapsulated flow-through column treatability system was designed and manufactured using a combination of commercially available components and custom fabrication. Custom component design was performed using SolidWorks 3D design software (Dassault, Systèmes, Waltham, MA). To fit within the constraints of common 10-cm inner diameter (ID) groundwater wells, many components of a standard laboratory column study needed to be miniaturized. A modified multi-channel peristaltic pump was devised to meter groundwater from the well, through the fluid train, and into sediment-filled, flow-through columns. An optional amendment injection unit was included for time-release amendment delivery, which mimics subsurface field injections. To preserve monitoring well water quality, the ISMA design included effluent capture inside the unit. Component materials, including gas and fluid conduits, were selected for compatibility with a variety of chemical contaminants, and designed modularly for quick assembly, reliable functionality, and rapid decontamination. The instrument shell was designed as separate tubular stainless-steel sections, for easy in transport, and connection in the field. Section connectors were designed to be load bearing, waterproof, and transmit all necessary fluid and electrical lines. To offer versatility, power requirements include alternating or direct current (AC, DC), with the option to couple to solar panels for long-term deployment in remote locations.

2.2. Trichloroethylene, hexavalent chromium demonstration

2.2.1. Site description

The first field demonstration site was an active military instillation, Naval Air Station North Island (NASNI), in San Diego County, California. The facility has legacy trichloroethylene (TCE) and hexavalent chromium [Cr(VI)] plumes (ppm) related to the manufacture and repair of fiberglass components, and a metal plating shop. A TCE groundwater plume, with several contributing sources, extends downgradient (northeast) approximately 800 m, while the Cr(VI) plume originates at Building 2 and extends downgradient approximately 215 m (Figure 1) [18, 19]. Groundwater is part of the Coronado hydrologic basin (Unit No. 10.10), which is a subunit of the Otay Hydrologic Unit. A lens-shaped body of freshwater (variable thickness) sits atop seawater; depth to water ranges from 1 to 8 m below ground surface (bgs). Aquifer transmissivity ranges from 0.05 to 100 m² per min across the entire site, with an approximate value of 39 m² min⁻¹ adjacent to the deployment location [20]. Sediments underlain the study site are predominantly hydraulic fill, consisting of medium-grained to coarse-grained, poorly graded sands and silty sands, with native organic silts and clays at depth. Hydraulic conductivity averages 9.4 ± 5.2 m d⁻¹, calculated from a series of 16 slug tests, and the hydraulic gradient is 0.001 m/m [20].

Figure 1.

Plume isoconcentrations and deployment locations at the trichloroethylene/hexavalent chromium (A) and the perchlorate sites (B). Images modified from Kalinowski et al., 2013 and McClellan et al., 2013.

2.2.2. Column preparation

Sediment materials from multiple drilling events were collected and composited, and used for column construction. Sediments were air-dried, sieved, and grain sizes ranging from 1000 to 250 μm in diameter were used in column construction; resultant porosities were on average 40%. Flow rates in the columns were 16.6 μL/min, for a resultant annual linear velocity 0.43 to 0.58 m d⁻¹. In this demonstration, three remedial strategies were screened in triplicate columns including: i) natural attenuation (site sediment, no amendments), ii) biostimulation with the addition of 10% w/v sodium lactate (supplemental carbon source, electron donor), and iii) bioaugmentation with the addition of sodium lactate (10% w/v) and the commercially available dechlorinating culture KB-1^® (SiREM, Guelph, ON). Sodium lactate was injected by a secondary pump at a flow rate of 0.23 μl/min (50 μM concentration at columns influent). Columns were amended with 3 mL of KB-1^® at the base of each appropriate treatment column before instrument deployment. The remaining three fluid lines on the pump served as influent controls, where groundwater was captured but not passed through a sediment column.

2.2.3. Analytical Methods

Column effluent samples were analyzed for chlorinated ethenes, including TCE, cis−1,2-dichloroethene (cDCE), vinyl chloride (VC), and ethene by EPA method 8260B (gas chromotography mass spectrometry). Chromium was analyzed following EPA method 7196A (colorimetric), and inorganic ions, including chloride, were analyzed following EPA method 300.0 (ion chromotography). All analyses were completed by EMAX Laboratories (Torrance, CA).

