Light Nonaqueous Phase Liquid Recovery Research Articles

Temporal moment-based approach to understand the dissolved-phase LNAPL recovery and associated characteristics in the porous system under dynamic groundwater table conditions

Abstract The dissolved-phase hydrocarbon recovery can be the first step in decontaminating the soil–water system if spilled with light/dense non-aqueous phase liquid (L/D-NAPL). This study proposes a temporal moment-based approach to investigate the effectiveness of groundwater table manipulations for recovering dissolved-phase byproducts of light non-aqueous phase liquid (LNAPL) from the subsurface system. Temporal moments were computed utilizing experimentally observed and HYDRUS-simulated dissolved-phase toluene concentration data, representative of LNAPL, under stable and dynamic groundwater table fluctuation (GWTF) scenarios. Zeroth temporal moment (ZTM) showed that the hydrocarbon mass recovery varied from 1,804 to 5,190.6 mg/L × h, with the highest variation for the rapid GWTF scenario. An increase in the ZTM of hydrocarbon was observed with an increase in the rate of change of magnitude of the water table and pore velocity fluctuation as in the case of a rapid GWTF as compared to a stable GWTF case. The value of mean residence time for the stable groundwater table case was highest for the entire experimental duration, followed by slow, general, and rapid cases. Temporal moment analysis revealed that the high dissolved-phase hydrocarbon recovery could be achieved by manipulating groundwater table conditions. The present study provides a powerful technique to improve dissolved hydrocarbon remediation in mineral aquifers using hydrological restorations.

Bioventing revisited: efficacy of enhanced biodegradation for sites with mobile LNAPL

Bioventing is a remediation technology that enhances aerobic biodegradation of petroleum-affected soil in the vadose zone by introducing oxygen to the subsurface. Bioventing was historically considered effective for decreasing petroleum hydrocarbons concentrations in soil but discounted for sites where mobile light non-aqueous phase liquid (LNAPL) accumulated in wells. While the science behind bioventing has not changed, the conceptual understanding of LNAPL depletion processes and framework for evaluating the efficacy of LNAPL remediation technologies has changed markedly since the 1990s. This shift leads to a new perspective on the utility and effectiveness of bioventing compared to other LNAPL remedial alternatives. A case study is used to illustrate that mass depletion by bioventing often outperforms hydraulic recovery over time. Biodegradation processes enhanced by bioventing deplete LNAPL constituents in mobile and residual LNAPL in the LNAPL smear zone, which means that bioventing can address a larger mass of LNAPL and may induce a beneficial composition change. Hydraulic LNAPL recovery technologies only access the mobile LNAPL and do not induce a composition change. Furthermore, for low LNAPL recoverability (transmissivity), bioventing typically uses less energy and produces less waste per unit mass of hydrocarbon removed than hydraulic recovery, making bioventing a more sustainable remedial technology.

Impact of water table fluctuations on the seasonal effectiveness of the pump-and-treat remediation in wet–dry tropical regions

The pump-and-treat (P&T) method is a common remediation approach that is used to mitigate occurrences of hydrocarbon contamination. In a well-studied site that had been contaminated by a large volume of jet fuel, continuous monitoring of the water table and floating phase thickness revealed the most significant process governing the effectiveness of the site’s active remediation system. The floating phase thickness recorded in the wells monitored varied greatly (> 0.50 m) and was negatively correlated with the water table level. Although the dependence of light non-aqueous phase liquid (LNAPL) thickness on water table fluctuations is widely recognized, LNAPL recovery, which is governed by alternating cycles of LNAPL release and entrapment in pore spaces, has been poorly described. Thus, we present a specific case in which LNAPL recovery is expected only episodically, when the water table falls sufficiently. In the period spanning from 2006 to 2008, LNAPL remediation recovered nearly 180 m3 of oil. In later years, the volume of recovered LNAPL declined and ceased between 2010 and 2014, when the water table rose. Importantly, our research demonstrates that the P&T remediation approach is solely effective during a period of 4 months in dry years. Thus, cleanup methods and contaminated site management strategies should be reconsidered.