2.3. Perchlorate demonstration

2.3.1. Site description

This field demonstration of the instrument was conducted at a small explosives-manufacturing facility located in the southwestern U.S. Legacy disposal practices at the site (since the 1960s) released ammonium perchlorate into sediment and groundwater, resulting in groundwater concentration maximums of 1000 μg/L (Figure 1). The groundwater table is approximately 53 m below ground surface with flow to the southeast at approximately 0.5 m/day (hydraulic gradient 0.005 m/m). Sediment in the area is characterized by low organic carbon content and mostly consists of silty sands and gravels, poorly-and well-graded sands, clayey sands and clayey gravels.

2.3.2. Column Preparation

Columns were packed with sieved sediment (0.5 - 1 mm grain size) obtained from drill cuttings as a product of well installation activities at site HPA-1. The mean effective porosity was 35%, and columns were operated at an effective flow rate of 15 μL/min, The following remedial strategies were tested in the ISMA in four replicates including natural attenuation (NA), and bioaugmentation with a seed culture containing perchlorate-reducing bacteria and the addition of sodium acetate. The seed culture, a facultative anaerobic microbial consortium, was obtained by enriching a sewage sludge composite sample from five different U.S. wastewater treatment plants. The bioaugmentation treatment columns received 1 mL of seed culture at the beginning of the experiment by injection at the influent port of each column. Sodium acetate was supplied to the columns during the entire deployment, at 1100 mg/L (8.1 mM) influent concentration. The three remaining fluid lines on the pump where used as influent controls.

2.3.3. Analytical Methods

Perchlorate was quantified using a Dionex ICS-2000 ion chromatography system with conductivity detector (Thermo Fisher Scientific, Sunnyvale, CA, USA) following EPA method 300.0; Anions were analyzed following EPA method 314.0. Details of the analytical methods have been published previously [21]. DNA was extracted from sediment using a PowerSoil DNA extraction kit (MoBio Laboratories, Inc., Carlsbad, CA) in combination with the DNeasy Blood and Tissue kit (Qiagen Inc., Valencia, CA). DNA was also extracted from the composite column effluent using the UltraClean Water DNA kit (MoBio Laboratories, Inc., Carlsbad, CA), according to the manufacturers’ protocol. Characterization was completed by quantitative polymerase chain reaction (qPCR) with a Mastercycler® ep realplex instrument (Eppendorf, Hamburg, Germany). The DNA target was the perchlorate reductase gene (pcrA), a proxy of perchlorate reducing bacteria in each sample [22, 23]. Plasmids containing the target DNA fragments used here were the same as those developed for previous studies [24].

2.4. Cost Assessment

A cost assessment was performed around the ISMA deployment at NASNI in San Diego. Materials costs were directly transcribed from the deployment as they were incurred. Energy costs were estimated based on maximum pump wattage and electricity costs for that given state [25]. Sample analysis costs reflect actual billing by the selected commercial analytical laboratory. Labor costs were calculated by estimating the effort required from personnel to conduct the deployment, in tandem with the results from discussions with remediation professionals. Salary estimates were collected from the Bureau of Labor Statistics [26], and travel estimates were based on actual incurred costs. Facility and administrative charges were estimated from pricing of commercial laboratory space rentals. Expenditures related to traditional laboratory treatability studies were based on actual or estimated costs from commercial laboratory partners. Pilot field-scale activities were transcribed from costs incurred from a previously executed field study by environmental professionals.

3. Results and discussion

In this study, a flow-through column treatability system designed for field deployment was fabricated and utilized to test prospective remediation strategies at two field locations. The goal was to develop a technology that creates a fast-track from lab-to-field, thereby enabling more candidate in situ remediation technologies to be evaluated in the field at a low operational cost and a low risk of causing adverse effects from testing. The ISMA technology was demonstrated at two hazardous waste sites, one containing TCE and Cr(VI), and the second, perchlorate.

3.1. Embodiment of the ISMA

The ISMA was designed to fit within a common U.S. groundwater well (10-cm) (Figure 2). The instrument’s exterior shells (ASTM A312) were 2.5 m in length with an 8.9-cm outer diameter (OD) (Eagle Stainless, Warminster, PA). An off-the-shelf pump, Ismatec MiniClick 6 Reglo-E peristaltic pump unit (IDEX Health & Science, LLC; Oak Harbor, Washington, USA), was modified to fit shell dimensions. Custom modifications included re-design of the motor mounting plate and addition of Ultem® (polyetherimide) cassettes for securing peristaltic pump tubing. A Reglo-E Digital control unit by the same manufacturer provided pump control, with adjustable flow rates to achieve desired column residence times. Sediment microcosm vessels consisted of custom glass columns (250 mm length, 14 mm ID) (Chemglass Life Sciences, Vineland, NJ) with Teflon^® screw caps and Viton^® O-rings (Cole Parmer, Vernon, Hills, IL) for a waterproof seal. Up to ten pre-packed columns are arrangeable on custom designed carousels inside the unit. The instrument fluid train consisted of Viton® tubing, and Teflon® fittings (Cole Parmer, Vernon, Hills, IL) for connections between components, and the groundwater intake was screened at 100 μm pore size. The treatment delivery system functioned independently of peristaltic pumps, and was comprised of a Silverpak 17C integrated motor, driver and controller (Lin Engineering, Morgan Hill, CA) designed to actuate (5) 10 mL syringes (Becton Dickinson [BD], Franklin Lakes, NJ) per unit. The treatment flow rate was controlled through a software interface developed in LabVIEW (National Instruments, Austin, TX). Power was provided by a custom 24-V DC system. Effluent collection occurred in collapsible Teflon® fluid capture vessels, containing microbial preservative (Kathon®, [5-chloro-2-methyl-4-isothiazolin-3-one, and 2-methyl-4-isothiazolin-3-one]. Gases produced by chemical or microbial activity were vented to sample capture lines, and volatiles of interested collected with an activated carbon cartridge (Anasorb CSC, SKC Inc., Fullerton, CA), when applicable. Custom instrument shell joints and end caps were fabricated from stainless steel. The top cap contained flow-through ports for power and gas vent lines, and was equipped with a welded eyebolt for instrument suspension by steel cabling. Images from multiple deployments are shown in Figure 3.

Figure 2.

Schematic of ISMA components and flow paths through the instrument. Groundwater from the contaminated aquifer is pulled directly into the ISMA unit through the intake, by onboard multichannel peristaltic pumps. Groundwater flow is divided into multiple fluid channels via manifold, and pumped through columns filled with sitesediment and selected amendments for contaminant remediation. After groundwater exits the columns, it is collected in effluent capture bags inside the unit for subsequent transport to the laboratory for further analysis. An amendment injection module is an optional, if slow-release amendment addition to the columns is desired.

Figure 3.

Images from ISMA field deployments. ISMA removal from transport container (A), Individual ISMA shell lengths containing various instrumentation (B), Treatment injection module and sediment columns (C), ISMA segment is lowered by boom truck into a well (D), Two shell lengths to be connected on the well box (E), Deck box containing instrument control box, batteries, inverter, and cable spool (F), ISMA fully deployed in well, suspended on well casing (vent line and power cables exiting instrument) (G), Transfer of column effluent to Volatile Organic Analysis (VOA) vials for analysis (H), Effluent capture vessels and carousel attachment (I).

3.2. ISMA demonstrations

3.2.1. In situ TCE/Cr(VI) treatability study

Grab samples of groundwater from the deployment well were collected immediately before instrument deployment; concentrations of TCE and Cr(VI) were 10 and 24 mg/L, respectively. Dissolved oxygen (DO) was 10 mg/L and oxidation-reduction potential 83.1 mV. The ISMA was deployed for 35 days, at a depth of 8 meters below ground surface, approximately 5 meters below the water table. Columns ran in upflow mode at a rate of 16.6 μL/min.