LNAPL Recovery Endpoints: Lessons Learnt Through Modeling, Experiments, and Field Trials

AbstractIt is important to estimate what light nonaqueous phase liquid (LNAPL) recovery can be practicably achieved from subsurface environments. Over the last decade, research to address this included a broad field program, laboratory measurements and experimentation, and modeling approaches. Here, we consolidate key findings from the research in the context of current literature and understanding, with a focus on a well‐validated, multiphase multicomponent modeling approach to achieve estimates of reasonable endpoints for LNAPL recovery. Simple analytical models can provide approximate saturation distributions and estimates of LNAPL recoverability via transmissivity approximation, but are insufficient to predict LNAPL saturation‐ and composition‐based recovery endpoints for various recovery technologies. This is because they cannot account for multiphase, multicomponent fate and transport and key processes such as hysteresis. Recent advances to improve estimates of the fraction of recoverable LNAPL and its transmissivity are summarized. These advances include further development and application of a well‐validated model to characterize active LNAPL recovery endpoints. We present key factors that affect the determination of LNAPL recovery endpoints, and outline how recovery endpoints are affected by natural source zone depletion (NSZD—currently gaining acceptance as a LNAPL remediation option). Major factors include geo‐physical characteristics of the formation, magnitude of an LNAPL release and partitioning properties of the key LNAPL constituents of concern. Based on the capabilities of the validated model, the paper also provides a basis to optimize LNAPL recovery efforts.

Investigation of the Source Relationship between Benzene Vapor and Alkylate-Containing Light Non-Aqueous Phase Liquid at Adjacent Industrial Facilities

This study describes the investigations undertaken at two adjacent properties for known contaminants of concern and for resolution of other “unexpected” contaminant sources. One property is a manufacturing facility with impacts from chlorinated solvent compounds releases at the site. During site investigations, benzene was found present in subslab and multi-depth soil vapor probes installed beneath the basement floor of on-site buildings. Additional investigations at the manufacturing site were initiated in response to a recent, unexpected appearance of an unknown light non-aqueous phase liquid (LNAPL) in three on-site perched water zone wells as observed during sequential quarterly monitoring events and in a soil vapor extraction well. The other property is an adjacent petroleum facility that has ongoing LNAPL recovery operations from past releases at its site. A claim was made that benzene vapor at the manufacturing property was solely related to the LNAPL that migrated onto the facility. Multiple lines of evidence, including LNAPL migration pathways and chemical fingerprints, were used to show that the LNAPL was sourced from the petroleum facility. Further, benzene vapor present in subslab vapor probes installed for vapor intrusion purposes was believed to be sourced from the LNAPL and not from historical operations at the manufacturing facility. However, detailed analysis of vapor distributions, the subsurface stratigraphy, and LNAPL characteristics indicate that spills during past operations at the manufacturing facility itself were likely the source of the benzene vapor found in the subslab vapor probes.

Impact of heterogeneous properties of soil and LNAPL on surfactant-enhanced capillary desaturation

This paper investigates low-concentration (<1wt%) surfactant flushing when used as a follow-up technology for multiphase vacuum extraction on heterogeneous sites. Challenges posed by soil permeability, pore-size distribution, mineralogy, light non-aqueous phase liquid (LNAPL) weathering and groundwater hardness were quantified through batch and soil column tests. Compatibility issues between the mixed mineralogy soils, hard groundwater, mixed LNAPL and usual anionic surfactants were observed. The selected solution was a Winsor type I system promoting an interfacial tension of 0.06mN/m between the site LNAPL and the amphoteric surfactant CAS in aqueous solution at pH12. Surfactant loses to adsorption and pore media plugging were observed in the fine soil fraction. The capillary desaturation curves (CDC) obtained with the column tests suggested mixed-wettability behavior. The soil permeability strongly influenced LNAPL recovery, as expressed by the relationship obtained between capillary numbers (NCa) and hydraulic gradients. In this case, the critical NCa, marking the onset of capillary desaturation, could only be obtained with realistic hydraulic gradients in the coarse soil fraction. At those gradients, potential LNAPL recovery was 30% at the most. Unlike previously published CDCs, the relationship between NCa (log-scale) and LNAPL recovery was not linear but dependant on residual LNAPL saturation.