Results showed significantly reduced amounts of TCE (homoscedastic 2-tailed student t-test, p < 0.05), and elevated levels of TCE byproducts including cDCE (p = 0.08), VC (p < 0.05), and ethene in bioaugmented sediment microcosm effluent, as compared to natural attenuation (biotic control) (Figure 4). These results provided evidence of successful conversion of aerobic site groundwater (DO 10 mg/L) to anaerobic conditions, which facilitated the reductive dehalogenation of TCE by the strict anaerobic bacteria culture added to the sediment. The application of in situ biological reduction of TCE necessitates treatability studies because the stepwise reduction of TCE to ethene may exhibit stalling, commonly at cDCE and VC [27]. Problematic is VC accumulation because the compound is flammable, volatile, carcinogenic, and toxic. Consequently, it is important to demonstrate successful dechlorination beyond VC to convince stakeholders that in situ bioremediation is suitable for site remediation. This is particularly important when co-contaminant mixtures are present, which often create unforeseen negative circumstances [28].

Figure 4.

Subsurface chemistry captured in the ISMA. Column containing materials from the subsurface formation (A), Chromium and TCE reduction reactions observed at Site 1 (NASNI) (B), Contaminant concentrations in composite effluent collected over the duration of ISMA deployment, comparing three remediation technologies: NA (natural attenuation), BS (biostimulation), and BA (bioaugmentation). Error bars represent standard error (n = 3) (C), Perchlorate reduction reaction observed at Site 2 (E), Concentrations of perchlorate, and perchlorate reductase (pcrA) in composite effluent and column sediment; INF (influent control) (F).

Although the bioaugmentation columns out-performed the natural attenuation and biostimulation treatment, TCE remained above the maximum contaminant level of 5 μg/L (0.04 μM), as did the chlorinated byproducts cDCE (70 μg/L, 0.72 μM) and VC (2 μg/L 0.03 μM) [29]. TCE reduction did not occur in the biostimulation columns, suggesting the addition of a dechlorinating culture was necessary for reduction at this site. This result is not uncommon to chlorinated solvent contaminated environments [30].

A notable secondary outcome of this study is the reductive dechlorination of TCE in the presence of high concentrations of Cr(VI) (>5 mg/L). It is commonly believed that Cr(VI) concentrations of >5 mg/L necessitate treatment with injection of a chemical reducing agents to treat Cr(VI) first, before biological reductive dechlorination of TCE to ethene is possible [31, 32]. This is largely due to Cr(VI) toxicity. The observed biological removal of TCE in the presence of 24 mg/L of Cr(VI) here, extends the reported spectrum of conditions conducive for reductive dechlorination of chloroethenes via bioaugmentation. Biostimulation and bioaugmentation columns showed an approximate 20% reduction in Cr(VI) concentrations (as compared to natural attenuation), suggesting that stimulation with sodium lactate facilitates chromium reduction. Cr(VI) has been shown to be reduced under reducing conditions [33], and the results here are also consistent with the site-specific laboratory treatability studies (not shown). Since the ISMA has the ability to run different treatment columns in replicate, under the same environment conditions (spatially and temporally contiguous), this reduces inadvertent variability between the treatments, and in combination with replicates, provides a more reproducibly and comparable assessment of remedial success. Additionally, with flowthrough columns, it is also possible to determine amendment lifetimes by tracking contaminant rebound in column effluent. These amendment lifetimes provide an additional metric for estimating cleanup times [34].

3.2.2. In situ perchlorate treatability study

Groundwater grab samples collected in the target well immediately prior to instrument deployment showed perchlorate concentrations at 228 μg/L. Dissolved oxygen and nitrate (competing electron acceptors) were 0.1 - 4.5 mg/L and 12 mg/L, respectively. The ISMA was deployed for 21 days at a depth of 58 m below ground surface, with column flow rates of 15 μL/min. Results show perchlorate was reduced from 228 ± 1 μg/L to 30 ± 37 μg/L in the columns supplied with the bioaugmentation treatment, although still above the preliminary remedial goal of 15 μg/L [35] (Figure 4).

DNA analysis of column effluent and sediment showed that bioaugmentation with nutrient addition led to an increase in gene copy numbers of perchlorate reductase (perA) (690-fold on average), indicators for perchlorate-reducing bacteria. Results further showed that perchlorate reducers mainly settled onto column sediment (concentration 2 - 3 orders of magnitude higher in sediment [copies/g] than in aqueous phase [copies /mL]). By its design, the ISMA allows analysis of microbial communities in column effluent and sediment and examination of their spatial distribution across the columns. The column sediment was sectioned in three equal sections (inlet, middle, and outlet) and DNA copy numbers were analyzed. Results show that the majority of bacteria in all columns resided in the inlet portion of the sediment columns (77 ± 20%). This was even more pronounced for the columns that were bioaugmented, where around 90 ± 10% were found in the inlet portion of the columns. The reasons for this likely are two-fold: in MNA columns to which no nutrients were added, different sediment filtration mechanisms [36, 37] straining the bacteria from the incoming groundwater most likely caused the high DNA copy numbers found near the inlet. In addition to sediment filtration, nutrient concentrations (carbon source and electron acceptors) in bioaugmented columns are highest at the inlet of the columns, and therefore provide ideal growth conditions for bacteria leading to higher numbers near the inlet. This has been found in several flow-through column studies [38-41]. Sampling of both the aqueous and sorbed fraction of contaminants has been recognized as essential to provide a complete picture of the microbial community [42, 43].

3.3. Cost Assessment

Costs associated with the NASNI (TCE/Cr[VI]) deployment are shown in Table 1. Illustrated are actual incurred costs and those with an included markup of 3x for labor, 2x for consumables and 3rd party equipment rentals, and an ISMA rental fee of $100/day. This was done to provide a more realistic cost of the ISMA technology if licensed to environmental consulting firms for use. It also aids in the analysis when comparing to commercially available treatability studies that have these costs embedded.

Table 1.

ISMA deployment costs. Present costs include expenses incurred during the NASNI study. A cost markup column was added for a realistic comparison to traditional treatability studies. Profit markups included: 3x for labor, 2x for consumables and equipment rental, and an ISMA leasing fee of $100/day.

Cost Element	Cost	Cost + Markup
Labor costs	$25,326	$75,978
Consumable and equipment costs	$3,749	$10,998
Laboratory analysis	$6,790	$6,790
Travel	$4,000	$4,000
Facility and administrative costs	$3,500	$3,500
Total	$43,365	$101,266

The primary cost driver for an ISMA deployment was labor costs. Labor costs were estimated based on the recorded efforts expended by personnel, and tasks were differentiated into four distinct skill sets and levels of expertise (Table S1). These included: i) Project Manager, ii) Technical Advisor (oversee selection/dosing of chemical and biological amendments), iii) Environmental Engineer (coordinate/oversee field activities), and iv) ISMA Technician (constructs and deploys the ISMA). Labor costs can be classified as either laboratory- or field-related. Laboratory labor includes column microcosm assembly (sediment processing [drying, homogenizing, crushing, sieving], and packing), and ISMA assembly (fluid train build-out, pump calibration, reagent loading). Field labor includes groundwater grab sample collection pre- and post-deployment, ISMA installation and retrieval, and sample handling (delivery or shipment to the certified laboratory for analysis). These costs are expected to decrease with deployment experience and efficiency; we estimate labor costs to decline by approximately two-thirds.

Materials costs included consumable ISMA components and field equipment rentals (Table S2). Consumables are projected to remain similar between deployments, however equipment rentals will varying depending on site needs. At the NASNI deployment location, it was necessary to utilize cable ramps (i.e., high traffic conditions), which increased overall rental costs. Additionally, boom truck (and operator) rental is a necessary reoccurring expensive. Currently the full ISMA requires a mechanical means for deployment and retrieval. Sample processing and analytics will also vary between deployment sites. Costs are a function of the selected certified commercial laboratory used for the analysis, and the number and type of analytical procedures performed. At NASNI, 14 samples were sent to the lab (i.e., 10 column effluents, 2 influent controls [no column], and two grab samples from the well [pre- and post-deployment]), and analyzed using 7 different methods (Table S2). Energy costs calculated for the duration of the deployment were minimal. Assuming maximum power needed to run the three pumps (30 Watts) over the 35 day deployment, and taking into account industrial electricity costs for California (15.50 cents per kilowatt-hour) [25], total cost for the deployment was just under $4.00.