Transmissivity as a Primary Metric for LNAPL Recovery—Case Study Comparison of Short‐Term vs. Long‐Term Methods

This article presents a case study and comparative analysis of light nonaqueous phase liquid (LNAPL) transmissivity estimated using short‐ and long‐term test methods at an active petroleum refinery. LNAPL transmissivity (Tn) is a recognized direct indicator of LNAPL recoverability with increasing acceptance by regulatory agencies. Historical releases at a refinery resulted in widespread LNAPL accumulations across the site and, as such, a focused approach is being implemented to enhance recovery, shorten remedial timeframes, and prioritize areas for recovery. Groundwater pumping systems operate continuously to maintain hydraulic containment of impacts, along with 12 LNAPL recovery systems. Transmissivity has been established as a primary metric and management tool for LNAPL recovery at the refinery. In this case study, estimated transmissivity values from short‐term data (baildown testing) and long‐term data (LNAPL skimming operations) from the same locations are analyzed and compared. Overall results are presented with respect to variations in transmissivities between the short‐ and long‐term tests, significance of data collection and quality, and consideration factors affecting transmissivity including fluid properties, soil types, hydrogeology, saturation levels, tidal effects, migration rates, and receptor risks. Additionally, the application of transmissivity as a metric for monitoring progress toward LNAPL recovery endpoints as part of the LNAPL remediation program development is discussed. ©2015 Wiley Periodicals, Inc.

USING THE STOMP SIMULATOR TO ASSESS BIOSLURPING RECOVERY AND REMEDIATION PROCESSES OF LIGHT HYDROCARBONS IN CONTAMINATED AREAS

Research documented in this thesis focuses on the multiphase flow and transport of Light Nonaqueous Phase Liquid (LNAPL) type contaminants as mediated by application of bioslurping remediation, one possible configuration of multiphase extraction (MPE). With the cooperation of the hydrology team of PNNL (Pacific Northwest National Laboratory, WA, USA), the numerical simulator STOMP (Subsurface Transport Over Multiple Phases) was modified to include bioslurping processes in 3-D assuming the presence of a fixed tube tip extraction position and using the well model concept. A base model was built using data from a hypothetical LNAPL recovery case study presented in the literature, combined with data from actual MPE field systems. The model was used to carry out a range of sensitivity analyses. These analyses were used to evaluate the effects of different elevations of the bioslurping tube tip, the progressive lowering of the tube tip, the extraction radius of influence, and the effects of vacuum pressure variations on oil recovery. The modeling results were found to be consistent with the expected physical behavior. A field case study of bioslurping was also conducted with the modified simulator. LNAPL recovery was adequately reproduced, with the predicted mass of extracted water being smaller than the field observations. In terms of free phase removal, the simulator adequately represented the multiphase extraction process intended with the bioslurping technique.

Estimating Remediation and Contaminant Respiration Emissions for Alternatives Comparisons at Petroleum Spill Sites

Greenhouse gas emissions assessments for site cleanups typically quantify emissions associated with remediation and not those from contaminant biodegradation. Yet, at petroleum spill sites, these emissions can be significant, and some remedial actions can decrease this additional component of the environmental footprint. This article demonstrates an emissions assessment for a hypothetical site, using the following technologies as examples: excavation with disposal to a landfill, light nonaqueous‐phase liquid (LNAPL) recovery with and without recovered product recycling, passive bioventing, and monitored natural attenuation (MNA). While the emissions associated with remediation for LNAPL recovery are greater than the other considered alternatives, this technology is comparable to excavation when a credit associated with product recycling is counted. Passive bioventing, a green remedial alternative, has greater remedial emissions than MNA, but unlike MNA can decrease contaminant‐related emissions by converting subsurface methane to carbon dioxide. For the presented example, passive bioventing has the lowest total emissions of all technologies considered. This illustrates the value in estimating both remediation and contaminant respiration emissions for petroleum spill sites, so that the benefit of green remedial approaches can be quantified at the remedial alternatives selection stage rather than simply as best management practices. ©2015 Wiley Periodicals, Inc.