For comparison purposes, costs estimates for traditional laboratory batch bottle and flow-through column treatability studies, and pilot field-scale remediation costs are shown in Figure 5. Lab-scale cost estimates were generated by industry partners for an identical batch and column study performed in the laboratory. Pilot-scale estimates were calculated from previous in situ injections completed at the site. Incurred costs of the ISMA deployment were comparable to laboratory batch bottle studies. With the ISMA markup included, total deployment costs are estimated at $101K, bracketed between the laboratory batch ($54K) and column ($198K) studies. The in situ pilot field-scale costs were substantially higher than laboratory studies ($559K), and are based on amendment injection and monitoring activities employed at a subsection of the site.

Figure 5.

Cost comparison of treatability study methods, including the ISMA deployment. An ISMA deployment with a profit markup was included for a more realistic comparison to other studies with these markups embedded.

3.4. Limitations

The diameter of the ISMA limits it applicability to wells 10-cm ID or larger, however 10-cm ID wells are one the most commonly employed monitoring well sizes at hazardous waste sites [44]. Additionally, the length of the device requires a saturated thickness of 8.3 m to fully submerse the unit during incubation in the well. If the water depth in the well does not meet these requirements, this issue can be overcome by relocation the groundwater intake location on the instrument body. Additionally if the total well depth is too shallow, the number of shell segments used in the deployment may be reduced. Currently the full ISMA is capable of capturing 8.4 L (total) of column effluent, which correspondingly determines the maximum deployment duration at a given pump rate. This constraint can be overcome by decreasing the number of columns below the maximum (i.e., 10), and diverting column effluent into fewer capture receptacles.

The length of the glass column microcosms (25 cm) limits column residence times, which becomes important for groundwater systems with high flow velocities. In these instances, slower reactions (including natural attenuation) may not go to completion or may not be visible within the microcosm. Conversely, if only a comparatively minor change is observed over the 25 cm column, then the implications of that result are substantial when the results are extrapolated to the subsurface environment. Microcosms in the ISMA are packed with previously collected site sediment that has been processed, similar to laboratory column studies. Consequently only cm-scale heterogeneities exist in the columns and may not represent the macroscale heterogeneities such as clay lenses and differential flow paths, which can impact remediation success [45, 46]. These factors can be assessed to some degree during a thorough site investigation, however even with field-scale pilot testing these subsurface heterogeneities may not be uncovered. In this study, subsequent sediment analyses for sorbed contaminants were not conducted. However, sediment analyses would be of particular interest if sorbents were included in column construction as the selected remediation technology to be tested.

Peristaltic pumps are advantageous because they meter fluid accurately, require little maintenance, and offer a broad range of flow rates. However peristaltic pumps cannot be deployed greater than 6 m below the water table due to limitations inherent to all peristaltic pumps: they cannot control a differential pressure above 1 bar (corresponds to 10 m depth). This challenge may be overcome with the addition of a pressure reducer at the groundwater intake.

4. Conclusions

The successful development and field demonstration of this remedial design tool, the In Situ Microcosm Array, shows its suitability for screening multiple, mutually exclusive treatment technologies in a groundwater monitoring well. Removing the need to collect, transport and store groundwater in a laboratory, the ISMA helps to reduce sources of error and uncertainty in the analysis, by preserving the chemical and biological signature of the groundwater. The IMSA combines the benefits of laboratory studies (tight system control, more complete mass balances, and no release of agents into the aquifer) with the greater realism of studies conducted in situ. It utilizes site groundwater extracted directly from the aquifer, ensuring minimal disturbance of groundwater chemistry and the indigenous microbial community. Use of the device requires access to a single groundwater well, which is left unimpaired upon ISMA removal. The ISMA provides a conduit for accelerated translation of remediation technologies from lab to field, by reducing the risk associated with generating in situ performance data. This technology has been validated by a third party through the ESTCP certification program [47]. This device may also find additional applications in bioprospecting in extreme environments (e.g., hot springs), risk assessment of genetically modified microorganisms, and fate and transport studies of new materials prior to mass production and environmental release.

Highlights.

• Study showcases a novel tool for screening site remediation technologies

• Autonomous device conducts treatability studies in groundwater wells

• Multiple treatment technologies are tested simultaneously in replicates

• Fully-encapsulated instrumentation does not impair the well or aquifer

• Testing is conducted prior to transfer of groundwater to the surface