222 Rn as Natural Tracer for LNAPL Recovery in a Crude Oil-Contaminated Aquifer

The objective of this study was to investigate whether 222Rn in groundwater can be used as a tracer for light non-aqueous phase liquid (LNAPL) quantification at a field site treated by dual-phase LNAPL removal. After the break of a pipeline, 5 ha of soil in the nature reserve Coussouls de Crau in southern France was contaminated by 5100 m3 of crude oil. Part of this oil seeped into the underlying gravel aquifer and formed a floating oil body of about 3.9 ha. The remediation consists of plume management by hydraulic groundwater barriers and LNAPL extraction in the source zone. 222Rn measurements were performed in 21 wells in and outside the source zone during 15 months. In uncontaminated groundwater, the radon activity was relatively constant and remained always >11 Bq/L. The variability of radon activity measurements in wells affected by the pump-and-skim system was consistent with the measurements in wells that were not impacted by the system. The mean activities in wells in the source zone were, in general, significantly lower than in wells upgradient of the source zone, owing to partitioning of 222Rn into the oil phase. The lowest activities were found in zones with high non-aqueous phase liquid (NAPL) recovery. LNAPL saturations around each recovery well were furthermore calculated during a period of high groundwater level, using a laboratory-determined crude oil–water partitioning coefficient of 38.5 ± 2.9. This yielded an estimated volume of residual crude oil of 309 ± 93 m3 below the capillary fringe. We find that 222Rn is a useful and cheap groundwater tracer for finding zones of good LNAPL recovery in an aquifer treated by dual-phase LNAPL removal, but that quantification of NAPL saturation using Rn is highly uncertain.

Optimized prefabricated vertical wells for light nonaqueous phase liquid recovery

Work presented herein examines optimization of prefabricated vertical well (PVW) performance during the extraction of light nonaqueous phase liquids (LNAPLs) in liquid and vapor phases in terms of ...

Multiphase Extraction of Light Non-aqueous Phase Liquid (LNAPL) Using Prefabricated Vertical Wells

Work in this paper describes the results of a field and modeling study conducted to investigate the extraction of LNAPL (jet propellant 4, i.e., JP-4, jet fuel) using Prefabricated Vertical Wells (PVWs) at a former air force base. Field testing consisted of 185 operating hours on 25 rows, each consisting of 7 or 8 PVWs. A total of 133 L of free-phase liquid is removed from the subsurface in addition to 467 kg of organics in the vapor phase as a result of volatilization due to the application of vacuum heads exceeding the JP-4 vapor pressure. Monitored field data are used to calibrate a numerical model for investigating the effect of irreducible water content, LNAPL, and gas contents on system performance. Analysis results show that increasing the irreducible water content from 5 % to 20 % decreases the free-LNAPL specific volume from 0.08 m to 0.067 m, which led to reduction in removal efficiency. Under a continuous saturation condition, more wetting fluid (water), traps the nonwetting fluid (LNAPL) in the soil pores, and reduces its mobility, which eventually impedes the LNAPL recovery. Given the model parameters that simulate field conditions, the PVWs show best removal rates in the gas phase. The model results indicated that 292 L of JP-4 can be recovered in 19.5 days (based on 8 h of operation per day) with an effective rate of 1.87 L/h. At the end of 4.25 years of system operation, the maximum benzene concentration is 0.002 ppm. At a gas extraction rate of 10,273 L/min, a maximum concentration of 0.38 ppm in the vapor phase is achieved after 2.5 years of vapor extraction.

Solar‐powered in situ soil washing with surfactant to collect LNAPL

AbstractThis article presents the findings of a sustainable, surfactant‐enhanced, product recovery pilot‐scale study (PSS) completed between January 2010 and May 2010 at the Hydrocarbon Burn Facility located at the John F. Kennedy Space Center in Florida. The goal of this study was to implement a unique, simple, and sustainable light nonaqueous‐phase liquid (LNAPL) recovery process and evaluate site‐specific volumes and rates of LNAPL that could be collected and the degree of soil and groundwater cleanup that could be achieved. The recovery process was a combination of groundwater recirculation at a rate of approximately 2.9 gallons per minute (11.0 liters per minute), soil washing via LNAPL mobilization, and collection of LNAPL via a hydrophobic LNAPL skimmer. A biodegradable surfactant, ECOSURFTM SA‐15, was added to the recirculation line to lower the interfacial tension and facilitate LNAPL recovery via mobilization. All equipment (submersible pump, LNAPL skimmer, surfactant feed pump, controls, and various other equipment) used was powered by a solar panel array. Approximately 60 gallons (227 liters) or 429 pounds (195 kilograms) of LNAPL were collected at the recirculation site over approximately three months during the PSS. The data suggest that surfactant amendments greatly enhanced free product collection. The maximum rate of free product collection was approximately 1 gallon (3.8 liters) per day. © 2012 Wiley Periodicals, Inc.

Characteristics of light nonaqueous phase liquid recovery in the presence of fine‐scale soil layering

The effects of fine‐scale layering of soil properties on light nonaqueous phase liquid (LNAPL) recovery using pumped groundwater drawdown were investigated using a Monte Carlo approach. Three‐dimensional simulations were performed to study the characteristics of a variety of measures describing LNAPL recovery in heterogeneous sandy aquifers. Heterogeneity models included both uncorrelated and correlated vertical permeability distributions, with both permeability‐scaled and unscaled retention characteristics. Lognormal permeability variances of 0.594 and 1.188 were considered. The Monte Carlo ensemble mean LNAPL recovery rates and cumulative recovery volumes differ significantly from homogeneous soil models. The ensemble variances are large, and confidence intervals on recovery measures are wide. Generally, lognormality is not preserved; recovery rates and volumes display asymmetric densities with standard deviations of the order of the means. Overall recovery efficiency depends critically on the layer permeability at specific elevations. Practical outcomes for field practitioners are discussed including the degree to which substantially increased times may be required for LNAPL recovery and the increased occurrence of situations where larger than expected volumes will remain unrecovered. Guidance is also given for the key measurements of layer permeabilities and their expected influence on recovery performance. It is concluded that moderate vertical layering of soil properties is sufficient to render homogeneous recovery models inappropriate.

LNAPL in fine‐grained soils: Conceptualization of saturation, distribution, recovery, and their modeling

AbstractA simple conceptual model is presented that leads to a quantitative description of the behavior of light non–aqueous phase liquid (LNAPL) in fine‐grained soil (FGS). The occurrence of large (15 feet) (4.6 m) LNAPL accumulations in observation wells in FGS and of LNAPL located below the water table is explained by macropore theory and capillarity of the FGS. Using soil capillary data, fluid property data, and a simple spreadsheet model, the LNAPL saturation in a soil profile and LNAPL recovery were predicted for a field study site. The predicted LNAPL distribution, saturation, and recovery matched the field observations and actual LNAPL recovery. Measured LNAPL saturations were &lt;2%, while model‐predicted values were &lt;3%. The model predicted recovery of ∼530 gallons (2009 L). After 1.5 years of continuous operation, a three‐phase, high‐vacuum extraction system recovered 150 gallons (568 L) of LNAPL. Application of a model that assumes homogeneity of the soil that is heterogeneous at a small scale may seem to be a misapplication; however, conceptualizing the model domain at a sufficiently large scale (3 to 6 feet; 0.9 to 1.8 m) allows for the FGS to be viewed as a homogeneous medium with small effective porosity.

A Modeling and experimental study of light nonaqueous phase liquid (LNAPL) accumulation in wells and LNAPL recovery from wells

A formulation is presented for numerical modeling of light nonaqueous phase liquid (LNAPL) accumulation in observation wells and LNAPL removal from recovery wells. The formulation includes fluid gravity segregation, changing water and LNAPL levels in the well, and the corresponding changes in fluid storage in the well bore. The method was added to a three‐dimensional finite difference model for hysteretic multiphase flow. The model predictions are compared to three‐dimensional pilot scale experiments of LNAPL infiltration and recovery. Model predictions matched experimental results well, demonstrating the validity of the model formulation. Model results indicated that soil layering and lack of mechanical equilibrium had a significant influence on LNAPL thicknesses in wells. Thus empirical approaches based on assumptions of soil homogeneity and mechanical equilibrium would not have been appropriate for prediction of LNAPL volume from toluene thicknesses measured in wells.

Soil Properties and Design Factors Influencing Free-Phase Hydrocarbon Cleanup

Light nonaqueous phase liquid (LNAPL) saturation and movement in the subsurface are controlled by soil capillary characteristics, permeability, and fluid properties. Where free-product occurs in monitoring wells, LNAPL saturations in the formation vary significantly as a function of the observed thickness in the monitoring well and soil type. Fine-grained soils generally exhibit lower LNAPL saturations than coarse-grained material for the same observed thickness in a monitoring well. MAGNAS3 [MAGNAS3. HydroGeoLogic, Inc., Herndon, VA, 1992; Huyakorn, P. S.; Panday, S.; Wu, Y. S. J. Contam. Hydrol. 1994, 16, 190−130], a three-dimensional, finite-element model that can simulate movement of three active phases (air, water, and LNAPL), was used to investigate LNAPL recovery in three different soil types and using several recovery designs to examine the effect of these factors on LNAPL recovery. The results of this analysis show that, because the relative mobility of LNAPL decreases with decreasing saturation ...